1. History Of Sanskrit Poetics Sushil Kumar De 2 Volumes bound in One

HISTORY OF SANSKRIT POETICS

SUSHIL KUMAR DE M.A. (Calcutta), D.LIT. (London) Formerly Professor of Sanskrit in the University of Dacca and in the Postgraduate Research Department, Calcutta Sanskrit College ; Professor Emeritus, Jadavpur University, Calcutta ; and Honorary Fellow, Royal Asiatic Society

IN TWO VOLUMES

Second Revised Edition

EPI.KANAIAAA MOYKenAAIAIA

Firma K. L MUKHOPADHYAY CALCUTTA

1960

C Second Edition 1960

Firma K. L. Mukhopadhyay, Publishers 6/IA, Bancharam Akrur Lane, Calcutta-I2

357- Price, Rs. 30.00A

Printed by J. C. Sarkhel at the Calcutta Oriental Press Private Ltd. 9, Panchanan Ghosh Lane, Calcutta 9

To

The Memory of My Father

PREFACE The first edition of this work was published in two separate volumes in 1923, 1925. Almost simultaneously in 1923 appeared the second edition of the Sāhitya-darpana by Mahāmahopādyāya Dr P. V. Kane, to which was prefixed an introduction of 177 pages dealing with the history of Alamkāra literature, subsequently elaborated in the third edition (1951) into 423 pages. Referring to these facts Dr Kane writes: "Both of us were thus practically pioneers in this field. Our books led many scholars to study the numerous works on Alamkāra, to point out defects and mistakes, to produce papers dealing with several as- pects of the Alankāraśāstra and to publish several important texts." This will make clear the necessity of revising the present work, which was written and published independently, but which, not yet superseded, had been out of print for more than thirty years. The scope of the present work was sufficiently explained in the preface to the first edition. An attempt has been made to approach the subject from the historical point of view, instead of presenting a bare account of the works and authors or a mere epitome of the different topics of Alamkāra. Presuming a general knowledge and taking the representative writers (and, as far as necessary, the neglected commentators and so-called minor writers), the development of Sanskrit Poetics has been traced through its fairly long and varied course of history, which covers an extensive literature of more than a thousand years. The first volume of the work, therefore, deals with the preliminary but important question of a working Chronology and original Sources, on the basis of which the second volume proceeds to set forth the history through divergent Systems and Theories. It will be seen that, barring some

( ii )

rearrangement and division into chapters, the general form of the first edition has not been changed ; but for conve- nience the two volumes are now published in one, divided into parts. In investigating the growth of Sanskrit Poetics it was, of course, not possible to ignore its content; but it has been thought more useful to lay stress on the essentials of doctrines and omit minor details, the object of this work being more historical than expository. It was not possible, for instance, to give a detailed technical analysis of indivi- dual rhetorical "figures", with which the Alamkāra-sāstra is traditionally and popularly associated ; but the general doctrine of poetic figures has been dealt with, in so far as they are not mere tickets of nomenclature but positive agents in the production of artistic beauty. The study of analytical Rhetoric, apart from its value as a formal dis- cipline, may be regarded as pedantic and futile ; but Sanskrit Alamkāra-śātra, in spite of its name, possesses a speculative interest by involving, besides mere Rhetoric, a great deal of what is known to-day as Criticism or Aesthetic. It was almost impossible for the Ālamkārikas, concerned that they were with form and technique, not to be interested in the general phenomenon of literature or theorise on general principles. For a similar reason of historical and general treatment greater emphasis has been laid on the earlier writers who cover the creative periods of the discipline; of later writers who show in the main an excessive dependence on their predecessors, only typical names have been selected with a due regard to their historical as well as intrinsic importance. The author has confined himself to Sanskrit Poetics, and has not taken into account the numerous vernacular works on the subject, as well as the allied but distinct theme of Dramaturgy ; for they deserve more room for a detailed and separate treatment. The Bibliographies and references are perhaps not exhaustive; but they are

( iii )

meant to supply such notable information as might lead to a further study. A word of explanation may be deemed necessary for the use of the term Poetics to designate a half-theoretical and half-practical discipline which goes by the name of Alamkāra-śāstra, Its scope has been sufficiently explained by H. Jacobi, who seems to have been the first to make this term current, in ZDMG Ivi, 1902, p. 393, fn 1. It may be pointed out that the commonly used term Rhetoric is inadequate in explaining the standpoint of a study which includes a great deal more than a mere normative treatment of rhetorical categories; while the expression Aesthetic in this connexion is misleading, inasmuch as the theoretical scope of Alamkāra literature is not co-extensive with what is understood by that term in modern philosophical studies. Comparative estimates, therefore, with reference to Western Aesthetic or European Critical literature have been generally avoided ; for that would not only exceed the limited scope of the present work but might lead to sweeping or mislead- ing generalisations. Those who are interested will find a short critical outline of Sanskrit Poetics in the light of modern Aesthetic in two essays by the present writer published in the Dacca University Studies, vol. i (1936), pp. 1-46, and in the New Indian Antiquary ix, nos. 1-3 (reprinted in the author's Some Problems of Sanskrit Poetics, Calcutta 1959, pp. 1-53 ). It is needless to say that, as the references will indicate, a careful use has been made of the accumulated but scattered labours of previous workers, but it has been supplemented by the author's own studies. Useless dis- cussion or controversy, however, has been avoided as much as possible, the author confining himself to stating his own view and leaving it to speak for itself. He is fully conscious that no one can claim he has said the last word on the subject. In the preface to the first edition the author recorded

• ( iv )

with sincere thanks the manifold help and encouragement he received in writing this work. But it is sad to recollect in this connexion the departed kalyāna-mitras, H. Jacobi, L. D. Barnett and F. W. Thomas. He has only to add that in the present edition he has profited by the suggestions of Protessor Dr V. Raghavan who took the trouble, in the midst of his manifold preoccupation, of sending useful bibliographical notes on some works and authors.

CONTENTS

VOLUME I

Chapter Page

I. Beginnings 1

II. Bharata 18

Commentators: Mātrgupta. Udbhața, Lollața. Śankuka. Bhatța Nāyaka. Harșa. Kīrtidhara. Abhinavagupta. Nānyadeva.

III. From Bhāmaha to Ānandavardhana ... 46

Bhāmaha. Daņdin. Udbhața. Mukula. Pratīhārendurāja. Rājānaka Tilaka. Vāmana. Rudrața and Rudrabhatta. The writers on Alamkāra in the Vișnu-dharmottara and Agni-purāņa.

IV. The Dhvanikāra and Ānandavardhana ... 101

Abhinavagupta

V. From Rājaśekara to Mahimabhațța ... 115

Rājaśekhara Dhanañjaya and Dhanika. Kuntaka. Kşemendra, Bhoja. Mahima- bhaṭṱa.

VI. Mammața and Allața ... 145

Commentators on Mammața ... 156

VII, From Ruyyaka to Vidyānātha ... 178 Ruyyaka. Hemacandra. The Vāgbhatas. Jayadeva. Vidyādhara. Vidyānātha

( vi )

Chapter Page

VIII. From Viśvanātha to Jagannātha ... 212

Viśvanātha. Keśava Miśra and Śauddhodani. Appayya Dīkşita. Jagannātha. Nāgojī Bhațța.

IX. Later Writers on Rasa and on Kavi-śikşā 238.

Šāradātanaya. Singa-bhūpāla. Bhānudatta. Rūpa and Jīva Gosvāmin. Viśvanātha Cakravartin, Kavi-karņapūra. Kavicandra. Arisimha and Amaracandra. Rājaśekhara. Deveśvara. Other writers.

X. Minor Writers on Alamkära ... ... 263

Conclusion 321 ... ...

Abbreviations 331 ... ...

Additions and Corrections ... 336 ...

Index of Authors and Works on Alamkāra cited in vol. i 337

CONTENTS

VOLUME II

Chapter Page . I. From Beginnings to Bhämaha 1 II. Bhāmaha, Udbhața and Rudrața 32 III, Daņdin and Vāmana ... ... 75 IV. Lollata, Śańkuka and others ... 108

V. The Dhvanikāra and Ānandavardhana ... 139

VI. Abhinavagupta and the Reactionary Systems 176

VII. Mammata and the New School 213 VIII. Some Later Writers of the New School 243

IX. Later Writers on Rasa 259

X. Writers on Kavi-śikşā ... 283

Subject-Index to vol. ii 299 ...

VOLUME I

CHRONOLOGY AND SOURCES

CHAPTER I

BEGINNINGS

( 1 ) There is an interesting passage in the Kāvya-mīmāmsā in which Rajaśekhara, while giving us a somewhat fanciful account of the divine origin of Poetics, mentions at the same time the names of the supposed original exponents of the discipline. It is said that the Spirit of Poetry (kāvya-purușa), born of the Goddess of Learning (sarasvati), was set by the Self-existent Being to promulgate the study of Poetics in the three worlds ; and he related it in eighteen adhikaranas to his seventeen will-born pupils These divine sages, in their turn, are said to have composed separate treatises on the portions respectively learnt by them. Thus Sahasrāksa wrote on kavi- rahasya, Uktigarbha on auktika, Suvarņanābha on rīti, Pra- cetāyana on anuprāsa, Citrāngada on yamaka and citra, Śeșa on śabda-śleşa, Pulastya on vāstava, Aupakāyana on upamā, Pārāśara on atiśaya, Utathya on artha-śleșa, Kubera on ubhayālamkāra, Kāmadeva on vainodika, Bharata on rūpaka. Nandikeśvara on rasa, Dhișaņa on dosa, Upamanyu on guņa and Kucamāra on aupanisadika. This tendency on the part of a Sanskrit author towards glorifying his science and there- by investing it with an ancient unalterable authority is not unusual, and such legendary accounts are often fabricated where the actual origin is forgotten ; but it is curious that we do not meet with them elsewhere in Alamkāra literature, al- though they find expression in Bharata and in Vātsyāyana with regard to the origin of the allied disciplines of Drama- turgy and Erotics respectively. The historical value of this passage of Rājaśekhara may, indeed, be well doubted ; but it is possible that this unique account, apart from its obviously mythical garb, embodies a current tradition, implying the actual existence, at some remote and forgotten period, of

2 SANSKRIT POETICS

early expounders of poetic theory, some of whose names are still familiar, but most of whose works have apparently perished. Thus, Suvarņanābha and Kucamāra (or Kucumāra) are also cited with reverence by the author of the Kāma-sūtra (i. 1. 13, 17), both of them as authorities on Erotics, but the latter especially as well-versed in the particular subject of aupanisadika, which is thus included in Erotics as well as in Poetics.1 The present-day text of the Natya-śastra, which goes by the name of Bharata, deals in an encyclopaedic manner with various topics, but Bharata is well known, as described by Rājaśekhara, as an authority on rūpaka. No work of Nandikeśvara on rasa has yet been discovered ; but his name is associated with a number of works, mostly late compilations, on erotics, music, histrionic art, grammar and Tantra.2 This traditional account may lead one to make the tempting suggestion of a very early systematic investigation of rhetoric- al issues but for the serious difficulty that there is hardly any material in the ancient literature itself to enable us to trace the origin of Poetics to a very remote antiquity. The Alamkāra-śāstra is never mentioned among the orthodox disciplines which constitute the so-called Vedangas, nor do we meet with any passage in the Vedic Samhitās, Brāhmanas or the earlier Upanisads in which we may find a real basis for a system of Poetics. The word upama, for instance, is found as early as the Rg-veda (v. 34. 9 ; i. 31, 15), and Sāyana ex- plains it in the sense of upamāna (as in Pāņini ii. 3. 72) or drstanta ; but there is nothing unusual in this use of the general idea of similitude, which need not be interpreted as having a particular speculative significance. It is conceded, on the authority of Yaska and Pānini, that the conception of upama or similitude considerably affected the Vedic language as well as its accent ; but beyond this grammatical or philo-

1 Cf. JDL iv p. 95. 2 For Nandikeśvara see below under ch. ii.

BEGINNINGS 3

logical interest, there is no indication of a dogma, much less of a theory, of Poetics in Vedic times. Nor should undue emphasis be laid on the use of poetic figures in Vedic litera- ture ; for between this unconscious employment of figures of speech and the concious formulation of a definite system, there must necessarily be a long step.1

( 2 )* The first evidence of a definite, if somewhat crude, activity in this direction is traceable in the Nighantu and Nirukta From the investigation of the peculiarities of the general form of language, which began early, attention was apparent- ly directed to the analysis of the poetic forms of speech; but the question was still regarded entirely from the linguistic point of view. The term alamkāra in the technical sense does not occur in the Nirukta, but Yäska uses the word alamkarisnu in the general sense of 'one in the habit of adorning', which Panini explains in iii. 2. 136 and which is apparently the meaning of the word occurring in the Satapatha Brāhmaņa (xiii. 8. 4. 7; iii. 5. 1. 36), and in the Chandogya Upanişad (viii. 8. 5). But in the Nighantu iii. 13, a list is given of particles of comparison relating to the Vedic upama, comprising twelve varieties, which are illustrated in the Nirukta i. 4, iii. 13-18 and ix. 6. Six of these varieties, viz., those indicated by the particles iva, yathā, na, cit, nu and a, are discussed in connexion with Yāska's treatment of upamārthe nipātas or particles indicative of comparison (i. 4), and partly included in what he designates as karmopamā in iii, 15. Then Yāska mentions bhūtopamā, in which the upamita becomes the upamana in character, and rūpopamā, where the upamita resembles the upamāna in point of form. The fourth variety occurs where the

1 P. V. Kane (HSP, 3rd Ed. Bombay 1951, pp. 315-19 ; also IA xli, 1912, p. 120f) argues at some length that the Rgvedic poets have had some ideas about a theory of Poetics, as well as of Drama and dramatic representation. Also B. N. Bhattacharya in JDL ix, 1923, p. 100f.

4 SANSKRIT POETICS

particle yatha is used. Then comes siddhopamā, in which the standard of comparison is well established and known to surpass every other object in a particular quality or act, and is characterised by the suffix vat. The last variety is lupto- pamā or suppressed simile, also called arthopamā (equivalent to the rüpaka of later theorists), which is illustrated in iii. 18 (also ix. 6), where the example is given of the popular appli- cation of the terms simha and vyāghra in a laudatory and śvan and kāka in a derogatory sense. The term upamāna itself is used by Yaska, but only to denote these particles of comparison (vii. 31). The significance of comparison in general is also referred to in i. 19, ii. 6, iii. 5, iv. 11, v. 22 and vii. 13. Incidentally Yāska quotes (iii. 13) the gramma- rian Gärgya's definition of upama,1 which is important from our point of view. As explained by Durgācārya, it lays down that upama occurs when an object which is dissimilar is reckoned, through similarity, with an object having similar attributes.2 It also states as a general rule that the standard of comparison should be superior in merit and better known than the object of comparison; but the reverse case is also admitted and illustrated (iii. 14-15) by two examples from the Rg-veda (x. 40. 2 ; x. 4. 6). The definition, too wide as it is, recalls Mammata's similar dictum, and undoubtedly estab- lishes a very early, but more or less definite, conception of the poetic upamā. By the time of Pānini this conception of upamā seems to have been tacitly recognised and we find him using in this connexion the technical terms upamāna,3 upamita4 and

1 athāta upamā yad atat tat-sadrśam iti gārgyah, tad āsām karma jyāyasā vā guņena prakhyātatamena vā kanīyamsam vāprakhyātam vopamimīte'thāpi kanīyasā jyāyāmsam. 2 evam atat tat-svarūpeņa guņena guņa-sāmānyād upamīyate ity evam gārgyācāryo manyate. 3 ii. 1. 55 ; iii. 1. 10, 2. 79, 4. 45 ; v. 4. 97, 137 ; vi. 1. 204, 2. 2, 72, 80, 127, 145, 169. 4 ii. 1. 56.

BEGINNINGS 5

sāmānya1 as well as general expressions like upamā2 (used in the sense of the rhetorician's upamāna), aupamya,3 upamārthe4 and sūdrśya.5 It is noteworthy that in nearly fifty sūtras distributed all over his work, Panini incidentally discusses, from the grammarian's point of view, the influence of the conception of comparison on the language in the varied dominion of affixes, including case and feminine suffixes, krt, taddhita and samāsānta terminations,6 in the making of compounds7 and in accent.8 The same influence is also traceable in the idea of atideśa, a term which is not used by Panini himself but which is made clear by his commentators, and which may be translated as 'extended application by analogy or similarity'. Kātyāyana, in several vārttikas,9 follows Pänini in noting the same influence of the idea of similitude, while Sāntanava in his Phițsūtra discusses it in connexion with accentuation.10 In the Mahābhāșya on ii. 1. 55, Patañjali has defined and illustrated Pānini's use of the term upamāna. A māna or measure, he says, is that which is employed in ascertaining a thing unknown; upamāna is approximate to the mäna and determines the thing not abso- lutely (but approximately), e.g. when we say 'a gavaya is like a cow'.11 Strictly speaking, a writer on Poetics will not accept the example adduced by Patañjali as an instance of

1 ii. 1. 55, 56 ; viii. 1. 74. 2 ii. 3. 72. 3 i. 4. 79 ; iv. 1. 69; vi. 2. 113. 4 viii. 2.101. 5 ii. 1. 6-7; vi. 2. 11. 6 i. 4. 79 ; ii. 3. 72 ; iii. 1. 10, 2. 79, 4. 45 ; iv. 1. 69, 4. 9; v. 1. 115-16, 2. 39, 3. 96, 106, 4. 97, 137. 7 ii. 1. 7, 31, 55-6 ; vi. 2. 11. 8 v. 1. 18 ; vi. 1. 204, 2. 2, 11, 72, 80, 113, 127, 145, 169 etc. 9 on i. 3. 21 ; ii. 1. 55, 2. 24, 4. 71 ; iii. 1. 10 etc. 10 e.g. ii. 16, iv. 18. 11 Mānam hi nāmānirjnāta-jnānārtham upādīyate'nirjnātam artham jñāsyāmīti, tat-samīpe yan nātyantāya mimīte tad upamānam, gaur iva gavaya iti, ed. Kielhorn i, p. 397.

6 SANSKRIT POETICS

poetic upamā,1 inasmuch as the characteristic charmingness essential in a poetic figure is wanting in such a plain expres- sion ; but this grammatical analysis of the general idea of comparison is an early and near approach to the technical conception of Poetics.2

( 3 ) A special interest attaches to these rules of Panini and the dicta of early grammarians, inasmuch as they form the basis of what may be termed the grammatical sub-division of the figure upamā into direct (śrautī) and indirect (ārthī) simile as well similes based on krt and taddhita suffixes, recognised as early as Udbhata's time. Thus, the authority for the śrauti upama, in which the notion of comparison is conveyed by particles like yathā, iva, vā or by the suffix vat, when vat is equivalent to iva, is supposed to be based on two rules of Pāņini (v. 1. 115-16), which lay down that the suffix vat is applied to the standard of comparison in the locative or genitive case and takes the place of the case-ending and iva, as well as to a noun which should otherwise be in the instrumental case in the sense of tena tulya (like that), if the similarity consists in an action and not in quality. Thus, we get the forms mathurāvat (=mathurāyām iva) pātaliputre prākāraḥ, caitravat (=caitrasya iva) maitrasya gāvaḥ, as well as brāhmaņavad (=brāhmaņena tulyam) adhīte, but not caitravat krśah. In the same way, we have compounded simile like kumbhāviva stanau, according to the vārttika (ivena samāso vibhaktyalopaḥ) on Pāņini ii. 4. 71, which is, therefore, taken as an instance of compounded direct simile (samāsagā śrautī upamā). Similarly, the ending kyac, accord- ing to Pāņini iii. 1. 10, is applied to a noun in the objective case, which is expressive of upamana, in the sense of

1 go-sadrśo gavaya iti nopamā, Citr. mīm. p. 6. 2 The conception of upama appears to be fully established in

BEGINNINGS 7

'behaviour' (ācāra), and gives us a simile in phrases like pauram janam sutīyasi; while the next rule of Pāņini lays down that the suffix kyan may be applied to a noun in the no- minative case in the sense of 'behaving like', and this forms the basis of a simile in such expressions as tava sadā ramaņīyate śrih. It is needless to cite more examples, for this will be sufficient to indicate that some of the speculations on poetic speech can be traced back to the early grammatical analysis of the same ideas, and they point to the fact that even in the age of Pānini, some of these conceptions appear to have been well established and to have considerably influenced his enquiry. It should be admitted that these tentative sallies of the grammarians are not definite enough to indicate the exis- tence of a system, but even the limited stock of notions, thus indirectly relating to Poetics, throws an interesting light on the genesis of later speculations on poetic speech. If any deduction is permissible from the name 'Alamkāra' (lit. embellishment)1 given to the discipline as well as from the contents of the earliest existing works on the subject, it will appear that the science started a posteriori out of the very practical object of analysing poetic embellishments of speech with a view to prescribe definite rules of composition; but it cannot be doubted that it received a great impetus from the highly developed enquiry into the forms of language made by the grammarians. From internal evidence as well as from the testimony, which admits of little doubt, of some of the ancient authorities on Poetics, it is clear that the theoretical background of the discipline was, to some extent, founded on the philosophical speculations on linguistics, so that Grammar, one of the oldest and soundest sciences of India, was its god-father and helped it towards ready

1 The theory of V. Raghavan (JOR ix, pp. 264-67; and Some Concepts of the Alamkāra-sāstra, Adyar 1942, pp. 258-67) that the original name was Kriyā-kalpa lacks definite proof (see Kane's criticism in HSP, pp. 329-31). The name Sähitya is of course late (Kane, op. cit. p. 328-29).

8 SANSKRIT POETICS

acceptance. Anandavardhana speaks of his own system as being based on the authority on the grammarians, to whom he pays an elegant tribute as the first and foremost thinkers (prathame hi vidvāmso vaiyākaraņāh, vyākaraņa-mūlatvāt sarva-vidyānām, p. 47); while Bhāmaha, one of the earliest known formulators of poetic theory, not only devotes one whole chapter to the question of grammatical correctness- a procedure which is followed by Vāmana-but also proclaims openly in vi. 63 the triumph of the views of Pānini. Apart from such details as the linguistic analysis of the idea of comparison referred to above, it can be easily shown that some of the fundamental conceptions of poetic theory, relating to speech in general, are avowedly based on the views of the grammarians to the exclusion of other schools of opinion. Thus, the question regarding the convention (samketa), where- by the expressed meaning of a word (abhidhā) is to be under- stood, is settled by a reference to the views of the gram- marians on this point. The grammarians hold, in opposition to the Naiyāyikas, Saugatas and Mīmāmsakas, that the import of a word is either jāti, dravya, kriyā or guņa, as expressed by the dictum catustayī śabdānām pravrttiḥ, cited from the Mahā- bhāşya1 by Mukula (p. 4) and Mammata (Śabda-vy. p. 2). In- deed, the whole analysis of the two functions of word and its sense, called abhidhā and lakșanā2, is borrowed from the gram- matico-philosophical ideas already elaborated by the gram- marians ; and even the new aesthetic system of Anandavar- dhana, in establishing the third function of vyanjana attempts to seek an authority for its theory on the analogy of the quasi-grammatical theory of sphota, which is associated with the name of the pre-paninian grammarian Sphotāyana, and which we find fully developed in the Vākyapadīya.

1 Mahābhāsya ed. Kielhorn, p. 19, 1. 20. Cf. also Kumāra-sambhava ii. 17, where this view is clearly mentioned, although Mallinātha would explain the catustayi pravrtti with reference to the four vivartas of Vac. 2 Cf. Mahābhāșya on viii. 1. 12.

BEGINNINGS 9

( 4 ) Some of these ideas, again, are more or less recognised in the different philosophical systems, which concern themselves with śabda or speech in general, in connexion with the ques- tion of scriptural revelation and interpretation. The theory of sphota, which, however, has only a far-fetched relation to Poetics, has also its significance in some systems of philosophy; and the idea of manifestation, implied in the vyañjanā-vrtti, which consists not in the expression of something new but in the manifestation of something already existing, is not al- together unfamiliar to Indian speculation. A similar train of thought meets us in the Sāmkhya idea of causation (i.117-18), in which the effect is not produced but is already comprehend- ed in the cause and therefore can only become manifest. We find another analogy in the general idea of the Vedāntin's moksa or liberation, which consists of a condition, not to be produced but to be made manifest or realised, by the removal of enveloping obstacles.1 The theory relating to the other two functions of words (abhidhā and lakaņā), which play such an important part in poetic speculation, had already engaged the attention, not only of the philosophical gram- marians, but also of the philosophers themselves, especially the Naiyāyikas and the Mīmāmsakas. The Naiyāyikas, for instance, hold that by denotation (abhidha) of a word, we understand not only the individual (vyakti) but also the genus (jāti) and the quality (guna) 2 ; while the Mīmāmsakas main- tain that it signifies primarily jāti, and the vyakti is under- stood by implication (aksepa) through its inseparable con-

1 It is noteworthy that the Vedānta-sūtra makes a direct use of the term rūpaka in the technical sense (i. 4. 1), upon which Appayya Dīkșita remarks : Bhagavatā bādarāyaņena 'nānumānikam apy ekeşām iti cen na, sarīra-rūpaka-vinyasta-grhīter darśayati ca' iti śārīraka-sūtre rūpakam angīkrtam (Citr. mīm. p. 54, ed. Kāvyamālā 1907). Cf. also Vedānta-sūtra iii. 2. 18. 2 Nyāya-sūtra ii. 2. 68.

10 SANSKRIT POETICS

nexion with jāti.1 The Nyāya-sūtra, again, gives an exhaustive list of the relations through which a word may be used in a secondary sense, the idea of secondary sense, variously called gauņa, bhākta, lāksaņika or aupacārika artha, having been tacitly admitted in almost all the systems.2 Indeed, the theories of Poetics on these points are considerably mixed up with the doctrines of the Nyaya and Mīmāmsa schools ; and even Bhämaha's early work on Poetics devotes a considera- ble space to the discussion of the logic of poetry and the expressive functions of words. We shall note in its proper place the influence of the Nyāya, Sāmkhya and Vedānta doctrines on the poetic theory of Rasa; but it may be pointed out here that the conception of upamā (here termed upamāna), by which is implied the general idea of analogy or comparison, plays an important part in the different philosophical systems in the discussion of the nature and criterion of knowledge. The Maitri Upanișad, for instance, treats of three pramānas (v. 10, 14), viz., perception, (drsta or pratyakșa), inference (linga or anumāna) and analogy (upamā or upamāna). Kanāda and Kapila, no doubt, reject analogy as an independent and authoritative evidence, but the Naiyāyikas admit it, the purpose served by it being, in their opinion, the percep- tion of a likeness in an object not perceived before. Vātsyäyana, commenting on i. 1. 3, defines it in terms almost similar to those used in the already quoted passage from the Mahābhāşya3 as sāmīpya-mānam upamānam. According to the Naiyyikas, therefore, the object attained by analogy or upamāna consists in the recognition of the connexion between the appellation and the thing designated by it, and thus it forms the very foundation of expression. The idea of

1 Pūrva-mīmāmsā i. 3. 33ff. 2 Vedānta-sūtra ii. 3. 16, iii. 1. 7; Nyāya-sūtra ii. 2. 64, also i. 2. 11, 14, 15 ; Sāmkhya-sūtra v. 67 etc. 3 P. 5, fn 11 above.

BEGINNINGS 11

atideśa-vākya reappears as the means of analogical cognition, i.e., as a helper of the actual perception of similarity between the well-known and the novel object at the first sight of the latter. The co-operation of the upamāna is also maintained to be essential in syllogistic reasoning, where it appears in the form of the syllogistic factor, called upanaya (i. 1. 32) or statement of the minor premise. A relic of this idea of upa- māna survives in Poetics in the treatment of Bhoja (Sk iii.50), who distinguishes the figure upamā from the figure upamāna, although this view finds no supporter except Appayya Dīksita, who adopts upamāna as a separate poetic figure.1 It is needless to dilate more on this point; but the idea of upamāna, together with that of atideśa, is similarly dealt with by the Mimamsakas. They, however, consider that the upamāna refers to an object, already familiar to us, as being similar to another object which is seen for the first time ; or in the words of Upavarsa quoted by Sabarasvāmin, the upamāna or analogy, being similarity, produces an idea respecting an object that is not present, just as the sight of a gavaya revives the memory of a cow.2

( 5 ) Although these speculations have an indirect bearing on Poetics and might have suggested and helped to develop some of its fundamental theories, they cannot be taken as a definite basis of any deductions as to the antiquity of the discipline itself. It is striking indeed that we have no direct or indirect reference to Poetics as a Sästra in early texts, although at the end of the 9th century Rājaśekhara speaks of a tradition which makes it the seventh anga. Poetics is omitted in the

1 Kuvalay. ed. N. S. P. 1913 p. 174. Nāgeśabhatța takes it as included in the figure upama itself, and Vämana appears to deal with one of its varieties (viz. ananubhūta-vișaya) as an instance of what he calls tattvākhyānopamā or descriptive simile (iv. 2. 7). 2 upamānam api sādrsyam a-sannikrste'rthe buddhim utpādayati, yathā gavaya-darśanam go-smaraņam, on i. 1. 15.

12 SANSKRIT POETICS

enumeration of the different branches of study in the well- known passage in the Chāndogya Upanisad (vii. 1. 2. 4, ed. Böhtlingk). Āpastamba (ii. 4, ll), speaks of the usual six angas, but Yājñavalkya (i. 3) enumerates in all fourteen Sastras, to which the Visnu-purāna adds four more. But in them Poetics is not mentioned at all. In a similar list in the Lalita-vistara1, there is a reference to kāvya-karaņa-grantha and nätya, which may be taken to imply Poetics and Dramaturgy respectively ; but the designation 'Alamkāra' is not met with until we come to Sukra-niti which includes it, in company with Artha-Sāstra, Kāma-śāstra and Šilpa-śāstra, among thirty-two different Sastras. It has been pointed out by Rhys Davids2 that in the old Pali texts, Anguttara (i. 72, iii. 107) and Samyukta (i. 38, ii, 267), there are references to a similar study. These passages are interesting historically as being opposed to the science, but they do not expressly or by implication mention it as Alamkāra-śāstra.3 It seems likely, therefore, that Poetics as a technical discipline must have been of comparatively late origin, and probably began to develop in the first few centuries of the Christian era.4 With the flourishing of Sanskrit learning and literature in the 4th and 5th centuries under the Gupta emperors, its development probably proceeded apace.

1 Ed. Lefmann i, p. 156. 2 In a letter to the present writer dated 24. 2. 1921. Cf. Wijesekera in IHQ, xvii, pp. 196f. 3 In Kautilya's Artha-sastra there is a chapter devoted to the procedure of writing sasanas, where mention is made of artha-krama, paripūrņatā, mādhurya, audārya, and spastatva as excellences which should be attained. These may correspond to the Gunas defined in earlier Alamkära-works, but perhaps they represent the common-sense view of the matter. 4 Patañjali refers to a large number of poetical works (ed. Kielhorn i, 283, 340, 426, 444; ii 34, 102, 119, 167, 313, 315; iii 143, 338 etc.), which fact apparently indicates a considerable poetical activity in his age, preluding a systematic investigation of poetic principles ; but there is no reference to Alamkāra literature in his time.

BEGINNINGS 13

Bühler's epigraphical researches1 have confirmed the indica- tion, first given by Lassen, regarding the development of Sanskrit kāvya-style in this period, and it may be taken as the prelude to a rhetorical activity which bore rich fruit in the 8th and 9th centuries in systematic investigations of poetic principles. His examination of the early inscriptions not only proves the existence of a body of highly elaborate prose and metrical writings in the kävya-style during the first five centuries A.D., but it also establishes the presumption that most of these prasasti-writers were "acquainted with the rules of Indian poetics".2 Bühler attempts to show that the manner in which these writers conform to the rules of Alamkära, crystallised in the oldest available manuals like those of Bhamaha and Dandin, would go to indicate "the existence of an Alamkāra-śāstra or some theory of poetic art".3 From this period of the time, we get more or less definite indications of the existence of doctrines of Poetics in the Kāvya-literature itself. In the two earlier Epics, no doubt, some of the more general terms of Alamkāra (like upamā, kāvya, nāțaka, kathā and ākhyāyikā) are used, but no chrono- logical conclusion is safe from the admittedly composite character of the present-day texts. But in the Buddha-carita of Aśvaghosa, as Cowell notes, we have the use of common

1 Die Indischen Inschriften, trans. IA xiii, 1913, pp. 29f. 2 Ibid, p. 146. 3 Ibid., p. 243. This conclusion is, to some extent, supported by a passage in the Girnar inscription itself (2nd century A. D.), which con- tains a reference to "prose and verse embellished (with poetic figures) and rendered attractive by poetic conventions and expressions which are clear, light, pleasing, varied and charming" (sphuta-laghu-madhura- citra-kānta-śabda-samayodārālamkrta-gadya-padya), EI viii, p. 44. The Sphuța, Madhura, Kānta and Citra qualities may correspond to Prasāda, Mādhurya, Känti and other excellences discussed, e.g. by Dandin. The inscription itself is written in prose having long compounds, and contains alliteration, repetition of sounds and other tricks in the approved style.

14 SANSKRIT POETICS

poetic figures like upamā, utprekşā and rūpaka, as well as of elaborate ones like yathā-samkhya and aprastuta-praśamsā in an ingenious way, which presumably betrays an acquain- tance with the teachings of Poetics.1 The very first verse uses the word upama in a somewhat technical sense, and we have also in iii. 51 the use of the term rasāntara to indicate a counter emotion which cancels an already prevailing one. Aśvaghoșa uses the terms hāva and bhāva (iv. 12) in the sense they have in dramaturgic Rasa-systems. Cowell, there- fore, is justified in remarking that a peculiar interest attaches to such poems "from their importance in establishing Prof. Bühler's views as to the successful cultivation, in Northern India, of artificial poetry and rhetoric-kāvya and alamkāra -in the early centuries of our era." This remark applies with greater force to the works of Kalidäsa, which appear with all the polish and perfection imparted to them by a trained and careful artist, To the later theorists they supply an inexhaustible store-house of quotations for the illustration of different poetic figures, expressions and principles. This conscious employment of varied and elaborate poetic figures and general observance of poetic rules in these early Kāvya- poets are not without their significance, and we may reason- ably presume from them a general diffusion of the know- ledge of Poetics in this age. The same tendency towards artificial or factitious composition is shown also in the prose romances of Subandhu and Bāna, Subandhu prides himself on his skill in the use of śleșa in every letter of his composition,2 and his Vāsava- dattā justifies this boast as a tour de force of extraordinary verbal jugglery. In the work itself Subandhu speaks of

1 E. H. Johnston in his revised ed. of the work (Punjab Univ. Publication, Calcutta 1936) appears to agree with this view (Pt. ii, Introd. p. lxxxix f). 2 pratyakşara-śleşamaya-prapañca-vinyāsa-vaidagdhya-nidhim pra- bandham | sarasvatī-datta-vara-prasādaś cakre subandhuḥ sujanaika- bandhuh, ed. Srirangam, 1906, p. 357-8.

BEGINNINGS 15

poetic compositions adorned with poetic figures like śleşa, divided into ucchvāsas,1 and displaying skill in the employ- ment of vaktra metre. He specifies also two important poetic figures, viz., utprekşā and ākșepa.2 Equally definite is Bāna's references to rhetoric in his mention, in one of the introductory verses of his Kādambarī, of the poetic figures upamā, jāti (=svabhāvokti), dīpaka and śleșa as well as of poetic rasa and śayyā. Bāņa also refers to verbal puzzles, such as akşara-cyuta, bindumatī, gūdha-caturthapāda and prahelika,3 and he seems to be aware of the Ālamkārika distinction between kathā and ākhyāyikā.4 In Harșa-carīta iii, para 5, Bāņa refers to Bharata-mārga-bhajana-gītam, and in ii. 4 speaks of actors acting in the Ārabhati Vrtti (discussed by Bharata xx. 54 f),

( 6 ) From all these indications the inference is not improbable that with the growth of a body of highly finished prose and poetical literature, the science of Poetics or at least the

1 Cf. Bhāmaha i. 25-26; Daņdin i. 26-27. 2 satkavi-kāvya-racanām ivālamkāra-prasādhitām, p. 303; dīrgho- cchvāsa-racanākulam su-śleşa-vaktra-ghatanā-patu sat-kāvya-viracanam iva, pp. 238-39; utprekşākşepau kāvyālamkāreşu p. 146. The reading of the first of these passages in the Calcutta edition is bauddhasamgatim ivālamkāra-bhūșitām, and the commentary of Sivarāma (18th century) explains it as alamkāro nāma dharmakīrti-kṛto grantha-viseşaḥ. No work of Dharmakīrti's called Bauddhasamgatyalamkāra has yet been found. Undue reliance need not be placed on an unauthenticated state- ment of a very late commentator; and Lévi is probably correct in denying that Subandhu makes any allusion to Dharmakīrti's literary activity (Bulletin de l'E'cole d'Extrême Orient 1903, p. 18). 3 ed. Peterson, p. 7. Subandhu (p. 146) refers to Śrnkhalā-bandha. 4 Ibid, p. 7 and Harşa-c. p. 7. -On the meaning of the word Alamkāra see J. Gonda in Volume of Eastern and Ind. Studies presented to F. W. Thomas (Bombay 1939) pp. 97-114; but this has hardly any direct connexion with the meaning it bears in Sanskrit Alamkāra literature. A similar attempt is made by A. B. Keith in Commemorative Essays presented to K. B. Pathak (Poona BORS, 1934), pp. 311-14.

16 SANSKRIT POETICS

investigation of rhetorical rules made considerable progress by the end of the sixth century A. D. The earliest known writers on Poetics, who lived somewhat later, themselves refer to still earlier authorities. Thus, Bhamaha speaks of Medhävin and others, whom he cites and whose work he avowedly utilises. Similarly Dandin refers to earlier writings, and one of his commentators mentions in this connexion two theorists before Dandin, viz. Kāśyapa and Vararuci, who are otherwise unknown to us as writers on Alamkāra. Apart from this fact of their own citation of earlier views, it cannot also be maintained with any cogency that the relatively developed style and treatment of even these early writers on Poetics could have been evolved by themselves in the absence of earlier tentative works, the existence of which may be presumed, for instance, by the employment by these writers, of certain technical words and formulas (e.g. vakrokti, rīti, guņa etc) without a previous explanation. As a cognate branch of study, however, which probably supplied Poetics with a model and the poetic theory with the important content of Rasa, Dramaturgy (Nātya-śāstra) appears to have established itself a little earlier. Compartively early texts, both brahmanical and buddhistic, speak of some kind of dramatic representation; and we have a very early reference in Pāņini to Krśāśva and Silālin as authors of nața-sūtras (iv. 3. 110-11).1 The early existence of treatises on the dramatic art is also evidenced by the fact that all the early authors on Poetics, Bhamaha, Dandin and Vamana, omit a discussion of this subject and refer their readers for informa- tion to such specialised works. The older specimens of these are perhaps lost ; but Bharata's Nātya-śastra, which is cited as the oldest known and most authoritative, cannot possibly be put, even in its present version, at a date lower

1 It is interesting to note that both Amara and Sāśvata in their lexicons do not explain the technical terms of Alamkāra, although they have distinct references to dramaturgic technicalities and to Rasa.

BEGINNINGS 17.

than the 6th century A. D. Bharata himself, however, devotes a whole chapter to the treatment of poetical gunas and alamkāras as decorative devices of dramatic speech. It seems, therefore, that the study of Alamkara was older than Bharata ; and the tradition of opinion, followed by Bhāmaha and Dandin, may have been post-Bharata in date, but was probably pre-Bharata in substance. Indeed, the different schools of opinion, represented by the alamkāra- rīti- and dhvani-theories, probably flourished some centuries before their views became crytallised in the present works of Bhāmaha, Dandin and the Dhvanikāra respectively, none of whom, as we shall see, can be taken as the absolute founder of the system he represents. This process of crystallisation must have covered a tentative stage whose productions, if they had been extant today, would have shown Bhamaha, Dandin and the Dhvanikara in the making. We cannot, therefore, start with the works of these writers as the absolute beginnings of the science, although with them we enter upon the historic and most creative stage of its existence Taking this fact into consideration, we may presume without dogmatising that the Alamkāra-sāstra started as a separate technical discipline from about the commencement of the Christian era and probably flourished in a relatively deve- loped form in the 5th and 6th centuries A. D. The course of this development is unfortunately hidden from us, until it emerges in a more or less self-conscious form in some chap- ters of Bharata and in the Kāvyālamkāra of Bhāmaha.

CHAPTER II

BHARATA

( 1 ) Although Indian tradition glorifies Bharata, the reputed author of the Natya-sastra, with the title of muni and places him in a mythic age, the widest possible divergence of opinion exists among scholars as to his actual date ; and he has been variously assigned to periods ranging from the 2nd century B.C. to the 2nd century A.D.1 That he is the oldest writer on dramaturgy, music and kindred subjects, whose work has survived, is generally admitted ; but at the same time the question arises as to how far the extant version of his work re- presents his original text. Abhinavagupta in the second intro- ductory verse of his commentary on Bharata informs us that Bharata's text, as known to him, consisted of thirty-six chapters (şattrimśakam bharata-sūtram idam); and he is aware of two recensions (dvividhah pātho drśyate on ch. xv) of some chapters. A comparison of the different printed editions men- tioned in our Bibliography below, as well as available MSS, would go to show that they do not agree about the number and sequence of chapters, nor about the number of verses in each chapter. The text is, thus, very uncertain and unsatisfactory,2

1 Regnaud in Annales du Mus. Guimet ii p. 66, also introd. to Grosset's ed. ; Pischel in GgA, 1885, pp. 763f; P. R. Bhandarkar in IA xli, pp. 157f; H. P. Sastri in JASB, v. p. 352f, and Cat. Sansk. MSS ASB, vi 1931, p. clxxviii ; S. Lévi in IA xxxiii, p. 63; Sten Konow in Ind. Drama p. 2; P. V. Kane in IA, xlvi (1917), pp. 171 83, and HSP pp. 39f; Manomohan Ghosh in JDL xxv, 1934, pp. 59. 2 For instance, Deccan College MS no. 68 (or 69) of 1873-74 contains. 38 chapters .- For some of these discrepancies see P. V. Kane, HSP, pp. 10-14. It should be noted that neither the printed text nor any MS. contains the full text of Abhinava's commentary. Abhinavagupta in many places refers to the views of others with kecit or anye, and discuss- es various readings (pp. 50, 93, 96, 226, 241, 269, 340 etc).

BHARATA 19

It is clear from these indications that it had been subjected to considerable rehandlings in later times before it assumed its present shape, and this fact has an important bearing on the date of the supposed author. There are several passages in the present-day text which probably throw some light on this process of gradual inter- polation and recasting. The curious colophon at the end in the Kāvyamālā edition, which appears to have puzzled its editor, designates the latter portion of the work as Nandi- bharata.1 Rice mentions2 a work called Nandi-bharata on music ; while a chapter, apparently from a work on dramatic gesture, is referred to as nandibharatokta samkara- hastādhyāya in a manuscript of a treatise on music and Abhinaya, noticed in Madras Catalogue xii, no. 13009. These works, probably late compilations, are named after Nandi or Nandikeśvara, whom tradition acknowledges as an ancient authority on music, erotics and histrionic art. A Nandin is quoted by Vātsyāyana (i.1. 8). Aufrecht is inclined to identify him with Nandikeśvara cited as a writer on Erotics in the Pañca-sāyaka i. 13 and Rati-rahasya i. 5.3 Nānyadeva mentions him as Nandin. Again, a work on histrionic art, attributed to Nandikeśvara, is known as Abhinaya-darpaņa 4; as this work refers to Bharata and his views several times (e.g. st. 12, 128, 149, 159, 162) it must be a later compilation. Räjaśekhara, as we have seen above, mentions Nandikeśvara as a writer on Rasa. But Nandikeśvara is better known as an authority on music and is cited as one of his sources by Sārngadeva (13th century) in his Samgīta-ratnākara (i. 1. 17) and by his commentator Kallinatha (p.47). Besides the references given

1 samāptaś cāyam [granthah] nandibharata-samgīta-pustakam. 2 Mysore and Coorg Catalogue, p. 292. 3 Pañca-sāyaka, ed. Sadananda Sastri Ghiladia, Lahore, 1921; Rati- rahasya, ed. ibid Lahore, no date. See Schmidt, Ind. Erotik, 1911, pp. 46, 59. 4 Ed. Manomohan Ghosh, Calcutta 1934 (about 330 verses) ; trs. A. Coomaraswamy and G. K. Duggirala, Cambridge Mass. 1917.

20 SANSKRIT POETICS

above, works on music attributed to Nandikeśvara are: Nandikeśvara-mate Tālādhyāya (Weber 1729), and Bharatā- nava1 supposed to be a condensed version of Nandikeśvara's work by Sumati, dealing with dramatic gestures and Tāla. A Nātyārņava of Nandikeśvara is cited in Allarāja's Rasa- ratna-pradīpikā. Abhinavagupta (comm. on Bharata, ed. GOS, ch.xxix) says that he had not himself seen (sākşān na drstam) Nandikeśvara's work, but relying on Kīrtidhara (yat tu kīrti- dhareņa darśitam ...... tat-pratyayāt) he would briefly refer to Nandikeśvara-mata. But he knows (p. 171) a work called Nandi-mata from which he quotes a verse on the Angahara called recita or recaka Elewhere he tells us that by Nandi- mata is understood the views of Tandu; for the names, Nandi and Tandu are, in his opinion, identical. The designa- tion, therefore, of the latter part of Bharata's text. a part of which deals, among other things, with music probably implies that it was compiled or recast at some later period in accordance with the views of Nandikeśvara. Similarly, we hear of a work called Matanga-bharata2 (of uncertain date) by Lakşmaņa Bhāskara, which apparently sets forth the views of Matanga. This is another old authority cited by Abhinavagupta (as Matanga-muni) who quotes (ch. xxx) two of his Anustubh verses ; by Śarńgadeva and his conementator (on i. 3. 24-25 ; i. 4. 9 ; i. 8. 19 etc.); by Śingabhūpāla (i. 51); and by Aruņācalanātha on Raghu (p. 100) with tathā ca matange. A work called Brhad-deśi ascribed to Matanga has been published3.

1 In Catalogue of MSS, BORI, xii, pp. 460-63; Madras Cat. xxii, 13006-08. 2 The term bharata appears in course of time to mean the dramatic or histrionic art generally, as it also came to mean the actor .- Raghavabhatta on Šakuntalā expressly refers to Adi-bharata by whom he might have meant our author, in contradistinction to these later Bharatas. On this question see S. K. De, The Problem of Bharata and Adi-Bharata in Our Heritage i, pp. 193.207 (reprinted in Some Problems of Skt. Poetics, Calcutta 1959, pp. 156-76). 3 In Trivandrum Skt. Ser. 1928.

BHARATA 21

The last chapter of the Natya-sastra, to which the colophon mentioned above is appended, contains a prediction that the rest of the topic will be treated in detail by Kohala1 (who apparently belonged to the same school),2 plainly shewing that the rewriting of the portion in question was done some time after Kohala, as well as Nandikeśvara, had spoken on the subject. Nandikeśvara's date is un- known; but Kohala, side by side with Bharata, is recognised as an ancient authority as early as the end of the 8th century A.D. in Dāmodara-gupta's Kuțțanī-mata (śl. 81). It is interesting to note in this connexion that Abhinavagupta, commenting on Bharata vi. 10, says that although Nātya is usually said to consist of five angas, the enumeration of eleven angas in the text is in accordance with the view of Kohala and others,3 to whose opinions the commentator makes many other incidental references mostly on the topics of Nātya and Geya.4 Kohala is cited also by Maņikya- candra (p. 65) on Mammața, by Śārngadeva (i. 1. 15), by Sāradātanaya who frequently quotes his views (pp. 204, 210, 236, 245, 251), by the authors of the Nātya-darpana pp. 25, 38, 132), while Śingabhūpāla (i. 51) acknowledges him as an authority on drama and allied arts. Hemacandra, with reference to the classification of the drama, says (p. 329 ; also p. 325): prapañcas tu bharata-kohalādi-śāstrebhyo'- vagantavyaḥ. Kohala is credited by most writers on Drama- turgy with the introduction of Uparūpaka. Mallinãtha on Kumāra vii. 91 quotes Kohala on the subject of Tala. A work on music called Tāla-laksana, probably a late compila-

1 Śeşam prastāra-tantreņa kohalah kathayisyati, xxxvii. 18. 2 See xxxvii. 24. Abhinaya-trayam gītātodye ceti pancangam natyam ........ anena tu ślokena kohalādi matenaikādaśāngatvam ucyate (on vi. 10). 4 For references see P. V. Kane, HSP, p. 24, 54-55. See also his Fragments of Kohala in Proc. A-I.O.C. (Patna), 1930, pp. 577-80 .- Abhinavagupta, however, holds that the entire text of the Nātya-sāstra as known to him was composed by Bharata himself.

22 SANSKRIT POETICS

tion, is attributed to him,1 and a Kohalīya Abhinaya-śāstra,2 purporting to embody his views, is also known. A work, called Kohala-rahasya3 in at least thirteen chapters, dealing with musical modes, is ascribed to the sage Kohala (described as son of Bharata) who reveals the subject on being requested by Matanga. The description that he was son of Bharata is perhaps due to the indication4 in Nātya-śāstra i. 26 (ed. Chowkhamba 1929, ed. GOS, Baroda 1956; but not included in the NSP ed. 1894) that Kohala, along with Sāndilya, Dhurtila and others, are spoken of as sons of Bharata. Since Abhinavagupta says (Abh. bh. p. 25) that Kohala describes the verse jitam udupatinā from Ratnāvalī (i. 5) as an example of Nandi in accordance with the rules of Bharata, P. V. Kane is of opinion that Kohala was later than Ratnāvalī, i.e. than 650 A.D. But this conclusion cannot be very well reconciled with the fact that Bharata and Kohala are already recognised as ancient authorities in the 8th century A.D. From Abhinavagupta's references and citations Kohala's lost work appears to have been mostly in verse. A work on music, entitled Dattila-kohalīya, is mentioned by Burnell (p. 606), apparently a compendium of the opinions of Kohala and Dattila. The latter, whose name occurs variously as Dantila and Dhūrtila, is mentioned by Dāmodara- gupta (śl. 123) and is cited as an old authority by Abhinava- gupta (as Dattilācārya, chiefly on music) who quotes (on Bharata xxviii, also p. 205) a verse of his in Anustubh ; by Śārngadeva (i. 1. 16) and his commentator Kallinātha (p. 49) ;-

1 IOC 3025, 3089 ; Madras Cat. with a Telugu comm. 12992. 2 Madras Cat. 12989. with a Telugu comm. 3 Madras Trm. I, C, 787 (the 13th chapter only). 4 The editions (GOS) of M. R. Kavi, and Chowkhamba Skt. Ser. include this verse (i. 26), but they note that it does not occur in the MSS consulted. It is, however, commented upon by Abhinava- gupta (p. 18)-Kohala's views are referred to by him p. 25, 103, 173, 182, 266 etc.

BHARATA 23

by Singabhūpāla (i. 51), as well as in various works on music. A work called Dattila on music is also published.1 In the same way Sāņdilya is mentioned as Nātya-śāstra- kāra by Singabhūpāla (i. 51). Kāśyapa or Kaśyapa as a muni who preceded Bharata and his opinion on Rāgas are cited by Abhinavagupta (on xxix, p. 394), as well as by Nānyadeva2. Sātakarņi is cited as a writer on dramaturgy by Sāgaranandin (on Sūtradhāra, 1. 1101). Višākhila is also cited as an authority on Kalā-śāstra by Vāmana (i. 3. 7), by Abhinavagupta (on xxviii, xxix. 31-33) and by Nānyadeva as an authority on music. Parāśara or Pārāśara, mentioned by Rājaśekhara, is also cited as a Bharata-putra in Nātya-śāstra (i. 32) ; his views on Nāndī and Totaka are quoted by Sāgara- nandin (ll. 1091, 2770 3202-3). Similarly, another Bharata-putra is Nakhakutta, who is also cited by Sagaranandin (11. 2668, 2994). Even the mythical Narada is regarded as pro- . pounder of Gāndarva-veda; and the Bhāva-prakāśana says that Narada taught Bharata the subject of evolution of Rasa, having learnt it from Brahma himself. From these indications it is likely that between Bharata's original text and its existing version, there came "Kohala and others" whose views found their way into the compendium, which goes by the name of Bharata and which indiscriminat- ing posterity took as genuine and unquestionable. The text- problem of the Nātya-śāstra, therefore, cannot be solved until the works of Kohala and other early writers, some of which appear to have been available to Abhinavagupta, are recovered. The process of incorporation must have occurred very early and was apparently complete by the end of the 8th century, when the work assumed more or less its present shape. Udbhata, about this time, actually appropriated

1 In Trivandrum Skt. Series, 1930. 2 On Käśyapa see below under Dandin. Abhinavagupta's re- ference to Kāśyapācārya at p. 239 shows that his work was, at deast partly, in verse.

24 SANSKRIT POETICS

(iv. 4) the first-half of the verse vi. 15 of the Nātya-śāstra, and makes only enough verbal change in its second half to admit Sänta as the ninth Rasa in the category of eight recognised by Bharata.1 Abhinavagupta, who commented on the existing text at the end of the 10th century, himself mentions several other previous commentators, of whom Lollata and Sankuka in all probability belonged to the 8th and 9th centuries. These indications will make it clear at any rate that the text existed in its present form in the 8th century A.D., if not earlier.

( 2 ) We have, on the other hand, the tradition as well as the statement of Bhavabhūti, who refers to the mythical Bharata as the tauryatrika-sūtra-kāra2, that Bharata's work existed originally in the sutra-form ; and this is also made likely by Pāņini's early reference to such nața-sūtras in his own time. Reminiscences of the sutra-style may indeed be presumed in the Nātya-śāstra vi and vii, which deal with rasa and bhāva;

1 It should be noted that Abhinavagupta immediately after the passage cited (see p. 21, fn 3 above) goes on to say: anena tu ślokena kohalādi-matenaikādaśāngatvam ucyatc, na tu bharate, tatsamgrhītasyāpi punar atroddeśāt, nirdeśe caitat krama-vyatyāsanāa ity udbhatah neti bhatta lollatah ...... vayam tvatra tattvam agre vitanişyāma ity āstām tvat (on Bh. vi. 10). This difference of opinion between Udbhata and Lollata on a question of textual interpretation supports the conclusion that Udbhata was probably familiar with the text of Bharata as Abhi- nava knew it and as it has come down to us. The tradition is recorded by Sārngadeva (i. 1. 19) that Udbhata was one of the early commen- tators on Bharata. Abhinava himself refers to Lollata, Śankuka and Bhatta Nāyaka as principal commentators whose views on Rasa he thought worth refuting in detail. Sarngadeva omits from his list Bhatta Nayaka but adds Udbhata. as well as Kirtidhara who is earlier than Abhinava having been cited by him (see above p. 20). For other reputed commentators on Bharata see below. 2 Uttara-carita, ad iv. 22 (ed N. S. P. 1906 .. p. 120). Abhinava- gupta in his commentary refers to Bharata's work as a Sūtra-work.

BHARATA 25

for in ch. vi we find a dictum1 on the genesis of Rasa, put in concise form of a sūtra, to the elucidation of which, after the manner of a bhäsya or vrtti, the rest of the chapter (written in prose with verse-quotations) is apparently devoted. It should also be noticed that a preliminary explanation is added at commencement of the chapter to reconcile this curious por- tion of the text with the rest of the work. Bharata, we are told, being requested by the sages, explains the characteristics of a samgraha, kārikā. and nirukta, and incidentally gives an illustration of a sūtra-grantha by putting a part of the text in that form. This discussion of the nature of a samgraha, kārikā, nirukta and sūtra would not have been relevant to the subject in hand but for this somewhat flimsy explanation, which, however, affords a device, far-fetched as it is, to intro- duce into the kārikā-text some vestiges of the older sūtra-form. It is not maintained that a sutra-text is necessarily older than a text in the kārikā-form ; for in our sūtra-text itself there are quotations in the vrtti of anubaddha or anuvamśya ślokas2, betokening the existence of earlier teachings on the subject, and disproving at the same time the orthodox belief that Bharata was the earliest teacher of the Nātya-veda. But if the tradition that Bharata's original work was composed in the sutra-form be accepted, then this portion of the existing version may be presumed to have been a survival of the original form. Similar fragments of the sūtra-bhāsya style

1 tatra vibhāvānubhāva-vyabhicūri-samyogād rasa-nispattiḥ, ed. Grosset p. 87, 1. 8; ed. Kāvyamālā p. 62, 1. 6. This dictum is cited as a Sūtra by all later writers, including Abhinavagupta and presumably his predecessors Lollata and others. For other in- stances of the Sūtra-bhäsya style in the work see P. V. Kane, HSP, pp. 15-16 .- The references to the text of the Natya-sastra here are generally to the Kāvyamālā ed. 2 The anubaddha verses, apparently of earlier writers, are those related to the subject in hand. By anuvamsya verses (verses of this kind occur also in the Mahabhūrata) are meant, according to Abhinava- gupta, those handed down traditionally from teacher to pupil (anu- vamsa-bhavam śisyācūrya-param parūsu vartamānam) vi, p. 25-26.

26 SANSKRIT POETICS

are seen in ch. xxviii-xxxi in such passages as beginning with ātodya-vidhim idānīm vakşyāmaḥ (xxviii. 1); also in xxxiii. 212 vādya-vidhānam vakşyāmi; in xxiv. 93 atra sūtra- dhāra-guņān vakşyāmaḥ etc. If we get the lower limit to the date of Bharata's work at 8th century A.D., the other limit is very difficult to settle, when we consider that there were apparently two versions, either independent or one based on the other. But it is not clear what weight should be placed on the testimony of Bha- vabhūti ; for if in the first quarter of the 8th century Bharata was known to him as a sūtra-kāra, it is not intelligible how at the end of the same century Udbhata makes use of Bharata's kārika, and Lollața and others, immediately following, ap- parently comment on the same text. The short space of less than half a century is not enough to obliterate all signs of the older version and replace it entirely with a new kārikā-text which, to all intents and purposes, is taken as the only author- itative version in later times, and in which, strangely enough, we find still lingering traces of the earlier sutra-text. The only possible explanation of Bhavabhuti's reference is that the historical Bharata, who was the sutra-kara on the three arts of dancing, singing and instrumental music, had already in Bhavabhuti's time become identified with the mythical Bharata; for the passage in the Uttara-carita gives an obvious- ly mythical account, through the mouth of Lava, that bhaga- vān Vālmīki, having composed his story of Rāma, gave it to bhagavan Bharata (the sūtra-kara on the three arts and ap- parently the nātyācarya of the gods) who revised it and got it acted through celestial nymphs. But this does not exhaust all our textual difficulties. Independent prose-passages also survive, in the midst of kārikās, in chs. xvii, xxviii, xxix, xxxi and xxxiv which, form- ing an integral part of the text, cannot be taken as mere vrtti, 'but which resemble, in some respects, the prose Smrti-frag- ments or more closely, the prose-fragments in the apocry- phal Bhela-samhitā. Again, the anubaddha or anuvamśya

BHARATA 27

ślokas, referred to above, correspond to the parikara- or samgraha-ślokas in later writings, and certainly indicate the probability of earlier speculations on the subject. These verses are generally taken from two distinct sources ; for some of them are in arya, while others are in anustubh metre. On the Aryā-verses Abhinava remarks (on vi. 85, p. 328) ; tā etā hy āryā eka-praghațțakatayā pūrvācāryar laksanatvena pathi- tāh, muninā tu sukha-samgrahāya yathā-sthānam viniveśitāh. In his opinion former teachers composed these Āryas and Bharata inserted them in proper places. From the facts adduced above, we are confronted with the problem of the inter-relation of these apparent survivals in our text, which contains vestiges of (1) independent prose- fragments (2) anuvamśya ślokas in āryā and anustubh metres and (3) passages in sūtra-bhāșya style, as well as (4) the present systematic kārikās. Space is too limited to dilate upon the ques- tion here, but an examination of these passages will reveal that these different styles do not possibly belong to the same period, but they probably indicate several stages in the growth of par- ticular forms of composition of dramaturgic works in general, each stage betraying its own partiality towards a particular form. Taking the present kārikā-text as the starting point, we find in it traces of earlier passages in sūtra-bhāsya style, of which it is presumably a recast. In the sūtra-bhāya, again, there are fragments of metrical passages which indicate, in their turn, another and still earlier kārika-stage; while the in- dependent prose-fragments perhaps represent the earliest form taken by such technical treatises. We can, therefore, distin- guish in their order of development (1) a stage of prose- treatises (2) a tentative period of kārikā-writing (3) the sūtra- bhāsya stage and (4) the final period of compilation of compendiums, which reverts again to the karika-form.1 This

1 This conclusion does not apply to the more or less imitative periods after the 10th century in which we find the kārika- and the sūtra-style, existing almost side by side.

28 SANSKRIT POETICS

conclusion perhaps finds some support in the repetition,. more or less, of a similar phenomenon in the sphere of the Dharma-śāstra, Artha-śāstra, Vaidya-śāstra and probably Kāma-śāstra, The loss of earlier treatises makes it difficult to dogmatise; but if this conclusion is correct generally, then our text may be supposed to contain remnants of all these styles and forms. It is not argued here that Bharata's work itself passed through all these stages or forms, from a rudimentary prose-version into a systematic metrical manual1 ;. but our text contains enough to betray the existence of pre- vious speculations in prose and in verse, as well as indicate the fact that it might have itself been once written in the sūtra-bhāsya form, which was recast, with considerable additions from other sources, into a convenient metrical. compendium.

( 3 ) Taking the substance of the work, apart from the vexing. question of different versions, the portion of the Nātya-śāstra,. which deals principally with music, has been conjectured2 on internal evidence to have been compiled about the 4th century A.D .; and it appears likely that the other portions were. also put into their present shape about the same time. Pischel's argument, however, on the date of the work, derived from the reference to Sakas, Yavanas, Pahlavas and Bāhlikas (e.g. xxxii. 103, Ch. ed.) in a text of such composite character is of doubtful value in determining the question finally ; but

1 Kane believes it possible (HSP, p. 16) that the original nucleus. of the Natya-sastra was in mixed prose and verse. He is also of opinion that the first five chapters were comparatively late additions. Abhinavagupta knows Bharata's text as consisting of 36 chapters, al- though he comments on the 37th ch. of what M. R. Kavi speaks of as the Northern recension (according to Kavi the Southern is the earlier text and consists of 36 chapters). 2 IA xii p. 158 f.

BHARATA 29

-it perhaps makes it probable that the upper limit of its date -cannot be put too early. We are in a position, however, to infer that the substance of Bharata's work is probably much older than that of Bhämaha,1 who may be assigned to the last quarter of the 7th century. Bhamaha, in his treatment of poetic figures (kāvyālamkāras), groups them in a curious but suggestive way, which probably indicates the different periods in the growth and multiplication of such figures.2 At the outset, he names and defines only five poetic figures (ii. 4) recognised, as he says, by other writers, viz, anuprāsa, yamaka, rūpaka, dīpaka and upamā, This represents the first stage ; but in course of time, six other figures appear to have been added, and Bhamaha mentions and deals with them next in ii. 66. Then he goes on to enumerate, two (or three, including svabhāvokti) more figures admitted by writers like Medhavin (ii. 88), who also appears to have dealt with upamā etc. (ii. 40). Finally, Bhāmaha defines and illustrates a further long list of twenty-three more figures in a separate chapter (iii, 1-4). The differentiation and multiplication of poetic figures with the progress of speculation is a familiar fact in Alamkāra literature; and the way in which Bhmaha success- ively enumerates and groups these figures probably shows that to the original five mentioned by him at the outset, others were added in course of time as the study itself advanced. Now Bharata, in his treatment of Alamkāras names (xvi. 41) only four such figures known to him, viz., yamaka, rūpaka, dīpaka and upamā, These four in reality correspond to the five mentioned by Bhamaha ; for anuprāsa may be taken as falling in the same class as yamaka, the one being varnābhyasa and the other padabhyasa. At the same time

1 Besides showing himself conversant with some theory of Rasa (ii. 281, 283 f), Dandin mentions the dramaturgic technical terms samdhi, anga, vrtti and laksaņa and refers to āgamāntara for their discussion (ii. 366). 2 Cf Jacobi in Sb. der Preuss, Akad. xxIV, 1922, pp, 220 f.

30 SANSKRIT POETICS

the very fact that anuprasa is thus differentiated from yamaka may indicate further refinement in these poetic figures. It is clear, therefore, that Bharata's work belonged to a period when the number of figures had not yet multiplied ; and one, . if not more, stages must have intervened between it and Bhāmaha's Kāvyālamkāra in which the number had already swelled into nearly forty in all.1 To this intermediate stage belonged Medhavin and others, whom Bhamaha cites, and the loss of whose work makes it difficult for us to trace the development thus indicated by Bhāmaha. There are also indications that Bharata's teachings are probably older than Kālidāsa, who generally adheres to Bharata's dramaturgic prescriptions.2 Kālidāsa refers, in Vikramorvaśīya ii. 18, to Bharata as the mythical nāțyācārya. In Raghu xix. 36, again, Kālidāsa speaks of anga-sattva- vacanāśraya nrtya which, as Mallinātha rightly points out, agrees with Bharata's dictum3: sāmānyābhinayo nāma jñeyo

1 By the end of the 6th and beginning of the 7th century, Bhatti illustrates as many as 38 different species of poetic figures, indicating that the process of refinement had proceeded very far indeed even in his time. 2 The discovery of the dramas, ascribed to Bhasa, does not invalidate this argument, for it is possible that they follow a tradi- tion or a system of opinion of which all traces are now lost; and. the date of Bhäsa itself is uncertain. 3 The Amara-kośa only mentions āngika and sāttvika abhinaya. This lexicon, which is said to have been translated into Chinese about 561-66 A. D. (see Nandargikar's introd. to Meghadūta, 1894, p. 73), admits after Bharata the eight Rasas, and explains some of the dramaturgic technicalities, giving three synonyms of the actor (śāilālin, krsāvin and bharata) from the names of the three well- known teachers of dramaturgy. Panini refers to the formation of the first two terms, but not that of the third; but this silence of the grammarian does not prove anything. The Jaina Anuogadāra- sutta (ed. N. S. P. 1915, fol. 134-145, also quoted in Weber ii 2, pp. 701-02) which, Winternitz thinks, was probably put together by the middle of the 5th century, mentions nine Rasas; and the enu-

COMMENTATORS ON BHARATA 3₺

vāg-anga-sattvajaḥ; while in Kumāra vii. 91 mention is made of samdhis, as well as of lalitangahara mentioned in Nātya-śāstra xx. 17 (ed. Chowkhamba xxii. 17). The lower limit of the date of Bharata's work, therefore, can be provisionally shifted back to the fourth or fifth century A.D., while it is almost certain that it existed in its present shape in the 8th century A.D.1 The upper limit cannot be put too early, because of the mention of Sakas, Yavanas, Pahlavas and other tribes, and probably does not go beyond the commencement of the Christian era ; but we have already pointed out (p. 28 above) that their mention in a com- posite text is hardly of a conclusive value. It is difficult to settle the relative age of the sūtra- and the kārikā-texts ; but if the tendency towards sūtra-bhāsya style may be presumed to have been generally prevalent in the last few centuries B.C., then the presumed sūtra-text of Bharata belongs ap- parently to this period2. It was certainly much earlier than the present kārika-text, in which Bharata is already a mythical sage as an expounder of the nātya-veda.

COMMENTATORS ON. BHARATA No commentatary on Bharata's Nātya-śastra exists today except that of Abhinavagupta. But the names of some reputed

meration is interesting from the inclusion of praśānta (not men- tioned by Bharata), apparently from religious motives. 1 With this view Kane (HSP, p. 19, 22) generally agrees. 2 It will be shewn later that the tradition that Bharata was the author of a Kāvya-laksana, which forms in substance of the kārika-verses of Mammata's Kav. prak., is entirely erroneous, as is also the statement made use of by Lévi that these kārikās are abridged from the Agni-purāna. Somadeva in his Yaśastilaka (959-60 A. D.) refers, indecd, to a bharata-pranīta kāvyādhyāya (Peterson ii. p. 45) which, considering Somadeva's date, could not have alluded to this tradition of Bharata's authorship of Mammata's kārikās, but possibly, from the term of reference, to ch. xvi of the Natya-sastra, which deals with kāvya-laksanas, kāvyālamkāras kāvya-guņas anď kāvya-doșas as embellishments of dramatic speech.

32 SANSKRIT POETICS

as well as actual commentators on Bharata are known from Abhinavagupta,1 Sarngadeva2 and other writers. They are: 1. Mātrguptācārya 2. Udbhața 3. Lollata 4. Śańkuka 5. Bhatta Nāyaka 6. Harsa 7. Kirtidhara 8. Abhinavagupta 9. Nānyadeva. We have also the views of several other writers cited by Abhinavagupta.3 They are: Bhatta Yantra (on Nātya and Ntta p. 208), Priyātithi (on Lāsyānga), Bhatta Vrddhi (on Tāla), Bhatța Sumanas (on Tāla), Bhatța Gopāla (on Tāla), Bhatta Śamkara (Vrtta-prakaraņa) and Ghaņțaka (on Nāțikā- bheda). Rāhula or Rāhala (pp. 115, 172, 197 etc), also men- tioned by Sārngadeva (i.1.17), is quoted several times. As in one of his verses Bharata is named (bharatenoditam, Abh.bh, i, p. 72) he must have been, as most of the writers mentioned above, later than Bharata. From his name, as well as from Hemacandra (p. 316), who mentions him as Sākyācārya, he appears to have been a Buddhist teacher, whose view Hema- candra pointedly ignores. It cannot be determined if all these writers composed commentaries on the whole or a part of Bharata's work: but from the references it seems

1 See P. V. Kane, Gleanings from Abhinava-bharatī in K. B. Pathak Comm. Vol. Poona 1934, pp. 385-400; V. Raghavan, Writers quoted in Abhinava-bhārati in JOR, vi, 1932, pp. 149 f, 199 f. 2 Sārngadeva mentions: vyākhyātāro bhāratīye Iollatodbhata- sankukūh/ bhattabhinavaguptas ca śrīmat-kīrtidharo'parah. He men- tions Rahula and Matrgupta elsewhere as two of his authorities, apparently on music. 3 The references to Abhinava-bharati by vols. and pages are to M. Ramkrishna Kavi's ed., Baroda 1926. Where the vol. is not indicated by a number but only by page, vol. i is meant.

COMMENTATORS ON BHARATA 33

that most of them wrote generally on music, and some specially on topics of dramaturgy as well. Bhatta Tauta, Abhinava's Guru, and Utpaladeva, his Parama-guru, are frequently quoted in his commentary. Tauta is known as having written a work called Kāvya- kautuka, which is now lost ; on this work Abhinava appears to have written a Vivarana (Locana p. 178). At the beginning of his commentary Abhinavagupta pays an elegant tribute to'Tauta as having expounded to him the Nātya-śastra, and references1 to Tauta's view on various topics confirm this. But it need not mean that Tauta actually composed a com- mentary on Bharata. Utpaladeva, author of Iśvara-pratya- bhijña and other works, is well known in the history of Kashmir Saivaism. He is quoted chiefly in the chapters on music. It is not clear if he wrote a commentary on these chapters or an independent work on music. Similar remarks apply to Sakalīgarbha who is cited once along with Udbhata (see below). Mention is also made of Tandu2 in Nātya-śastra (iv. 17-18) as one who instructed Bharata in the representation of Angahāras together with various karanas (postures) and recakas (gestures). An unknown Țīkākāra or Țītākrt is frequently cited by Abhinava throughout the text.

Mātrguptācāryu Rāghavabhatta on Sakuntalā3 and Vāsudeva on Karpūra- mañjarī4 quote Mātrguptācarya as a writer on Dramaturgy;

1 See Kane in the work cited p. 388; also in his HSP, pp. 209-12; V. Raghavan as cited above pp. 153-62. Three metrical lines from Tauta are cited by Abhinava p. 291-92. 2 With regard to Tandu cited on p. 90 (vol. i) we are told that Nandi is another name for Tandu, as the word Muni designates Bharata (tandu-muni-śabdau nandi-bharatayor āpara-nūmūni). Hence Nandi-mata (cited on p. 171). in Abhinava's opinion, stands for the views of Tandu. See above p. 20. 3 Ed. NSP, 1922, pp. 5, 6, 7 (on Rasa), 8 (definition of Nātaka), 13 (Vithyanga), 15 (languages to be employed by various characters), 20 (Bhūsaņa), 57 (Sañcārikā), 62 (Senāpati), 74 (Hasita), 110 (Pātakā- sthanaka), 123 (same), 126 (Sanskrit employed by Nica Patra), 154 (Kañcukin), 156 (Pratīharī). 199 (Paricārikā), p. 230 (Phala-yoga), etc. 4 Ed. NSP, 1900, p.' 5 (Sūtradhāra); cf Aufrecht i, 448a. 3

34 SANSKRIT POETICS

and Sundaramiśra in his Nāțya-pradīpa (dated 1613 A.D.), commenting on Bharata's remarks on Nāndī says : asya vyākhyāne matrguptācāryaiḥ ... iyam udāhrtā.1 This has been taken by S. Lévi to imply that Mātrgupta wrote a vyākhyāna or commentary on Bharata, and that assuming him to be the poet who lived under Harsa-Vikramāditya (Rājatarang. iii. 125, 252) we get in him a very early commentator (7th cen- tury) on Bharata. But the available evidence does ont appear to be conclusive. Our Mätrgupta, as the profuse metrical quotations on dramaturgic topics by Rāghavabhatta and others show, probably wrote an independent metrical work on Dramaturgy, in which he might have in the usual course commented on Bharata's precepts; and the word vyākhyāna need not be construed to mean a commentary. Mātrgupta of Kahlana was a king and poet; how is it that he is cited in these works with the designation of Acarya which signifies a teacher? Mātrguptācārya is known to Abhinavagupta who quotes him on music (ch. xxix). Sāradātanaya in his Bhāva- prakāśana quotes his view on Națaka-vastu; Sāgaranandin in his Nāțaka-lakşaņa-ratna-kośa quotes several verses of his (pp. 5, 14, 20, 21, 23, 50) ; and Sārngadeva mentions him as an authority on music.2

Udbhata As already noted above (p. 32, fn 2), Sārngdeva in his Samgīta-ratnākara (i. 1. 19) informs us that Udbhata was one of the early commentators on Bharata's text. This is very likely, although Udbhata's commentary has not yet been recovered.

1 Quoted in IOC iii, p. 347. Mätrgupta is also cited by Ranga- nātha on Vikramorvašīya (dated 1659 A. D.). ed. NSP, 1914, p. 5 (on Nāndī); by Sarvānanda on Amara, p. 145 (Adbhuta Rasa), 147 (Bībhatsa Rasa), 150 (Vyabhicāri-bhāvas), 161 (Anubhāva in Srn- gāra), 16 (on Tāla). 2 The citations from Mätrgupta are collected together by T. R. Chintamani in his Fragments of Matrgupta in JOR ii, 1928, pp. 118-28.

COMMENTATORS ON BHARATA 35

Sārngadeva's statement appears to be confirmed by several citations of Udbhata's views by Abhinavagupta. One of these references, already quoted above (p. 24, fn. 1) occurs in Abhinava's commentary on vi. 10 (p. 266-67) in which it is stated that certain views of Udbhata on textual interpretation were not accepted by Lollata, another commentator on Bharata. On ix. 182 (vol. ii, p. 70) and xviii. 76 (vol. ii, · p. 441), again, Abhinava quotes certain readings of the text made out differently by Udbhata. In two other passages (on xxi. 17 and xxi. 42 on Samdhis) Abhinava takes exception to the interpretation of Udbhata as laksya- or agama-viruddha. In still another passage on the Vrttis (on xviii. 110 vol. ii, pp, 451-52) Abhinava informs us that Udbhata accepted only three Vrttis (and not the usual four of Bharata), namely, Nyāya-ceștā, Anyāya-ceșțā and Phala-samvitti. In this connexion Abhinava further refers to a certain writer, called Śakalīgarbha, who would admit five Vrttis (namely, the four of Bharata and another called Atma-samvitti in place of Udbhata's Phala-samvitti) ; but these views have been refuted by Lollata and others. Again, Kuntaka (pp. 113-15) disagrees with Udbhata's view of the Śva-śabda-vācyatā of Rasa as being opposed to Bharata's opinion. These detailed references to Udbhata's views or comments on topics dealt with in such far apart chapters as vi, ix, xviii and xxi, make it probable that Udbhata commented on the whole of the Nātya-śāstra. But this presumption is not applicable to the case of Sakalīgarbha who appears to have flourished between Udbhata and Lollata. He might have been a writer on certain topics of Dramaturgy; but whether he composed a commentary on Bharata is not clear.

Lollața Lollata is extensively cited by Abhinavagupta not only in ch. vi (on Rasa-sūtra), but also in ch. xii, xiii, xviii and xxi. Lollata is mentioned as rejecting Udbhata's views on the Vrttis

36 SANSKRIT POETICS

(see above) and on the eleven topics of Nātya (on vi. 10).1 Several other characteristic views of Lollata are also mention- ed; for instance, his view that Rasas are numerous (on vi, 45),2 although eight or nine are traditionally accepted for the stage; that Nātikā is Șatpadā (on xviii. 60)3 against Śankuka's opinion that it is Astapada. References are also made to Lollața's views on Dhruvā Tāla (on xii. 14)4; on Kaksyā (xiii. 1)5 ; on what he calls (on xxi. 29) Anusamdhi (dealing with the acts of a Patāka-hero) ; on his omission of xviii. 32 from the text6. These references to different parts of the text would go to confirm the tradition that Lollata was also a regular commentator on Bharata's text. We have no definite material to determine the date of Lollața ; but all later citations agree in supporting the tradi- tion that he was earlier than the commentator Śankuka whose theory of Rasa is said to have been directly levelled against that of Lollata. Judging from his name, Lollata was pro- bably a Kashmirian ; and if any chronological inference can be drawn from the Kashmirian Abhinavagupta's reforence (mentioned above) to an opinion of Udbhata being controvert- ed by Lollata, he was later than or contemporaneous with Udbhata, also a Kashmirian, whose latest date is 813 A.D. The theory of Rasa advocated by Lollata probably obtain- ed traditionally before he definitely formulated it and became its first noted advocate; for Abhinava in his commentary (on ch. vi) tells us that Dandin in his idea of Rasa follows the same view. Unless we presume Lollata's priority to Dandin, we should take this to indicate that some theory or dogma similar to it was already known to Dandin, even before Lollata brought it into prominence. Lollata is also taken to be what is technically described as Dīrgha-vyāpāra-vādin with reference to the controversy about the function of Abhidha ; for he is said to have main- tained that the primary function of Denotation of a word is

1 Vol. i, p. 266. 2 Vol. i, p. 299. 3 Vol. ii, p. 436. 4 Vol. ii, p. 134. 5 Vol. ii, p. 196. 5 Vol. ii, p. 423.

COMMENTATORS ON BHARATA 37

so far-reaching that it is competent in itself to express all other implied or suggested sense. But it is doubtful if this view is directly ascribed to Lollata by any standard Sanskrit theorist of importance, although it is criticised without the name of the promulgator by Mammata (p. 225), Mahimabhatta (p. 27), Hemacandra (p. 215), Vidyānātha (p. 43) and others. Govinda in his Kāvya-pradīpa (p. 149) thinks that adherents of this view are followers of Bhatta-mata. Abhinava attributes a similar view (Locana p. 188) to the Bhatta or Prābhākara school ; this might have been the source of Govinda's state- ment. The Mimamsakas and grammarians had already discussed the question of Abhidha ; and it must not be for- gotten that several attempts to explain the fact of Dhvani (suggested sense), including Rasa-dhvani, obtained before the Dhvanikāra himself. It is probable that Lollata was one of those who offered one of the several solutions to the question alluded to in the first verse of the Dhvanyāloka. We shall see that Lollata was probably a Mīmāmsaka in his view of Rasa, even if he was not a Dīrgha-vyāpāra-vādin. The only direct quotation from Lollata (and not mere consideration of his views) consists of two verses given by Hemacandra (p. 215). If he was a prose commentator, how do these metrical quotations fit in?1

Śaħkuka Abhinavagupta very frequently refers to Śankuka's views on different dramaturgic topics; eg. on Ranga-pitha (on iii, 21-22)2; on Rasa-sūtra (on ch. vi)3; on Nātaka (on xviii.10)4; on the king as a character (on xviii. 12)5; on Nāțikā-bheda (on xviii. 60)6; on Pratimukha and Vimarśa

1 V. Raghavan believes (Some Concepts pp. 207-8 : JOR vi. p. 169) that Lollata's other name was Āparājiti, son of Aparājita, be- cause a quotation from Āparājiti by Rājaśekhara (p. 45) is given by Hemachandra (p. 215) with the name of Lollata. 2 Vol. i, p. 75. 3 Vol. i, pp. 239, 298, 318. 4 Vol. ii, p. 411. 5 V. ii, p. 414. 6 Vol. ii, p. 436.

38 SANSKRIT POETICS

Samdhis (on xxi.40,42) etc.1 As the citations relate to matters covered by ch. iii to xxix. it is probable that Sankuka composed a commentary on the entire text of Bharata. Abhinavagupta informs us (p. 275) that his teacher Bhatta Tauta disapproved of Śankuka's views on Rasa. To Śankuka are also ascribed several verses in the autholo- gies of Sārngadhara, Jahlana and Vallabhadeva,2-which indi- cates that there was also a poet of the same name. Kahlana mentions (iv. 703-5) a poet Sankuka and his poem Bhuvanā- bhyudaya. The reference is to the time of Ajitāpīda, whose date is given as 813 A.D. by Cunningham and 816 A.D. by S. P. Pandit. If our Śankuka is identical with this poet, then he may be assigned to the first quarter of the 9th century.

Bhatța Nāyaka Basides referring to his view on the Rasa-sūtra of Bharata (ch. vi)3 Abhinavagupta (Locana p. 27: also Abh. bh. xvi. 4)4 quotes under the name of Bhatta Nāyaka a verse śabda- prādhānyam āśritya5, which Hemacandra (p. 3-4) gives as a quotation from a work, named Hrdaya-darpana6, and which is also cited without the name of the author by Mahimabhatta and his commentator. Jayaratha also speaks (p. 12) of Bhatta Nāyaka as Hrdayadarpaņa-kāra. It is probable, therefore, that some lost work of Bhatta Nāyaka's bore this title ; and the references also indicate the probable source

1 See P. V. Kane, HSP, pp. 50-51. for six other instances from ch. xxiv to xxix where Sankuka has been cited by Abhinavagupta. 2 In the first two anthologies Śankuka is called son of Mayūra, who is identified by some with the author of Sūrya-śataka, a con- temporary of Bāņa. 3 Vol. i, p. 278. 4 Vol. ii, p. 298. 5 This verse is also quoted by Jayaratha p. 9. Māņikyacandra (p. 4) also ascribes the verse to Bhatta Näyaka, whom he refers on p. 8 as the Hrdayadarpaa-kāra. 6 P. V. Kane (HSP p. 187) suggests that the proper name of work was Sahrdaya-darpana, but evidence is meagre to support this suggestion.

COMMENTATORS ON BHARATA 39

of the quotation which occurs immediately before the verse in question in Abhinavagupta. It is, however, not clear whether this Hrdaya-darpana is his alleged lost commentary on Bharata. Mahimabhatta's anonymous commentator tells us that this Hrdaya darpana, like the Vyakti-viveka, was composed with the special object of demolishing the Dhvani- theory1 formulated by Anandavardhana ; and this statement may explain why Abhinavagupta, as an adherent of the theory, takes so much pains to controvert Bhatta Nāyaka's views in his Locana 2. as well as in his Abhinava-bhāratī.3 Mahimabhatta, who had a similar object of combating the Dhvani-theory, claims entire originality for his own treatment by boasting that he had never looked into the Darpana at all.4 The citations from the Hrdaya-darpana also indicate that it was probably composed in metrical form, and apparently never took at all the shape of a prose commentary. A careful examination of the very few passages apparently referring to this work will shew that the topics dealt with in it centre round the question of Dhvani in poetry, in correla- tion with the theory of Rasa intimately connected therewith. Abhinava, for instance, while discussing (p. 16) the verse bhama dhammia vīsattho, which is given by Anandavardhana as an instance of suggestion with an expressed injunction implying a prohibition, refutes Bhatta Nāyaka's opinion as to this negative implication. In another place. Abhinava criticises (p. 21) the significance attached by Bhatta Nāyaka

1 darpaņo hrdaya-darpanākhyo dhvani-dhvamsa-grantho'pi, explain- ing the pun on the word darpana used by Mahimabhatta in i. 4. 2 Both Bhatta Nayaka and his Hrdaya-darpana are cited by name in Locana pp. 27, 28, 63. Other references occur on pp. 11, 12, 15, 19, 21, 29, 36, 67, 68-all of which consist mostly of direct cri- ticism in support of Dhvani-theory. 3 For instance, on Bharata p. 1: bhatta-nāyakas tu brahmaņā paramātmanā yad udāhrtam ...... iti vyākhyānam hrdaya-darpaņe pra- tyagrahit. 4 adrsta-darpanā mama dhiḥ i. 4.

40 SANSKRIT POETICS

to the word aham in the verse attā ettha nimajjai, which is discussed by Anandavardhana as an example of suggestion of a contrary kind where the expressed prohibition implies an injunction, It is evident from these references that Bhatta Nāyaka's work, like Mahimabhatta's, was designed not merely as a refutation of the general theory of Dhvani, but also as a special attack on Anandavardhana's exposition of the same. To take a minute point, the Dhvanikāra in i. 13 uses the verb vyanktah in the dual number with a special object in view, as Anandavardhana's (as well as Abhinava's) explana- tion rightly indicates. Bhatta Nāyaka seems to have attacked this use of the dual number, upon which Abhinavagupta remarks (Locana, p. 33): bhatta-nāyakena yad dvi-vacanam dūşitam tad gaja-nimīlikayaiva.1 It appears, therefore, that the Hrdaya-darpana was not a commentary on Bharata,2 but a metrical treatise in the anustubh with a running prose commentary, dealing with the ques- tion of Dhvani, and incidentally with the question of Rasa- dhvani. No doubt, Abhinava in his own commentary on Bharata, as well numerous other later writers taking their cue from Abhinava, criticises at some length Bhatta Nāyaka's. theory of Rasa, along with those of Lollata and Sankuka and with special reference to Bharata's particular sūtra on the subject in ch, vi (also Locana pp. 67-68) ; yet Bhatta. Nāyaka is nowhere mentioned directly as a commentator on the same text. Very rarely Bhatta Nayaka's interpretations of particular passages of Bharata are cited by Abhinavagupta, as they are with regard to those of Udbhata, Lollata and Śankuka.8 It is probable that Bhatta Nāyaka's peculiar theory

1 Mahimabhatta also refers (p. 19) to this discussion, quoting these words of Abhinava from the Locana. 2 As V. V. Sovani in Bhandarkar Comm. Volume, p. 390 (contra in JRAS, 1909, pp. 450-52) states. 3 T. R. Chintamani collected together 'Fragments of Bhatta Nāyaka' in JOR i, 1927, pp. 267-76; also in Proc. A-I. O. C., Allahabad 1929, ii, pp. 155. 193.

COMMENTATORS ON BHARATA 41

of Rasa (which, however, bears a resemblance to Abhinava's own) called for a special refutation in the hands of this champion of the Dhvani-theory, because Bhatta Nāyaka denied the expressive function of Dhvani and attempted to explain the concept by postulating the function of Bhogī- karaņa. But there is no definite indication to shew that this theory of Rasa, being a corollary to Bhatta Nāyaka's general. theory of expression, was not incidental to his main thesis, which was directed towards the demolition of the new idea of Dhvani and establishment of another explanation of that concept. This may be the reason why Sārngadeva, in his enumeration of Bharata's commentators before his time, omits the name of Bhatta Nāyaka.1 There can hardly be any doubt that Bhatta Nāyaka was. familiar with the text of the Dhvanyaloka, including Ānanda- vardhana's Vrtti, and should, therefore, be placed in a period later than the date of Anandavardhana. The conclusion is supported by the statement of Jayaratha (p. 12) that Bhatta Nāyaka lived after the Dhvanikāra by whom Jayaratha, like many other later writers, invariably means Anandavardhana without distinguishing him from the so-called Dhvanikāra. On the other hand, the oldest writer to mention and cite Bhatta Nayaka is Abhinavagupta, from whom he does not appear to be chronologically very distant. Bhatta Nāyaka, therefore, flourished between the last quarter of the 9th and the last quarter of the 10th century; and it will not be wrong if we assign him to the end of the 9th and the beginning of the 10th century A.D. This date makes it likely that he is identical, as Peterson suggested, with Bhatta Nayaka who is mentioned by Kahlana (v. 159) as having flourished in the

1 See p. 32 above, footnote 2. With this view P. V. Kane (HSP, p. 214) agrees. Similarly Ruyyaka, while reviewing the different systems which obtained before his time, mentions Bhatta Näyaka not as a commentator but as an independent author who advocated a new system in common with other explanations of the Dhvani theory (p. 9, ed. N. S. P.)

42 SANSKRIT POETICS

reign of Samkaravarman, son and successor of Avantivarman of Kashmir.

Harsa

Harșa or Śrī-harșa is said to have composed a Vārttika on the Nātya-sastra. He is cited by Abhinavagupta sometimes by name (on v.7, 1801 ; xxix. 101) and sometimes as Vārttika-krt or Vārttika-kāra (i. 84 ; ii. 97-98 ; iv. 267-68)2, while Vārttika without the name of the author3 or Harsa-vārttika (iv, 331)4 is also cited. On v. 8-15 there is a prose passage from the Vārttika quoted5. The citations are more profuse from the first six chapters, but there is one reference to ch. xxix. Sāra- dātanaya (p. 238) refers to Harsa's view that the Totaka differs from Nātaka in having no Vidūşaka6. As the Vārttika is no longer available no definite conclusion is possible. These citations however, make it highly probable that it might not have been a regular commentary (in spite of its name)7, but dealt, mostly in Āryā metre (with occasional prose), with relevant parts of the Nātya-śastra.

Kīrtidhara Śārngadeva informs us (see above p. 32, fn 2) that Kīrtidhara was a commentator on Bharata's text. He must have been earlier than Abhinavagupta who tells us (ch.xxix) that not having seen Nandikeśvara's work himself, he is relying on Kīrtidhara's account of it (see above p. 20). There are several references8 in Abhinava's commentary to Kīrtidhara or

1 Vol. i. p. 211 and p. 251. 2 Vol. i, p. 31; i, p. 67; i, 172, 174 respectively. 3 Vol. i. p. 174. 4 Vol. i. p. 207 5 Vol. i, p. 212. 6 Śrīharșa Miśra quoted in Prabhākara Bhațța's Rasa-pradīpa (a prose passage) may or not be our author. 7 As in the case of Nānyadeva's Bharata-bhāsya or Bharata- vārttika. 8 See V. Raghavan in JOR, v, 1932, p. 198; Kane in Pathak Comm. Volume, p. 388.

COMMENTATORS ON BHARATA 43

Kīrtidharācārya on Nātya and Nrtta (ch. iv) 1 and on the Geyādhikāra section. These citations show that Kīrtidhara, like Sārngadeva, was interested mostly in music, but they are not sufficient to establish that he wrote a regular comm entary on the entire text of Bharata.

Abinavagupta Although Abhinavagupta contented himself with the writing of commentaries in the field of Sanskrit Poetics, his works have almost the value of independent treatises for their profound erudition and critical acumen. As his reputation in Poetics rests on his exposition of the Dhvani-theory, it would be better to take him up in connexion with the Dhvani- kāra and Ānandavardhana. The entire text of Abhinava's commentary, called Abhi- nava-bharatī, on Bharata's text which consisted of 36 chapters, is not available, either in the printed edition or in any MS. His comments on ch.vii (except the prose portion and the first few verses), ch. viii and ch. xxxiii-xxxiv are missing, and there are also short gaps (e.g. on the last verses of ch. v), As he refers to his Locana in this work, it was composed later.

Nānyadeva A work called Bharata-bhāşya (also Bharata-vārttika in some of its colophons) by Nanyadeva or Nānyapati is available in a unique MS (221 folios) in the BORI collection2. The author is called Mithileśvara (king of Mithila) in one of its verses, while the colophons describe him as Mahāsāmantā- dhipati. The work is unknown to Abhinavagupta who was earlier in date and who is utilised in it but rarely mentioned by name. Nanyadeva is known as the founder of the Karņā-

1 Vol. i, p. 208. 2 BORI Cat, of MSS, xii, no. 111 of 1869-70, pp. 377-83. The work also goes by the name Sarasvafi-hrdayalamkara. Dr. C. P. Desai of Tarapur, Thana, Bombay, is understood to be editing the work for Khairagarh Music University, M. P.

44 SANSKRIT POETICS

taka dynasty of Mithila who ruled from 1097 to 1147 A.D1. The author mentions another work of his called Grantha- mahārņava. Although it is called a Bhäsya, it is not a direct commen- tary on Bharata's text, It was ambitiously planned in four Amsas, each devoted to one of the four kinds of Abhinaya; but the extant portion, itself extensive, deals only with one kind, namely, Väcika, and relates chiefly to ch. xxviii to xxxiii of the Natya-sastra, which deal with music. The MS, though old, is defective, wanting in ch. v, xvi and xvii (the total number of promised chapters being seventeen). Bharata is profusely quoted, but other old authorities like Nārada, Sātātapa, Dattila, Kāśyapa (also Brhat-Kāśyapa and Vṛddha-Kāśyapa), Matanga, Bṛhad-deśī, Nandi-mata, Yastika (otherwise unknown), Kīrtidhara, and Viśākhila are frequently cited. Sārngadeva appears to be the only author who cites Nānyadeva.

BIBLIOGRAPHY

Nāțya-śāstra and Abhinava-bhāratī

Editions and Translations. (1) Ed. Sivadatta and K. P. Parab, NSP, Kāvyamāla 42, 1894 (Adhy. 1-37). (2) Traité de Bharata surle Théâtre, Texte Sanskrit, édition critique par Joanny Grosset, t. i. pt. i, Paris 1898 (ch. i-14 only, not completed), in Annales de l'Université de Lyon. In Roman characters. (3) Adhy. 18, 19, 20 and 24, published by F. Hall in his ed. of the Daśa-rūpaka, in Bibl. Ind. Calcutta 1865. (4) Le 17me chapitre de Bhāratīya Nātyaśāstra, intitulé Vag-abhinaya, par P. Regnaud, in Annales Musée Guimet I, 1880. In Roman characters. (5) La Métrique de Bharata, Texte Sanscrit de duex chapitres (15-16), suivi d'une interprétation francaise, par P. Regnaud, in Annales du Musée Guimet II, 1881. In Roman characters. (6) Textes

1 EI, i, 395 at p. 364; IHO, vii, pp, 679-87.

COMMENTATORS ON BHARATA 45

Sanscrits des 6me et 7me chapitres, in Rhétorique Sanscrit par P. Regnaud, Paris 1884. In Roman characters, (7) Adhy. 28 (in Roman characters) in Contribution à l'étude de la musique hindoue by J. Grosset, Paris 1888, in Bibl. de la Faculté de Lettres de Lyon; also B. Breloer, Grund- elemente der altindischen Music nach dem Bhāratīya Nātya- śāstra, Text, Uebersetzung und Erklaerung (ch. 28). Diss. Bonn, 1922. Both in Roman transliteration. (8) A part of Adhy. 6 (On Rasa-sūtra), ed. S. K. De as an Appendix to his Theory of Rasa in Asutosh Mookerjee Comm. Vol., Orientalia, pt. iii, 1922, p. 240f, now revised and reprinted in his Some Problems of Skt. Poetics, Calcutta 1959, pp. 219-35. (9) Adhy. 6 (Rasādhyāya) with Abhinava's comm., with Eng. trs. of the original text of the chapter, ed. Subodh Chandra Muker- jee, Calcutta 1926 (Thesis, Paris Univ.). (10) ed. Batuknath Sarma and Baladev Upadhyay. Chowkhamba Skt. Series, Benares 1929. This ed. gives 36 chapters. (11) With Abhinava- bhāratī, ed. M. Ramakrishna Kavi. In four volumes. Gaekwad's Orient. Series, Baroda 1926, 1934 etc. (On this ed. see P. V. Kane, HSP, pp. 14, 16 "fundamentally uncriti- cal"; S. K. De in IHQ iii, pp. 859-68). (12) Prakrit verses in Adhy. xxxii, ed. Manomohon Ghosh in IHQ, viii, 1932. {13) English translation by Manomohan Ghosh, Bibl, Ind. vol. i (Adhy. i-xxvi), 1959.

CHAPTER III

FROM BHAMAHA TO ANANDAVARDHANA

BHAMAHA

The earliest citation of Bhamaha in later Alamkāra literature is to be found in two passages in Anandavardhana's Vrtti on the Dhvanyāloka (pp. 39, 207); and at p. 236 Ānanda- vardhana quotes anonymously Bhamaha iii. 27. The next interesting reference occurs in the commentary of Pratīhāren- duraja, who informs us (p. 13) that his author Udbhata composed a work, presumably a commentary on Bhāmaha, which is described as Bhāmaha-vivarana. This state- ment is confirmed by Abhinavagupta (Locana pp. 10, 40, 159, vivaranakrt) and Hemachandra (Comm. pp. 17, 110) ; while Ruyyaka cites the commentary generally as bhāmahīya udbhata-lakşaņa (p. 183) and Samudrabandha discribes it as kāvyālamkāra-vivrti (p. 89). There are also numerous pas- sages in Udbhata's independent work, Kāvyālamkāra-sam- graha, which unmistakably copy some of the definitions of poetic figures directly from Bhämaha, and do not hesitate to repeat the very language of the earlier work.1 Vāmana, Udbhața's contemporary, also appears to betray an acquaintance with Bhamaha's text.2 Bhamaha, for

1 See, for instance, the definitions of the figures rasavat, atiśayokti, sasaņdeha, sahokti, apahnuti, utpreksā, yathāsamkhya, aprastuta-praśamsā paryāyokta, ākșepa, vibhāvanā, virodha and bhavika .- Bhamaha is quoted extensively by Abhinavagupta and other later writers. 2 Bhāmaha's work called Bhāmahālamkāra (but Kāvyālamkāra is the first verse) consists of six Paricchedas or chapters and about 400 verses. The topics covered are: I. Purposes of Kāvya, its. definition and divisions from different pomnts of view; Sargabandha, Kathā and Ākhyāyikā; reference to Vaidarbha and Gauda modes

BHĀMAHA 47

instance, defines the figure Upamã (ii. 30) as viruddhenopamā- nena ...... upameyasya yat sāmyam guņa-leśena sopamā; and Vāmana seems to paraphrase this definition in the concise form of a Sūtra: upamānenopameyasya guņa-leśataḥ sāmyam upamā (iv. 2. 1).1 Again, speaking of effective implication (atiśayavān arthaḥ) to be found in Upamā, Bhāmaha lays down (ii. 50): yasyātiśayavān artha katham so'sambhavo mataḥ/. iştam cātiśayārthatvam upamotprekşayor yathā/ / Reading together Vamana iv. 2. 20 and 21 (anupapattir asambhavaḥ and na viruddho'tiśayah), we find that Vāmana is apparently repeating the same view ; and in his Vrtti on the first Sūtra, he adds upamāyām atisayasyestatvāt, making it clear in the next Sūtra that an effective implication (atisaya), which is contradictory, should be avoided. Vämana also reproduces anonymously a verse of an unknown poet whose name is given by Bhämaha (ii. 46) with the same verse as Sākhavardhana. Such repetition of views in more or less. standardised phraseology in a technical treatise, or the quota- tion of the same illustrative verse in a similar context need not be taken as conclusive ; but Vāmana, in his Vrtti on v. 2. 38, actually though not accurately, quotes a part of a verse from Bhamaha ii. 27, and comments on the peculiar usage of the word bhanguram employed therein.2

some Doșas pertaining generally to the Kāvya. II-III. The three Guņas (Mādhurya, Prasāda and Ojas); treatment of Alamkāras which ends with ch. iii (for a list of the poetic figures see vol. ii. ch. ii(i). IV. Eleven Dosas, with illustrations. V. Eleven Dosas again which arise from a faulty Pratijña, Hetu or Drstanta, VI. Sauśabdya or grammatical correctness (elaborated later by Vämana in fifth Adhi- karņa of his work). 1 Cf Bharata xvi. 41. 2 The verse is quoted with Bhamaha's name in Jayamangala on Bhatti x. 21; ananymously in the Vakrokti-jivita (along with other verses from Bhämaha) and in Lacana p. 40 anonymously.

48 SANSKRIT POETICS

This will justify us in placing Bhamaha chronologically before Udbhata and Vamana who, as we shall see, flourished in the last quarter of the 8th century A. D., and will give us one terminus to the date of Bhamaha. With regard to the other terminus, controversy has been keen and busy. Pathak finds in the mention of a nyāsakāra in Bhämaha vi. 36 a clear reference to the Buddhist Jinendra- buddhi, author of a commentary (ed. Varendra Research Society, Rajshahi 1913, 1919-25) on the Kāśikā, and comes to the conclusion that "as the nyāsakāra (meaning Jinendra) lived about 700 A. D., Bhämaha must be assigned to the 8th century".1 Against this K. P. Trivedi has demonstrated2 that the allusion to the opinion of the nyāsakāra cannot be taken as an unmistakable reference to Jinendrabuddhi's views, and that the existence of some other nyāsakāras is also made probable by the citations of Mädhava, as well as by a punning passage in Bāna's Harsa- carita.3 Jacobi4 has joined issue by adding a doubt on the correctness of the date assigned by Pathak to Jinendrabuddhi who, on the authority of Kielhorn, was probably later than Haradatta (d. 878 A. D.). No fresh light is thrown on the question by the conjecture5 that Bhamaha in i. 42 refers to the Megha-dūta by his con- demnation of the poetical device of employing clouds, among other things, as messengers ; nor by Pathak's other equally fanciful supposition that Magha ii. 86b refers to Bhamaha

1 IA xli p. 232 ff, at p. 235; see also JBRAS xxiii pp. .25-26. 2 IA xlii pp. 204 f, and at pp. 260-1. 3 krta-guru-padanyāsāh (ed N. S. P. p. 96), explained by Samkara as krto'bhyasto guru-pade durbodha-sabde nyaso vrttir vivarano yaih. 4 Sb. der Preuss. Akad xxiv (1922), pp. 210-11. 5 Haricand, L'Art Poétique de l'Inde p. 77; J. Nobel in ZDMG lxxiii, p. 192.

i. 16.1 Nothing, again, is gained by the controversy over the question whether Bhämaha, the son of Rakrilagomin and worshipper of Särva, was a Buddhist, as indicated by the opening and closing verses of his work.2 Jacobi however, has shewn3 that Bhamaha has made considerable use of the teachings of Buddhist philosophers in ch. v, and that the upper limit to Bhamaha's date should be determined with reference to that of Buddhist Dharmakirti, some of whose philosophical doctrines Bhamaha has utilised even to the repetition of Darmakirti's actual phraseology. Dharma- kīrti is placed by Jacobi between the sojourn in India of Yuan Chwang and Yi-tsing respectively (630-643 and 673-695 A, D.), as he is not known to the former, while the latter refers to him among those of late years.4 The upper limit, therefore, of Bhamaha's date should be fixed at the third quarter of the 7th century A. D. Sāntarakșita in his Tattva-samgraha (ed. GOS, 1926, p. 219, verses 912-14), who is said to have flourished about 705-762 A. D., cites three verses of Bhämaha (vi. 17-19) criticising the Buddhist Apoha-vada. This would also go to establish that Bhamaha cannot be placed much later than the seventh century. We can. therefore, place Bhamaha approximately in the period between the last quarter of the 7th and the middle of

1 But see Dandin i. 10; Vāmana i. 1. 1 (vrtti); Rudrața ii. 1 and Anandavardhana p. 5, for the same idea of sabda and artha as constituents of poetry. 2 On this controversy, see JRAS 1905, pp. 535 f; JRAS, 1908, pp. 543f; Trivedi introd. to Prataparudra; Haricand op. cit. p. 71; Pathak in 1A 1912, p. 235. 3 op cit pp. 211-12. G. Tucci (Bhamaha and Dinnāga in IA, June 1930) is of opinion that Bhamaha's logical theories allude to Dinnaga, and not to Dharmakīrtī who was posterior to Bhamaha. 4 See. Takakusu, Record of the Buddhist Religion, 1896, p. 181; cf p. Iviii. Taranātha in his Geschichte (tr. Schiefner pp. 184-5) makes him a contemporary of the Tibetan king, Strong-bstan-sgam- po, who died about 650 A.D. Cf. Kern, Manual of Indian Buddhism, p. 130. 4

50 SANSKRIT POETICS

8th century. As it is probable that he might have been a younger contemporary of Dharmakīrti's and also presumably lived some time before his commentator Udbhata, it would not be wrong if we place him towards the end of the 7th and the commencement of the 8th century A. D. ( 2 ) We have already discussed the relation which Bharata's treatment of Alamkäras might have borne to Bhamaha's much fuller and later disquisition. What we find in Bharata con- stitutes the earliest speculation on the subject that we possess ; but Bhamaha himself tells us that he had predecessors whose works he apparently utilised. While referring to these prede- cessors (or contemporaries) generally as anye,1 apare 2 or kecit,3 Bhāmaha cites twice by name one Medhāvin in ii. 40, 88. One of these passages is referred to by Nami-sādhu on Rudrata xi. 24, where (as well as in two other places on i. 2, ii. 2) the full name is given as Medhävirudra, which form also occurs in Rajasekhara (p. 12)4. This writer was thus earlier than Bhamaha but probably later than Bharata. Bhämaha's work is divided into six chapters with a total of about 400 verses (mostly in Anustubh) (see above p. 46, 1 i. 13, 24; i1. 4, 57; iii. 4; iv. 12 etc. 2 i. 14, 31 ; ii. 6, 8; iii. 4; iv. 6 etc. 3 ii. 2, 37, 93; iii. 54 etc. He also cites one Rāmaśarman in ii. 19, but from ii. 58 this author appears to be a poet; the name of his work is given as Acyutottara. The Rajamitra cited in ii. 45 appears from iii. 10 to be a Kāvya. Besides Nyāsa (vi. 36), Pāņini (vì. 62-63) and Kanabhaksa (v. 17), we have mention of Sākavardhana (ii. 47) and a work called Aśmaka-vamśa. These citations are not of much use for chronological purposes. 4 Rajasekhara couples Medhāvirudra's name with that of Kumaradāsa and adds the information that he was a born-blind poet The name does not constitute the names of two different poets, Medhavin and Rudra, as some writers suggest, nor need we take it on the late authority of the Trikānda-sesa as a name of Kālidāsa .- Hultzsch (ed. Megha-dūta p. xi) states that Vallabha- deva in his comm. (xi. 6) cites Medhävirudra, but this is not found in the printed text.

BHĀMAHA

fn 2) ; it is smaller in extent than Dandin's work (about 660 verses). At one time it was believed1, on the indication given by the Jayamangala on Bhatti, that the Alamkāra- chapters in that Kāvya2, especially canto x, was meant to illustrate the rhetorical teachings of Bhamaha in particular ; but the date now assigned to Bhämaha will readjust his relation to Bhatti in a new light. Bhatti tells us in xxii. 35 that he composed his pcem in Valabhī ruled over by Śrīdharasena3. It appears that no less than four Sridharasenas ruled at Valabhi roughly between 500 and 650 A.D., of whom the last flourished, as his latest grant shows, in 651 A.D. Bhatti, therefore, at the latest, lived in the first half of the 7th century ; and if, as his editor con- cludes,4 he may be assigned to the end of the 6th and the beginning of the 7th century, he was certainly older than Bhamaha by almost a century, Bhāmaha probably knew his work and therefore remarked, while dismissing verbal juggleries like prahelikā (ii, 20): kāvyāny api yadīmāni vyākhyā-gamyāmi śāstravat| utsavaḥ sudhiyām eva hanta durmedhaso hatāḥ / /. with a pointed reference to Bhatti's self-boasting in xxii. 34 .: , vyākhyā-gamyam idam kāvyam utsavaḥ sudhiyām alam/, hatā durmedhasaś cāsmin vidvat-priyatayā mayā/ / The treatment of Alamkāras in Bhatti may, therefore, be

1 Jacobi in ZDMG lxiv, p. 130f. 2 The three cantos of this Kavya (x-xii) comprising what is called Prasanna-kāņda are supposed to illustrate matters concerning Poetics. Thus, canto x (75 verses) illustrates Alamkāras, xi (47 verses) Mādhurya Guna and xii (87 verses) the figure Bhävika which is called a Prabandha-guņa. The Bhatti-kūvya consists of 22 cantos, chiefly illustrating rules of Sanskrit Grammar. 3 kāvyam idam vihitam mayā valabhyām | śrīdharasena- narendra-pālitāyām. The Jayamangalā reads śrīdhara-sūnu-narendra in the second line, but this cannot be supported in view of the fact that we do not hear of any prince of the name Narendra, son of ·Śrīdhara, in the list of Valabhi princes known to us. Mallinatha and Bharatamallika do not comment on this verse. 4 ed. B.S.S. Incrod. p. xxii.

52 SANSKRIT POETICS

presumed to supply one of the missing links in the history of rhetorical speculations anterior to Bhāmaha. A remarkable coincidence of treatment, which probably started the theory of Bhatti's appropriation of Bhama ha's teachings, is at once noticeable not only in the order, number and presumed naming and characterisation of different poetic figures; but a detailed examination will at the same time shew that beneath this general agreement there are enough discrepancies which will indicate that neither of them follows scrupulously the views of the other. The agreement apparent- ly shows that the two authors were not chronologically distant from each other by such a considerable length of time as might betoken a material difference in the number, order or definition of the poetic figures ; while the discrepancies may be reasonably explained as indicating that they did not probably draw from the same source. The special object of the particular canto in Bhatti being that of illustrating the various forms of poetic figures pre- valent in his time, we may presume that it was probably based on a particular treatise on Alamkāra to which the poet generally adheres. He is said to have mentioned in all 38 such independent figures, along with 39 subspecies of some individual figures. He does not himself give the names of these figures, but they are indicated by the Jayamangalā, as well as in some MSS which apparently preserve the tradi- tional nomenclature1. These, with one trifling exception (udāra=udātta), correspond to the particular names given to them in Bhämaba. As to the order or sequence of treatment, a comparative table will show that Bhāmaha gives the first 23 figures (up to viseşokti) in the same order as in

1 The commentators, however, differ among themselves in the naming of the poetic figures in several stanzas .- The Jayamangalā is printed in the NSP ed. of the Kāvya (1887), while the commentary of Mallinätha is given in the BSS ed. (in 2 vols. 1898). The com- mentary of Bharatamallika along with Jayamangalū has been edited in 2 vols, Calcutta 1871-73.

BHÃMAHA 53

Bhatti with the exception of the pairs, rūpaka, and dīpaka, arthāntara-nyāsa and āksepa, which are given in an inverse order. The rest of the figures appear with a slightly different arrangement, because Bhāmaha admits aprastuta- praśamsā omitted by Bhatți, and adopts a somewhat different order in mentioning the five figures here treated in common, until we come to virodha. From here, again, the order is the same, excepting that Bhāmaha mentions bhāvika (which is separately illustrated in another canto by Bhatti), while Bhatti admits an unknown figure nipuņa, and adds hetu and vārtā which are expressly rejected by Bhämaha. As the exposition of Jayamangla shows, Bhatti generally follows the definitions of Bhämaha where the figures are common (even in the cases of subspecies of these figures1), with only a few exceptions. These exceptions, though few, are yet significant. They refer in particular to the figures yamaka (of which Bhatti mentions 20, while Bhamaha only 5 subspecies), upamā (where the treatment of subspecies is slightly divergent), rūpaka (of which the four subspecies of Bhatti do not corres- pond to the two of Bhāmaha), aprastuta-praśamsā omitted by Bhatti, and nipuna omitted by Bhāmaha. At the same time, Bhāmaha mentions but rejects prahelikā, hetu, sūkşma, leśa and varta, of which Bhatti admits only hetu (probably as an afterthought) and vārtā. Bhatti does not recognise svabhā- vokti, which is mentioned but apparently disfavoured by

1 e.g. the figure äksepa, of which the two subdivisions ukta- vişaya and vakşyamāna-vişaya are found in both Bhāmaha and Bhatti, they being unaware of the different interpretation of Vamana and the somewhat fine differentiations of Dandin. The same remark applies to dipaka and its three subspecies, which do not agree with the exposition of Bharata, Dandin or Vamana. Cf also the three subspecies of ślesa viz. sahokti-śl°, upama-śl° and hetu-śl°, illustrated by Bhatti and mentioned by Bhamaha in iii. 17, although later writers. like Dandin and Udbhata, speak of ślesa as coming with many other figures. Pratīhārendurāja distinctly alludes (p. 47) to this division admitted by Bhamaha: bhamaho hi "tat sahoktyupamā-

54 SANSKRIT POETICS

Bhamaha. It is possible that Bhatti's original ended naturally with asis, as Bhamaha's work itself does; but he tacked on hetu and nipuna1 as two supplementary figures popular in his time. The bhāvika, which both Bhāmaha and Dandin call a prabandha-guna, is said to have been illustrated by Bhatti in a separate canto (xii), entitled bhāvikatva-pradarśana. But by far the greatest divergence is noticeable in the treatment of the subspecies of yamaka, rūpaka and upamā. No two writers are indeed agreed with regard to the treatment and classification of yamaka, and Bhatti on this point is scarcely in agreement with any of the known writers on the subject, such as Bharata, Dandin Rudrata, the author of the Agni-purana and Bhoja among earlier authorities. Probably he is drawing upon some old author whose work is not known to us.2 In the classification of rūpaka, which Bhāmaha subdivides into samasta-vastu- vişaya and ekadeśa-vivarti, Bhatți seems to follow a different tradition which mentions four subspecies, respectively 1 This figure is included in udāra or udātta by Jayamangalā, while Bharatamallika and Mallinatha take it as an illustration of preyas on the authority of Dandin and Devānātha (the latter pro- bably a commentator on Mammata having the same name). 2 The names of some of these subspecies of yamaka are now lost but for the naming of them in Bhatti, and later authors speak in altogether different terminology. Some of these are apparently preserved in Bharata, who mentions as many as ten subspecies, but in most cases they are differently defined. For instance, the samudga of Bhatti may be the same figure as defined by Bharata, but the yukpūda of Bhatți x. 2 is called vikrnta by Bharata and is known as sandasta in Rudratța. Similarly the pādānta illustrated in x. 3 is called āmredita in Bharata; while cakravāla of Bharata is different from the figure so named in Bhatti and seems to coincide with the kūñci of the latter, the kāñci of Bharata being an alto- gether different subspecies. It appears that names like vrnta, mithuna, or vipatha cannot be traced in any of the existing works, but some of the kinds illustrated by Bhatti under these strange names may be found under different designations in other writers later than Bhatti. In naming these in Bhatti, the Jayamangala is probably

BHÂMAHA 55

designated kamalaka (viśiştopamā-yukta), avatamsaka (śeşarthānvavasita or khaņda-rūpaka1), ardha-rūpaka and lalāmaka (anvarthopamā-yukta). In the subspecies of upamā, Bhatti illustrates upama with iva and yatha (in common with Bhamaha); and his luptopamā and taddhitopomā probably correspond to some extent to samāsopamā and upamā with vat mentioned by Bhamaha. But Bhatti does not illustrate prativastūpamā of Bhāmaha nor does he refer to nindoo, praśamso°, ācikhyāso° and mālo°, criticised by Bhāmaha but recognised by Dandin.2 At the same time, Bhatti's saho® and samo° have nothing directly corresponding to them in Bhāmaha. It will be clear from this brief exposition3 that, leaving aside the subspecies, there is a general agreement between the treatments of Bhatti and Bhämaha with regard to the independent poetic figures. It may be noted that Bhāmaha agrees with Bhatti in taking ananvaya, sasamdeha, upamā- rūpaka and utprekşāvayava as self-standing figures, while Dandin includes the first two in the sub-species of- upamā, and the last two in those of rūpaka and utpreksa respectively. Bhämaha also agrees with Bhatti in rejecting prahelikā, sūkşma and leśa; but vārtā and hetu, also similarly rejected by Bhämaha, are admitted by Bhatti. Dandin expressly recognises all these, excepting varta in place of which he probably admits the more comprehensive svabhāvokti, which

1 mentioned in Vāgbhatālamkāra iv. 66. 2 Bharata (xvi. 49-50) mentions nindo° and prasamso°, while his kalpito° probably corresponds to ācikhyāso°. The kalpito° is admitted by Vāmana (iv. 2. 2) but apparently defined in a different sense. 3 See also H. R. Divekar in JRAS, 1929, pp, pp. 825-41 for a com- parison and contrast of treatment made respectively by Bhāmaha and Bhatți. Dandin is followed in this view by all later writers, except Vämana, who still regards these as independent figures. It seems, therefore, that Vamana vi. 3. 33 is a criticism of Dandin ii. 358, and

56 SANSKRIT POETICS

is disfavoured by Bhämaha and not illustrated by Bhatti. The most material discrepancy with reference to independent figures occurs in the remarkable omission in Bhatti of apra- stuta-praśamsā (which, like svabhāvokti, is a recognised figure in later times) and in the occurrence of nipuna unknown in later literature. Coming to the subspecies, however, the discrepancies are more striking. Admitting that some of the fine differentiations, as in the case of Dandin's innumerable subvarieties of independent figures, may have been invented by the ingenuity of the theorist himself, this argument does not seem to apply very well to Bhatti, who was himself no theorist but only professed to illustrate the poetic figures popu- lar in his time and presumably based his treatment on some standard treatise. The conclusion, therefore, is likely that Bhatti made use of a text unknown to Bhämaha but not materially differing from Bhamaha's own sources; and that the interval between these two authors did not witness much change in the discussion of poetic figures, except what is apparent in the simplification of the treatment of yamaka and rūpaka, in the dropping of a figure like nipuna and adding an important figure or an important sub-figure like apra- stuta-praśamsā or prativastūpamā respectively. The progress is not so remarkable as that indicated by the enormous stride made in the interval between Bharata. who mentions only four independent figures, and Bhatti, who mentions thirty-eight.1

1 Although the name Bhamaha is not a common one in Sans- krit, it attaches itself (besides two verses in Subhāsitāvalī 1644-1645 that are also found in our text ii. 92, iii. 21) to a commentator on Vararuci's Prūkrta-nrakaśa, who is probahly a different author. The Kāmadhenu comm, on Vāmana also cites several verses from a trea- tise apparently on the kalās by Bhāmaha (p. 29, ed. Benares); but as our Bhamaha, as well as his Bhamahālamkara (p. 39), is also. cited in several places in the same commentary, it is possible that these verses occurred in some lost chapter of his work where he mentioned the names of the kalās (atra kalānām uddesah krto bhā-

DANDIN 57

BIBLIOGRAPHY Edition. Text (i) by K. P. Trivedi as Appendix viii to his ed. of Pratapa-rudra° in BSS, Bombay 1909. The edition is based on Madras MS no. 12920 (Cat. xii, p. 8675). The work is named Bhāmahālamkāra. (ii) by P. V. Naganatha Sastry, with Eng. trs. and notes, Tanjore 1927. Also separately text only, Tanjore 1927. (iii) by B. N. Sarma and B. Upadhaya, Chowkhamba Sanskrit Series, Benares 1928. None of these printed texts can serve as a critical edition. The MS material is meagre and citations from Bhāmaha in later writers have readings which have not been con- sidered. The text at many points is unsatisfactory. Commentary. The only known comm. is Bhāmaha-vivarana by Udbhata, which is now lost. See above p. 46.

DANDIN

( 1 ) The date of Dandin. author of the Kāvyādarśa, is one of the most difficult problems in the chronology of Alamkāra literature. Anandavardhana does not directly cite him, as he cites Bhamaha, and the earliest mention of Dandin's name occurs in Pratīharenduraja (p. 26). Dandin's own work gives us hardly any clue. His references to the Brhatkathā written in bhūta-bhāșā (i. 38), or to the Setu-bandha known to him in māhārāstrī Prakrit (i. 34) throw little light on the question ; and no definite chronological conclusion is deducible from the verses ii. 278-79, which express, under the form of the figure preyas, the supreme gratification of a certain king, Rājavarman (or Rātavarman), on the occasion of his obtaining the much-coveted beatific vision of his adored deity.1 The solution proposed to the

on the Vrtta-ratnākara (pp. 5-6) quotes long passages from Bhāmaha which, if authentic, indicate that Bhämaha might have written also on Metrics.

58 SANSKRIT POETICS

prahelikā in iii. 114 (also @f. iii. 112) by Taruņavācaspati and other commentators that it refers to the Pallava kings of Kāñcī 1 only supports the Tamil tradition that Dandin was probably a South Indian author. The allusion to Dandin i, 1, again, in a verse attributed by Sārngadhara (no. 180) to Vijjā or Vijjakā (whose date is unknown but who is tentatively supposed by some to be Vijaya, wife of Candrāditya and daughter-in-law of Pulakeśin II, about 659 A.D.),2 implies merely a pleasant raillery at the expense of Dandin by some later boastful poetess. The only definite terminus to Dandin's date is obtained from references in South Indian vernacular works on Alam- kāra, belonging in all probability to the 9th century A.D., which cite him as an established authority. The Sinhalese treatise Siya-bas-lakara, which Barnett thinks cannot "in any case be later than the 9th century A.D."3 cites Dandin in v. 2 as one of its authorities. The Kanarese work Kavirāja- mārga (in three chapters), attributed to the Rāstrakūta prince Amoghavarsa Nrpatunga (who flourished in the first half of

8 ; also Agashe's preface to Daśakumāra ed. B. S. S. pp. lxii f) to be Rājasimhavarman, otherwise known as Narasimhavarman II of Kañci (end of the 7th century), one of whose birudas (viz. kālakāla which is also a name of Siva) Dandin is supposed to have alluded to in iii. 50; while iii. 25 is presumed to imply a pun referring to the royal token (mahāvarāha) of Cālukya Pulakeśin II. But the passage under discussion looks like a reference to a legendary rather than a contemporary prince; and, as Pischel suggested, the entire verse 278 may have been taken directly from a work referring to his story. Cf Jacobi op. cit. p. 214. 1 The phrase asta-varna occurring in the prahelikā is also found, as G. K. Sankara points out, in the Mamandur Inscription of Mahendravarman I. Premachandra on Dandin interprets pundraka instead of pallava in the verse, which fact indicates that it is capa- ble of a different interpretation .- There are references to Kaveri, Cola and Kalinga in iv. 43-44. 2 See Agashe op. cit. pp. lix f. 3 JRAS, 1905, p. 841. The work has been edited by Hendrick

DAŅŅIN 59

the 9th century), gives six verses1 which are exact transla- tions. of corresponding verses in Dandin. Pathak, in the introduction to his edition (1898) of this work (p. 19), further adds that in ch. iii most of the verses "are either translations or adaptations from the Kāvyadarsa," and that there are also convincing indications of Dandin's "influence on other parts of the work" as well. This will give us the 9th century as the lower limit to Dandin's work, a conclusion which may also be established by showing that Dandin was probably earlier than Vāmana, who may be assigned to the beginning of the same period. We need not enter into this point in detail here, but there are several unmistakable indications which show that .Vämana's work betrays a further progress in the elaboration of some of the fundamental ideas which are dealt with by Dandin. The stress which Dandin puts on the theory of Rīti (which he calls Märga) is carried to its furthest extreme by Vämana, who elevates Rīti to the rank of the very essence of poetry. While Dandin mentions two types of Mārga, Vāmana adds an intermediate third Rīti; and from Mammata ix. 4 we learn that Vämana was the first to suggest this threefold division.2 Again, while Bhamaha and Dandin apparently engage in a controversy over the classification of Kāvya into Kathā and Ākhyāyikā, Vāmana peremptorily brushes aside all discussion and refers the curious reader to the wor ks of "others."3 Dandin is also anxious to show, in the course of a long digression, that the word iva is indicative of utprekşā (which figure itself is admitted by Bhāmaha ii. 88 only in

1 viz., those defining asadhūraņopamā, asambhavopamā, anu- śayākșepa, viśeşokti, hetu and atisayokti respectively. 2 It is noteworthy also that Dandin is unaware of the more or less technical term riti, made so familiar by Vāmana, but uses the almost synonymous expression marga, also used by Vamana in iii. 1. 12.

60 SANSKRIT POETICS

deference to the views of Medhavin) ; but to Vāmana (iv. 3. 9, vrtti) it is already an established fact. Such instances can be easily multiplied, but what is given here will be enough to indicate Dandin's priority to Vämana,1 and fix the lower limit of his date at the end of the 8th and the commencement of the 9th century2.

1 It is supposed by Kielhorn (with whom Peterson in his pref. to Daśakumāra agrees) that Dandin ii. 51, in which some of the upamā-doșas are justified, is directed against Vāmana iv. 2. 8f, im- plying thereby that Dandin is later than Vämana. But if we take the texts of Bhamaha, Dandin and Vamana together on this point, we can only make out the following facts. Bhämaha, in accordance with the opinion of Medhävin, brings forward (ii. 39-40) seven upamā-doșas, viz., deficiency (hīnatva), impossibility (asambhava), dis- parity of gender (linga-bheda), disparity of number (vacana-bheda), contrariety (viparyaya), excess (adhikatva) and non-sımilitude (asādrśya.) Dandin, tacitly asuming these, only remarks about two pairs of them (viz. disparity of gender and number, excess and deficiency) that they do not necessarily disturb comparison if they do not wound the cultivated sensibility. In this he is substantially following. Bhämaha, who says generally that the upameya cannot in every res- pect be similar to the upamāna (ii. 43), a dictum which is implicitly accepted by most later writers, who define upamā as bhedābheda- pradhane upama. Therefore, deficiency etc. become faults only when they disturb the sense of appreciation of the man of taste. Vämana, on the other hand, mentions six upama-dosas instead of seven, including viparyaya in adhikatva and hīnatva (iv. 2. 11 vrtti), with the final pointed remark : ata evāsmūkam mate sad dosāh. It appears, therefore, that Dandin ii. 51f is a link in the chain bet- ween Bhamaha ii. 39f and Vamana iv. 2. 8f 2 Pischel's argument (pref. to Srng. til.) that Dandin is identical with the author of Mrcchakatika on the ground that Dandin ii. 362 (st. limpafiva, ed. Bibl. Indica) occurs also in that drama (ed. N.S.P. 1916, i. 34) lands us, apart from other objections, in the absurdity of identifying Dandin with Bhäsa as well, inasmuch as the same verse is also found in the so-called Bhasa-damas, Carudatta (i. 19) and Bāla- carita (i. 15). The attribution, again, of the same verse in Sārngadhara 3603 and Vallabhadeva 1890 to Bhartrmeņtha and Vikramāditya further discredits Pischel's theory. The occurrence of the verse in

DAŅŅIN 61

(2) The upper limit is not so easy to settle. Peterson, follow- ing Maheśacandra Nyāyaratna, points out1 that Dandin ii. 197 is a reminiscence of a passage in Bāņa's Kādambarī p. 102, 1. 16. (ed. BSS), and Jacobi is inclined to accept this view. Bāna lived about 606-647 A. D. in the reign of king Harsa, whose biographer he was. Jacobi also points out a resem- blance between Dandin ii. 302 and Mägha ii. 4. Pathak, again, remarks2 that Dandin's threefold classification of karman into nirvartya, vikārya and prāpya (ii. 240) is taken from Bhartrhari's Vākyapadīya iii. 45f. Bhartrhari, according to Yi-tsing died about 651 A.D., while Magha probably belonged to the second half to the 7th century3. Thus Bana, Bhartrhari

on this point) only shows that Dandin did not disdain to borrow well-known verses for purposes of illustration and criticism, as he himself admits in a general way in i. 2. It should also be noted that in the Bibl. Indica ed. of the text, the verse is given twice (1) as a half-verse quotation in ii. 226 and (2) in full ii. 362. But this reading, on which apparently Pischel's theory was based, is doubtful, and is contrary to readings in other MSS. In the Tibetan version of the text (JRAS, 1903), as well as in the Madras edition, the verse occurs only once as a half-verse quotation in ii. 226, the full verse being omitted in the text and given in the Madras ed. only in the accompanying commentary. Pischel is hardly accurate in stating that Pratītārendu attributes this verse to Dandin; for the commentator, in the course of his discussion on utpreksā simply says (p. 26) that Dandin has already discussed at great length that the verse limptiva is an illustration of utpreksā containing afīśaya. 1 Pref. to Daśakumara°, new ed. 1919, p. ix. Other such re- miniscences are presumed in Dandin i. 45 (= Sakuntalā i. 20, ed. M. Williams; Cf. JRAS, 1905 p. 841f), ii. 286 (=Raghu viii. 57), iì. 129 (=Sakuntalā i. 26) etc. Taruņavācaspati is of opinion (on i. 2) that Dandin consulted the usages of poets like Kālidāsa. See other parallel passages' collected together by Agashe (preface to Dašakumūra, pp. liv f). 2 1A xli, 1912 p. 237. 3 See Kielhorn in GN, 1906, p. 143-46. Cf Māgha ii. 83 where he shows himself fully conversant with Poetics; also ii. 8, 86, 87,

62 SANSKRIT POETICS

and Mägha probably all belong the same age and flourished in the first half or the middle of the 7th century. These evidences, although suggestive, do not in their nature appear to be decisive; and we are ultimately thrown upon the question of Dandin's relation to Bhämaha, which might support these evidences and with reference to which indeed the chronology of Dandin should be settled. If Bhamaha's. priority to Dandin can be definitely established, then we arrive with this at a more or less satisfactory limit to the date of the latter. The question is, no doubt, beset with many difficulties ; but so far as a comparative study of their res- pective texts indicates, the presumption is strong in favour of Bhamaha's priority ; because, while Dandin criticises Bha- maha's innovations, Bhamaha apparently never does so in cases of Dandin's innovations which are indeed much more numerous. The materials for such a critical study (apart from a consideration of their general theories) consist of several passage, occurring in their respective texts, which are either (1) identical or very similar in phraseology, or (2) so- closely related to each other that the one author appears to be criticising the other. As the question has already engaged a great deal of controversy1, which has thrashed out almost all the details we will here discuss it very briefly. As instances. of the first group of passages, we may cite Bh i. 20ab and D i. 7cd ; Bh. i. 17cd and D i. 29ab (definition of Mahākāvya) ; Bh ii. 66ab and D ii. 4cd (enumertion of certain Alamkäras) ;

xiii. 69, xiv. 50, xix. 37, xx. 44 (where he refers to Bharata) .- Om Mägha's date, see S. K. De, Hist. Skt. Lit. Calcutta 1942, pp. 88- 89 and references contained therein. 1 M T. Narasimhiengar in JRAS. 1905, pp. 53f; K. B. Pathak in JBRAS xxiii, p. 19: R. Narasimhachar in IA xli, 1912, p. 90; p. 232; Trivedi, introd. to Pratāparuara p. 32 and IA, xlii, 1913, p. 258-74; H. Jacobi in ZDMG, lxiv, p. 134, in SBAW, xxiv, 1922 (Bhämaha und Dandin. ihr Alter etc), p. 210-226. and xxxi, 1928 (Zur Frühgeschichte d mnd. Poetik) ; J. Nobel in ZDMG, lxxiii, 1919, p. 190f and his Beitraege zur aelteren Geschichte des Alamkārasāstra, Berlin 1911, p. 78 ; P. V. Kane in HSP. pp. 96-108, etc.

DAŅŅIN 63

Bh ii. 87ab and D ii. 244ab (illustration of Vārtā) ; Bh iii.1ab and D ii. 5cd (enumeration of certain Alamkāras) ; Bh iii. 53 and D iii. 363 (the figure Bhavika) ; Bh iii. 5 and D ii. 276 (illustration of Preyas) ; Bh iv. 1-2 and D iv. 2-3 (enumeration of Dosas) ; Bh iv. 8ab and D iv. 5ab (definition of Apartha). The verbal coincidence in these passages is so striking that there can be no doubt that it should be taken as something more than merely accidental. It does not, however, preclude the possibility of their being taken from a common source, or being standardised definitions or enumerations common enough in such technical treatises. The second group of passages, also betraying enough verbal similarity, is more interesting and important ; because they certainly express contradictory views of their respective authors, if not actually meant as direct mutual criticism. In two of these passages, Bhämaha and Dandin are, each in his turn, rejecting an illustration which is adduced by the other, but both citing the illustration in question in exactly identical phraseology. Thus, Bhāmaha cites and rejects (ii. 87), after the figure hetu, the illustration gato'stam arko bhātīndur yāntī vāsāya paksinah, characterising it as bad poetry, to which. he says, some writers give the name of vārtā. Dandin does not mention vārtā, but approvingly cites (ii. 244) the same half-verse under the figure hetu, pointedly remarking that the illustration under discussion is good. Similarly, the half- verse himāpahāmitra-dharair is given as an instance of the fault avācaka by Bhāmaha (i. 41), but Dandin gives the verse in full in a different context (iii. 120) as an example of a variety of prahelikā; Bhāmaha apparently condemning it as faulty, while Dandin taking it as a piece of ingenious con- struction. Taking the examples in their contexts as quoted from a common source, the passages apparently indicate that Dandin is not in agreement with Bhämaha (who condemns these) but expressly justifies their propriety. A closer contact of views and similarity of expression are to be found in those passages in this group, which relate to

64 SANSKRIT POETICS

. (1) the discussion of the comparative merits of the gaudīya and vaidarbha mārgas (Bh i. 31-35 and D i. 40 f) (2) the distinction between prose kathā and ākhyāyikā (Bh i. 25f and D i. 23f), and (3) the enumeration of the ten dosas (Bh iy. 1-2 and D iv. 2-4) ; and those who maintain Dandin's priority to Bhämaha hold that in these cases the latter is undoubtedly criticising the former. In the first of these instances, Bha- maha's remarks merely show that he is more or less indiffer- ent to the literary value of Märga or Rīti as modes of com- position, and laughs at the distinction which some writers make between gauda and vaidarbha types, himself giving preference, if any, to the former. In his opinion, as he says in the next verse (i. 36), what is important in poetry is not Rīti but Vakrokti. It appears that the view which Bhamaha is criticising was traditional or referred to as a matter of common controversy, as he himself says in this connexion: gatānugatika-nyāyān nānākhyeyam amedhasām. Jacobi points out that the Gauda Märga, long before Dandin, could never establish for itself a good reputation, and Bāna had already condemned it as akşara-dambara (Harşa-carita i. 7). Dandin, on the other hand, attaches great importance to Riti in poetry which under the name mārga, occupies a considerable part of his treatment; but he acknowledges, in spite of his own distinguishing of two such extreme types as gauda and vaidarbha, that there are other intermediate modes finely differentiated (i. 40), and that the types admitted by him are not capable of exact definition (i 101f), although he himself prefers the vaidarbha. It may also be added that Bhamaha is unaware of the peculiar analysis of mārga given by Dandin with reference to the ten essential gunas, but he mentions casually (and not in connexion with Rīti) only three gunas which may be admitted in all good composition. The respective characterisation, again, of the two Rītis has hardly any point of contact, and Bhämaha's remarks, if supposed to be levelled against Dandin in particular, are certainly off the mark; for each of them approaches the

DAŅŅIN 65

subject from the standpoints of entirely different schools of opinion1. Similar remarks apply to the other two cases, in which one hardly finds any direct reference by Bhamaha to Dandin. Dandin does not accept as characteristic or essential those marks of distinction between a Kathā and an Ākhyāyikā which Bhamaha enumerates, and apparently quotes in this connexion the half-verse i. 29ab from Bhamaha i. 27cd. The distinction, denied by Dandin, is admitted (along with Bhämaha) by earl- ier as well as later writers; for Bāņa designates his Harsa- carita as an akhyāyikā and his Kādambarī as a kathā2, and some such distinction is also implied by the Amara-kośa. With regard to the other passages which enumerate the Dosas, it appears that Bhamaha, following the traditional recounting of ten orthodox Dosas (cf Bharata xvi. 84)3 mentions the same number of Doşas, but adds that pratijnā-hetu-drstānta-hīnatva is not desirable in poetry (iv. 2). At the same time, this ele- venth defect is interesting to him from the standpoint of logical exposition, and he deals with it in v. 1f,4 apparently considering that defective logic is also to be looked upon generally as a notable flaw in a composition. Dandin enume- rates the same ten Dosas (iv. 2-3) in exactly similar phraseo- logy, and conservatively maintains the view (iv. 4) that the so-called eleventh fault is difficult to judge and unprofitable to discuss5. 1 This point will be discussed in detail below in vol. ii. 2 See Taruņavācaspati's remarks on Dandin i. 25. 3 The definitions, however, do not agree. See Jacobi op. cit, p. 222f. 4 For Bhämaha's exposition of logic in relation to other philo- sophical writers see K. B. Pathak in ABORI, xii. pp. 372-87. 5 Emphasis is also put on some verbal resemblance between Bh i. 22 and D i. 21-22. In these passages, however, the standpoints of the two theorists are quite distinct, although they use somewhat similar phrases. Bhämaha here expresses his disapproval of a dis- astrous ending, perhaps in conformity with a similar conventional prohibition in the drama. Dandin, on the other hand, takes the ulti- mate triumph of the hero for granted, and does not trouble him- 5

66 SANSKRIT POETICS

From the above discussion, che conclusion is very probable: that Dandin was familiar with the text of Bhamaha whom, as a notable predecessor expressing contrary views, he could hardly ignore, On this point we have the almost unanimous testimony of Dandin's commentators1, who expressly state that in most of these disputed passages Dandin controverts. the earlier opinions of Bhamaha. It will not be necessary,. therefore, to enter here into the details of their respective theories, which not only indicate some fundamental and im- portant differences, as one should expect in writers belonging to two different schools of opinion, but also the fact that. Dandin, in dealing with most of the topics, has gone into greater details and finer distinctions, apparently betokening that in his age the study was more advanced and fraught with. greater complexity than in that of Bhamaha2.

self about the admittedly forbidden tragic ending. He appears to express the view that it is artistically more effective if the rival of the hero is set forth at the outset in all his glory and then his downfall is secured through the superior virtue of the hero himself. One does not, also find any point in Bhämaha ii. 37-38, which cri- ticises the classification of the figure upamā into many subvarieties (like mindo°, prasamso° and ācıkhyāso°) but which is taker by some to imply an attack on Dandin's elaborate subdivision of the same figure. The tri-prukaratvam cannot possibly refer to Dandin, who mentions not three but thirty-two subvarieties : while nindo°, praśamso° etc. are also mentioned by Bharata. 1 é. g. Taruņavācaspati on i. 23-24, 29; ii. 235, 237, 358 ; iv. 4 etc ; Harinātha on i. 15 (cited in ABod 206b); Vādijanghāla on i. 21. 2 See, for instance, their respective views on rīti, guna and dosa, on alamkāra (which last element Dandin does not distinguish funda -- mentally from gunas, ii. 3.), on vakrokii (Bh ii. 85 and D ii. 362); their respective order of treatment of alamkaras (which Bhamaha. deals with in successive groups, while Dandin's thirty-five indepen- dent poetic figures are given as if they are well recognised); Dandin's minute and fine differentiation of infinitel sub-species of indi -. vidual figures ; their respective treatment, of yamaka, upamā, utprekșā, ananvaya and sasamdeha, upamā-rūpaka and utpreksāvayava (which last four Dandin does not accept as independent figures) etc. These points will be discussed in detail in the next volume.

DAŅŅIN 67

If this conclusion of Bhamaha's probable priority is accepted, then we get his date as the upper limit to that of Dandin, the lower limit being, as already discussed, the same as that of Bhamaha, namely the date of Udbhata's contem- porary, Vāmana. Dandin, therefore, flourished probably in the first half of the 8th century.1

(3) There cannot be any doubt that Dandin, like Bhāmaha, must have been indebted to his predecessors ; and if he does not mention any one of them by name, he gives enough evi- dence of his having utilised their works, including that of Bhämaha. Dandin, however, makes a general acknowledg- ment in i. 2, and refers to the opinions of "others" and of "learned men" (e.g. i. 9, 10 ; ii. 2, 7, 9, 54; iii. 106) ; while he makes no secret of his having "observed" and probably bor- rowed his illustrations from earlier poets, to whom reference2 is made in i. 30,100 ; ii. 65, 223, 225, 363 ; iv.7, 32, 42, 57. The Hrdayangama commentary on i. 2 mentions in particular two authors, named Kāśyapa and Vararuci (vi. 2 ; ii. 7), whose works Dandin is supposed to have utilised. Similarly in Vādijanghāla's Śrutānupālinī commentary Kāśyapa, Brahma- datta and Nandisvāmin are spoken of as Dandin's predecess- ors. These may be mythical or traditional names; but Kāśyapa is also mentioned by another admirer of Daņdin's who composed the Sinhalese rhetorical work already referred to. He is known to Abhinavagupta as a Muni who preceded Bharata, and his opinion on Rägas is cited in Abhinava- bhāratī. Kallinātha on Samgīta-ratnākara (ii. 2. 31) quotes

1 The question whether our Dandin is identical with Dandin, author of Dasakumara-carita, does not concern us here; on this see S. K. De, History of Sansk. Lit. Calcutta 1947, pp. 207-9. Of his personal history nothing is known, unless we hold that the two Dandins are identical and are prepared to accept the Avantisundari- kathā (ed. M. R. Kavi, Madras 1924) as a work ol Dandin. See S. K. De, Aspects of Skt. Lit., Calcutta 1959, pp. 296-308. 2 See this point discussed in Agashe op. cit. pp. liii f.

68 SANSKRIT POETICS

three verses of Kasyapa . and among old authorities on music he is mentioned by Nänyadeva (11th-12th century) who mentions also Brhat-Kāśyapa and Vrddha-Kāśyapa. The Pañca-sayaka, on the other hand, cites him (iv. 19) as an authority on Erotics, and the Agni-purana as an authority on Metrics. One Kāśyapa is cited by Pāņini in viii. 4. 67, and a grammarian Kāśyapa, as Aufrecht notes, is quoted by Mādhava. Pischel1 has already negatived the suggestions of Prema- chandra Tarkavagish2, Peterson3 and Jacobi4 that Dandin in i. 12 refers by the word chando-viciti to a treatise of his own, so named, on the subject of prosody. The word, however as indicated by Dandin himself in the same verse, by his reference to it as sā vidyā, does not necessarily mean any particular treatise but the science of prosody in general5 ; for which, in addition to the references given by Pischel, one need only cite Kautilya's Artha-śāstra (i, 3.1) and Āpastamba Dharma-sūtra ii. 4. 8, where the word chando-viciti occurs ; also Rājaśekhara p. 6 and Hemacandra, Comm. p. 5. In iv. 49, again, Dandin refers to a kalā-pariccheda, which Peterson takes to be a clear reference to another work of Dandin's ; but the reference is more likely to an additional or supple- mentary chapter to his Kāvyādarśa, as Taruņavācaspati suggests (p. 282). It is noteworthy that the Kāmadhenu com- mentary on Vamana similarly quotes from a lost work or chapter of Bhamaha's on the kalās6. Daņdin's Kāvyādarśa consists three Paricchedas or chapters (four in M. Rangacharya's ed.) and about 660 verses. The topics comprehended are: I. Definition and division of Kāvya; the two Mārgas (Vaidarbha and Gauda) and ten Gunas

1 Pref. to Srng. til. p. 14 f. 2 On Dandin 1. 12. 3 Introd. to Daśakumāra p. ix-x. 4 Ind. Stud. xvii p. 447. 5 See P. V. Kane in IA, 1911, p. 177. 6 See above p. 57 footnote.

DAŅDIN 69

pertaining to them ; the essential requirements of a good poet (Pratibhā, Śruta and Abhiyoga). II. Definition of Alamkāra, and enumeration and description of 35 Arthālamkāras viz. svabhāvokti, upamā, rūpaka, dīpaka, āvṛtti, ākşepa, arthā- ntara-nyāsa, vyatireka, vidhāvanā, samāsokti, atiśayokti, utpreksā, hetu, sūkşma, leśa or lava, yathāsamkhya or krama, preyas, rasavat, ūrjasvi, paryāyokta, samāhita, udātta, apah- nuti, śleşa, viśeşokti, tulyayogitā, virodha, aprastuta-praśamsā, vyājokti, nidarśanā, sahokti, parivṛtti, āsīḥ, samkīrņa and bhāvika. III. Elaborate treatment of Śabdālamkāras, namely yamaka, citra-bandha and 16 varieties of prahelikā; ten Doşas (in ch. iv in Rangacharya's ed,).

(4) Commentators On Dandin The commentaries on Dandin, as the following Bibliogra- phy will show, are numerous. Most of these are compara- tively modern, excepting perhaps that of Tarunavācaspati as well as the anonymous commentary called Hrdayangama, both printed in the Madras edition. With this exception, they are hardly useful for an historical or critical study of Daņdin.

BIBLIOGRAPHY Editions. (1) Bibl. Ind. 1863, by Premachandra Tarkavagish with his own commentary, reprinted by Bhavadeva Chat- topodyaya, Calcutta 1881. (2) by Jivananda Vidyasagar 1882 ètc. (3) edited and translated by O. Böhtlingk, Leip- zig 1890. (4) with two comms. (one of Taruņavācaspati and the other entitled Hrdayangama) by M. Rangacharya, Madras 1910. (5) by S. K. Belvalkar and Rangacharya B. Raddi with Sansk. comm. and English notes, BSS. Bombay pt. i. (1919). and pt. ii. (1920). (6) Sansk. text and Eng. trs. by S. K. Belvalkar, Poona 1924 (Advance issue, ch. i. and ii in 1920). (7) Sansk and Tibetan text, ed. Anukul Chandra Banerji, Cal. Univ. 1939. (9) ed. with Ratnaśrī

70 SANSKRIT POETICS

commentary of Ratnaśrijñana who follows the Tibetan version of the text, by Anantalal Thakur and U. Jha. Mithila Institute, Darbhanga 1957. (8) ed. D. T. Tatācharya with Hrdayangama, and comms. of Vādijanghala and Taru- ņavācaspati, Bombay(?) 1941. (10) ed. V. Krishnamachari, with the commentary of Vadijangāla, Srinivasa Press, Tiruvadi 1936. Our references are to the Madras edition of M. Rangacharya unless otherwise indicated: this edition divides the work into four chapters instead of three. Commentaries. (1) by Taruņavācaspati, Madras Cat. xii, 12834. Printed as noted above. It appears to be an old comm. But as it cites (on i. 40) Bhoja ii. 28, as well Daśarūpaka i. 8 (in comm. on i. 31), it cannot be placed very early. This commentary also refers (on i. 30) to a poet Hastimalla who may be the Jaina Hastimallasena who wrote dramas and poems. Keśava Bhatțāraka, son of Taruņavācaspati, was Guru of Mahārājādhirāja Rāma- nātha (the Hoysala Vīra Rāmanātha) who came to the throne in 1255 AD. See V. Raghavan in JOR, xiii. at p. 305. So Tarunavăcaspati's date would be the end of the 12th and first half of the 13th century. In the colophon to some MSS he is called 'Sadhu'. (2) Hrdayangama by an anonymous author. Madras Cat. 12833. Printed as above (ch. i-ii only). The comments on Dandin ii. 284, 286, 288, 289-91 are quoted verbatim but anonymously by Bhoja in his Srngāra-prakāśa (ch. xi). (3) Mārjanā by Harinātha, son of Viśvadhara and younger brother of Keśava. ABod 206b ; Peterson vi p. 30 (extract). MS in BORI (Catalogue xii, no. 124) copied in Samvat 1746=1690 A.D. Hārinātha also wrote a comm. on Bhoja's Sarasvatī-k.° He must be later than Keśava Miśra whose work on Alamkāra he cites, (4) Muktāvalī by Narasimha-sūri, son of Gadādhara and grandson of Krsna-śarman. Mitra 2394 (Aufrecht i. 102b).

DANIN 71

(5) Candrikā by Triśaraņatațabhīma. Mentioned in Hall's Index, p. 63. (6) Rasika-rañjanī by Viśvanātha. Oppert 4112 (Aufrecht i. 103a). (7) Vivrti or Kāvya-tattva-viveka-kaumudī by Krsņa- kiņkara Tarkavāgīśa Bhatțācārya of Gopālapura in Bengal. IOC pt. iii, no. 1128/1497, p. 321. (8) Śrutānupālinī by Vādijanghāla (or °ghanghala). Printed as noted above. Stein pp. 61, xxviii, extract no. 1179. The Report of Peripatetic Party of Madras MSS Lib. 1917-19 mentions an acquisition of this comm. for the Library. MS in BORI, Cat. xii, no. 125. The comm. mentions Kāśyapa, Brahmadatta and Nandisvāmin as predecessors of Dandin. (9) Comm. by Bhagiratha. Aufrecht i. 102b, (10) Comm. by Vijayānanda. MS in BORI, Cat. xii, no. 123 (incomplete). (11) Vaimalya-vidhāyinī by Mallinātha, son of Jagan- nätha. Aufrecht ii. 20a. This is perhaps the same Mallinatha as is referred to by Viśveśvara in Alamkaus°, p. 69 as a commentator on Kāvyādarśa; and he should be distinguised from the better known Kolācala Mallinātha. (12) Comm. (incomplete) by Tribhuvanacandra, otherwise called Vadisimha, a Jaina. HPS iii, no. 57. (13) Comm. by Yamuna or Yāmuneya. MSS in BORI, Cat. xii, no. 126. It is probably a South Indian work which divides the Kāvyādarśa into 4 chapters, as in Madras ed. of Rangacharya. (14) Ratnaśrī by Ratnaśrījāna. Ed. as descibed above. The author was a Ceylonese monk who wrote under the patronage of some Rāstrakuța king, named Tunga, under the overlordsip of Rājyapāla of Gauda and Magadha (c. 908 A. D.). Authors quoted, besides Aśvaghoșa and Kālidāsa, are Mātrceța, Āryaśūra, Kohala, Rāma-śarman, Medhāvirudra, Kambala, Harivrddha, Bhāmaha, Bhar-

72 SANSKRIT POETICS

trmeņtha, Guņādhya, Cāndra vyākaraņa, Mallanāga, and Dharmakīrti. (15) Anoymous Comms. in Mitra 297, Oppert 7903; . SCC vii 21. A comm. by Dharmavācaspati in Oppert 2581 is probably a mistake for Taruņavācaspati. Regnaud (Rhétorique, p. 367 fn) also mentions a commentary by Väcaspati and refers to Taylor ii 501 ; which work pro- bably refers also to this commentary. UDBHAȚA

(1 ) Udbhata, who wrote a commentary named Bhāmaha- vivarana or °vivrtil on Bhamaha, as well as utilised the latter's work2 in his Kāvyālamkāra-samgraha, certainly lived before the final exposition of the Dhvani-theory by Ananda- vardhana3 who, in the middle of the 9th century, actually cites Bhatta Udbhata twice at pp. 96, 108. Udbhata's name indicates that he was probably a Kashmirian. Kahlaņa (iv. 495) mentions a certain Bhatta Udbhata who was a sabhāpati of king Jayāpīda of Kashmir (about 779-813 A. D.); and Bühler. to whom we owe the discovery of Udbhata's work in Kashmir, identifies him with the author of the Kāvyālamkāra- samgraha (or °sara-samgraha). Accepting this identification, 1 The Bhamaha-vivarana is frequently mentioned and cited by later writers; e.g. Abhinavagupta in his Locuna (p. 10, 40, 134, 159), Hemachandra (Comm. p. 17, 110), Manikyacandra (Samketa, ed. Mysore p. 289), Samudrabandha (on Alam. sarvasva, p. 89). Pratī- hārendurāja (p. 13) etc. 2 Udbhata secms to have taken over almost verbatim from Bhāmaha the definitions of āksepa, vibhāvanā, atiśayokti, kathāsam- khya, utprekşā paryūyokta, apahnuti, virodha, aprastuta-praśamsā sahokti, sasamdeha and ananvaya. The poetic figures are enumerated almost in the same order as that of Bhamaha. But Udbhata omits a few Alamkāras defined by Bhāmaha (e. g. vamaka, upamā-rūpaka, utpreksvayava) and adds a few not defined by the latter (e. g. punar- uktavad-ābhūsa, karyalinga, drstānta and samkara). 3 Cf the opinions of Pratīharendurāja (p. 79), Ruyyaka and Jayaratha (p. 3) and Jagannātha (pp. 414-5). 4 Kashmir Rep p. 65.

UDBHATA 73

we should,. however, place the most flourishing period of Udbhata's activity, as Jacobi points out, in the first part of Jayāpīda's rather long reign; because this sovereign in the latter part of his career appears to have alienated the Brah- mans by his oppression of the people. Udbhata, therefore, should be assigned to the end of the 8th century, and he may have lived into the beginning of the 9th. Besides the lost Bhāmaha-vivaraņa, Pratīhārendurāja tells us (p. 15) that Udbhata wrote a poem. called Kumāra- sambhava, from which are taken most of the illustrations in the text. We have already mentioned above that Udbhata probably wrote also a commentary on Bharata's Nātya-śāstra, but it is no longer available. The followers of Udbhata or the Audbhatas are mentioned by Abhinavagupta in his com- mentary on Bharata. Udbhața's Kāvyālamkāra-samgraha consists of six chapters. (called Vargas), contains 75 Kārikās in Anusțubh with 95 illustrations, and deals with 41 Alamkāras. Some verses are assigned to Udbhata in the Subhāșitāvalī 498, 1463, 3453, of which no. 498 is a well known gnomic stanza found also in such collections as the Cānakya-sataka. Although closely following Bhamaha in the treatment of Alamkāras, Udbhata has certain views peculiar to himself, which are either absent in Bhämaha or in which he differs from his predecessor. For instance, Bhāmaha speaks of three kinds of lesa while Udbhata mentions two kinds, and the basis of classification is different; Udbhata's three Vrttis, on which the classifica- tion of Anuprāsa proceeds, are absent in Bhāmaha )

( 2 ) MUKULA AND PRATIHĀRENDURĀJA Pratīhārendurāja, Udbhata's commentator, was, as he himself tells us, a native of Konkana and a pupil of Mukula. 1 For more instances see P. V. Kane, HSP, pp. 127-28 .- For a résumé of the topics covered by Udbhata's work see vol. ii. ch. ii(2).

74 SANSKRIT POETICS

Mukula is known to us as the author of Abhidhā-vrtti- mātrkā, a work on the grammatico-rhetorical question of Abhidhā and Laksaņa, consisting of 15 Kārikās with Vrtti. From the last verse of this work we learn that the author's father was Bhatta Kallata who lived, according to Kahlana v. 66, in the reign of Avantivarman of Kashmir (855-884 A.D.)1 and was, therefore, a contemporary of Ratnākara and Anandavardhana. Accepting Kahlana's statement, Mukula should be placed roughly towards the end of the 9th century and the beginning of the 10th. His pupil Pratīhārendurāja, therefore, belongs approximately to the first half of the 10th century.2 In his commentary on Udbhata. ralled °Laghu-vrtti, Pratīhārendu quotes from Bhāmaha, Daņdin, Vămana, the Dhvanyāloka and Rudrata, actually naming most of them, and appears to be fairly familiar with the Dhvani-theory, as explained by Anandavardhana, to which, however, he does not subscribe. Peterson appears to suggest3 the identification of Pratī- hārenduraja with Bhattenduraja, whom Abhinavagupta refers to as asmad-upādhyāya in his Locana (pp. 25, 43, 116, 160, 207, 223) as well as in his commentary on Bharata, where this teacher is quoted sometimes simply as upādhyāya. At the commencement and close of his Locana,4 Abhinavagupta indicates his immense indebtedness to this teacher, and in one place (p. 160) we find in his praise the somewhat grandilo- quent epithet vidvat-kavi-sahrdaya-cakravartin; which toge- ther with the fact that Abhinava also indicates that he learnt Kāvya from Bhattenduraja will go to support the conjecture

1 Bühler op. cit. pp. 66, 78 2 Cl. Pischel, Pref. to Śrñgtil. p. 12. 3 Introd. to Subhāș°, p. 11, but contra in Aufrecht i 59a. Banhatti's more recent advocacy of identity is hardly based on sub- stantial and convincing reasons. The only instance where the two Indurājas are confused or identified occurs in Samudrabandha p. 132; but this is no strong evidence. 4 ed. Kävyamala p. 1; and ch. iv in JDL. 1922, p. 42 (reprinted m S. K. De, Some Problems p. 245 f).

MUKULA AND PRATIHĀRENDURĀJA 75

that this preceptor was apparently well versed in both the theory and practice of poetry. Although chronology does not stand in the way, there are several reasons which might induce one to distinguish the two Indurajas. From Abhinava's commentary on the Bhagavad-gita1, we learn that Bhattendu was son of Śrībhūtiraja and grandson of Saucuka of the Kātyāyana gotra; but of Pratīhārendu's genealogy or personal history we know nothing, except that he was a Kauńkaņa and a pupil of Mukula.2 Bhattendu appears chiefly as a poet who wrote, as Abhinava's quotations show, in Sanskrit and Prakrit, and whose verses apparently supplied a ready source of apt poetic illustrations to his pupil's works, probably inspired by himself. If some of his opinions on Rasa and allied topics are quoted by Abhinava in his com- mentary on Bharata, they bear no kinship to Pratīhārendu's views, as expressed in the latter's commentary on Udbhata. Although the prefixes Bhatta and Pratīhara, being mere honorific titles, need not make any serious difference, Abhi- nava's citation of his teacher always as Bhattendurāja (and never as Pratīhārendurāja) is somewhat remarkable; and in view bf the fact that these two Indurājas were probably contemporaries, might this not indicate that Abhinava meant to imply a difference? The two writers are never confused even in later anthologies, for the poet it always designated Indurāja3 or Bhattendurāja. The conjecture, therefore, is not unlikely that Abhinava's teacher may have been the poet Bhattendurāja, who is quoted under this designation in

1 Bühler, op. cit. pp. 80 and cxlxii-viri. 2 It is curious that Abhinava, who takes care to refer to most of his teachers and "teacher of teachers" (parama-guru) should have omitted a reference to Mukula, whose work, if he was a parama- guru, should have been important to him, inasmuch as it dealt with the quasi-rhetorical question of the functions of word and its sense. 3 Two verses of Induraja are cited at 287 and 306 (Kavi's ed. of Nātya-śāstra.)

76 SANSKRIT POETICS

Ksemendra's two works1 as well as in the poetical anthologies2 of Särngadhara, Vallabhadeva and Jahlaņa. The commenta- tor Pratīharenduraja, on the other hand, was never known for his poetical pretensions, and was chiefly a writer on Poet- ics, who obviously belonged in his views to the older system of Udbhata and did not, as Abhinava did, believe in newly established doctrine of Dhvani, with which, however, he ap- pears to be fully conversant. Referring to this new theory of Anandavardhana, Pratīharendu states in one place (p. 79) that what is known as Dhvani and taken to be "the soul" of Poetry by some thinkers is included implicitly by his author Udbhata in the treatment of some of the poetic figures under discussion, and therefore need not be separately considered. Bhattenduraja, on the other hand, appears to have favoured the new theory of Dhvani ; for Abhinavagupta (Locana p. 2) tells us that this teacher of his explained to him the Mangala- śloka of Ananda's Vrtti in the light of the Dhvani-theory. The standpoints of Pratīhārendurāja and Abhinava are so divergent that it is difficult to admit any spiritual relationship between the two ; for the former was in no way an adherent. of the Dhvani-theory, of which Abhinava was a recognised advocate.

( 3 )

RĀJĀNAKA TILAKA Jayaratha, commenting on Ruyyaka's (or Rucaka's) Alamkāra-sarvasva, refers in more than one place (ed. NSP 1893, pp. 15, 124, 205) to an Udbhata-viveka or Udbhata-

1 Aucit, vic. under sI. 25, 31 ; Suvrtta-til° under śl. 2, 24, 29, 30. 2 Subhūşitvalī 918. The verse parārthe yah pīllām, ascribed to Induraja in the Paddhati of Sarngadhara (1052), is quoted anony- mously twice by Anandavardhana (pp. 53, 218), a fact which, how- ever, is not decisive ; because Abhinava's commentary is silent as to the authorship of this verse, which occurs in Bhallata-śataka 56 and is ascribed to another poet Yasas in the Subhas° 947. The Sadukti- karņāmrta attributes it to Vākpati, while Hemachandra (Comm. p. 257) and Jayaratha (p. 108) cite it anonymously.

RĀJĀNAKA TILAKA 77

vicāra by Rājānaka Tilaka, and states that Ruyyaka generally followed the views propounded by Tilaka. We know that a Rājānaka Tilaka was Ruyyaka's father. An anonymous commentary, entitled Vivrti, has been published in the Gaekwad's Series (see below under Bibliography)1, along with the text of Udbhata commented upon. It has been claimed by the editor of this publication that the Vivrti is identical with the Viveka or Vicara cited by Jayaratha. Attention was drawn to this commentary in a notice of its unique MS, existing in the Madras Govt Oriental MS Library by the present writer2, as well as by Banhatti about the same time in his edition cf Udbhata's work. Banhatti is rather cautious in his discussion of the question of identity, but he appears to be inclined to the view that the Vivrti is "evidently a distinct work from the Udbhata-viveka or -vicāra of Rājānaka Tilaka mentioned by Jayaratha." The editor of the Vivrti, however, makes an elaborate attempt to demonstrate that the Vivrti cannot but be taken as the lost Viveka or Vicāra of Tilaka3. The evidence adduced is plausible indeed, but does not appear to be conclusive, and in the absence of more definite data it would be better to leave the question open. That this anonymous commentary is late is evidenced not only by its content but also by its citation of Rājaśekhara's Viddhaśāla- bhañjikā, as well as by its obvious appropriation of Mam- mata's standard work. Its value as an exegetical work cannot be placed too high.

BIBILOGRAPHY

Udbhata Editions. (1) Text ed. in Roman characters by G. A. Jacob in JRAS, 1897, pp. 829-53 ; (2) Text with Pratīhārendurāja's comm. by M. R. Telang, Nir. Sag. Press 1905, 1915 (this edition is useful for its comm., but some verses occurring

1 See a review by the present writer in JRAS, 1934, pp. 173-74. 2 In BSOS, iv, 1926, p. 279. 3 P. V. Kane (HSP p. 130) agrees with this view.

78 SANSKRIT POETICS

in the comm. are given mistakenly as Kārikā-verses) ; (3) ed. N. D. Banhatti, with the comm, of Pratīhārendūrāja, Bombay Skt. Series, Poona 1925. Our references are to Telang's edition, unless otherwise indicated. Commentaries. (i) comm. by Pratīhārendurāja (as printed in the above editions), (ii) an anonymous commentary called Vivrti (ascribed by the editor to Rājānaka Tilaka), ed. K. S. Ramaswami Sastri. Gaekwad's Orient. Series 1931.

Mukula Edition. By M. R. Telang, N. S. P., Bombay 1916.

Pratīhārendurāja Edition. With the text of Udbhata as described above.

Rājānaka Tilaka See above under Commentaries.

VĀMANA ( 1 ) The upper limit to Vamana's date is given by his own quotations (IV. 3. 6 and i. 2. 12) from the Uttara-rāma-carita (i.38) and Mahāvīra-carita (i.54) of Bhavabhūti, who is known to have flourished under the patronage of Yasovarman, king of Kanauj, in the first quarter of the 8th century1. The lower

1 Raja-tarang' iv. 144; Bhandarkar, pref. to Mālafi-mādhava pp. xiiif; JBRAS xxiii, p 92f; S. P. Pandit in prer. to Gaudavaho p. lxviif; WZKM ii 332f. Reference is made in an argument on an illustrative passage in the Vrtti on iii. 2. 2 to Subandhu (v. l. Vasubandhu) a minister of Candragupta. There has been a good. deal of controversy over the identity of the king (bhūpati) who is said to have been a patron of Subandhu or Vasubandhu (see IA x1,. 1911, p. 170f, 312; xli, 1912, p. 1, 15; IHQ i, p. 261). V. Raghavan (IHQ xix, 1943, pp. 70-72) has shewn that it is unnecessary to suppose that Vamana refers to Subandhu, the well known author of the prose Kathā Vāsavađattā; Vāmana is speaking of a minister named Subandhu of Candragupta Maurya and Bindusāra-whose drama is cited by Abhinavagupta as Vāsavadattā Nāțya-dhārā of Mahākavi

VĀMANA 79

limit is given by Rājaśekhara's quotation (p. 14) from Vāmana i. 2. 1-3, and his reference to the Vāmaniyas, which indicates. that by end of the 9th century Vamana had a respectable number of followers going by his name. We learn also from Abhinavagupta (Locana, p. 37)1 that Vāmana was probably known, in the middle of the 9th century, to Anandavardhana who, however, never directly cites Vämana but seems to refer, in a manner not to be mistaken, to the latter's Rīti theory in his Vrtti on iii. 52. Like Bhāmaha, Dandin and Udbhata, Vamana probably lived before the Dhvani-theory, under Anandavardhana, came into prominence ; and Pratīhārendu- rāja, who professes a great reverence for Vāmana's views, expressly states (p. 81), while discussing alamkāra-dhvani, that in such cases Vämana has employed the term vakrokti (iv. 3. 8)2. We will not be wrong, therefore, if we fix the lower limit of Vämana's date at the middle of the 9th century3.

Subandhu, the term Nātya-dhāra signifying that the entire drama was developed by a series of acts within acts. 1 The Lacana quotes from Vamana at pp. 8. 10, 180; also Abhinava-bharati vol. i, p. 288 (=Vamana i. 3. 30-31),-Vamana quotes (iv. 3. 10 ubhau yadi vyomni) from Māgha iii. 8; also Vāmana v. 2. 9= Magha i. 25. The words vo bhartr-pindasya cited as un- grammatical - in v. 2. 28 occurs in Pratijñā-yaugandharūyaņa iv 3; and the verse saracchaśānka-gaurena cited in iv. 3. 25, occurs in Svapna-vāsavadatta iv. 7. 2 Even supposing with Jacobi that Vamana was contemporaneous with the ananymous Dhvanikara, he cannot yet be shown to have been influenced in any way by the opinions of that school. The remarks of Ruyyaka (p. 7) and Jayaratha apparently support the trend of Pratīhārendurāja's opinion, and Jayaratha expressly says, with reference to these old writers, that they were unaware of the views of the Dhvanikāra (dhvanikāra-matam ebhir na drstam, p. 3), the Dhvanikara being, in Jayaratha's opinion, the same as Ananda- vardhana himself. 3 Cappeller's thesis propounded (in Vamana's Stilregeln pp. iii f; also pref. to his ed. pp. vii f) that Vamana should be placed later than 1000 A.D. is disproved entirely by the quotations given above. Cf. Pischel, op. cit. pp. 23 f. The mention of Kavirāja need not, as Pischei shows, of itself place Vamana as late as 1000 A.D.

80 SANSKRIT POETICS

These considerations make it probable that Vamana lived between the middle of the 8th and the middle of the 9th cen- tury, at about 800 A.D., and justify Bühler's identification, in deference to Kahlana iv. 497 and "the tradition of Kashmirian Pandits", of our Vāmana with the Vämana who was a minis- ter of Jayāpīda of Kashmir (779-813 A.D.). This conclusion makes Udbhata and Vāmana contemporaries and rivals ; and the way in which Rājaśekhara, Hemacandra and Jayaratha refer to the two rival schools of Vamanīyas and Audbhatas lends colour to such a supposition. The Vrtti (with illustration) on the Sūtras, called Kavi- priyā, is composed, as its mangala-śloka indicates, by Vāmana himself (cf iv. 3. 33). This is confirmed by the fact that later writers ascribe both parts of the work to Vämana1. The illustrations, as he himself informs us, are both svīya and parakīya (iv, 3, 33). Vāmana's work, after the sūtra-style, is divided into five Adhikaranas, each of which is divided again into Adhyayas. The first and fourth Adhikarana have three Adhyayas, the rest two each, the total number of Adhyayas being twelve. The subject-matter of the Adhikaranas is in- dicated by their naming: (i) Sārīra (ii) Doşa-darśana (iii) Guņa-vivecana (iv) Ālaņkārika and (v) Prāyogika. The last Adhyāya deals with Sabda-śuddhi or grammatical correctness as an aspect of Prayogika. The number of Alamkāras defined and illustrated is thirty-six.

( 2 ) It has already been noted that Vamana, in many respects, attempts to improve upon the system of Dandin. Vamana does not claim entire originality with regard to the illustra- tions he cites, and many of them may be traced to well known sources. The Riti-theory itself, which Vamana for the first time clearly and systematically enunciates, is probably older that Bhamaha, who alludes to the classification of the gaudi

1 Eg. Pratīhārendurāja (p. 17, 76, 81, 84); Locana (p. 37).

MANGALA 81

and vaidarbhī; and Vāmana himself cites from unknown expositors of the past, e.g. in his Vrtti on I. 2. 11, 12-13 ; 3. 15. 29, 32; II. 1, 18; 2. 19 ; III. 1. 2. 9, 25 ; 2, 15 ; IV. 1, 7. etc .. with atra lokāḥ or tathā cāhuḥ. While Daņdin supplies an important link between these unknown authors and Vāmana, we find the theory in its completely self-conscious form in the latter. But it appears to have languished after Ananda- vardhana came into the field, in spite of the fact that Vāmana's influence apparently created a school known in later times as the Vāmanīya.

MANGALA

To this school probably belonged Mangala, who must have been a comparatively early writer, having been cited by Rājaśekhara (pp. 11, 14, 16, 20). Mangala, we are told by Hemacandra (Comm. p. 195), agrees with Bharata in his definition of Ojas, and maintains with Vamana that Dandin is not right in emphasising it in the gaudī riti, inasmuch as it is common to all the Ritis.1 This is all we hear about this writer, but it would indicate that in his views he leaned towards the system of Vämana who was probably his pre- decessor. A poet Mangala is quoted in the Sadukti-karņāmṛta.

( 3 )

COMMENTATORS ON VĀMANA

The existing commentaries on Vamana are mostly late, and are therefore hardly acceptable to a critical and historical student. The Kama-dhenu by the South Indian, Gopendra Tippa Bhūpāla, who was governor under Devarāya II (1423- 46 A. D.) of the Vijayanagara dynasty, is a lucid exposition of the text, and its popularity is indicated by its frequent publication in India.

1 Cf. Mānikyacandra (Samketa, Mysore ed. p. 292).

82 SANSKRIT POETICS

BIBLIOGRAPHY

Editions and Translation. (1) Vämana's Lehrbuch der Poetik ed, by Carl Cappeller, Jena 1875 ; (2) ed. by Durga- prasad and K. P. Parab, Nir. Sag. Press, Bombay 1880, also 1889 ; (3) ed. by A. Borooah, Calcutta 1883 ; (4) ed. Grantha- pradarśinī, 1895 ; (5-6) ed. with Kāma-dhenu Comm. Ben. S. S. 1908, and Srivani-vilasa Press, Srirangam 1909 ; (7) ed. N. N. Kulkarni with extracts from Kama-dhenu. Poona 1927 ; (8) Vāmana's Stilregelen, bearbeitet v. C. Cappelier (Trans. of the last chapter), Strassburg 1880 .- Eng. Transl. by Ganganatha Jha, Allahabad (no date), published originally in the Indian Thought, vol iii-iv 1912; reprinted Orient. Bk. Agency, Poona 1928 (2nd revised ed.). Our references are to the Benares edition of 1908, unless otherwise indicated. Commentaries. (1) Kāma-dhenu by Gopendra (or Govinda) Tippa (or Tirpurahara as he himself renders the name, introd. v. 8) Bhūpala, about whom see above. He cites, among numerous other authors, Vidyādhara, Vidyānātha, Bhatta Gopāla (the commentator on Mammata ?), Ghaņtāpatha (apparently of Mallinatha) and Dharmadāsa Sūri's Vidagdha- mukha-mandana (on ii. 2. 21). He is thus later than the 14th century. He cites a work called Kavi-gajākuśa. Ed. with the text in Grantha-pradarini 1895; in the Benares Sansk. Series and in the Srivani-vilasa Press, etc. Our references are by pages to the Benares Ed. (2) Sāhitya-sarvasva by Maheśvara (IOC 566 ; ABod 207b). See Śrīvatsalāñchana (under Commentators on Mammata below). (3) A Commen- tary by Sahadeva quoted in the notes to the Gaekwad edition (Baroda) of Kāvya-mīmāmsā p. 5.

RUDRAȚA AND RUDRABHAȚȚA

( 1 ) The lower limit of Rudrata's date is furnished by the citation of Rudrata by name and reference to his kāku- vakrokti figure (ii. 16) by Rājaśekhara (p. 31) at the end of the

RUDRAȚA 83

9th and beginning of the 10th century. This conclusion is supported by two considerations. Vallabhaceva who, as we shall presently see, flourished in the first half of the 10th century, quotes Rudrata by name in his commentary on Māgha (ed. Kashi Sansk. Series 1929, ii. 44) and mentions (on iv. 21) that he himself also composed a commentary on Rudra- ța's treatise on Alamkāra, where he had discussed in detail the points in question ; while in the same commentary Hultzsch notes numerous references, mostly anonymous, to Rudrata1. Again, Pratīhārendurāja, about the same time, quotes anony- mously (pp 42, 49) Rudrata's Kārikā-verses vii. 35 and xii. 4, as well as cites (p. 43) the illustrative stanza in Rudrața vii. 362. Abhinavagupta also quotes anonymously (Locana, p. 45) a Kārikā-verse of Rudrața's (vii. 38). This sets aside altogether the conjectural date, viz. the second half of the 11th century, assigned by Bühler3, as well as his revised date4, viz, the middle of the 10th century, which Peterson5 first put forward. The upper limit cannot be so definitely settled ; but it seems probable that Rudrata was younger than Bhamaha, Dandin and Vämana, with regard to whose date his own time is to be settled. We need not go so far as to hold with Jacobi6 that Rudrata derived his idea of vakrokti from Ratnākara's well known poem Vakrokti-pañcāśika, and therefore was

1 See Hultzsch's pref. to his edition of Megha-dūta with Valla- bha's comm., London 1911, p. x-xi. These references are not all found in the printed text. This Vallabhadeva must be distin- guished from the compiler of the Subhas° who bears the same name. 2 Cf Pischel in GgA, 1885, p. 764. Other quotations by Prati- hārendurāja are: p. 11=R viii. 40; p. 31=R viii. 89; p. 34=R viii. 95. Also Dhanika on iv. 35=R xii. 4. 3 Kashmir Rep. p. 67. 4 IA xii. 30. 5 Peterson. i (Detailed Report 1883) p. 14; also introd. to Subhāș° r. 105. Their arguments are based on the date of Nami-sadhu's. commentary on Rudrata, which is now known to be dated in 1069 A.D. See below on Nami-sādhu. 6 WZKM ii. 151 f.

84 SANSKRIT POETICS

later than Ratnākara, son of Amrtabhānu, who lived under Brhaspati and Avantivarman of Kashmir ; but it is clear that if this new idea of vakrokti did not originate with Ratnäkara or even with Rudrata, it was defined for the first time by the latter and illustrated by the former as a particular poetic figure. This verbal figure is described as resting on śleşa (paronomasia) or kāku (intonation) and is based on a deliberate misunderstanding of one's words for the purpose of making a clever retort (Rudrata ii. 14-17). Bhāmaha (ii. 85), on the other hand, had taken vakrokti, not as a particular poetic figure, but as a certain strikingness of expression which characterises all poetic figures ; while Dandin had limited the range of vakrokti and made it a collective name for all poetic figures with the exception of svabhāvokti (ii. 362 and comm. thereon). Vamana was the first to regard vakrokti as a special poetic figure (arthālam- kāra), but he too used the expression in a more or less general sense to denote a particular mode of metaphorical expression based on laksanā or transferred sense (iv. 3. 8)1. From this it appears that (1) the term vakrokti travelled through all these writings from a very broad sense as the distinguishing characteristic of all poetic figures to the precise and narrow signification of a specific verbal poetic figure in Rudrata's definition ; a definition which, however, unquestionably established itself in all later writers (except in Kuntaka who developed his idea directly from Bhāmaha); (2) the order of development points apparently to the conclusion that Rudrata was probably later than Bhamaha, Dandin and Vamana ; for in his time the broader and older connotation of vakrokti was out of date, and it came to be looked upon as a defined species of śabdālamkāra; and (3) its illustration by Ratnākara indicates its existence, independently of Rudrata, in the 9th century A. D. These indications make it probable, apart from a detailed examina- tion of Rudrata's other theories in relation to those of

1 Cf Jacobi in ZDMG lxiv, p. 130f.

RUDRAȚA 85

Bhāmaha, Dandin and Vāmana, that the substance of Rudrata's teaching was probably later than that of these older writers. If this conclusion is accepted, then Rudrata should be placed after Vämana, who is the latest member of this group ; and this gives us the upper limit to his date. It seems probable, that Rudrata should be placed between the first quarter of the 9th century and its end; for he appears to be unknown to Anandavardhana. It will not be wrong if we accept the most plausible date assigned to him by Pischel1, viz. the middle of the 9th century2. The date makes him a contemporary of Ānandavardhana, who never cites or refers to Rudrata, as he. does to other well- known predecessors, and by whom this peculiar ālamkārika connotation of vakrokti, if known, was not apparently recognised. Rudrața's Kāvyālamkāra consists of 16 Adhyāyas and 734 Kārikās in Āryā metre (excepting the concluding verses) and comprehends almost all the topics of Poetics. This enumera- tion excludes 14 Kārikās after xii. 40 (dealing with eight kinds of Nāyikā) which are declared to be interpolated passages.3

( 2 ) What is said here about Rudrata does not apply to Rudra or Rudrabhatta, although Pischel4, Weber5, Aufrecht6 and

1 Pref. to Śrng. til. pp. 12, 26. 2 The suggestion of Jacobi that Rudrata, whose name implies that he was a Kashmirian, was a contemporary of Samkaravarman of Kashmir, successor of Avantivarman, doez not make any essential difference to our conclusion, although it is not certain that the ex- ample of vakroktt given by Rudrata in ii. 15 was at all prompted by Ratnakara, whose work contains similar railleries between Siva and Gaurī. 3 For summary of topics in Rudrata's work set vol. ii, ch. ii (3) beiow. 4 Prcf. to Srng. til .; ZDMG xlii, 1888, pp. 296-304, 425. 5 Ind. Stud. xvi. 6 ZDMG xxvii pp. 80-1, xxxvi p. 376; Cat. Bod. 209b; Cat. Cat. pp. 528b, 530a.

86 SANSKRIT POETICS

Bühler1 take the two authors to be identical. The identity is declared doubtful by Peterson2 and is not admitted by Durga- prasada3 and Trivedi4. Finally, Jacobi has set at rest this controversy by showing elaborately, from an examination of their respective texts, that these two writers were in all prob- ability different persons5, From v. 12-14 of the Kāvyālumkāra of Rudrața, as inter- preted by Nami-sadhu, it appears that Rudrata, also called Satānanda, was son of Bhatta Vämukha and a follower of the Sama-veda. Rudrabhatta's genealogy or personal history is unknown. But much has been made of the apparent similarity of the two names. The last verse6 of the Srngāra- tilaka, however, expressly gives the name of its author as Rudra, with which description most of the MSS agree7 ; while both Nami-sadhu and Vallabha call the author of the Kāvyā- lamkāra by the name of Rudrata. The two authors belong, again, apparently to two different religious persuasions, Rudra being a worshipper of Siva, and Rudrata omitting a reference to this deity and mentioning instead Bhavānī and Murāri (besides the usual Ganeśa).

1 Kashmir Rep. p. 67. 2 Rep. i, p. 14; pref. to Subhās° pp. 104-5; but contra in Report ii, p. 19 footnote. 3 Footnote to the ed. of Srng. til. p. 1. 4 Notes to his ed. of Ekāvalī p. 3. 5 WZKM ii, 1888, pp. 151-56; ZDMG xlii pp. 425 f. 6 The last verse, however, is not found in some 'MSS. 7 With the curious exception of a Kashmirian MS in Sāradā characters (Bühler's Kashmir Rep. no. 264) where the name in given as Rudrata. This unique testimony raises a legitimate suspicion, but it can be explained as a piece of not unnatural confusion made by a Kashmirian scribe, to whom the more famous name of Rudrata must have been more familiar. The same remark applies to the India office MS no. 1131 (Cat. vii, p. 321) and the South Indian MS in Madras Cat. xxii (1918) no. 12955, in which, however, the last verse gives the name as Rudra. This evidence, however, of colophons of MSS is not decisive; and it is well known that even later anthologies and writers of note make a similar confusion between the two authors.

RUDRABHAȚȚA 87

Taking the works themselves, Rudrata's text of sixteen chapters covers a much larger ground than Rudra's much shorter work of three chapters, and presents a distinctly different outlook. Rudrata puts a greater emphasis on the kāvyālamkāras or poetic figures which supply, as Nami-sādhu points out, the name of the work itself, and which absorb its eleven chapters, leaving only five concluding chapters for a brief supplementary treatment of Rasa, the cognate topic of nāyaka-nāyikā and the general problems of poetry. The keystone of Rudra's system, on the other hand, is the idea of Rasa, having special reference to śrngāra (ch. i. and ii) with just a summary description of the other Rasas (ch. iii) ; and the chief value of his smaller text consists in his minute poetical treatment of the theme of śrngāra-rasa and nāyaka- nāyika. It will appear, therefore, that while Rudrata's scope and method are more ambitiously theoretical and comprehen- sive, Rudrabhatta merely singles out a part of the whole subject, and not troubling himself about definitions and rules (which appear almost word for word as they are in Rudrata) gives us, in his apt and finely composed illustrative verses, a practical poetical manual on the subject of love and other sentiments. Jacobi, therefore, rightly remarks that "Rudrata appears as an original teacher of poetics, while Rudra, at his best an original poet, follows, as an expounder of his śāstra, the common herd." With regard to the common topics, there is however, a general agreement, even to minute details, which has misled some scholars to attribute the two works to the same author. But beneath this general agreement, the two works reveal many points of difference which affect some of the funda- mental conceptions of their respective authors. Taking, for instance, their treatment of Rasa, we find that while Rudra (i. 9) follows the general tradition, prevailing from Udbhata's time, of mentioning nine Rasas, Rudrata adds one more, viz. preyas (xii. 3), treating them in an order somewhat different from that followed by Rudra. Rudra enumerates and dis-

88 SANSKRIT POETICS

cusses at some length the bhāvas (i. 10-19), which are sum- marily referred to by Rudrata in one verse only (xii. 4). A similar difference will be noted in the treatment of Vrttis, of which Rudra (i. 12) mentions, after Bharata (xx. 24f), the usual four (viz. kaiśikī, ārabhațī, sāttvanī and bhāratī), origin- ally taken as styles of dramatic composition but borrowed here apparently with a similar purpose from dramaturgy to poetry (cf Bharata loc. cit. 21). Rudrata, on the other hand, speaks (ii. 19f), after Udbhata,1 of five vrttis (viz, madhurā, praudhā, paruşā, lalitā and bhadrā), which have nothing to do with the above four, but being comprised under allitera- tion (anuprasa) refer primarily to suitable sound-adjustment. by special arrangement of letters. With regard to the cognate topic of nāyaka-nāyikā, similar material discrepancies can be detected. While Rudra describes at some length the eight. conventional avasthās (conditions or situations) of the nāyikā (i. 131-32), Rudrața mentions only four (viz., abhisārikā, khaņditā, svādhīna-patikā and proșita-patikā, xii. 41f), al- though to make up for this unwonted divergence there is in some MSS a long passage (between xii. 40 and 41), describing. the usual eight conditions but irreconcilable to its context, and rightly stigmatised by Rudrata's editor as interpolated. The third class of heroine, again, viz., the courtezan (veśyā), appears to be favoured by Rudra (i. 120-30), while Rudrata dismisses her in two verses only (xii. 39-40) with an apparent note of condemnation. The tenfold state of a lover, beginning with desire and ending in death, is mentioned in passing by Rudrata (xiv. 4-5), but Rudra defines and illustrates each of these states in detail (ii. 6-30). While the trespasses in love, according to Rudra (ii. 53), depend on time (kāla), place (deśa) a nd circumstances (prasanga), Rudrata thinks (xiv. 18) that a fourth condition, viz. the person concerned (pātra) should: be added.

1 Udbhata mentions only three vrrtis in connexion with anuprāsa, viz. paruşā, upanāgarikā and grāmyā (i. 4-7).

RUDRABHAȚȚA 89

All these indications make it highly probable that Rudrata and Rudra were two different persons ; but if this is so, how are we to explain the fact most of the verses in the Śrngāra- tilaka are, but for their difference in metre (anustubh and arya respectively) identical almost word for word with the corresponding verses in the Kāvyālamkāra? This point has been emphasised with some plausibility by the advocates of the identity of the two writers. But it should be noticed that this extraordinary verbal coincidence does not extend beyond those verses which give the rules and definitions ; for the illustrative stanzas in the Srngāra-tilaka, composed in a variety of metres, and forming a distinctive feature of this more poetical work, do not occur in the Kāvyālamkāra at all. It is not unusual to find similar treatment and terminology in technical treatises, abounding in standardised and conventional rules and definitions ; but this is not enough to explain this extraordinary plagiarism tempered, it is true, by the presence of highly poetical and presumably original1 stanzas composed to illustrate these dry rules and definitions. Nor is this explanation, which is based on the supposed identity of the two writers, at all free from consider- able doubt in view of the discrepancies noted above. The real explanation probably lies in the supposition that Rudra, apparently a later writer and chiefly a poet, and never pre- tending to be an original teacher of Poetics, found in these ready-made rules of Rudrata enough poetic possibilities, as well as an opportunity of displaying his own poetic powers, and proceeded forthwith to furnish the missing poetic illustra- tions.2 As he did not apparently aspire to write an original

1 Some illustrative stanzas in the Srng. til. are taken from earlier works like the Amaru-sataka. 2 These chapters in Rudrata are purely expository, and are not fully illustrated as the preceding chapters are. One may be led to suppose that Rudrata himself composed the Śrngāra-tilaka to supply this deficiency, but this hypothesis does not sufficiently explain the divergence of views on the points noted above, which apparently indicates that the task of supplementing could not have been under-

'90 SANSKRIT POETICS

thesis on the topics concerned, he did not trouble to alter the wording of the fixed canons and made only enough changes to suit the metre. While Rudrata is concerned directly with rules and prescriptions, Rudra is more practical in his object and treatment and intends his treatise on love apparently to serve as a psychologico-poetical guide to the gay science, furnishing it, as he does, with an elaborate analysis of the various moods and sentiments, which belong to the province of Erotics as well as Poetics1.

taken by himself but by some other author, who held some views different from his own. It cannot be argued that chapters xii-xvi in question, which contain this deficiency, is a later addition by -some other hand, on the ground that the closing verse of ch. xi reads like the end of the work itself; for a similar remark applies to the closing verses of chapters iii, iv and v. 1 The quotations from these two writers in later literature are unfortunately mostly anonymous, and do not materially belp us in deci- ding this question. Among the very few cases where the author's name is cited along with his verses, Mammata. himself a Kashmirian, cor- rectly refers (ad ix. 8a) the verse sphutam arthālamkūram to Rudrata (iv. 32). Samudrabandha (p. 6) and Hemacandra (p. 286 Comm.) refer rightly to Rudrata vii. 38-40 and vii. 27 respectively. Ruyyaka, while discussing Rudrata's views does not refer to his treatment of the Rasas at all. On the other hand, Baladeva Vidyābhūsaņa (p. 35) cites vāmatā durlabhatvam from Šrngāra-til. (ii. 41), and correctly refers to the author as Rudra. Vallabhadeva in his anthology cites a number of verses from the Kāvyālamkāra (421=ii. 17; 730=iii. 57; 1387=vi. 10; 1667=ii. 30; 2047=vii. 71; 2061=vii. 33; 2234=vii. 41; 2409 =vii. 32), and with two exceptions, gives the name of the author correctly as Rudrata. Similarly, Sarngadhara quotes eight verses from the Srngāra-til. (3409=i. 95; 3567=i. 35; 3568=i. 81; 3679=ii. 107; 3578=ii. 12; 3579=ii. 50; 3675=i. 51; 3754=i. 30), and with two exceptions, again, refers them correctly to Rudra. The two excep- tions of Särngadhara (viz. 3773 and 3788) are wrongly attributed to Rudra; but Vallabhadeva quotes the very same verses (2234 and 1667), and correcting the mistake refers them rightly to Rudrata. Vallabhadeva's two exceptions (2247 and 3122) cannot be found in Rudrata to whom they are attributed, nor in Rudra. Jahlana attributes 21 verses to Rudra, of which 5 cannot be traced in his work. Of these, kim gauri mam occurs in Rudrata ii. 16; ambā sete'tra occurs

COMMENTATORS ON RUDRAȚA 91

( 3 ) Rudrabhatta's date is uncertain; but as Hemacandra (p. 110) is the oldest writer to quote anonymously and criticise his mangala-verse (i. 1), we should place him between Rudrata and Hemacandra, i.e. later than the 9th century but earlier, probably not much earlier, than the 12th. One verse, however, of Rudra, which cannot be traced in his Srngāra- tilaka, is quoted by Dhanika' (iv. 60, ed. NSP 1917. p. 103); and if this Rudra is the same as our author, he should be placed before the enc of the 10th century.1

( 4 ) COMMENTATORS ON RUDRAȚA

Vallabhadeva Vallabhadeva's commentary on Rudrațālamkāra, referred to by himself in his commentary on Māgha (on iv. 21, vi. 28) is the earliest known commentary on Rudrata and is still to be recovered. Vallabhadeva, who bore the surname of Paramaratha-cihna and who describes himself as son of Rājānaka Ānardadeva2, is a well-known commentator on several standard poetical works, including those of Kālidāsa, Māgha, Mayūra and Ratnākara. He was apparently a Kashmirian and probably belonged to the first half of the 10th century ; for his grandson Kayyata, son of Candrāditya, wrote a commentary on Anandavardhana's Devī-śataka3 in

in Kavi. vacana- samac .; 505 = Subhās° 2247 (attributed to Rudrata) = Sadukti-karņā° ii. 73 (attributed to Bhațța); ekākinī yad abalā is attributed wrongly to Rudra (as in the Paddhati 3773) but correctly assigned by Vallabhadeva to Rudrata. Hemacandra quotes anony- mously three verses from Rudra (i. 1=p. 110; i. 44=p. 304; i. 68= p. 305). 1 Dhanika also quotes anonymously from Rudrata (xii. 4) in his comm. on iv. 35, ed. N.S.P. Cf. Rudrabhatta i. 16. 2 From the closing verse of the comm. on Vakrokti-pancūsikā Änandadeva seems to have held some high appointment in Kashmir (colophon in Kāvyamālā Gucchaka i, p. 114: Mitra x, no. 4064). 3 See foctnote to the ed. of Vakrokti-pañcāśikā in Kāvyamālā,

92 SANSKRIT POETICS

977-78 A. D., during the reign of Bhīmagupta of Kashmir (977-82 A. D.). Vallabha's preceptor was Prakāśavarsa who, Hultzsch thinks, is perhaps identical with the poet Prakāśa- varsa, who is quoted in Subhāşitāvalī and Sārngadhara-paddhatī and who wrote a commentary on Bhäravi (Aufrecht i. 347). Our author must be distinguished from Vallabhadeva, the compiler of the anthology Subhāsitāvalī, who is assigned by Aufrecht to the 16th century1. From Hultzsch's list of Vallabha's quotations in his commentary on Mägha it is interesting to note that he cites Medhāvirudra, Bhāmaha, Udbhața, Bhatți and the Vișamabāna-līlā (apparently the Prakrit poem of Anandavardhana). Nami-sādhu After Vallabhadeva, comes Nami (known as Nami-sādhu or Nami-pandita), who is described as a Šveta-bhikșu, indicating that he was a Śvetambara Jaina, He describes himself as "the bee that sucked honey from the lotus- feet of Śrī-śālibhadra Sūri, the ornament of the gaccha of the city of Tharapadra". We know that Jinabhadra Sūri, who was a pupil of Salibhadra, wrote in Samvat 1204=1148 A. D.2 Sālibhadra is also referred to as Śrī-śāli-sūri in another work of Nami's called Sadāvaśyaka-fikā, of which the date is given in the work itself as Samvat 1122=1063 A. D.3 In one of the concluding verses of his commentary on Rudrata, Nami states that his commentary was composed in Samvat 1125=1069 A. D.4 Among writers on Poetics

Gucchaka i, pp. 101-2: and to ed. of the Devi-sataka in ibid, Gucchaka ix. p. 1. Cf. Hultzsch's pref. to Megha-dūta p. ix. 1 Bühler (Kunst Poesie p. 71) thinks that anthologist Vallabhadeva flourished between 1400 and 1350 A.D. The date now requires revision,. for this anthology is quoted directly (with the names of the work and author) by Vandyaghatiya Sarvananda in 1160 A.D. in his commentary on the Amara-kośa. See on this question S. K. De in JRAS, 1927, pp. 471-91 and BSOS, v, pt. i, p. 27f; v, pt. iii, p. 499f. 2 Peterson i p. 68. 3 Peterson iii p. 13. 4 pañca-vimsati-samyuktair ekādasa-samāsataih / vikramāt sama-

COMMENTATORS ON RUDRAȚA 93

cited by Nami on Rudrata, we find the names of Bharata (p. 150, 156, 164), Medhāvirudra (pp. 2, 9, 145), Bhāmaha (p. 2, 116), Dandin (pp. 2, 5, 169), Vāmana (pp. 11, 100, 116), Udbhata (pp. 69, 82, 150) and the Arjuna-carita by Ānanda- vardhana. He also cites Tilaka-mañjarī (on xvi. 3) apparently of Dhanapāla, and one Jayadeva on prosody (on i. 18, 20).1 It is interesting to note that Nami quotes a Prakrit verse from one Hari (ii. 10), presumably a writer on Poetics, which mentions eight Vrttis instead of five of Rudrata. Āśādhara Peterson (ii, p. 85) brought to our notice another Jaina com- mentator on Rudrața, called Āsādhara, son of Sallakșaņa and Ratni. He was a Jaina teacher who lived till Samvat 1296= 1240 A.D. He should be distinguished from Āsādhara, son of Rāmaji Bhatta, a very late writer who composed a com- mentary on Appayya's Kuvalayānanda2. In the Praśasti at the end of his Dharmamrta, Asadhara's personal history is given. He belonged to the Vyaghreravala family and was son of Sallakşaņa (or Lakşana) and Ratni. He was born in the fortress of Mandalakara situated in the country round the Sākambharī (Sambhar) Lake. He had by his wife Sarasvatī a son named Chahada, who was a favourite of Arjunavarman of

tikrantaih pravrsidam samarthitam. The reading panca-vimsati° in this verse is doubted on the authority of Kielhorn's palm-leaf MS (Report, 1880-81, p. 63) which reads instead sat-saptati°. This new reading will place Nami much later; but it is not supported by other MSS (e.g. Mitra 3102; Stem 61; Peterson i p. 16), as well as by the evidences adduced above from other sources. Kielhorn's reading is obviously incorrect in itself, because it makes the verse deficient in metre. 1 For this author see H. D. Velankar, Jayadaman, Bombay 1949. This Jayadeva and his Chandah-śastra are also cited by Abhinavagupta in his comm. on Bharata. For his date (before 950 A.D.) see P. K. Gode in Poona Orientalist, pp. 33-38 (reprinted in his Studies in In- dian Lit. Hist. 1, pp. 138-43). 2 The two are confused by Aufrecht i. 54b and, following him by Harichand Sastri p. 18; but Aufrecht expresses doubt about their identity.

94 SANSKRIT POETICS

Malava (first quarter of the 13th century). After the invasion of Sāhibadīna, king of the Turuskas(apparently Shāhābu-d-din Ghūr, Sultan of Delhi, who vanquished Prthurāja in 1193 A.D.), Asādhara emigrated to Mālava and lived in Dhārā. where he learned the doctrines of Jaina faith and Jinendra- vyākaraņa from Paņdita Mahāvīra, pupil of Dharasena. Āsādhara was reputed for his learning, being praised by the sage Udayasena, by the poet Vihlana (sic) who is not the Kash- mirian poet Bihlana who lived about 1070-90 A,D., but minis- ter of Vijayavarman king of the Vindhyas, and by the great yati Madanakīrti. He wrote more than 15 works, of which he gives a list, referring also to his commentary on Rudrata1. His Trişasti-smrti-śāstra, which gives stories of 63 great men mentioned in the Jaina Purānas, was composed in 1236 A.D.2

BIBLIOGRAPHY Rudrata Edition. Kāvyālamkāra Ed. Durgaprasad and K. P. Parab, Nir. Sag. Press, Bombay (Kāvyamālā 2) 1886, 1909 with Nami's comm. Our references are to the second edition. of 1909. Nami-sādhu Edition, Kāvyamālā 2, 1886, 1909, with the text (see above) .. Āśādhara Mss. Aufrecht i 103a, 779a. No MS of Vallabhadeva's commentary on Rudrata is. known. Rudrabhatta Editions. Śrngāra-tilaka (1) by Pischel with an introd .. Kiel 1886 (2) Kāvyamālā Gucchaka iii 1887, 1899. Our re- ferences are to Pischel's edition.

1 Peterson ii p. 85 and fn; Bhandarkar Rep. 1883-84, pp. 103-4 .. 2 Aufrecht i 54b. Winternitz (Geschichte, ii, p. 354, fn 5) assigns 1228-1244 A.D. as the date of Asadhara's literary activity. See also Kielhorn in EI. ix. 1908, p. 107f.

VIȘŅU-DHARMOTTARA 95

Commentary. Rasa-tarangiņī by Gopāla Bhatța, son of Harivamśa Bhațța Drāvida. See under Commentators on Mammața and Bhānudatta, below. Mentioned in Kāvyamālā ed. of the text (p. 111). He calls his author Rudra. If the Gopāla Bhatta, cited by Kumārasvāmin, be the same person, then he is earlier than the beginning of the 15th century. As Gopāla Bhatta, appears to be a South Indian writer, this. conjecture is probable. Oppert's entry of Vana-tarangiņī (ii 2711, 1787) on Rudrata is apparently a mistake for this work, and the name Rudrata is a confusion for Rudra.

THE WRITER ON ALAMKARA IN THE VIŞŅU- DHARMOTTARA AND AGNI-PURANA

The Vişnu-dharmottara The Visnu-dharmottara1, counted as an Upapurāna, is one of the later Puranic works of an encyclopaedic character which,. in the course of its rapid treatment of multifarious subjects, devotes a few chapters to the topics of dancing (Nrtta), singing (Gīta) and music (Ātodya), Poetics (Alamkāra), Dramaturgy (Natya) and allied subjects. The work consists of three Kāndas of more than eight hundred chapters, the distribution and number of chapters in each Kanda being as follows: i-ch. 269; ii-ch. 183; iii-ch. 355. We are concerned here only with a few chapters of Kanda iii, which begins with what is called Citra- sūtra or the art of painting and proceeds to the allied topics of dancing, singing and music. Ch. 14-15 of this Kānda is devo- ted generally to some topics of Poetics. They name and define 17 kinds of poetic figures (alamkāras), which are: Anuprāsa, Yamaka (two types Samdasta and Samudga distinguished), Rūpaka, Vyatireka, Śleșa, Utpreksā, Arthāntara-nyāsa, Upa- nyāsa (probably a kind of Vyājokti), Vibhāvanā, Atiśayokti, Svabhāvokti (or according to some MSS Vārta), Yathā-samkh-

1 For a detailed account of the content of this huge compila- tion, chapter by chapter, see R. C. Hazra, Studies in the Upapurānas,. i, Calcutta 1958, pp. 147-218.

96 SANSKRIT POETICS

ya, Viśeşokti, Virodha, Nindā stuti, Nidarśana and Ananvaya. In this list may be included Upama which, however, is not explicitly mentioned or defined. In some of these verses reference is made to older authorities (purātana or purāņa). Ch. 15 speaks of Kävya and distinguishes it from Sāstra and Itihāsa. The Mahākāvya is then defined and characterised ; it should contain nine Rasas, namely, Śrňgāra, Hāsya, Karuņa, Raudra, Vīra, Bhayānaka, Bībhasta, Adbhuta and Sānta; there should be description of towns, countries etc ; it should have a Nāyaka and Pratināyaka, but it should not describe death of the Nāyaka. Ch. 16 names, defines and classifies Prahelikās, Ch. 17-31, based mostly on Bharata's Nātya-śāstra, deal generally with dramaturgy. The Rūpakas are classified into 12 varieties, namely, Nāțaka, Nāțikā, Prakaraņa, Prakaraņī, Utsṛsțikānka, Bhāņa, Samavakāra, Īhāmrga, Vyāyoga, Vīthi, Ņima and Prahasana. Eight types of Nāyikās are distinguished, namely, Vāsaka-sajjā, Virahotkaņțhitā, Svādhīna-bhartrkā, Kalahānta- ritā, Khaņditā, Vipralabdhā, Proşita-bhartrkā and Abhisārikā. Ch. 18-19 deal with the different kinds of vocal (Gīta) and instrumental (Atodya) music, appropriate to the nine Rasas. Ch. 20-29 define Nātya, classify Nrtta, deal general- ly with the construction of the stage and its attendant religious rites and describe the four modes of Abhinaya, namely, Āngika, Sāttvika, Vācika and Āhārya. Ch. 30 deals in 28 verses with the nine Rasas (including Santa) and their mode of delineation. Ch. 31 describes 49 Bhävas, including Nirveda. This brief résumé will give a rough idea of the topics inclu- ded in the 28 chapters of the Upapurāna on Alamkāra, Nātya and allied subjects. But the section is essentially a rapid com- pilation and presents no definite system. The date of the Upa- purāna, however, must be early. It is later certainly than Bhara- ta's Nātya-śastra which is used as a source, but it is earlier probably than Bhatti and the works of Bhamaha and Dandin, who enumerate a much larger number of poetic figures. Its internal evidence as a whole indicates that it cannot be

THE AGNI-PURAŅA 97

earlier than 400 A. D. and later than 500 A. D. It was com- posed either in Kashmir or in the northernmost part of the Panjab; but its date is too early for any reference, direct or indirect, to the Kashmirian Dhvani school.1

The Agni-purāna The writer on Alamkara in the Agni-purana (chs. 336-346) attempts to treat the subject of Poetics in what may be described roughly as an eclectic fashion, but the account is much fuller and more systematic. It appears to follow at the same time a tradition which is distinct from and probably older than that of the Kashmirian school of Anandavardhana, although the work itself may not have been old. The date of this apocryphal work, which is counted as one of eighteen Mahāpurāņas is uncertain ;2 but there is enough evidence to show that the Alamkara-section in it is chiefly a compilation by a writer, who was himself no great theorist but who wanted to collect together and present a workable epitome, conforming in essentials to the teachings of no particular system, yet gathering its material from all sources. This is apparent not only from its independent, if somewhat loosely joined and uncritical treatment, but also from the presence of verses culled from Bharata, Bhāmaha, Dandin and probably other old unknown writers.3 Bharata is cited by name in Agni 339. 6, and a large part of its treatment of nātya, nrtya, abhinaya and rasa follows Bharata's exposition, even to the literal

1 On the question of the date and provenance of the Upapurāņa see R. C. Hazra, op. cit. pp. 205-12. 2 The present Agni-purāna is said (IHQ, xii, pp. 683-89) to be a recast of an earlier and original Agni-p. which is the text described by the Matsya-p. and Skanda-p. and quoted by some Dharma-sāstra writers. 3 Thus, the verse abhidheyena samıbandhat (Agni 344. 11-12). which is also cited anonymously by Mammata (Sabda-vyāpā° p. 8) and in the Kūma-dhenu on Vamana iv. 3. 8, is attributed to one Bhartrmitra by Mukula (p. 17). The verse is not traceable in Bhartrhari. Bhartrmitra is cited also by Someśvara in his comm. on Mammata (p. 16). 7

borrowing and paraphrase of some of his well known verses, Thus, Agni 337. 11-12=Bharata xxii. 28-29 ; Agni 338. 12= Bharata vi. 36; Agni 338, 7-8=Bharata vi. 39 ; Agni 342. 15-17=Bharata xvii. 62-65. The definition of poetry (Agni 336. 6) and of poetic figure (Agni 341. 17) are copied literally from Dandin i. 10 and ii. 1 respectively. Cf also Agni 336. 13=Daņdin i. 29=Bhāmaha i 27; Agni 336. 23, 25,26= Daņdin i. 12, 15, 17=Bhämaha i. 20. This will be enough to indicate not only the general nature of the work, but also the probability that these chapters of the Purāna were compiled later than Dandin.1 Two verses again (apāre kāvya-samsāre and śrngārī cet kaviḥ) occur in Agni 338. 10-11 and in the Dhva- nyāloka of Ānandavardhana (p. 222), who flourished in the middle of the 9th century.2 As the authorship of one of these verses (śṛngārī cet kavih) is ascribed expressly by Abhinava- gupta to Anandavardhana himself (Abh. Bh. ed. GOS, i, p. 295) we can assume that the Agni borrowed this verse from the Dhvanyāloka. We cannot draw any definite inference from the Agni-purana's omissson of a direct reference to Vamana's teachings, but the definition of the term vakrokti bears some resemblance (341. 33) to Rudrata's novel characterisa- tion of the same figure (ii. 14-16).3 On the other hand, this section of the Agni-purāna (like the Alamkāra section of the Visnu-dharmottara is not cited as an authority in the sphere of Poetics (if we except Bhoja's anonymous appropria-

1 Some of the definitions of poetic figures given by the Agni (E.g. Rūpaka, Utpreksā, Višeșokti, Vibhāvanā, Ākșepa, Aprastuta- praśamsā, Samāsokti and Paryāyokta) occur in almost identical lan- guage in Dandin and Bhāmaha. 2 As the tradition of opinion embodied in the Agni-purāna appears. to have been followed and further developed by Bhoja in his Sarasvatī-k°, its date is presumabiy earlier than the 11th century. It is not suggested that Bhoja built up his elaborate system on the confused verses of the Agni, bnt the tradition of opinion is not essentially different. On a post-Bhoja date see IHQ x, pp. 767-79. 3 This point need not be emphasised; for this idea of vakrokti: was prevalent, as Ratnakara's poem shows, in the 9th century.

THE AGNI-PURĀŅA 99

tions) until we come to the time of Viśvanātha (14th century) who quotes Agni 336. 3-4 (=i. 2) and 337. 7. Whatever may be the date of the Purana as a whole, which is a kind of an ambitious cyclopaedia, incorporating sections on various departments of knowledge, we may, from what has been said, be justified in assigning the Alamkāra-section to a period later than the middle of the 9th century.1 The concept of dhvani is casually included in the figure āksepa,2 after the manner of most old authors, who flourished before that theory came into prominence.3 With regard to the content of the Alamkāra-section, the arrangement of chapters and topics are as follows. Ch. 336 defines and classifies Kāvya. Ch. 337 deals with the topics of Dramaturgy (12 kinds of Rūpaka and Uparūpaka, 5 Artha- prakrtis and 5 Samdhis). Ch. 338 considers the Rasas (with Sthāyi-bhāva, Vibhāva and Anubhāva), the types of Nāyaka-Nāyikā and their qualities. Ch. 339 speaks of four Rītis (Pāñcāli, Gaudī, Vaidarbhī and Lațī) and four Vrttis (Bhāratī, Sāttvati, Kaśikī and Ārabhațī). Ch. 340 is devoted to dancing. Ch. 341 is concerned with four kinds of Abhinaya (Sāttvika, Vācika, Āngika and Āhārya). Ch. 342 defines and classifies Sabdālamkāras, including seven varieties of Citra and sixteen kinds of Prahelikā. Ch. 343 deals with Arthālamkāras ; Ch. 344 with what are called Sabdārthālamkāras (but it

1 P. V. Kane (Hist. of Dharma-śūstra i, pp. 170-73) gives c. 900 A.D. as the date of the Smrti-chapters; to the Alamkāra-section he assigns the same date. Also see IHQ xii. p. 689-90. 2 The word dhvani is also used in the opening verse of this section (336. 1; cf Bhoja i. I); but apparently it alludes to the grammatical word, which reveals the sphota and which is indicated by the same term in the Vākyapadīya. Bhoja, makes the same omission; but of course he was more fully aware of the views of Anandavardhana and his followers. The Purana was probablv aware of the Dhvani- theory as propounded by Anandavardhana (one of whose verses is. appropriated), but it did not apparently subscribe to it 3 Cf Ruyyaka Dp. 3f.

100 SANSKRIT POETICS

includes Ākşepa, Samāsokti and Paryāyokta). Ch. 345-46 are devoted to Guņas and Doșas.

BIBLIOGRAPHY Editions. The Visnu-dharmottara was printed in the Puthi form by Venkatesvara Press, Bombay 1912 .- The Agni- purana has been printed frequently in India, of which the Bibiotheca Indica ed. of Rajendralal Mitra, in 3 vols, 1873, 1876 and 1878, is most useful. Our references are to this edition (ch. 336-346 of this ed .= ch. 337-347 of the Ānandāśrama ed.). Translation into English in two vols. by Manmathanath Datta, Calcutta 1903-4. Other eds. by Ānandāśrama, Poona 1900 and by Venkatesvara Press (in Puthi form), Bombay 1901. The Vahni-purāņa is a different work.

CHAPTER IV

THE DHVANIKĀRA AND ĀNANDAVARDHANA

( 1 ) nandavardhana has been assigned by Bühler and Jacobi to the middle of the ninth century, on the strength of Rāja- taranginī v. 34, which makes him one of the ornaments of the court of Avantivarman (855-84 A. D.) We are pretty certain of the time of Abhinavagupta, Anandavardhana's commen- tator ; for, as he himself states, his Brhati Vrtti on the Iśvara- pratyabhijñā was written in 1015 A. D., while his Krama- stotra was composed in 990-991 A.D. and his Bhairava-stotra or Iśvara-stotra in 992-93. From Abhinavagupta's remarks at the end of his Locana commentary on Uddyotas i and iii of the Dhvanyāloka, it appears that the study of this famous work was traditional in his family, and that his own commentary was composed as a rejoinder to another, called the Candrikā (p. 60), written by one of his predecessors in the same gotra1; and four times in his ° Locana (pp. 123, 174, 185, 215) he discusses or controverts the views of this earlier com- mentator, who is specifically referred to as the Candrikā- kāra at pp. 178 and 185.2 This should certainly allow some generations to lie between Anandavardhana and Abhinava- gupta, and negative completely Pischel's contention that in three passages Abhinavagupta speaks of Anandavardhana as one of his teachers. These passages occur at pp. 37, 183, and 214 of the printed text, but a perusal of them with re-

1 candrikā-kārais tu pathitam ...... ity alam pūrva-vamśyaih saha vivādena babhunā, p. 185: ity alam nija-pūrvaja-sagotraih sākam vivā- dena, p. 123, etc .- The Candrikā-kāra is also cited by Someśvara (p. 55) in his comm. on Mammata. 2 This Candrika is also referred to in a punning verse at the beginning of Mahimabhatta's Vyakti-viveka (i. 5): dhvani-vartmany ati-gahane skhalitam. vūnyāh pade pade sulabham/rabhasena yat pra- vṛttā prakāsakam candrikūdy adrstaiva, on which the commentator remarks: candrikā jyotsnā dhvani-vicūrana-grantho'pi (p. 1).

102 SANSKRIT POETICS

ference to their context will show that the honorific word guru, if it at all refers to Anandavardhana, must refer to him, not literally but figuratively, as parampara-guru, whose work was held in esteem in his family1; or (which is more likely) the reference is to one or other of Abhinavagupta's teachers, such as Bhatta Tauta or Bhattenduraja, the former of whom is cited as asmad-upādhyāyāh or asmad-guravah very often in Abhinava's commentary on Bharata. Again, Kayyata states that he wrote his commentary on Änandavardhana's Devi- śataka (ed. Kāvyamālā, Gucchaka ix) at about 977 A.D., so that by the end of the tenth century Anandavardhana was well enough established in fame to have two such learned com- mentators. Finally, Rājasekhara, who lived about the end of the 9th and the beginning of the 10th century, mentions and cites Ānandavardhana by name in his Kāvya mīmāmsā (p. 16), and this should certainly clear up any doubt as to the authenticity of the date assigned by Kahlana and accepted by Bühler and Jacobi.

( 2 ) The celebrated work on Poetics known as Dhvanyāloka (also called Kāvyāloka or Sahṛdayāloka),2 of which or a part of which Anandavardhana is reputed to be the author, may be distinguished ino two parts, viz. (1) the Kārikā, consisting of verses and treating of dhvani, and (2) the Vrtti, or ex- position, generally in prose with illustrative. verses, of the Kārikā. Now the question has been raised whether the Kārikā and the Vrtti are of the same authorship or should be attribu- ted to different authors.

1 Jacobi, WZKM iv, pp. 237-38. 2 The work is called Kāvyāloka by Abhinava in the concluding verses of Uddyota iit and iv. But in his comm. on Bharata (vii, vol. i, p. 344; xvi. 5, vol. ii, pp. 299-300) he refers to his own comm. on the Dhvanyāloka as Sahrdayāloka-locana .- For a brief summary of the contents of the work see Kane in HSP, pp. 190-91 .- On Dhvanvāloka and the text of the Dhvani-kārikās see S. P. Bhatta-

THE DHVANIKĀRA AND ĀNANDAVARDHANA 103

Abhinavagupta. who is followed in this respect by several later writers on Poetics, carefully distinguishes between the Kārikākāra and the Vrttikāra, by directly opposing them, and also by using the term vrtti-grantha in contradistinction to the kārika1. In three of these passages (pp. 123, 130-1; ch. iv p. 29) Abhinavagupta expressly tries to reconcile the con- flicting views expressed by the Kārikākāra and the Vrttikāra. Bühler2 first drew attention to this point; and Jacobi3, relying on Abhinavagupta's testimony, put forward the sug- gestion that the Dhvanikara, the supposed author of the Kāri- kā, was a different and older writer who should be distinguished from Anandavardhana, the author of the Vrtti. In support of this, it has been pointed out that one does not find complete agreement of opinion between the two parts of the work, although the one is an exposition of the other. On the other hand, it seems that the system as given in its bare outline by the Kārikākāra in his concise verses has been considerably expanded, revised, and modi- fied by the Vrttikāra; and many problems not discussed or even hinted at by the former are elaborately treated of by the latter. In one place, for instance (p. 123), Abhinava- gupta clearly points out that the classification of dhvani according to vastu, alamkāra, and rasādi is not expressly taught in any Kārika ; while in another place in ch. iv, Abhi- navagupta states that the question as to the source of the endless variety of artha in poetry is mentioned by the Vrtti-

1 pp. 1, 59-60, 71, 78, 85, 104, 123, 130-1 ; ch iv pp. 25, 29, 37, 38, 39, 40 in JDL., ix, 1923, Calcutta University. One of these pas- sages from Locana on p. 123 would imply that the earlier com- mentator, the Candrikākara, probably made a similar distinction between the Karikakara and the Vrttikara. For these passages col- lected together, see S. K De in the BSOS i, pt. 4, p. 3 (reprinted in S. K. De, Some Problems of Sanskrit Poetics, Calcutta 1959, pp. 80-90, where the whole question is discussed) and Harichand Sastri op. cit. pp. 86-87,

104 SANSKRIT POETICS

kāra but not touched upon by the Kārikākāra. Indeed, it seems that Anandavardhana in his classical Vrtti attempted to build up a more or less complete system of Poetics upon the loosely joined ideas and materials supplied by the brief Kārikās ; and his success was probably so marvellous that in course of time the Kārikākāra receded to the background, completely overshadowed by the more important figure of his formidable expounder ; and people considered as the Dhvani- kāra not the author of the few memorial verses but the com- mentator Anandavardhana himself, who for the first time fixed the theory in its present form. The term "Dhvanikāra" itself came gradually to be used in the generic sense of "the creator of the Dhvani School", and therefore indiscriminately applied by later writers to Anandavardhana, who might not have been the founder of the system, but who came to receive that credit for having first victoriously introduced it in the struggle of the schools. It is not surprising, therefore, that in the apocryphal verse ascribed to Rājaśekhara in Jahlaņa, Änandavardhana is regard- ed as the founder of the dhvani-theory. Similarly, Samudra- bandha (p. 4), passing in review the five schools of Poetics before Ruyyaka, mentions Anandavardhana as the founder of the fifth or last Dhvani School. This would also explain the two groups of apparently puzzling citations from the Dhvany- äloka met with in the works of later writers, in which they either confuse or identify Anandavardhana with the Dhvani- kāra. On the one hand, we have several Kārikās cited under the name of Anandavardhana, while on the other, several passages which occur in the Vrtti are given under the name of the Dhvanikara. This confusion was so complete in later writers that even in the latter part of the eleventh century Mahimabhatta, who professed to demolish the new theory by his fierce onslaught in the Vyakti-viveka, quotes from the Kārikā and the Vrtti indiscriminately under the generic ap- pellation of the Dhvanikara. In the same way Ksemendra,

THE DHVANIKARA AND ANANDAVARDHANA 105

in the first quarter of the twelfth, make Anandavardhana responsible for Kārikās iii. 24 and i. 4 respectively, while still later writers like Jayaratha, Viśvanātha, Govinda and Kumārasvāmin regard Anandavardhana himself as the Dhvanikāra, to whom the the Kārika, as well as the Vrtti, is indiscriminately attributed1. Mammata, generally a careful writer, distinguishes Anandavardhana from the writer of the Kārikās, whom he styles dhvanikāra or' dhvanikrt (pp. 213 and 214), but in one place (p. 445) he apparently falls into confusion and ascribes to the Dhvanikāra a verse which un- doubtedly belongs to the Vrtti. The question, however, of the differentiation of the Kārikākāra and Vrttikāra cannot yet be taken as finally settled.

( 3 ) If the Dhvanikara, however, is distinguished from Ananda- vardhana, the question naturally arises-who was this. Dhvanikāra, and what date should be assigned to him? Abhinavagupta does not give us any information on this point. Jacobi, in the learned introduction to his translation of the Dhvanyāloka, poses the question very ably without, however, furnishing a precise solution. Sovani's hypothesis2 that the name of the unknown Kārikākāra was Sahrdaya, is hardly convincing; for his grounds for this presumption are that (1) one of the alternative names of the work itself is

1 Kşemendra, Aucitya-vicūra, p. 134 = Dhıvanyūloka iii. 24; Hemacandra, Comm. p. 26 = Vallabhadeva, Subhās° 157 = Dhvanyā- loka i. 4; Govinda Țhakkura, p. 16 = Dhıvanyāloka p. 221; Viśva- nātha, p. 114 = Dhvanyāloka, p. 130; Jayaratha, p 119 = Dlıvanyā- loka p. 111; Kumārasvāmin, p. 64 = Dlıvanyāloka iii. 3. Rājaśekhara (p. 15) in his only quotation from Änandavardhana really cites a Parikara-sloka of the Vrtti at p. 137. Kuntaka, on the other hand, quoting the Prakrit verse talā jaamti (Dhv. p. 62), which is Ananda- vardhana's own (from his lost Visamabūņa-lilā), appears to designate Änandavardhana as the Dhvanikāra (see introd. to Vakrokti-jīvita, 2nd ed. p. xi).

106 SANSKRIT POETICS

Sahrdayāloka and that (2) the use of the words sahrdaya and kavi-sahrdaya at the end of chapter iv of the Dhvanyāloka and in the beginning of Abhinavagupta's commentary is significantly corroborative. It is well known, however, that the word sahrdaya (lit. a man with a heart) is used in innu- merable places in Alamkāra literature, as in the verses in question, to designate a man of taste, a judge of literary beauty, a connoisseur of Rasa. Anandavardhana himself discusses sahrdayatva at some length in his Vrtti (p. 160), and Abhinavagupta arrives at a concise definition of a sahrdaya thus (p. 11): yeşām kāvyānušīlanābhyāsa-vaśad viśadībhūte mano-mukure varņanīya-tanmayī-bhavana-yogyatā, te hṛdaya- samvādabhājah sahrdayāh, a definition which became so much standardized that Hemacandra does not scruple to copy it literally (Comm. p. 3) 1 In the absence of materials it is very difficul to decide the question finally. Jacobi maintains, on the indication of a passage in Abhinava, that this unknown Dhvanikāra was a con- temporary of Manoratha, who is placed by Rāja-tarangiņī (iv. 497 and 671) in the reign of Jayāpīda and his successor Lalitāpīda i.e at the third quarter of the eighth and the first quarter of the ninth century (about 780-813 A.D.); but there are difficulties which seriously stand in the way of our arriving at a definite decision on this point. While discussing the various theories which deny the existence of dhvani, Anandavardhana quotes a verse anonymously with the remark: tathā cānyena krta evātra ślokah, upon which Abhinavagupta in his gloss remarks: tathā cānyena iti. granthakrt-samāna-kāla-bhāvinā manoratha-nāmnā kavinā. If we suppose that by granthakrt Abhinavagupta means Anandavardhana, then Manoratha, who is thus made a contemporary of the latter, lives in the middle or second part of the ninth century, i.e. somewhat

1 Mammata begins his work (p. 10) with a reference to kavi and sahydaya, who are etymologically distinguished by Vidyādhara (p. 21); and both Mammata and Viśvanātha declare that the sahrdaya

THE DHVANIKÄRA AND ĀNANDAVARDHANA 107

later than the date assigned to him by Kahlana, presuming of course that both the Manorathas are identical persons. If, on the other hand, we suppose that granthakrt refers, as Jacobi conjectures, to the anonymous Dhvanikāra, we are confronted with the fresh difficulty that by the term granthakrt Abhinavagupta invariably means Ānandavardhana (pp. 12, 37, 90, etc.). To remove this difficulty we must suppose either (1) that Kahlana is wrong, as Pischel argues, in assign- ing Manoratha to the reign of Jayāpīda and Lalitāpīda, (2) that the two Manorathas were not identical persons. or (3) that Abhinavagupta himself has confused the Kārikākāra with the Vrttikara in a manner not usual with him. As there are no definite means of deciding any one of these equally plausible propositions, the conjecture that the original Dhyanikāra was a contemporary of the Manoratha of Kahlana cannot be taken to have been definitely proved.1

( 4 ) It seems, on the other hand, that the Kārikās date back to an earlier time than the first quarter of the ninth century, in which the Dhvanikara is placed by Jacobi as a contem- porary of Manoratha. The allusion to Manoratha and the apparent discrepancy in Kahlana's statement need not trouble us, nor need we challenge the otherwise trustworthy testimony of Abhinavagupta ; for it is quite reasonable to suppose that the Manoratha under discussion is perhaps a poet who was, Abhinavagupta says, contemporaneous with Anandavardhana, and therefore quite a different person from the well known Manoratha of Kahlana. This is perhaps a much simpler explanation than straining the word granthakrt to mean the Kārikākāra in the face of Abhinavagupta's own distinct indication to the contrary ; and in this way we are not affected in the least by Kahlana's Manoratha, with whom we have nothing to do. If, on the other hand, we place the

1 T. R. Chintamani in JOR ii (1928), pp. 44-47 tries to reconcile the discrepancies.

108 SANSKRIT POETICS

Dhvanikara in the time of Kahlana's Manoratha, this would leave only a bare margin of one or two generations between the Kārikākāra and the Vrttikāra, which does not seem to be enough to make room for a period of scholastic exposition of the subject. But undoubted traces of such activity are preserv- ed to us in the few memorial verses-parikura-ślokas (pp. 34, 130, 137, 147, 163), samgraha-ślokas (pp. 87, 223), samkşepa- ślokas (pp. 44, 74, 243)-incorporated by Ānandavardhana in his Vrtti which itself, therefore, is not likely to be the first of its. kind. These Ślokas are a sort of recapitulation-stanzas which are adduced by the Vrttikāra from unknown sources, some- times to explain the meaning of the Kārikās, but more often to amplify and supplement them. But at the same time we need not suppose a very long intervening period between the original dogmatic formulator of a theory and its first thought- ful expounder ; for it is not necessary that a system should always require a long stretch of time in forming itself. The phenomenon is not unusual that if a literary or intellectual movement is already afoot and is, at it were, in a effervescent state, a few generations, or at most a century, are enough to bring it to the inevitable culmination, or at least to some preliminary completion. If we suppose that a system of dhvani had been in existence at a very early period, we should expect to find. as we do find to a certain extent in the case of the Rasa-theory, its influence work- ing, at least indirectly, on the earlier writers who prece- ded Anandavardhana, although this argument in itself does not carry with it a decisive force. It may be ad- mitted, on the other hand, that the Dhvanikāra apparently shows himself conversant with some theories of rasa, rīti and alamkāra. But this neither proves nor disproves his own antiquity or that of his system, for there is no evidence to show that he was aware of the particular views of Bhamaha, Dandin or Vamana who championed these theories; nor are these writers to be taken, like the Dhvanikāra himself, as the absolute founders of the systems they individually

THE DHVANIKĀRA AND ĀNANDAVARDHANA 109

represent. It only goes to establish that the theory, enunciated by the Dhvanikära, might have existed side by side with these systems, as we find them in the extant works; for it could not have been much later, inasmuch as such a supposition would bring it too near the time of Anandavardhana himself. If the Dhvanikāra was contemporaneous with Dandin or Vämana, he may be placed at most a century earlier than his commentator in the first half of the 8th century.

( 5 ) If Anandavardhana gave the final authoritative shape to the dhvani-theory (only the details of which were worked out by Abhinavagupta and others), the anonymous Dhvani- kāra was not its absolute creator. This is made clear by the first Kārikā, which tells us that the theory was already taught by earlier thinkers, and that it existed even at the time of the Dhvanikāra himself in various forms, handed down, as Anandavardhana explains, in unbroken tradition (paramparayā yah samāmnātah), although it may not have been explained, as Abhinava adds in his gloss, in particular books (avicchin- nena pravāheņa tair etad uktam, vināpi višişta-pustakeşu vivecanad ity abhiprayah, p. 3). This implies without doubt that the school existed from a very early time, but some unknown writer gathered together, summed up, and fixed the theory in a form which obtained considerable literary esteem for his work and the honoured but somewhat vague appella- tion of the Dhvanikara for himself. But his name and fame, in course of time, were eclipsed by those of his great Vrttikāra who succeeded in establishing the theory for all time and to whom posterity began to ascribe, not altogether undeservedly, all the honours of his predecessor, so that one of the latest writers on Alamkāra, Kumārasvāmin :(p. 288), glorifies him with the curious but significant epithet-Dhvanyācārya.

( 6 ) Very little is known of Anandavardhan's personal history. The colophon at the end of ch. iii of his work in the India

110 SANSKRIT POETICS

Office MS calls him noņopādhyāyātmaja, while the colophon to ch. iv gives the form jonopādhyāya. Of these two forms of the name of his father, the former seems to be correct, for referring to Anandavardhana's Devī-śataka, Hemacandra (Comm. p.225) cities its author as nona-sutah śrīmad-ānanda- vardhana-nāmā. Kayyata also, commenting on the last punning verse of the same work, refers to the author as the son of Nona, and mentions his two works, the Vişamabāna- līla and Arjuna-carita, supposed to have been punningly allud- ed to in that verse. Both these works are cited in Anandavar- dhana's Vrtti, by Abhinavagupta (pp. 152, 176, 222), by Hema- candra (pp. 15, 213); and the first work appears to be a Prakrit poem. Anandavardhana himself refers to another work of his own at p. 233, on which Abhinava adds the gloss : granthā- ntara iti viniścaya-fīkāyām dharmottamāyām yā vivrtir amunā granthakrtā krtā. This is apparently a work called Dharmottamā a commentary on the (Pramāna-) viniścaya of Dharmakīrti. Abhinava in Locana iv (p. 31) refers to another work of Anandavardhana's called Tattvaloka, in which the latter is said to have discussed, among other things, the relation between kāvya-naya and śāstra-naya.

( 7 ) '

ABHIN AV AGUPTA Abhinavagupta's fame rests chiefly on his philosophical works on Kashmir Saivaism, but he appears also to. have attained a considerable reputation in the realm of Poetics by his two remarkable commentaries on Bharata and on Ananda- vardhana, called respectively Abhinava-bhāratī and Kāvyāloka- locana. Since the Locana is cited several times by name, it seems to have been written before the other commentary. He also cites in his Locana (p. 179, also p, 29) another commentary (vivarana) of his own, now lost, on the Kāvya-kautuka by one of his teachers (asmad-upādhyāya) Bhatta Tauta. Nothing is.

ABHINAVAGUPTA 111

known of this Bhatta Tauta (also called Bhatta Tota) ; but it appears from its fourth introductory verse that Abhinava's commentary on Bharata was inspired by this teacher, who is cited there very frequently, just as his Locana was inspired by his other teacher Bhattenduraja. The exact scope and extent of Tauta's lost work is not known, but from Abhinavagupta's. reference (pp. 187, 275, 310) it appears to have dealt with the theories of poetry in general and Rasa in particular, and explained some relevant passages from the Nātya- śāstra.1 The Kāvya-kautuka is also referred to in the anonymous. commentary on the Vyakti-viveka (p. 13); and Hemacandra. (p. 316) quotes three verses from Bhatta Tauta in his text and reproduces (p. 59) in his commentary (appropriating the passage directly from Abinava on Bharata) an opinion of this teacher in connexion with the theory of Rasa. Ksemendra in his Aucitya-vicāra (under śl. 35) attributes to Tauta a frag- ment of a verse which is given in full but anonymously by Hemacandra (p. 3)2. Tauta is quoted also by Māņikyacandra (at p. 5), by Śrīdhara and by Candīdāsa in their respective commentaries on Mammata. Bhatta Tauta, together with Bhattenduraja, who is ex- travagantly praised in the Locana and whose relation to Abhinava we have already discussed, were probably his preceptors in Kāvya and Alamkāra. His references to his. instructors in philosophy, like Siddhicela3, Laksmanagupta and others, in his philosophical works possess no interest for us ; but it may be noted that Abinava, in his Pratyabhijña-vimarśinī Laghu-vrtti refers to Utpala as his parama-guru, the teacher

1 See above under commentators on Bharata p. 33, 101. Someś- vara (Comm. on Mammata p. 55) appears to have seen the work, as well as Abhinava's comm. on it (tac ca Bhatta-Totena Kāvya -- kautuke, Abhinavaguptaś ca tad-vțttau nirņitam). 2 This verse is ascribed, perhaps wrongly, to Māmaha (or Bhā- maha ?) in Kāma-dhenu on Vāmana, p. 4, ed, Benares. 3 Cited in Locana, concluding verses.

112 SANSKRIT POETICS

of his teacher. This description of Utpala (see above p. 33) is repeated in his Locana (p. 30), where Abhinava discusses the term pratyabhijña, used in the text (i. 8), and refers to what is said on this point by Utpala. This Utpala is well known in the history of Kashmir Saivaism as the author of the Iśvara- pratyabhijñā (on whose Sūtra and Vrtti Abhinava wrote a Laghu Vrtti and a Brhati Vrtti respectively), and is assigned by Bühler (op. cit p. 79) to the first half of the 10th century. From what Abhinava himself says in his numerous works on Kashmir Saivaism, we may indicate the line of spiritual succession (guru-paramparā) thus: Somānanda- Utpala-Laksmanagupta-Abhinavagupta; Somānanda being probably a pupil of Vasugupta who is taken as the earliest founder of the Pratyabhijña-sastra. The guru-parampara in his study of the Tantras may be given thus: Sumatinatha- Somadeva-Śambhunātha-Abhinavagupta. In the concluding portion of his Parātrimsikā-vivaraņa, Abhinava gives us an interesting personal and genealogical account, in which he tells us that he was son of Kāśmīraka Cukhala1 and grandson of Varähagupta, and had a brother named Manorathagupta. Abhinava's date is easily gathered from his relation to Utpala and Anandavardhana, as well as from his own dating of some of his works. As we have mentioned above, his Krama-stotra was composed in 990-91 A. D. and his Bhairava- or Iśvara-stotra in 992-93 A.D, while his Brhati Vrtti on Utpala's Pratyabhijña bears the date 1015 A.D. We can therefore, place him with certainty at the last quarter of the 10th and the first quarter of the 11th century2.

1 Bühler's MS has kāśmīraka viculaka (op. cit. p. clv) as well as °cukhala (p. clvii); the real name appears to bave been Narasimha- gupta, while his mother's name was Vimala 2 For the woiks of Abhinavagupta see K C. Pandey, Abhinava- gupta, ChSS, Benares 1935 pp. 122-24; V. Raghavan in JOR, xiv pp. 318-20 and New Cat. Cat. i, pp. 224-26. On Writers quoted in Abhinavabharatī see V. Raghavan in JOR vi pp. 153-62.

ABHINAVAGUPTA 113

BIBLIOGRAPHY

Ānandavardhana Edition and Traaslation. (i) With the Kārika and Locana ed. Durgaprasad and K. P. Parab, in Kāvyamālā 25, Nir. Sag. Press Bombay 1890, 1911. (ii) With Locana of Abhinavagupta and Kaumudī of Udayottunga, ed. Kuppu- swami Sastri; 1st Uddyota only, Madras 1944 .- Transla- tion into German with an introd. by Hermann Jacobi in ZDMG lvi-lvii, 1902-03 (reprinted separately, Leipzig 1903). Translated into English, i-ii only, by K. Rama Pisharoti in Indian Thought ix-x (1917-18); not completed. Our references are to the Kāvyamālā ed. of 1911. The text has been printed with modern commentaries (none of these utilise any fresh MS, and can be called a critical ed.)-(i) by Madhusudan Misra with his own Avadhāna Com. Calcutta 1939. (ii) by Badarinath Sarma with. his own Dīdhiti Com. Haridas Sansk. Series, Benaras 1937, etc.

Abhinavagupta Editions. (1) In Kāvyamālā 25, 1890, 1911 with the text (first three uddyotas only) as above. (2) Fourth uddyota without the text by the present writer in the Journal of the Department of Letters, Calcutta Univ., vol. ix, 1923 (reprinted in S, K. De, Some Problems pp. 236-267). The full title is Kāvyāloka-locana. (3) First Uddyota with the text, ed. Kuppusvami Sastri, as above. (4) with the text, ed. Pattabhirama Sastri, Kashi Sans. Series, Benares 1940. Commentaries. (1) Locana-vyākhyā-kaumudī by Parame- śvarācārya. Oppert 2694. (2) Añjana, anonymous in Madras Cat. xii, 12895, extract (only on the first Uddyota). The author's name appears to be Dāśarathi (K, Kunjunni Raja, Contribution of Kerala, Madras 1958, p. 244). The author erroneously describes Bhattendurāja as the parama- guru of Abhinava. (3) Locana-kaumudī (on Uddyota i only)

114 SANSKRIT POETICS

by Udayottunga or Udaya-rāja from Kerala, publ. as above, Madras 1944 by Kappuswami Sastri. The author probably flourished in the second half of the 15th century (K. R. Pisharoti in Journal of the Ganganath Jha R. Institute, i, pp. 445 52). He wrote a Dūta-kāvya named Mayūra-samdeśa (ed. C. Kunhan Raja, Poona Orient. Series 1944) ; on which see K. Kunjunni Raja, op. cit. p. 228. The Candrika comm. on Anandavardhana, referred to by Abhinava and others, is apparently lost. On Abhinava-bhāratī on Bharata see above under Bharata. No MS has been discovered of Abhinava's Kāvya-kautuka- vivaraņa, or of the Kāvya-kautaka itself. A list of citations of poetical and Alamkāra works by Anandavardhana and Abhinavagupta, is given by Jacob in JRAS, 1897, pp. 290 and 297. Only omit the name of Dhanika given by Jacob as an author quoted by Abhinava.

CHAPTER V

FROM RĀJASEKHARA TO MAHIMABHAȚȚA

RĀJAŚEKHARA

( 1 ) Rājaśekhara, son of mahāmantrin Durduka or Duhika and Sīlāvatī1 and great-grandson of the poet Akālajalada of the Yāyāvara family, is better known as a poet and dramatist than as a writer on Poetics. In his Bāla-rāmāyaņa i. 12, Rāja sekhara describes himself as the author of six works which must have existed even before this presumably early produc- tion of his.2 It is not known whether his other three well known dramas belong to this period, but it appears that he probably composed more than six works. Hemacandra (Comm. p. 335) cites a work of Rājaśekhara's, entitled Hara-vilāsa, as an example of a poem which bears the name of its author (sva-nāmānktā), and quotes two verses from the same (Comm. pp. 334-335) ; from which Ujjvaladatta also (ii. 28) gives a half-verse. In the Kāvya-mīmāmsa, again (xvii p. 98), there is a reference to another work of his own, called Bhuvana- kośa, for information on general geography.3 Ujjvaladatta (ii. 76) also quotes a line from Rājaśekhara on the synonyms

1 Bāla-rāmā° ad i. 7, 13 ; Bāla-bhā ad i. 8; Viddha-śāla° ad i. 5. Rajaśekhara calls himself a Kavirāja (Karpūra-mañj° i. 9; Viddha-śāla i. 5), which is, according to Kāv.mīm., the seventh, out of the ten stages, of poetic skill, one degree higher than that of a mahākavi. 2 In Karpūra-mañj° i. 9 we are told that Rājaśekhara began his. career as a bāla-kavi, so called apparently from his two works Bāla-rāmā° and Bāla-bhā° ; while in this Prakrit drama, as well as in Viddha-śāla°, he appears to have attained the distinction of a Kavirāja. 3 Kane (HSP, p. 207-8) believes that the Bhuvana-kośa was not a. separate work but formed a part of the Kav. mīm.

116 SANSKRIT POETICS

of Siva, which, if not occurring in the Hara-vilāsa, was pro- bably taken from an unknown lexicon by him. In the two anthologies of Vallabhadeva and Sarngadhara, we get a considerable number of verses ascribed to Rāja- śekhara. Of these, about 24 have been identified by Sten Konow in Rājaśekhara's four plays,1 but about 10 have not yet been traced in any of his known works, nor are they to be found in his Kāvya-mīmāmsā. These untraced verses, including most of the memorial verses on poets, probably belong to another and younger Rājaśekhara.2 There can be hardly any doubt that the Kāvya-mīmāmsā should be ascribed to the dramatist Räjasekhara, although it is not mentioned in these enumerations of Rājaśekhara's works.3 Our author gives his own name at the end of the first chapter of this work as yāyāvarīya Rājaśekhara, which agrees with the description given in the dramas and which makes later writers cite our author simply as yāyāvara4. The opinions of the yayavara family, to which he belonged and in which

1 ed. Karpūra-mañj° pp. 189-91. 2 This other Rajasekhara may or may not be the Jaina Rāja- śekhara, author of the Prabandha-kośa (1348 A.D.). Rice 282 men- tions a work called Karpūra-rasa-mañjarī by Bālakavi, which apparent ly refers to Rajasekhara and his well known Prakrit drama, and not to any work on Alamkāra. 3 Aufrecht notes (ABod 135a) that the Kāv. mim. is cited by name by Samkara in his commentary on Sakuntalā. 4 Bāla-bhā° 1. 6. 13; Viddha-śāla° i. 5; and Dhanapāla in Tilaka-mañj° ed. Kāvyamālā 85, 1903, śl. 33, and Māņikyacaņdra in Samketa comm. (ed. Mysore) p. 308. Also Hemacandra (p. 235) and Someśvara (ed. Jodhpur 1959, p. 224, yāyāvarīya). Nārāyaņa Dīkșita on Viddha-śāla. i. 5 quotes Devala to show that yāyāvara means a kind of a householder (dvividho grhasthah, yāyāvaraḥ śālīnaś ca; see Mitākșarā on Yajña° i. 128) according to which yāyūvara means a particular class of Brahmans who lead a plain life and do not accept gifts etc. Rājasekhara's wife came from Ksatriya family. but Anuloma marriage was permitted. See Kane, Hist of Dhamma-śāstra ii. pp. 641-42.

RĀJASEKHARA 117

were born poets and scholars like Surānanda,3 Akālajalada,2 Tarala3 and Kavirāja, are cited frequently by him under the general designation yāyāvarīya,4 as well as under the individual names of these famous members of the family who are enumerated in the Bāla-rāmāyaņa i. 13 and elsewhere. He also quotes with respect the views of his wife Avantisundari of the Cāhuna family (pp. 20, 46, 57), for whose pleasure, we know, he wrote his Karpūra-mañjarī (i. 11) and who seems to have been an accomplished authoress. The present work does not also omit a display the author's love for Prakrit dialects (pp. 34, 51) as well as his knowledge of geography (ch. xvii), of which he gives ample evidence in Act x of the Bāla-rāmāyana. These and other details, on which we need not dwell any further, show that our Rajasekhara is no other than the well known dramatist.5 ( 2 ) The published text of the Kāvya-mīmāmsā in eighteen chapters is apparently the first part of a projected extensive volume, of which a general summary or scheme is given in

1 An ornament of the country of the Cedis (cedi-mandala-nan- dana, Jalhaņa's Sūkti-muktāvali 88-89, p. 47). His patron Raņavigraha is supposed oy Bhandarkar (Report, 1887-91, p. xix) to have been the brother-in-law of Krsna II of the Rāstrakūta dynasty, whose dates range from 875 to 911 A.D. Quoted also in Kāv. mīm. p. 75. 2 Rājasekhara's great-grandfather. This is not his real name but sobriquet derived from the expression in Sr. Paddhati 777 = Subhās° 843 (dākșiņātya). Famous for his poetical jems, some of which were plagiarised by Kādambarīrāma (Jahlaņa Sūkti-muktāvali 83-84 (p. 46). Called mahārāstra-cūdāmaņi in Bāla-rāmā° i. 13; also see Viddha- śāla° i. 5. 3 Cited by Jahlana. Author of a work, called probably Suvarna- bandha. 4 pp 2, 3, 4, 10, 11, 13, 14, 16. 17, 20, 23, 26, 27, 30, 31, 35, 42, 43, 46, 50, 56, 58, 61, 62, 63, 78, 90. 91, 94, 99, 100. 5 See S. K. De, Hist. of Kāvya Lit. ch. viii for Rājaśekhara anđ his dramas .- In Karp. m. (on i. 6) he describes himself as sarva-bhāșā- catura. As a poet he claims that in his former births he was Val- mīki, Mentha and Bhavabhūti!

118 SANSKRIT POETICS

the first chapter ; this fact is also indicated by occasional remarks (like rītayas tisras tās tu purastāt p. 10 and tam aupaişadike vaksyamah p. 11) relating to topics to be dealt with in other succeeding parts. If the complete work, as projected, consisted of eighteen adhikaranas, we have now only one part surviving on the preliminary topic of kavi- rahasya alone.1 Keśava Miśra (pp. 32, 67) quotes three verses from an Alamkāra work by Rājaśekhara, which, if they be- long to our author, were apparently taken, as their contents indicate, from some lost chapters on ubhayālamkārika and vainodika respectively. The popularity of the Kāvya-mīmāmsa with later writers is indicated by the extensive use made of it by Ksemendra, Bhoja, Hemacandra and the younger Vāgbhata. Hemacandra, for instance, literally copies long passages from chs. viii, ix, xiii-xviii ; while Vägbhata borrows the same (as well as other) portions either directly from the same source or indirectly through Hemacandra.2 Rājaśekhara himself is indebted to many old writers and cites directly the opinions of Medhāvirudra (p. 12), Udbhata and Audbhatas (pp. 22, 44), Vāmana and Vāmanīyas (pp. 14, 20), Rudrata (p. 31), Mangala (pp. 11, 14, 16, 20) and Ānanda (p. 16), besides unnamed authors who are cited under the general designation ācārya.3 We also find the name of Āparājiti. One Aparajita is4 quoted in Subhāșitāvalī 1024, and mentioned as a contemporary poet and author of Mrganka- lekhā-kathā in Karpūra-mañjarī, ad i. 8. Rājaśekhara also

1 Cf. introd. to Kāv. mīm. p. xvii-xviii. 2 A comparative table is given of these wholesale borrowings at the end of the notes in the Gaekwad ed. of the text. 3 pp. 3, 9, 13, 16, 20, 23, 30, 35, 50, 51, 56, 57, 58, 61; 62, 78, 94, 99. 4 One Aparājita-rakșita is quoted in Kavīndra-vacana but as his name implies he was probably a Buddhist, and is possibly not identi- cal with Rajasekhara's contemporary Āparajitī. V. Raghavan (JOR. vi. p. 170) thinks that this Aparajiti of Rājasekhara is none else than Lollata. See above p. 37, for 1.

RĀJASEKHARA 119

mentions Surananda, one of his ancestors, his wife Avanti- sundarī,1 Pālyakīrti (p. 46), Śyāmadeva (pp. 11, 13, 17) and Vākpatirāja2 (p. 62), all of whom, as the citations show, seem to have expressed some opinion on the topics under discussion.

( 3 ) The date of Rajasekhara has been settled with some exactitude. We learn from his four extant plays3 that his ancestors lived in Mahārastra, and he himself spent much of his life in the midland as a teacher (upādhyāya) to a king named Mahendrapāla (otherwise known as Nirbhaya or Nibbhaya) ; Rājaśekhara was also patronised by his son and successor Mahīpāla.4 We also understand that one of his plays, the Bāla-bhārata, was performed at a place called Mahodaya, for which he shows a partiality also in his Kāvya-mīmāmsā (p. 94). Fleet has shown5 that this Mahīpala should be identified with the Mahīpāla of Asni inscription, dated 917 A. D., and he agrees with Pischel6 that Mahodaya is another name for Kanyakubja or Kanauj,7 with which place this king, as well as Mahendrapäla, is connected in the

1 A Sanskrit verse of Avantisundari is quoted in Kav. mīm. p. 46 ; but no work of hers has come down to us. Hemacandra in his Desī-nāma-mālā quotes three Prakrit verses of Avantisundarī (i. 81. and i. 157). 2 This Vākpatirāja (Bapai-rāa) is apparently the author of the Gaudavaho (middle of the 8th century; Kahlana iv. 144) and must be distinguished from Muñja-vākpatirāja, the 7th Paramāra king of Malava, who reigned from 947 to 995 A.D. See below under Dhanañjaya. 3 Viddha-śāla° i. 6; Bāla-rāmā° i. 5; Bāla-bhāra° i. 7, 11 ; Kar- pūra-mañj° i. 5, 9. 4 Bāla-bhāra° i. 9. 5 IA xvi. 175-78. 6 GgA, 1883, pp. 1217f. 7 Cf. Bāla-rāmā° x ad 87, 89, 90. Rājaśekhara's partiality for Mahodaya is also apparent in our text at pp. 8, 94.

120 SANSKRIT POETICS

Siyadoni inscription1. It has also been shown by Aufrecht2 and Pischel3 that Mahendrapala, whose dates appear to be 903-07 A. D. from Kielhorn's summing up of the names of the four sovereigns of Kanauj as presented by the Siyadoni inscription, went also by the biruda of Nirbhara or Nirbhaya (Nibbhara or Nibbhaya in the Prakrit form), a fact of which Fleet seems to have been unaware.4 Rājaśekhara appears to have become at some time of his life a protégé of Yuvarāja who has been identified with Yuvarāja I Kayūravarsa, the Kalacuri ruler of Tripuri (in the kingdom of Cedi), where the poet's relative Surananda had migrated. From these evidences, it is clear that Rājasekhara must have flourished in the beginning of the 10th century, and probably also lived to- wards the end of the 9th. This is also supported by the fact that the latest writers quoted by Rājasekhara are the Kashmirian Ratnākara and Anandavardhana, who belong to the middle or second half of the 9th century, while the ear- liest writer to mention Rājaśekhara appears to be the Jaina Somadeva, whose Yaśastilaka is dated 960 A. D.5 About the same time Abhinavagupta (in his Comm. on Bharata) expressly mentions Karpūra-mañjarī as a Sattaka, as well as Bāla-rāmāyaņa. In the Anthologies Rājaśekhara is extensively quoted from the 12th century onward.

1 EI i. 170 f. 2 ZDMG xxvii (on Sārngadhara-Paddhati). 3 op. cit. p. 1221. 4 Fleet further shows (op. cit. p. 175f) that this Mahendrapala must not be taken, as Peterson and Durgaprasada are inclined to take him, to be identical with the feudatory Mahendrapala, whose inscription from . Dighwa-Dubauli, dated 761-62 A.D., he has edited in IA xv. 105, and who is distinct from the pupil of Rajasekhara. 5 For other details about Rājasekhara see Sten Konow's edition of Karpura-manj' (Harvard Orient. Series 4, 1901). 175f, which gives a full bibliography .- It is difficult to summarise the diverse and somewhat diffuse contents of the Kāvya-mīmāmsā; but a general résumé will be found below in vol. ii. ch. ix (3). A good summary will be found also in Kane, op. cit pp. 199-201. 6 See F. W. Thomas, introd. to Kavindra-vacana, where most of these citations are collected together.

DHANAÑJAYA AND DHANIKA 121

BIBLIOGRAPHY Edition. With introd. and notes by C. D. Dalal in Gaekwad Orient. Series, Baroda 1916. The text has been printed with his own Comm. by Narayan Sastri Khiste in the Chowkhamba Sanskrit Series, Benares 1931; again, by Madhusudana Misra with his own Comm. and Hindi trs. in the same series 1931, 1932. Our references are to the Baroda ed .- Translated into French by Nadine Stchoupak and Louis Renou, Paris 1946.

DHANAÑJAYA AND DHANIKA

(1 ) The date of Dhanañjaya may be taken as approximately settled at the last quarter of the 10th century A. D. The author informs us (iv. 80) that he was son of Visnu and that he flourished in the circle of distinguished literary men surround- ing king Muñja, who himself seems to have been a man of taste and learning, as well as a patron of letters. We should not, with Peterson1, confound this Muñja, better known as Muñja-vākpatirāja, with Vākpatirāja (or Bappai-rāa), the author of Gaudavaho, who lived in the first half of the 8th century under king Yasovarman of Kanauj and is referred to by Kahlaņa (iv. 144) and Rājaśekhara. Our Muñja appears to. be the seventh ruler of the Paramara dynasty of Malava, who, as his own inscriptions record2, came to the throne in 974 A.D., succeeding his father Harşadeva Sīyaka, and reigned till about 995 A.D., when he was defeated, imprisoned and exe- cuted, as the Cālukya inscriptions attest3, by Cālukya

1 introd. to Subhā,4° p. 115. 2 Arch. Survey. Western Ind. iii. 100 = 1A vi. 48-51; IA xiv. 159-60. See Bühler, Das Navasāhasānka-carita (transl. in IA xxxvi, pp. 149-172), 1888, p. 116f. 3 IA xii. 270, xvi. 18, 23, xxi. 167-68 ; EI ii. 212 f. All the references (regarding Muñja's date) are collected together in Haas's introd. to his ed. of the Daśa-rūpaka (q. v.).

122 SANSKRIT POETICS

Tailapa II. Besides being known as Vākpatirāja1, owing perhaps to the fact that he was himself a poet, he had several other birudas, such as Amoghavarsa, Prthvī-vallabha, and Srī-vallabha; and one of his inscriptions calls him Utpalarāja2; a fact, overlooked by the editors of the Kāvyamālā Series (Gucchaka i, p. 131), made them confound him with Utpala, the Saiva philosopher of Kashmir, who was Abhinavagupta's parama-guru. This ruler is mentioned by Sambhu3, as well as by Padmagupta4, as 'a friend of poets' (kavi-bāndhava or kavi-mitra) ; and Bhoja, his nephew and successor, appears to have inherited these traits of his character. ( 2 ) The Daśa-rūpaka of Dhanañjaya, in its treatment of Dramaturgy, is apparently based on the time-honoured authority of Bharata; but as Bharata's huge compendium, both from the practical as well as theoretical point of view, is discursive and cumbersome with its load of histrionic and other matters, Dhanañjaya attempts to sift the mass of details, and, limiting himself only to Dramaturgy, restates the general principles in the form of a practical, condensed and system- atic manual. These features of the new contribution ap-

1 Dhanika quotes (on iv. 54-55) one of Muñja's verses twice, cit- ing him in the first instance as Muñja and in the other case as Vāk- patirāja-deva; while Dhanapala in his Tilaka-mañjarī uses both the names with respect to the same person. One of Muñja's descendants, Arjunavarman, who ruled in the beginning of the 13th century, re- produces one of Muñja's stanzas, with the remark that it was com- posed by one of his ancestors "Muñja, whose other name was Vāk- patirāja " (Comm. on Amaru-śataka, ed. Kāvyamālā 1916, p. 23). This verse is attributed to Muñja also by Jahlaņa p. 199. 2 Ksemendra quotes verses from Muñja in his three works (Aucit. vic. under śl. 16; Kavi-kantha° under ii. 1; Suyrtta-til. un- .der ii. 6) referring to him as Utpalarāja. See also Sārngadhara (126 vākpatirājasya; 1017 utpalarājasya), Vallabhadeva (3414 śrīharşadevāt- maja-vākpatirājasya), and Jahlaņa pp. 63 and 199 (Šrī-muñjasya). 3 Rājendra-karņa-pūra, śl 17, 36. 4 Navasāhasānka° i. 7, 8; ii. 93.

DHANANJAYA AND . DHANIKA 123

parently obtained for it such reputation and currency that in course of time it seems to have superseded not only all other treatises on the subject but also the basic work of Bharata himself. Viśvanātha, for instance, refers now and then to Bharata and gives one or two (mostly conventional) quotations from Nātya-śāstra ; but in the main he bases his treatment of dramaturgic topics on Dhanañjaya; while Vidyānātha admits, in the nātaka-prakarana of his own work, his indebted- ness to the latter, with the remark eśā prakriyā daśarūpokta- rīty anusāreņa (p. 131). The Daśa-rupaka, consisting of four chapters called Prakā- śas, deals almost entirely with the topics of dramaturgy, but the fourth and last Prakaśa contains a treatment of the theory of Rasa. The first Prakāśa distinguishes Nrtya, defines the five Artha-prakrtis and the Samdhis (with their Angas), and concludes with definitions of Viskambhaka, Praveśaka and other dramatic devices. · The second Prakāśa is devoted to the topic of Nāyaka and Nāyikā, their characteristics, their adjuncts, and considers the four dramatic Vrttis and their Angas. The third Prakasa is concerned with the Prologue and other requisites of the ten kinds of Rupaka. In the fourth Prakāśa we have an exposition of its peculiar theory of Rasa in which, not the relation of Vyangya-vyañjaka but that of Bhāvya-bhāvaka is posited, after Bhatta Nāyaka, between Rasa and Kāvya.

( 3 ) Dhanika, also described as son of Vișnu, and author of the Avaloka commentary on Dhanañjaya's work, was probably one of Dhanañjaya's numerous illustrious contemporaries ; for he may be assigned to the same period. Dhanika quotes from Padmagupta (also known as Parimala)1, who wrote about 995 A.D., as well as from Muñja, and is quoted in his turn by Bhoja in his Sarasvatī-kanthabharana in the first half of the

1 on ii. 37b = Navasāhasāńka° vi. 42.

124 SANSKRIT POETICS

11th century. He is also described in one of the MSS1 as hold- ing the office of mahāsādhyapāla under king Utpalarāja, who is apparently our Muñja-vākpati, the patron of Dhanañjaya. The suggestion that the author and the commentator of Daśa-rūpaka are one and the same person, chiefly on the ground of the apparent similarity of names and identity of patronymic, as well as the inadvertant attribution of a verse of Dhanañjaya's to Dhanika in some later works like the Sāhitya-darpana (ad vi. 64a=Daśa-rūpaka iii. 29), need not be seriously considered.2 Jacobi, however, supports this suggestion3 by pointing out that there is no separate mangalā- carana to the commentary. This hypothesis, however, cannot altogether get rid of the fact that Dhanañjaya and Dhanika are indeed distinguished by some later writers. For instance, Vidyānatha, in his numerous references to the Daśa-rūpaka, cites the Karika-verses and never from the commentary,4 although his commentator, Kumārasvamin, falls in one place5 (p. 29) into the error of attributing one of Dhanañjaya's

1 Wilson, Select Specimens, 3rd ed. I, xx, xxi, endorsed by Hall p. 3 notes. It is curious to note that Dhanika (on iv. 23, ed. Parab) quotes nidrardha° which occurs in the Caura-pañcasikā (ed. Solf no. 36) attributed to Bihlana, but this anonymous quotation (which also occurs in Kuntaka) does not of itself place Dhanika later than the middle of the 11th century, the date of Bihlana; for the authen- city of the verse is not beyond question, as it is attributed to Kalaśaka in Subhās° 1280 and Jahlaņa p. 152, and Bihlana's authorship is open to question. It is not safe. therefore, to base any chorono- logical conclusion on this quotation. Haas has not noticed the verse at all. 2 Haas (Pref. to ed. Daśa-rūpaka xxxiv) is inaccurate in stating that there is in the commentary "a number of indications of a difference of authorship," and in support of this he cites ii. 20b-21a, iii. 32b, iv 43c. Jacobi elaborately shows (GgA, 1913, pp. 304f) that Haas has entirely misunderstood these passages. 3 Op. cit. p. 303. Also Lévi in JA, 1886, p. 221. 4 p. 46. 101, 102, 104, 105, 114, 124, 131, 219, 221, 228. 5 In other passages the citation appears to be correct, pp. 47,

DHANAÑJAYA AND DHANIKA 125

verses (ii. 23b) to Dhanika. It may also be urged that a mangala-verse to the Avaloka, occurring in one of the MSS, is rejected by Hall as spurious, chiefly on the ground that its style is "too pedestrian for so ornate a stylist as Dhanika" (p. 4 note). This 'pedestrian' stanza is apparently the same as that which occurs at the outset of Aufrecht's Bodlein MS, noticed by him in his Bod. Cat. 203a. On the other hand, the absence of the mangala-verse need not in itself be taken as decisive; for while Mammata has no separate mangala- verse to his Vrtti, we find them in Vāmana and Ruyyaka. Särngadhara in his anthology attributes to Dhanika several verses (3417 and 3973) which the latter gives as his own in his commentary (on iv. 3a and ii. 10a). If, therefore, we suppose, as it is more likely, that the author and the commen- tator were not identical, then Dhanika may be taken as a brother of Dhanañjaya (a supposition which explains the apparent similarity of names and identity of patronymic), who collaborated in the production of the work1 by writing the commentary. From the Avaloka we learn that its author composed poems in Sanskrit and in Prakrit, and also wrote a treatise, entitled Kāvya-nirņaya (on iv. 35 ; seven verses quoted) which alluded to the Dhvanyaloka and apparently dealt with the general topics of Poetics. For other less known commentaries on Daśa-rūpaka, see Bibliography given below.2

128, 130, 221, 233, 235, 259. Ranganātha on Vikramor° (about 1656 A.D .; ed. N. S. P. 1914 p. 31) falls into the same mistake. Mallinātha on Kumāra° i. 4 and Sisu° vii. 11 quotes Daśa-rūpaka .correctly (ii. 36b and ii. 24a). 1 This supposition does not militate against the passage (on iv. 33), referred to by Jacobi, in which the commentator intimately identifies himself with the author saying asmabhih .. .... nişidhyate, meaning that the prohibition is made both by his author and himself. 2 For later dramaturgic works, e.g. Vasantarājīya of Kumāragiri, see under Minor Writers below ch. x.

126 SANSKRIT POETICS

BIBLIOGRAPHY

Dhanañjaya

Editions. (1) By Fitz-edward Hall (with Avaloka), Bibl. Indica 1861-65. (2) by Jivānanda (a mere reprint of the above) Calcutta 1897. (3) by K. P. Parab, NSP, Bombay 1897, 1917 (with Avaloka), 4th ed 1928 .- Transl. into English, with transliterated text. introd. and notes, by G. C. O. Haas in Columbia Univ. Indo-Iranian Series, New York 1912 (not very reliable, but exhaustive introd. and index). Our references are to Hall's ed. On Haas's ed, see criticisms of Jacobi in GgA, 1913, p. 302f, anđ Barnett in JRAS, 1913, p. 190f.

Dhanika Editions. Printed in Hall's and Parab's editions, with the text. Dhanika's Kāvya-nirņaya is probably lost,

Other Commentaries on Dhanañjaya and Dhanika (1) Comm. by Nrsimhabhatta. MS in Govt. Orient. MSS Library, Madras (see BSOS, iv, 1926, p. 280). It is really a Laghu-Țīkā on Dhanika's Comm. (Dasarūpasya yā vyākhyā Dhanikena samāhitā tasya Bhatta-Nrsimhena laghu-fīkā vidhīyate). It is interesting to note from this. verse that Dhanika is mentioned as a commentator on Daśa-rūpaka, thus recording the tradition that the com- mentator Dhanika was different from Dhanañjaya, author of the Dasa-rūpaka. Bhatta Nrsimha also commented on Bhoja's Sarasvatī-kaņthābharaņa. (2) Țīka by Devapāņi (cited by Ranganātha on Vikra- morvaśīya ed. N. S P. 1904, pp. 6, 31 ; cf AFI 444 and ABod 135b), No MSS discovered. This is the author who is wrongly called Pani by Wilson (Select Specimens) and Aufrecht. Being anterior to Ranganatha, his date should be earlier than 1656 A.D. Ranganātha also refers to a Sāhasānkīya-țīkā in the same context (p. 31).

KUNTAKA 127

(3) °Paddhati by Kuravirāma (MS in Hultzch 554; only three pages) is not a commentary on the Daśa-rūpaka, as Hultzsch's entry would imply. It is an independent work on Dramaturgy consisting of 110 verses. See Madras Trm II, A, 820 (c), Kuravirāma is a modern but fertile South Indian commentator who lived at the court of Zemindars of Kärvetinagaram in North Arcot District, and wrote comms. also on two well known poems, Campū- bhārata of Anantabhatta and Viśvaguņādarśa of Venkața. He mentions in his comm, on the last-named poem a commentary by himself on Appayya's Kuvalay°, as well on Dhanañjaya. See Hultzsch i, p. xi. (4) Comm. by Bahurūpa Miśra, for an account of which see V. Raghavan in JOR, Madras, viii, pp. 321-34. As Bahurūpa quotes from Bhoja's Śrngāra-prakāśa and Sāradātanaya's Bhāva-prakāśana he must be later than 1250 A.D.

KUNTAKA

(1) Kuntaka is better known in Alamkāra Jiterature under the descriptive designation of the Vakrokti-jīvita-kāra1 from the peculiar name of his work Vakrokti-jivita, which itself is- so called because of its central theory that vakrokti is the 'soul' or essence of poetry. The work had been known only through quotations and references until it was edited from two imperfect MSS by the present writer. Kuntaka's date2 is fixed approximately by his quotation from the dramatist Rājasekhara, on the one hand, and by

1 Ruyyaka, ed. Kāvymālā p. 8. with Jayaratha (also pp. 12,. 150 etc,) and Samudrabandha thereon (p. 4); Viśvanātha ed. Durga- prasad p. 14; Kāma-dhenu on Vāmana I. 1. 1, ed. Benares p. 6, etc. 2 This question has been dealt with in detail in the introd. to S. K. De's editio princeps of the Vakrokti-jivita, and is only. briefly referred to here.

128 SANSKRIT POETICS

Mahimabhatta's citation of Kuntaka and his work, on the other. Mahimabhatta flourished, as we shall see, towards the end of the 11th century. We may, therefore, place Kuntaka between the middle of the 10th and the middle of the 11th. As this date falls in with the known dates of Abhinavagupta (whose latest date is 1015 A.D.), we may take Kuntaka as a contemporary of this commentator on Anandavardhana. Although Abhinava refers to various views about vakrokti held before his time, it is remarkable that he never alludes to the Vakrokti-jīvita-kāra who, as his title rājānaka indicates, was probably a Kashmirian, and whose work, if written before Abhinava's time, ought not to have been, from its important nature and content, thus entirely ignored by a rival theorist.

(2) The first two and a part of the third chapter of the work, which have been published, give a general outline of Kuntaka's main theory ; but it is not known how many chapters his original treatise comprised. The fourth chapter in the Madras MS, however, which breaks off without completing the work, may be presumed to have formed its natural con- clusion, inasmuch as it deals with the last variety of vakratā enumerated by the author. The running prose Vrtti, accom- panying the Kārikā-ślokas, and forming an integral part of the work itself, appears to have been composed by Kuntaka himself; for not only the commentator expressly identifies himself with the author, but the citations of later writers1 indicate that the Kärikas should be taken en bloc with the Vrtti. Besides quotations from Kālidāsa, Bhavabhūti, Anangaharsa (author of the Tāpasa-vatsarāja), Hāla, Bāņa, Māgha, Bhāravi, Bhallata, Amaru, Mayūra, Śrīharșa, Bhațța Nārāyaņa, Rājaśekhara, Kuntaka mentions by name Sarvasena, Mañjīra, Māyurāja, and the Udātta-rāghava,

1 Comm. to Vyakti-viveka p. 16; Kāma-dhenu on Vāmana p. 6, etc.

KȘEMENDRA 129

and quotes from Bhamaha, Rudrata and the Dhvanikāra (=Anandavardhana). The work stands unique for its ex- position of the theory of vakrokti, which is apparently developed on the lines indicated by Bhamaha,1 as well as for its analysis of a poetic figure on its basis, which is implicitly accepted by all writers from Ruyyaka to Jagannātha.2

BIBLIOGRAPHY Edition. By S. K. De, in the Calcutta Oriental Series, 1st ed. from a single MS 1923 ; 2nd revised and enlarged ed. is based on two MSS (from Madras and Jesalmere respec- tively), containing Unmeșas i, ii and a part of Unmeșa iii, with a résumé of the unedited portion of Unmeșa iii and iv, and an introduction. Calcutta Orient. Ser, 1928.

KȘEMENDRA

(1) The industrious Kashmirian polygrapher Ksemendra, with the surname Vyāsadasa, is notable in Sanskrit Poetics for his two interesting treatises, Aucitya-vicāra-carcā and Kavi-kanthābharana He refers to another work of his,3 devoted to the treatment of poetic figures, entitled Kavi- karņikā. Kșemendra himself gives us an indication of his date. The concluding verses of his two works, as well as of his Suvrtta-tilaka (ed. Kāvymālā Gucchaka 2, 1886), state that he wrote in the reign of king Ananta of Kashmir, while the colophon to his Samaya-mātrka tells us that it was finished

1 For an exposition see S. K. De's Introd. to the 2nd ed. of Kuntaka's work. Generally speaking, Kuntaka's Vakrokti signifies a mode of expression, differing from and transcending the ordinary mode of speech, and resulting in a characteristic charmingness (vaicitrya or vicchitti), and depends on the imaginative activity of the poet (kavi-pratibhā-nirvartitatva). 2 See Jacobi, Ueber Begriff und Wesen der poetischen Figuren in GN 1908. 3 In Aucit. vic. śl 2.

9 6

130 SANSKRIT POETICS

in the reign of the same king in 1050 A. D. His Daśāvatāra- carita, on the other hand, is dated by himself in 1066 A. D., in the reign of Kalasa, son and successor of Ananta. Ananta reigned from 1028 to 1063 A. D. crowning his son Kalasa in 1063. Bühler1 is right, therefore, in fixing the period of Ksemendra's literary activity in the second and third quarters of the 11th century.2

( 2 ) Peterson proposed3, against the opinion of Bühler, who appears to have left the question open, the identification of Kşernendra with Ksemarāja, the Saiva philosopher of Kashmir, who was a pupil of Abhinavagupta, and who wrote, among numerous other works, a commentary on the Siva-sūtra and on Abhinavagupta's Paramārtha-sāra. Stein supports this identification, but Peterson himself appears to admit later on4 that his own theory is doubtful. In his Aucitya-vicara, Kşemendra pays homage to Acyuta or Visnu; but we know that he was, like his father, a Saiva in his youth but was converted afterwards into Vaisnavism, as he himself indicates, by Somācārya. This fact, as well as chronology, does not stand in the way of the proposed identification, but there

1 Kashmir Rep. p. 46. 2 Dhanika, who lived towards the end of the 10th and the beginn- ing of the 11th century, appears to quote (on i. 61) two verses, which occur in some MSS of Ksemendra's Brhatkathā-mañjarī (ii. 216, 217), and this fact apparently militates against this conclusion of Ksemendra's date; but we know that the Brhatkatha-manjari was composed about 1037 A.D., and as the four lines in question occur in one of the MSS only, it is generally admitted now, for this and other reasons, that they are later interpolations. Ksemendra (Aucit. vic. ad śl. 11. 16, 20) quotes Parimala (otherwise known as Padma- gupta) who was a contemporary of Dhanañjaya and Dhanika. 3 i (Detailed Report), 1883, p. 11, 85 and Bühler in IA xiii, 1884, p. 29. Bühler really proposed the identification of Ksemaraja, author of Sāmba-pañcāśikā, with Ksemendra who wrote Spanda-samdoha, but distinguished both from the poet Ksemendra Vyāsadāsa (see Kashmir Rep. p. 81 and fn). 4 iv p. xxiii.

KŞEMENDRA 131

is no direct evidence to support it. Kșemendra describes himself as son of Prakāsendra and grandson of Sindhu1, and the name of his preceptor is given as Gangaka.2 He was also father of Somendra, and preceptor of Udayasimha and rājaputra Lakșaņāditya3. We know nothing, on the other hand, of Ksemaraja's genealogy or personal history. But we are told at the end of the Brhatkathā-mañjari that Kşemendra learnt sāhitya from Abhinavagupta, while Ksemarāja at the end of his Svacchandoddyota4 (as well as in the colophon to his Stava-cintāmaņi) is described as śişya of the same great philosopher. It is worth noting, however, that while Kşemendra's surname Vyāsadāsa5 is given in all his works (with the exception of his Kalā-vilāsa), it does not occur in any of Ksemaraja's philosophical treatises. Ksemendra has taken care to let us know a great deal about himself, but Kșemarāja always hides his light under a bushel and is apparently free from this trace of natural vanity. The question, therefore, cannot be taken as definitely settled, and can be satisfactorily solved when, as Bühler long ago pointed out6, the name of Ksemaraja's father is found.7 A list of Ksemendra's numerous works is given below. BIBLIOGRAPHY a. Aucitya-vicāra-carcā Editions. (1) Kāvyamālā Gucchaka i, 1886. (2) With comm. Sahrdayatoșiņī of Švetāraņya-nārāyaņa. Madras 1906. Our 1 Concluding verse of the Daśāvatāra. 2 Aucit. vic. under sl. 39. He quotes also Bhatta Tauta. Among. other citations we find Bhatta Bhallata, Gauda-kumbhakāra and Kunteśvara-dautya of Kalidāsa. 3 Kavi-kanthāe under v. i (pp. 138.139). Of his pupil Udayasimha. Kșemendra quotes Lalitābhıdhāna-mahākāvya. 4 Bühler op. cit. App. ii p. clxix (extract). 5 Three stanzas are attributed to Vyāsadāsa in Subhās° (460, 1658, 3039). 6 IA xiii, loc. cit. 7 For a brief résumé of Ksemendra's two works, see below vol .. ii, ch. ix (i); on his didactic and satiric works see S. K. De, Hist .. of Sansk. Lit., Calcutta 1947, pp. 404-10.

132 SANSKRIT POETICS

references are to the former. Also ed. Chowkhamba Sansk. Series, Benares 1933. For an account of the work, see Peterson in JBRAS xvi pp. 167-180, where all the quotations in the work are collected together and discussed.

b, Kavi-kaņthābharaņa Editions. (1) Kāvyamālā Gucchaka iv, 1887, 1899 (2) Chow- khamba Sansk. Series, Benares 1933. A monograph on the work with analysis and German translation by J. Schönberg, Wien 1884 (in Sb. der Wiener Akad.). There is no trace of Kşemendra's Kavi-karņikā. The Works of Ksemendra. A list of the works of Kşemen- dra, published and unpublished, is given here. Those which are quoted in Aucitya-vicāra, Kavi-kanthābharaņa and Suvṛtta- tilaka are marked respectively with the signs (A), (K) and (S). 1. Amrta-taranga (or °turanga) (K). 2. Aucitya-vicāra. 3. Avasara-sāra (A). 4. Kanaka-jānakī (K). 5. Kalā-vilāsa (ed Kāvyamālā Gucchaka i). 6. Kavi-kaņțhābharaņa. 7. Kavi- karņikā (A). 8. Kșemendra-prakāśa (mentioned in ABod 38b). 9. Caturvarga-samgraha (A, K, and ed. Kāvyamālā Gucchaka 5). 10. Cāru-caryā (ed. Kāvyamālā Gucchaka 2). 11. Citrabhārata-nāțaka (A and K). 12. Darpa-dalana (ed. Kāvyamāla Gucchaka 6, 1891). 13. Daśāvatāracarita-kāvya (ed. Durgaprasad and K. P. Parab, NSP, 1891). 14. Deśopa- deśa (K; ed. M. Kaul, Srinagar 1923). 15. Dāna-pārijāta. 16. Narma-mālā (ed. M. Kaul, Srinagar 1923). 17. Nīti- kalpataru (may be the same as Nīti-lata quoted in A). 18. Padya-kādambarī (K). 19. Pavana-pañcāśikā (S). 20. Bṛhat- kathā-mañjarī (ed. Sivadatta and Parab, NSP 1901). 21. Bauddhāvadāna-kālpalatā (A ; with its Tibetan version, ed. Sarat Chandra Das, 2 vols. Bibl. Ind. 1888-1918). 22. Bhārata-mañjarī (ed. Sivadatta and Parab, NSP 1898). 23. Muktāvali-kāvya (A and K). 24. Munimata-mīmāņsā (A). 25. Rājāvali (mentioned in Kahlaņa i. 13). 26. Rāmāyaņa- mañjarī (ed. Bhavadatta and Parab, NSP 1903). 27.

BHOJA 133

Lalitaratna-mālā. 28, Loka-prakāśa (Text in transliteration, Ind. Stud. xviii, 1898, pp. 298-412 ; J. Bloch with trans. notes etc. P. Geuthner, Paris 1914) (A). 29. Lāvaņyavatī- kāvya (A and K). 30. Vātsyāyana-sūtra-sāra (A and quoted in the Pañca-sāyaka). 31. Vinaya-vallī (A). 32. Vetāla-pañca- viņśati (from the Brhatkathā-mañjarī, ed. H. Uhle, München 1924). 33. Vyāsāstaka mentioned in Bühler's Kashmir Report (1877) no. 154 ; see p. 45-46. 34. Śaśivaņśa-mahākāvya (K). 35. Samaya-mātrkā (ed. Durgaprasad and Parab, NSP 1888). 36. Suvrtta-tilaka (ed. Kāvyamālā Gucchaka 2; also ed. Chowkhamba Skt. Series 1933. 37. Sevya-sevakopadeśa (ed. Kāvyamālā Gucchaka 2). The Hastijanaprakāśa mentioned by Schönberg and Peterson is by Ksemendra, son of Yadu Śarman (see Kāvyamālā p. 115 fn and Aufrecht i. 765). The Navaucitya-vicara in Schönberg is probably the same work as Aucitya-vicāra. The Kalā-vilāsa has been translated into German by R. Schmidt in WZKM xxviii, 1914, p. 406-35 ; the Darpa-dalana by the same in ZDMG Ixix, 1915, pp. 1-51 (also ed. and transl. by B. A. Hirszbant, St. Petersberg 1892) ; Samaya-mātrkā, trs. by J. J. Meyer, Leipzig 1903. Parts of Bthatkatha-mañjarī has been translated by Sylvain Lêvi (1st Lambhaka with text in Roman) in JA vi, 1885, pp. 397- 479 ; by Leo v. Mankowski (Pañcatantra, with text in Roman), Leipzig 1892.

BHOJA ( 1 ) The earliest writer on Poetics who quotes Bhoja seems to be Hemacandra1 who flourished, as we shall see, in the first half of the 12th century; while Vardhamana, who however did not write till 1140 A.D., mentions Bhoja in the second verse of his Gana-ratna, the Vrtti on which explains this Bhoja as the author of the Sarasvatī-kanthabharana. The latest writer quoted by Bhoja appears to be Rājasekhara2,

1 p. 295 Comm., besides anonymous quotations. 2 From Karpūra-manj°, Bāla bhā°, and Viddhaśāla°. See Sten

134 SANSKRIT POETICS

whose latest date is the beginning of the 10th century, although some verses from the Caura-pancāsikā (no. 12, ed. Bohlen), attributed to Bihlana, occur in the Sarasvatī-k° (ad i. 152)1. Bhoja appears also to refer in one verse (ad i. 71, p. 22) to Muñja, apparently Muñja-vākpatirāja of Mālava, Jacob2 is misleading in putting down the name of Nami-sādhu (who did not write his commentary on Rudrata till 1069 A.D.) in the list of authors quoted by Bhoja; for the verses in question, though found in Nami, are not Nami's own but really quoted by him from previous authors3. Bhoja also quotes about sixteen times several verses occurring in Daśa-rūpaka and its commentary4, which belong to the time of Muñja, i.e. the end of the 10th and the beginning of the 11th century. The internal evidence of the text, therefore, places the author of the Sarasvati-k° in the period between the

Konow's ed. of Karpūra-mañj° pp. 198 f., for the quotations; also Jacob JRAS, 1897, p. 304f. 1 We have not based any chronological inference on this, because Bihlana's authorship of the work is not beyond question, and Solf tries to demonstrate the existence of a poet called Cora or Caura, whose date is not known. 2 op. cit. p. 304. 3 For instance, the verse ayam padmāsanāsīna° (Bhoja ad i. 51, p. 15) is found, no doubt, in Nami on xi. 24, but it is really a quotation, along with several other verses in the same context, from Bhāmaha ii. 55. Similarly the two verses sa marutā and sa pītavāsā, quoted by Bhamaha himself (ii. 41, 58) from some previous authors (one of whose names is given as Ramasarman) occur in Bhoja anonymously (ad i. I21 pp. 43, 44), but they are also quoted by Nami in the same context. There is no reason to suppose that Bhoja took these verses from Nami's Comm. instead of going directly to Bhämaha, from whom he quotes several other verses directly (e.g. ākrośan nāhvayan, Bhoja ad iii. 8, p. 144=Bhāmaha ii. 94; Bhāmaha ii. 92=Bhoja ad iv. 51, p. 226-7= Subhas° 1645 bhāmahasya). Similar remarks apply to the other supposed quotations given by Jacob, whose mistake is probably due to the fact that Bhamaha's text was not available to him. 4 One verse under Daśa-rūpaka iv. 66 (lakşmī-payodharotsanga-) which Dhanika quotes as his own (yatha mamaiva) is quoted by Bhoja as example of Anyokti (S. K. iv).

BHOJA 135

second and the fourth quarters of the 11th century; and as this date fits in easily with the known date of the ninth Para- māra ruler Bhoja of Dhara, one of the well-known princes of the 11th century, noted for his patronage of letters, the two may be taken to have been reasonably identified. Our Bhoja is frequently cited in later Alamkāra literature as Bhoiarāja, and sometimes simply as räjan1 which designation, like that of muni applied to Bharata, seems to mark him out par ex- cellence in this literature. ( 2 ) Kahlaņa states (vii. 259)2 that king Bhoja of Dhāra was the true friend of poets ; and it is possible that he had himself literary predilections. He was son and successor of Sindhu- rāja and nephew of Muñja-vākpatirāja who was also, as we have seen, a great patron of letters. The date of Bhoja is well known from his own and other inscriptions3. Alberuni4 mentions him as still reigning in 1030 A.D., while the date Saka 964= 1042 A.D. is given by the Rajamrgānka which is attributed to Bhoja. We know also that he fought with Cālukya Jayasimha III between 1011 and 1019 A.D., and with the latter's successor Someśvara (1042-1066 A.D.) who, ac- cording to Bihlana, took Dhara by storm and forced Bhoja to flee. Bihlana himself speaks of Bhoja as of a contemporary whom he did not visit though he might have done so'5. In

1 e.g. Vidyādhara pp. 98, 150, 192, 287, 304, and Mallinātha pp. 287, 304 etc. 2 sa ca bhoja-narendraś ca dānotkarşeņa viśrutau/sūrī tasmin kşaņe tulyam dvāvāstām kavi-bāndhavau. 3 IA vi, p. 53f (Ujjain Plate, 1021-22 A.D.); EI i, p. 230-33; EI ix, p. 182 (Banswara Plate, 1020 A.D); EI xviii, p. 320 (Betma Plate, 1020 A.D.) ; the Sarasvati Image Inscription in the Br. Museum (Rupam, 1924, p. 18 ; 1033 A.D.); Tilakwada Copper plate (Proc. of the Ist Orient. Conf. p. 319 ; 1047 A.D.) etc. 4 ed. Sachau i. 191. According to Merutunga, Bhoja succeeded Munja in Samvat 1078=1022 A.D. See, however, Bhandarkar, Rep. 1882-83, pp. 44-45. 5 Bühler's ed. Vikramānka° p. 23 fn ; also text xviii. 96.

136 SANSKRIT POETICS

Kahlana's assertion, referred to above, with respect to Bhoja and Kşitirāja, the phrase tasmin kşane is taken by Bühler to refer to the period when, after the nominal coronation of Kalasa in 1062 A.D., Kşitirāja became a samnyāsin and some- times visited king Ananta in order to console him. If this interpretation is correct, we get a limit to Bhoja's date at 1062 A.D. A copper-plate of his successor Jaysimha1, how- ever, is dated 1055 A.D., and throws doubt on Bühler's con- jecture. All this, however, will justify us in fixing Bhoja's date with great probability between 1010 and 1055 A.D .; i.e, roughly covering a part of the first and whole of the second quarter of the 11th century, and he may have lived in- to the third quarter of the same century. The exact dates of his accession and death are unknown ; but it seems that he died after long illness, in the midst of wars with Bhīma, king. of Gujarat and with Kalacuri Karņa, king of Tripurī2.

( 3 ) Besides his well-known Sarasvatī-kanthābharaņa, Bhoja appears also to have written a work called Śrngāra-prakāśa3, a MS of which exists in the Government Oriental MSS Library, Madras4. It is composed in 36 prakāśas5, and is described as the largest known work in Sanskrit Poetics. It deals with both Poetics and Dramaturgy. The first eight.

1 EI iii, pp. 46-50 (Mandhata Plate). 2 Prabandha-cintāmaņi of Merutunga, Tawney's trs. p. 4. 3 This work is mentioned by Vidyādhara p. 98 ; by Kumārasvāmin p. 114, 221 ; by Rāyamukuța and Sarvānanda on Amara; by Hemādri on Raghu etc. 4 Mentioned in the Rep. of the Working of the Peripatetic Party of the Library, 1916-19. The work has not yet been published, except three Prakāśas (22-24) by Yadugiri Yatiraj of Melcote (Mysore 1926) and. extracts given by V. Raghavan in his Śrngara-prakasa (i, pt. 2, pp. 513-42). But V. Raghavan has made a detailed study of the work in the work cited (vol. i, pt 1 and ii, Bomby 1940, pp. 1-542). 5 The whole of ch. xxvi is missing, as also the end of ch. xxv and beginning of ch. xxvii, besides smaller gaps.

BHOJA 137

chapters are devoted to the quasi-grammatical question relating to word and its sense as the means of expression, and the theory of vrtti. The ninth and the tenth chapters describe the blemishes and excellences of expression (dosa and guna) ; while the eleventh and twelfth chapters deal respectively with the Mahakavya and the drama. The next twenty-four chapters. treat exclusively of the Rasas, of which the Śrngāra or love in its various aspects (in relation to the four Puruşārthas, Dharma, Artha, Kāma and Mokșa) is maintained, in the light of his novel theory of one Rasa of Ahamkāra-Abhimāna- Śrngāra, to be the principal and essential; and the work derives its name from Bhoja's theory that Srngara is the only one Rasa admissible1. As in the Sarasvatī-k°, this work, in the manner of a cyclopaedic compilation, gives a large num- ber of quotations to illustrate the rules and principles laid down. Śāradātanaya's Bhāva-prakāśana, which deals with the same subject, constitutes really a summary of the important. chapters of Bhoja.

( 4 ) The Sarasvati-kanthābharana, consisting of five Paricchedas, is not a very original work, but consists chiefly of a patient compilation in an encyclopaedic manner from earlier treatises,. especially from Dandin, from whom he takes, according to the calculation of Jacob2, no less than 164 illustrations. From the index of citations given by Jacob, we find that Vamana is quoted 22 times, Rudrata 19 times, the Dhvanyāloka more than 10 times (six of the kārikās being reproduced), while it is curious to note that Bhoja makes a good use

1 Cf Vidyādhara rājā tu śrngāram ekam eva śrngāra-prakāśe rasam urarīcakāra p. 98; Kumārasvāmin p. 221 śrūgāra eka eva rasa iti śrgāra-prakāśa-kārah. For a brief résumé of the work see below under vol. ii, ch. 6 .- Bhoja in four chapters (xviii-xxi) deals with what he · calls Dharma-śrngāra, Artha-śṛngāra, Kāma-śṛngāra and Mokșa-śṛngāra. But he devotes 16 chapters (xxii-xxvi) entirely to what may be called Laukika Śrngāra in its Sambhoga and Vipralambha aspects. 2 loc. cit.

138 SANSKRIT POETICS

of Bhatti's illustrations of the figure yamaka and its numerous subspecies. After dealing with general topics of Poetics, the work speaks somewhat symmetrically of 16 Dosas respectively of Pada, Vākya and Vākyārtha, and 24 Guņas respectively of Sabda and Vakyartha. In the second and third chapter 24 Sabdālamkāras and Arthālamkāras respectively are defined and illustrated. In the fourth chapter 24 Sabdārthā- lamkāras are similarly dealt with. It is noteworthy that the Rītis, mentioned as six in number, are regarded as Sabdārthā- lamkaras. In the fifth chapter we have a treatment of Rasas, Bhāvas, Nāyaka-nāyikā, the five Samdhis, and four Vrttis, etc. While the chief value of Bhoja's work consists in its abundant wealth of illustrations and examples, numbering more than 1500, to every rule and prescription, it is nevertheless interest- ing as embodying, in the main, a tradition of opinion, which is also represented in the Agni-purana, but which in many respects stands apart from the orthodox Kashmirian school. Bhoja is credited with having composed more than 80 works, most of which are voluminous. His work on Grammar (ed. Madras Univ. 1937 ; .also ed. Trivandrum Skt. Series, with Hrdayahāriņī Comm. of Nārāyaņa Daņdanātha, 1935-48) is also called Sarasvatī-kanțhābharaņa.

( 5 ) The commentators on Bhoja, as noted below, are numerous, but they are not of much importance. Ratneśvara's commentary has been published several times together with the text, but so far only three chapters of it have been printed.

BIBLIOGRAPHY a. Sarasvatī kaņțhābharaņa Editions. (1) by A. Borooah, Calcutta 1884. (2) by Viresvara Sastri, Benares 1888 (chs. iv and v). (3) by Jivananda Vidyasagar with Comm. of Ratneśvara (on. chs. i-iii), Calcutta 1894. (4) by Kedarnath Durgaprasad and Vasudev L. Panshikar, Nirnaya Sagar Press, with Comm.

BHOJA 139

of Ratneśvara (on i-iii) and of Jagaddhara (on iv), Bombay 1925, 1934. Our references are to ed. Boorah 1884. Commentaries. (1) Ratna-darpana by Miśra Ratneśvara. Ed. with the text by Jivananda, Calcutta 1894; ed. Benares and ed. NSP, as above. The nominal author Rāmasimha- deva, mentioned in the introductory stanza 2, is apparently the author's patron. In the colophon, the author's name is given as Miśra Ratneśvara ; and in Benares ed. of the text, the commentary is said to have been written at the command of Ramasimha-deva (of Tirhut?). In the Catalogues, the work is sometimes inaccurately given as by Rāmasimha-deva. The author refers to a comm. on the Kāv. prak. by himself, Only the first three chapters of this Ratna-darpana have been published in the editions noted ; and both the Madras and Bodleian MSS contain these chapters only. Ratneśvara appears to have flourished in the 14th century A.D. (2) Mārjanā by Harinātha, mentioned by himself in his Comm. on Dandin (A Bod 206b). See above p. 70. (3) Duşkara-citra-prakāśikā by Lakşmīnātha Bhatța. He may be identical with Lakșmī- natha who, according to Kielhorn Report 1880 81 p. 71, wrote his Pingala-pradipa in 1601 A.D. Kielhorn's MS of this latter work appears to have been copied in 1660, while Burnell's (Pingalārtha-dīpikā pp. 53b, 175b) in 1632 A.D. (4) °Tīkā by Jagaddhara, son of Ratnadhara and Damayanti. Extract given in Ulwar Cat. 1086 and Stein p. 275. The printed portion of the Comm. in the NSP ed. is on the 4th chapter. This work is probably earlier than the 17th century but later than the 14th (see Bhandarkar. Pref. to Mālatī-mādhava pp. xviii-xxi). Jagad- dhara's genealogy is given thus: Candeśvara->Vedeśvara (or Vedadhara)->Rāmadhara (Rāmeśvara)->Gadādhara-> Vidyādhara->Ratnadhara->Jagaddhara. He wrote several commentaries (Aufrecht i. 195) e.g. on the Megha-dūta, Vāsavadattā, Veņī-samhāra, Mālatī-mādhava etc. MS in

140 SANSKRIT POETICS

Stein (p. 276) is dated Śaka 1521= 1460 A.D. (5) Comm. by Harikrsna Vyāsa. SCB 34.

b. Śngāra-prakāśa The only known MS is in the Government Oriental MSS Library, Madras, mentioned in their Report quoted above. The work has not yet been published except in parts ; see above p. 136. Of other published works bearing the name of Bhoja, the Samarāngaņa-sūtradhāra (ed. T. Ganapati Sastri, 2 vols, GOS, Baroda 1924, 1925) deals chiefly with architecture and iconography; the Yukti-kalpataru (ed. Isvara Chandra Sastri, Calcutta 1917) with Nīti-Sāstra ; the Tattava-prakāsa, ed. T. Ganapati Sastri, with Tātparya-dīpikā comm. of Śrīkumāra, Trivandrum Skt. Series 1920 ; trs. E. P. Janvier in IA liv, 1925, pp. 151-56) with religio-philosophical topics ; while the Rāja- martanda commentary on the Yoga-sūtra (ed. Bibl. Ind. Calcutta 1883 ; ed. Chowkhamba Skt. Series, along with the text and five other comm. Benares 1930; ed Jivananda Vidya- sagar, Calcutta 1903 ; trs. Ganganath Jha, Bombay 1907) is devoted to an exposition of the Yoga philosophy.

MAHIMABHATȚA

( 1 ) Rājānaka Mahiman, Mahimaka or Mahimabhatta, who is cited generally as the Vyaktiviveka-kāra1 from the name of his work, was, as indicated by his title, probably a Kash- mirian writer, who describes himself as son of Śrī-Dhairya and disciple of mahākavi Šyāmala. He informs us at the outset of his work (i. 3) that his principal object is to consider the views of the Dhvanikara ; and as in the course of his. discourse he examines the text of the Dhvanyāloka, quoting

1 Viśvanātha, ed. Durgaprasad, NSP., 1915, p. 18, 249 ; Mallinātha on Kirāta iii. 21; Ruyyaka, ed. NSP., p. 12; Keśava Miśra p. 80-81 ; Jagannātha p. 13 etc. Keśava mentions his name as Mahiman.

MAHIMABHAȚȚA 141

from the Kärika and the Vrtti with a minuteness which can- not be mistaken, we may infer with certainty that he was later than Anandavardhana1. It is also probable that Mahima- bhatta was later than Anandavardhana's commentator, Abhinavagupta ; for in some places he betrays an acquain- tance with the latter's work. At p. 19, for instance, Mahima- bhatta quotes directly a long passage from the Locana (p. 33), and shows himself alive to the point in- volved in Abhinava's discussion by criticising it. The passage refers to Dhvanyāloka i. 13 where the Dhvanikāra uses the verb vyanktah in the dual number with the express purpose, as Abhinava explains, of indicating a duality of sense. Bhatta Nāyaka appears to have taken exception to this use of the dual number, upon which Abhinava concludes by remarking: tena yad bhatta-nāyakena dvi-vacanam dūşitam tad gaja-nimīlikayaiva. Mahimabhatta, referring to this discussion, quotes anonymously the remarks of Abhinava (not only the aboveline but the whole passage), with the state- ment: kecid vimāninaḥ ... yad āhus tad bhrānti-mūlam (p. 19). The terms of reference apparently indicate, as Narasimha Iyengar rightly points out,2 that Mahimabhatta is here referr- ing clearly to Abhinava as a theorist of a rival system who. if not contemporaneous, could not have flourished long before his own time. It should be noted that Mahimabhatta quotes and criticises (p. 28) certain views set forth by Kuntaka in his Vakrokti-jīvita (i. 7-8) and attempts to show that Vakrokti, like Dhvani, is to be included under Anumāna. He also quotes from Rājaśekhara's Bāla-rāmāyana (pp. 40, 50) and Viddha- śālabhañjikā (p. 85). This gives us one terminus to Mahima- bhatta's date. On the other hand, Ruyyaka who, as we shall see, flourished in the first half of the 12th century and pro- bably also wrote the anonymous commentary on Mahima- bhatta (printed in the Trivandrum edition of the text), is the

1 Cf Jayaratha p. 12; dhvanikārāntarabhāvī vyaktivivekakāra iti, the Dhvanikara being, to Jayaratha, Anandavardhana himself. 2 JRAS, 1908, pp. 65f.

142 SANSKRIT POETICS

earliest writer to quote and criticise Mahimabhatta.1 We may, therefore, assign Mahimabhatta to the period between Abhinava and Ruyyaka, i.e later than the first quarter of the 11th but earlier than the first quarter of the 12th century, and approximately fix his date towards the last half or the end of the 11th century. This date will be in harmony with the probable date of Śyāmala, who is mentioned by Mahimabhatta as his preceptor, if this Śyamala is the same poet as is quoted by Kşemendra.2

( 2 ) It is difficult to determine what relation Mahimabhatta bore to Śańkuka who was also, like Mahimabhatta, an anumiti- vadin3 in his theory of Rasa, for Sankuka's work has not yet been recovered. Our author claims for himself originality

1 Iyengar (op. cit.) and Harichand (op. cit. p. 105) think that Mahimabhatta is "quoted or criticised" by Mammata; but, as Kāv. prak. v, p. 252 (B. S. S. 1917) shows, Mammata does not at all cite Mahimabhatta or his work, but only criticises an anumana-theory which tries to explain the concept of dhvani by means of inference. No chronological conclusion can be based on this ; for Änandavardhana also refers to a similar theory long before Mahimabhatta wrote. 2 Aucit. vic. ad śl. 16 ; Suvrtta. til. ad ii. 31. Also Subhāș° 2292. Kşemendra's Šyāmala appears to be identical with Šyāmilaka, who. wrote the Bhāņa entitled Pāda-tāditaka (ed. Ramakrishna Kavi and Ramanatha Sastri, Madras 1922) ; for the verses, attributed to Śyāmala in the two works of Ksemendra noted above, occur as śl. 33 and 125 respectively in the printed text of the Bhana. The colophon describes. the author of the Bhana as son of Visveśvaradatta and an udicya (northerner), which makes it probable that he is the Kashmirian Šyāmala, Śyāmalaka or Śyāmilaka, also cited by Abhinavagupta. Both. Abhinava and Kuntaka quote anonymously verses from this Bhāņa. The verse ascribed to Syamalaka in the Subhās° (prāyascittam mrgayate priyā-pāda-tāditah | kşālanīyam śiras tasya kāntā-gandūşa- sidhubhih) refers unmistakably to this Bhana and the second line yah

occurs in a slightly modified form in the Bhāņa itself (ad śl. 132). Rājaśekhara cites a Śyāmadeva (pp. 11, 13, 17). 3 He is so called by Mallinātha (Taralā p. 85) and Kumārasvāmin (p. 219).

MAHIMABHAȚȚA 143.

of treatment and freedom from slavish imitation,1 and his omission of all references to his predecessor need not, there- fore, appear strange. The only testimony of Rāmacaraņa, an 18th century Bengal commentator on Viśvanātha,2 need not be seriously considered; but it is probable that the theory developed by Mahima did not originate in himself. Ananda- vardhana refutes at some length some theory of anumāna (pp. 201f) which attempted to explain that the suggested sense, posited by the dhvani-theory, can be arrived at by the process of logical inference. Mahimabhatta himself gives Antara- ślokas or Antarāyās (besides Samgraha-ślokas summarising a discussion), which add to the discussion and are probably adduced from extenal sources, indicating previous exposition of similar topics by other writers. Thus, Mahimabhatta pro- bably worked out systematically some such thesis (anticipa- ted, it may be, by Anandavardhana), as a direct rejoinder to Ananda's classical exposition ; but there is no evidence to. connect him with the theory accredited to Śańkuka by Abhinavagupta and others. The Vyakti-viveka, consisting of three Vīmarśas, is. essentially a vigorous piece of polemic writing, which does not propose to set forth any new theory or system, but whose only object is to demolish the theory of Dhvani by shewing that the- so-called function of Vyañjana posited by Ãnandavardhana is nothing more than the already recognised process of Anumana or logical inference. In the first Vimarśa he states and amplifies his own position by criticizing the definition of Dhvani. In the second Vimarśa he considers the question of Aucitya, relating to Sabda and Artha. In the third Vimarsa about forty examples are cited from the Dhvanyā-loka and shewn to be really cases of Anumāna and not of Vyañjanā.3 1 He says, for instance, that he has written his work without looking. into Candrikā and Darpana, which apparently had the same object in view as the demolition of the dhvani-theory (i. 4, 5). 2 ed. Durgaprasad p. 248, ed. Röer p. 121 note : śańkuka -- matānuyāyinām vyaktiviveka-kārādīnām matam dūşayati. 3 Mahimabhatta's views will be considered in detail below in vol. ii ..

144 SANSKRIT POETICS

( 3 ) From the Vyakti-viveka itself (p. 108) we learn that Mahimabhatta also wrote a work entitled Tattvokti-kośa, where he is said to have discussed what he calls pratibhā- tattva 1, in connexion with the poetic conception of an idea. Mahimabhatta's work which recognised the new con- cept of dhvani, but tried to explain it by the established process of anumāna (and not by the separate function of vyañjana explained by Anandavardhana), never found any recognition in the hands of later theorists, most of whom became partisans of the latter. Even his commentator does not appear to possess much sympathy for his somewhat extreme view,2 and Mahimabhatta is rather unique in having no followers in later literature. The commentator referred to has been identified with some reason3 with Ruyyaka, who has another commentary on Mammata to his credit, as well as several independent works on Alamkära. We shall deal with him hereafter as an inde- pendent writer on Alamkāra.

BIBLIOGRAPHY Edition. By T. Ganapati Sastri with introd. notes and an anonymous Comm. (attributed to Ruyyaka), in the Trivandrum Sanskrit Series 1909. The Comm. breaks off in the middle of the 2nd Vimarśa.

1 Cf Jacobi Sb. der Preuss. Akad. xxiv. 225 fn. 2 Mahimabhatta's views are vigorously criticised by Ruyyaka (Alam. sarvasva, pp. 12f), and Viśvanātha (Sāhitya-darpaņa, p. 248f). 3 For the arguments summarised see Kane, HSP, p. 245.

CHAPTER VI

MAMMAȚA AND ALLAȚA

( 1 ) A great deal of uncertainty exists with regard to the exact date of Mammata whose name,1 as well as the title rājānaka, indicates that he was probably a Kashmirian. The story relied upon by Hall2 and Weber3 that Mammata was the maternal uncle of the author of Naisadha may be relegated to the region of fantastic fables which often gather round celebrated names. The lower limit of Mammata's date, however, may be fixed with reference to one dated MS4 and two commentaries on the Kāvya prakāśa, of which the date can be ascertained. The commentary of Māņikyacandra is expressly dated in Samvat 1216=1159-60 A. D. The exact

1 Aufrecht (i. 432) notes that Mammata's original name was Mahimabhatta on the mistaken authority of Kesava Miśra's erroneous citation (p. 80-81). The passage in which Keśava cites Mahimabhatta clearly refers not to Mammata but to Mahimabhatta, author of the Vyakti-viveka, mentioned in the same context. The verse anaucityād rte, however, which is thus quoted and attributed to Mahiman by Kesava, occurs originally in the Vrtti of the Dhvanyāloka (p. 145) from which apparently it is also cited in the Vyakti-viveka (pp. 31, 114) with a sa evaha, along with many other verses similarly quoted from the same text. Keśava might have taken the verse directly from Vyakti- viveka's citation, without knowing the original source, and wrongly attributed it to Mahiman himself. It does not occur at all in the Kav. prak. Aufrecht's suggestion, therefore, that Mammata is a corruption of the name Mahimabhatta, like his other supposition that the name Rudrabhatta yields Rudrata, is unwarranted. Cf Peterson ii p. 19. 2 Introd. to Vāsavadattā p. 55. 3 Hist. of Sansk. Lit. (Eng. trans, 2nd ed.), p. 232 fn. 4 A Jesalmere Jaina Bhandar MS of Kav. prak. appears to have been copied in Samvat 1215 Āśvina 14(=1158 A.D. Oct. 8) at Anahila- pātaka while Kumārapala was still ruling. It is noteworthy that the 10

146 SANSKRIT POETICS

date of Ruyyaka's commentary is not known, but we know - from other sources that Ruyyaka flourished in the second and third quarters of the 12th century. The earliest dated MS from Jesalmere appears to have been copied in 1158 A.D. Mammata, therefore, cannot be placed later than the beginn- ing of the 12th century.1 The other limit cannot be settled so satisfactorily. It has. been maintained that Mammata in one verse (bhoja-nrpates tat-tyāga-līlāyitam under x. 26b ; B. S. S. ed. 1917, p. 684) eulogises Bhoja with whom he may be presumed to have been contemporaneous. This is sought to be supported by

colophon states that it is the joint work of Mammata and Alaka (krfī mammațālakayoḥ). See P. K. Gode in JOR, xiii, p. 46-53 (=his Studies. in Ind. Lit. Hist. i, p. 235f.) 1 Jhalakikara maintains, on the authority of Paramānanda Cakravartin and Nagojī on Mammata, that Mammata in several places. criticises Ruyyaka, who therefore must be placed earlier than Mammata. But the passages he cites do not support his contention. Thus, the verse rājati tațīyam (Kāv. prak. p. 758) is supposed to be directed against Ruyyaka p. 199. where the same verse is quoted in the same context. It appears, however, that Mammata gives this verse as an instance of sabdālamkāra-samkara without any comment but with the simple statement that here we have a commixture of yamaka and anuloma- pratiloma-citra dependent on one another. Ruyyaka, on the other hand, citing the same verse and referring to the opinions of "other authors" comments on it at some length. He remarks that though the verse is given by some as an example of śabdālamkāra-samkara, such commixture of śabdālamkāras, in his opinion, is not possible, and the example is faulty. The verse itself occurs in Ratnākara's Hara-vijaya (v. 137). Jayaratha and Samudrabandha also remark in this connexion. that the anonymous authors, referred to by Ruyyaka in his criticism, allude to "Mammata and others." Besides, Ruyyaka himself quotes (p. 102) Mammata's Kārikā iv. 15-16. Jayaratha expressly says that Ruyyaka wrote a commentary on Mammata called Kāvyaprakāśa- samketa (p. 102). In several other places, both Jayaratha and Samudra- . bandha point out that Ruyyaka is criticising Mammata (e.g. Jayaratha pp 77, 102, 107, 150, 163, 199, 204 ; Samudrabandha pp. 23, 25, 119, 156,- . 243, 249, etc.).

MAMMAȚA AND ALLAȚA 147

the story, related by a very late commentator Bhīmasena,1 that Mammata was the son of Jayyata and had two brothers Kayyata and Uvvata, of whom Uvvata (or Uvața) is taken to be the well known commentator on Vedic works, some of which, as he himself tells us, were composed in Avanti while Bhoja was still reigning (bhoje rājyam praśāsati). It is sugges- ted on this ground that Uvvata was probably the medium of the quotation referred to above relating to his royal patron; or, assuming it to have been composed by Mammata himself, it might have obtained for its author an introduction into the munificent court of Bhoja 2. But this theory is untenable ; for Uvvața tells us that his father's name was Vajrata and not Jayyata ; and it is not clear that the stanza in question, given anonymously as an instance of the figure udatta (which consists of a description of the wealth and prosperity of an exalted personage) was composed by Mammata himself, who certainly borrows similar illustrative verses from various sources. All that this anonymous verse may be taken to establish is that its allusion to king Bhoja indicates that Mammata was probably not earlier than Bhoja. We may, therefore, place Mammata between Ruyyaka on the one hand and Bhoja on the other, if we may assume, on the authority of the commentators, the identity of this Bhoja with the Paramāra Bhoja of Dhāra, the reputed author of the Sārasvatī-k°. In other words, Mammata probably belongs to the period between the middle of the 11th and the first quar- ter of the 12th century. Allowing two generations to intervene between him and Ruyyaka, we may assign Mammata's literary activity roughly to the last quarter of the 11th century. Mammata mentions Abhinavagupta who was still living in 1015 A.D. (see above), and quotes anonymously (under x. 131,

1 Introd. to ed. Kāv-prak. in B.S.S (3rd. ed. 1917) pp. 6-7: also extract in Peterson i, p. 94. 2 Ganganatha Jha in his introd. to his trans. of Kāv. prak. pp. vi-vii.

148 SANSKRIT POETICS

purāņi yasyām) from Navasāhasānka-carita (canto i), which was composed about 1005 A. D.

( 2 ) Although well-known for his Kāvya-prakāśa, which helped to establish finally and exclusively the doctrines of the Kashmirian school of Anandavardhana, Mammata is also the author of a less known work entitled the Sabda-vyāpāra- paricaya which, as its name implies, is a short dissertation on the expressive functions (Vrtti) of words, a topic which he discusses also in the second Ullāsa of Kav. prāk. Mammața, like most writers on Poetics, was also well-versed in the allied science of grammar, proficiency in which he also dis- plays in the larger work.1 On a summary examination of the contents of the Kāvya-prakāśa,2 it will appear that the work is care- fully planned and systematically worked out. Peterson, however, on the indication given in Rājānaka Ananda's Nidarśana3 commentary, first called in question the unity of

1 A work on music called Samgita-ratnāvalī is attributed to Mammața by Gajapati Nārāyaņadeva in his Samgīta-Nārāyana (see V. Raghavan in ABORI, xvi (1934-35), p. 131 ; and references therein). 2 The Kāvya-prakāśa, in ten Ullāsas, consists of Kārikā, Vrtti and illustrations. The topics in brief are: I. Purpose, source and definition of Kāvya, and its division into Uttama, Madhyama and Adhama. II. Explanation of Abhidhā, Laksaņā and Vyañjanā, and subdivisions of Lakşaņā and Vyañjanā. III. Vyañjakatva of all kinds of senses. IV. Division of Dhvani into Avivakşita-vācya and Vivakșitānyapara-vācya. Nature of Rasa. V. Guņībhūta-vyangya and its eight subdivisions. VI. Citra-kāvya. VII. Doșas of Pada, Vākya, Artha and of Rasa. How a Doșa may become charming. VIII. Distinction of Guņa and Alamkāra. Only three Gunas (Madhurya, Ojas and Prasada) admitted. Combina- tion of letters conducive to Gunas. IX. Figures of Sabda. Vakrokti (of Śleşa and Kāku), Anuprāsa (Cheka-, Vrtti- and Lāța-), Yamaka and its varieties, Śleșa, Citra and Punar-uktavadābhāsa. Figures of Artha, enumerated and defined as 61. 3 The name of this comm. is Sitikantha-vibodhana as well as Kāvyaprakāśa-nidarśana.

MAMMAȚA AND ALLAȚA 149

-the work, although his first erroneous impression1, corrected afterwards by himself2, was that the Kārikā-text was composed by Mammata, while the running prose Vrtti was added by some other hand. There is enough evidence now to show that Mammata composed nearly the whole work (Kārikā and Vrtti), and only a small portion of the last chapter, left incomplete by him, was completed by another author, whose name is given by Ananda as Alata or Alaka. That the fact of joint-authorship is skilfully concealed is supposed to be alluded to in the last verse, given in some of the MSS3, which apparently says that "this way of the learned, though different yet appearing identical, is not strange, for here the only cause is a properly constructed (plan of) combination." This may be explained, no doubt, as meaning that the author here claims the credit of having skilfully removed, in his systematic work, all conflict of opinions held by different authors on Poetics ; but most commentators agree in finding here a hint implying that the work left incomplete by Mam- mata was completed by some other person, and the traces of joint-authorship are ingeniously obliterated. Māņikyacandra Sūri, one of the earliest commentators, comments on this verse ; atha cāyam grantho'nyenārabdho'pareņa ca samarthita iti dvi-khaņdo'pi samghațanā-vaśād akhaņdāyate. Ruyyaka remarks in his Samketa commentary: esa grantho grantha- kṛtānena katham apy asamāptatvād apareņa ca pūritāvaśeşatvād dvi-khaņdo'py akhaņdatayā yad avabhāsate tatra samghatana- iva hetuh. In this view Ruyyaka is followed by Jayanta Bhatta, Someśvara, Narahari Sarasvatītīrtha, Kamalākara, Ânanda, Jajñeśvara and other early as well as late commen- tators on Mammata. Rājānaka Ānanda, however, is more

1 Rep. i p. 21 f. 2 Rep. ii p. 13 f. Cf Bühler in IA xiii p. 30. 3 ity eşa mārgo viduşām vibhinno'/ pyabhinna-rūpah pratibhāsate yat/ na tad vicitram yad amutra samyag/ vinirmitā samghațanaiva hetuḥ.

150 SANSKRIT POETICS

explicit and quotes a traditional verse1 in his Nidarśana com- mentary to show that Mammata composed the work up to the treatment of the figure parikara (x. 32), while the rest, consisting of a small portion of the concluding chapter, was completed by Alaka, Alata or Allața2. This statement about the joint-authorship of the Kāvya- prakāśa receives confirmation from an independent source. Commenting on Amaru-śataka (ed. Kāvyamālā 18, 1916, śl. 30), Arjunavarman, who flourished in the first quarter of the 13th century, quotes from the Kāvya-prakāśa under vii. 14 (the verse prasade vartasva3 cited therein) with the remark: yathodāhrtam doşa-nirņaye mammațālakābhyām. In the same chapter on Doşa in the Kāvya-prakāśa, Amaru 72 is quoted as instancing the fault technically known as jugupsā- ślīla (vulgarity causing disgust), because the word vāyu in the verse is supposed to connote vulgar associations. Arjuna- varman defends Amaru from this fastidious criticism with the pointed remark: kim tu hlādaikamayī-vara-labda-prasādau kāvyaprakāśa-kārau prāyeņa doşa-drstī, yenaivamvidheșvapi paramārtha-sahrdayānanda-padeşu sarasa-kavi-samdarbheşu doşam eva sākşāt akurutām. Both these passages, which mention the dual authorship of the Kāvya-prakāśa, refer in particular to ch. vii where the dosas or faults of composition are discussed. Unless the remarks be taken to imply a general

1 kṛtaḥ śrī-mammatācārya-varyaih parikarāvadhih/ prabandhaḥ pūritaḥ śeso vidhāyālata- (°laka or °llața- ) sūriņā//. 2 This is perhaps the reason why in some MSS of the work the colophon puts down the names of Mammata and Allata (or Alaka) as the authors, e.g. Bodleian MS (Hultzsch Collection 172), which is a Kashmirian MS in Śāradā characters, reads: iti kāvyaprakāśābhidhānam kāvya-lakşaņam samāptam, krtiḥ śrī-rājānaka-mammațākālakayoḥ. Also Stein, Jammu Cat. MS no. 1145 (cf introd. p. xiii f), 1173. See also colophon of an early MS mentioned above p. 145-46, fn. 4. The dual authorship of the Kāv. prak. is accepted by V. S. Sukthankar in ZDMG, xvi, p. 477-90. 3 This verse is ascribed to Candraka in Sārngadhara 3565 .- On this question, see Kane in IA, 1911, p. 208.

MAMMAȚA AND ALLAȚA 151

reference to the fact of joint-authorship without particularly meaning collaboration of any special chapter, one may be led to the conclusion the Allata (here mentioned as Alaka) had a hand not only in the 10th, as the tradition makes it out, but also in the 7th chapter.1

( 3 ) Of the three forms of the name, Alaka, Alata and Allata, the last, which is given in Stein's Jammu MS, seems to be the most authentic. The ta is a well-known suffix to Kashmirian names, and Stein says that "this form of the name is the only one known to the tradition of Kashmirian Pundits, to whom the double authorship of the Kāvya-prakāśa is otherwise perfectly familiar."2 But Alaka is as good a Kashmirian form of the name. This Allata or Alaka is supposed to be the same as Rājānaka Alaka who wrote a commentary on Ruy- yaka, and is quoted by Ratnakantha as such.3 If this identi- fication, which was suggested by Peterson but disfav oured by Stein, is correct, then we must also ascribe to him the Vişama- padoddyota commentary4 on Ratnakara's Hara-vijaya, where Alaka is described as son of Rājānaka Jayānaka. It appears strange, however, that Allata the continuator of the Kāvya- prakāśa should also be the commentator of Ruyyaka, who in his turn commented on the same work. This will måke the two writers commentators on each other's text; and if this were so, we may naturally expect a reference to this fact

1 H. R. Divekar in JRAS, 1927, holds that Mammata composed only the Kārikas up to the figure Parikara and that the remaining Kārikās and the whole of the Vrtti were composed by Alaka. But his arguments are hardly convincing. 2 See Jammu Cat. pp. xxiii f. Stein notes that the form Allata of the name is "found also in the fine birch-bark codex of Kāvyaprakāśa- samketa written by Pandit Rājānaka Ratnakaņtha in the Saka year 1570(A.D. 1648)." 3 Peterson ii p. 17f. 4 i pp. 13, 17. Cf. Bühler. Kashmir Rep. p. 45. The work, extending Over 50 cantos, has been printed in Kāvyamālā 22.

152 SANSKRIT POETICS

by Ruyyaka, who otherwise alludes to the dual authorship of the Kāvya-prakāśa, but does not mention the name of Allata as the continuator.1

( 4 ) A tradition, chiefly obtaining in Bengal, as we find it in Baladeva Vidyābhūșaņa and Maheśvara Nyāyālamkāra, two very late Bengal commentators on the Kāvya-prakāśa,2 imputes the authorship of the Kārikās (here called sūtras3) to Bharata and the prose-Vrtti to Mammata, while Bharata him- self is said to have drawn upon the Agni-purana. While the last assertion about the Agni-purāna has no foundation in fact and is apparently prompted by the amiable but unhis- torical imagination of late writers, which delights in exalting the antiquity of the Puranas, the suggestion of Bharata's authorship of the Kārikas is too unauthentic and fanciful to be accepted. Mammata's authorship of the Kārikās has been

1 It is clear, however, that the combination of names in the colophon to a MS of the Kav. prak. (containing, in the same codex. the text and Ruyyaka's Samketa commentary), viz., iti śrīmad-rājānakāllata- mammata-rucaka-viracite nijagrantha-kāvyaprakāsa-samkete prathama ullasah, should not lead us to think, as Peterson and Stein do, that the Kāv. prak. is a joint-compilation of Allata, Mammata and Rucaka (or Ruyyaka), but it only indicates the names of the authors of the original work (viz. Mammata and Allata) as well as the name of the author of the °Samketa commentary comprised in the codex. 2 Vidyābhūşaņa's Sāhitya-kaumudī on Mammața, ed. Kāvyamālā 63, 1897, p. 2, and comm. ; also comm. p. 1. and text p. 189. (Cf. Peterson. ii p. 10f.); Maheśvara's comm. (ed. Jivananda, 1876) p. 1. This view is also endorsed by Jayarāma Pañcānana, another Bengal commentator on. Mammața (see Peterson ii pp. 21-22, 107). 3. The term sutra should not mislead us into thinking that the work might have been originally composed in that form, upon which the later kārika-verses were based; for it is not unusual for the commentators. to refer fo Mammata's kārikās themselves as sūtras; e.g. °Pradīpa, ed. Kāvyamālā 1912, p. 378 sūtre vibhāga upalakșana-parah ; p. 384 sūtram copalakşaņatayā yojyam ; °Prabhā p. 381 sūtrāşkarānusāratah ; "Uddyota ed. Chandorkar, x p. 123.

MAMMAȚA AND, ALLAȚA 153:

declared by Hemacandra (Comm. p. 109=Kāvya-prak. v. 1-2b) in the first quarter of the 12th century, as well as accepted by a succession of authors and commentators like Jayaratha, Vidyādhara, Mallinātha, Kumārasvāmin and Appayya. Vaidyanatha, commenting on ° Pradipa (i. 1), alludes to this. tradition and rejects it expressly1; and in this view most of Mammata's other commentators agree. Apart from this, the evidence of the text itself goes directly against such a hypo- thesis. The Kārikās iv. 4-5 are expressly supported in the Vrtti by a dictum of Bharata (vi. p. 87. ed. Grosset), and this. implies a distinction between the author of the Kārika and that of the Nātya-śāstra.2 The Kārikā x. 8b, again, says mālā tu pūrvavat, implying from the context that the figure mālā- rūpaka follows the rule laid down for the figure mālopamā, which, however, is not taught in any of the previous Kārikās, but explained in the Vrtti. This apparently indicates that the Kärika and the Vrtti form one block which should be attributed to one and the same author.3 The source of this tradition is probably the unquestioned. reverence paid to the sage Bharata, but it may also be due- to the fact that Mammata himself has made a considerable use of Bharata's Karikas. Thus Bharata vi, 15, 17-21= Mammața iv. 6-11. Mammața, however, has also made a similar use of Kārikas and illustrative verses of many of his predecessors. Thus, the Kārikā in Mammața vii. 10 karņāva-

1 °Prabhā ed. Kāvyamālā p. 2. 2 Cf Vaidyanātha on i. 1 ; granthakrd iti mammațabhattākhyasya kārikākartur nirdeśaḥ. ..... bharata-samhitāyām kāsāmcit kārikānām darśanāt sa eva granthakrd iti na yuktam; caturthe-"kāraņāny atha kāryāņi sahakārīņi" (iv. 4) ityādi kārikārthe "tad uktam bharatena" iti bharata-sammati-pradarśanasyāsamgatitvāpatteḥ. 3 To the same effect Vaidyanätha commenting on this passage, ed. Kāvyamālā 1912, p. 329: etad eva sūtram sūtra-vṛttikțtor ekatve- jñāpakam, mālopamāyāh sūtrāvanuktāyā vrttāveva kathanāt. Also cf other agreeing opinions quoted in Jhalakikara's comm. ed. B.S.S. 1917, p. 599. See also S. K. De, Mālā tu pūrvavat in ABORI, vi, 1925 (reprinted in Some Problems of Skt. Poetics, Calcutta 1959, p. 131f).

154 SANSKRIT POETICS

tamsādi-pade) appears as a samgrahaśloka in Vāmana's vrtti on ii. 2, 19 ; while the definition of the figure āksepa in Mammața x. 20 is taken from Bhāmaha ii. 67a and 68a, or Udbhata ii. 2a and 3a as found quoted in Abhinava's Locana p. 36. Again, Mammata iv.1 and 3 are clearly paraphrases from the Dhvanyāloka ii. 1 and 3. Mammata also makes a large use of Rudrata's illustrations.1 The Kāvya-prakāśa, consisting of ten Ullāsas, traverses the whole field of Sanskrit Poetics (with the exception of drama- turgy) in only 143 Kārikas and about 620 illustrations derived from various sources. As it combines the merit of fulness with that of conciseness, it became one of the classic works of Sanskrit Poetics and Rhetoric which has al- ways maintained a great authority and popularity throughout India. It sums up and explains in the succinct form of a brilliant text-book all the previous speculations on the subject, becoming in its turn the starting point of endless exegetic works and text- books. As such it occupies a unique position in the history of Sanskrit Alamkāra literature.2

BIBLIOGRAPHY

a. Kāvya-prakāśa Editions .. The work has been very often published, in part or as a whole, with or without commentaries. Only im- portant ed. are mentioned here. Text only or text with modern comm .-- (a) ed. Nathuram, Education Press Calcutta 1829 (perhaps the editio princeps). (b) ed. Mahesh Chandra Nyayaratna, with his own comm. Tātparya-vivaraņa, Calcutta 1866. (c) ed. Vamanacharya Jhalakikar with his own Bala-bodhini comm. Bombay

1 This has been shown by Sukthankar in ZDMG, lxvi, p. 477f, referred to above. 2 For a discussion of the various topics covered by the work see S. K. De, Some Problems, pp. 108-130.

MAMMAȚA AND ALLAȚA .155

Skt. Series 1889, 1901, 1917. (d) with a comm. by Hari- shankar Sarma Maithila, ed. D. R. Sastri, Chowkhamba Skt. Series, Benares 1926. (e) with Budha-manorañjanī comm. of Mallāri Lakșmaņa Sastri, Madras 1891. Text with old comm. (a) With Ādarśa comm. of Maheśvara Nyāyālamkāra, ed. Jivananda Vidyasagar Calcutta 1876 (the same, ed. Calcutta Skt. Ser. 1936). (b) With the comm. of Kamalākara Bhatta, ed. Papa Sastri, Benares 1866. (c) With Pradīpa of Govinda hakkura and Prabhā of Vaidyanātha Tatsat, ed. Nir. Sag. Press, Bombay 1891, 1912. (d) With Pradīpa and Uddyota of Nāgojī Bhatta (ch, i. ii, vii and x), ed. D. T. Chandorkar, Poona 1896, 1898, 1915. (e) With Pradīpa in Pandit x- xiii, 1888 91. (f) With Pradīpa and Uddyota (complete), ed. V. S. Abhyankar, Anandasrama Press 1911 .. (g) With Pradīpa, Uddyota, Prabhā, Samketa (of Rucaka) and Bāla- cittānurañjanī of Narahari Sarasvtītīrtha (ch. i, ii, iii, and x), ed. S. S. Sukthankar, Bombay 1933, 1941. (h) With Sāhitya- kaumudī of Baladeva Vidyābhūșaņa, ed. Nir. Sag. Press, Bombay 1897. (i) With the Samketa of Mānikyacandra, ed. V. S. Abhyankar, Anandasrama Press, Poona 1921 ; ed. R. Shama Sastry, Mysore 1922. (j) With the Dīpika of Candīdāsa, ed. Sivaprasad Bhattacharya, Sarasvati Bhavana Texts, Benares 1933. (k) With Sampradāya- prakāśinī of Śrīvidyā-cakravartin and Sāhitya-cūdāmaņi of Lauhitya Bhatta Gopāla, ed. H. Harihara Sastri, Trivan- drum Skt. Series, in 2 vols., 1926, 1930. (1) With Sudhā- sāgara of Bhīmasena Dīkşita, ed. Narayan Sastri Khiste, Chowkhamba Skt. Series, Benares 1927. (m) With Samketa of Rucaka, ed. Sivaprasad Bhattacharya in Calcutta Oriental Journal ii, 1935 ; (n) With Viveka of Śrīdhara, ed. Sivaprasad Bhattacharya, Sanskrit College, Calcutta 1959, pt. i (ch. i-iv). Translations. (a) into English by Ganganath Jha in Pandit xviii-xxi, 1896-99; reprinted and separately published, Benares 1899, 1918 ; 2nd ed., India Press, Allahabad

156 SANSKRIT POETICS

1925. (b) into English by. Pandurang P. Joshi (ch. i. iž and x). Bombay 1913. Our references are to the pages or by Kārikās of BSS ed. of Jhalakikar 1917. The commentaries are discussed and enumerated in detail below.

b. Šabda-vyāpāra-paricaya Edition. By M. R. Telang, N. S. P. Bombay 1916. The work is also called Sabda-vyāpāra-vicāra in a BORI MS noticed in Cat. xii, p. 343-44.

THE COMMENTATORS ON MAMMAȚA

( 1 ) There is hardly any other technical work in Sanskrit which has been so much commented upon as the Kāvya-prakāśa, and no less than seventy different commentaries and glosses will be found noticed in the various reports, catalogues and journals relating to Sanskrit MSS. They count as their authors not only independent and notable writers on Poetics like Ruyyaka and Viśvanātha, but also men having other literary interests, like the Naiyāyika Jagadīśa and Narasimha Țhakkura, the grammarian Nāgojī Bhatța, the Mīmāmsaka Kamalākara Bhatța, the Vaisņava Baladeva Vidyābhūșaņa, as well as the Tantrika Gokulanatha. Very few of these commentaries have yet been printed. We mention here the more important.and better known of these writers, noting their dates when known and supplying whatever information we can gather about them.

Rājānaka Ruyyaka or Rucaka His commentary is called Samketa. He is identical with Ruyyaka (q. v.), author of the Alamkāra-sarvasva ; middle of the 12th century. Ed. Sivaprasad Bhattacharya in Calcutta Oriental Journal

COMMENTAIORS ON MAMMAȚA 157.

ii, 1935 ; also partly (on i, ii, iii and x) ed. S. S. Sukthankar, Bombay 1933, 1941.

Māņikyacandra His commentary is also called Samketa. It is dated in Samvat 1216=1159-60 A.D.1 Māņikyacandra was a Jaina author of Gujarat, who belonged to the Kotika- gaņa, Vajra-śakha, Rājagaccha. The concluding verses of his commentary trace his spiritual genealogy to Śilabhadra, after whom came in succession Bharateśvara, Vairasvāmin (Vīra°), Nemicandra and Sāgarendu. Our author states that he was a pupil of Nemicandra, as well as of his successor Sagarendu, who is identified by Peterson2 with the Sagarendu who wrote out in the Samvat 1252 (=1196 A. D.) at Pattana the first copy of the Amamasvāmi-carita3 Our Mānikyacandra seems to be identical with Māņikya- candra, author of Parśvanātha-carita, which is said to have been completed on the Dewali of Samvat 1276(=1220 A.D.) in Devakūpa (Divbandar) by the sea (v. 36). In it the author gives a spiritual genealogy, traced up to Pradyumna Sūri and corresponding exactly to that given in the °Samketa4. Māņikya also appears to have written a Nalāyana or Kubera-purāņa5. Mānikyacandra, mentioned in Merutunga's Prabandha- cintamani as having flourished under Jayasimha of Gujarat, seems to be a different person.6 Edition with the text. (i) By Vasudeva Abhyankar, Anandasrama Press, Poona 1921 ; (ii) by R. Sharma Sastry, Mysore 1922.

1 Peterson iii, extr. p. 322, where the verse giving the date is incom- plete, but it is given in full in Jhalakikara's introd. to Kāv. prak. p. 22. 2 iv, p. cxxviii. 3 iii, App. p. 98. 4 See extract in Peterson iii, App. p. 157-63 ; also vi, p. xci. The verse sat-tarkī-lalanā-vilāsa°, describing his preceptor Nemicandra, occurs in Parśvanatha-carita also, as in his °Samketa. See extracts in Peterson iii, pp. 160 and 321. 5 Peterson iii, App. p. 357. 6 For a discussion of Manikyacandra's date, see R. C. Parikh's ed. of Someśvara's comm. pp. 12-13.

158 SANSKRIT POETICS

Narahari called Sarasvatī-tīrtha His commentary is called Bāla-cittānurañjanī. He also refers to two works, Smrti-darpana and Tarka-ratna (with its Dīpikā commentary), written by himself. Aufrecht notes that Narahari is also the author of a commentary on the Megha-dūta, of which there is a copy in the Cambridge University Library; and Stein notes (p.67) a Kumārasambhava- țīkā by Sarasvatītīrtha. His commentary on Mammata states that he was born in Samvat 1298=1241-42 A.D., in Tribhu- vanagiri in the Andhra country. He traces his own genealogy to Rāmeśvara of Vatsa-gotra, and describes himself as son of Mallinatha and Nagamma and granson of Narasimha, son of Rameśvara. He had a brother named Nārāyaņa. When he became an ascetic, he took the name of Sarasvatītīrtha and composed his commentary at Benares1. Edition. A part only (on i, ii, iii and x) in S. S. Sukthan- kar's edition of Kāvya prak. mentioned above, Bombay 1933, 1941 ; extract from MSS in Peterson i, 74 and IOC iii, pp. 325f. Jayanta Bhatta His commentary is called °Dīpikā or Jayantī. He gives: its date as Samvat 1350=1294 A.D. He calls himself son of Bharadvaja who was the family-priest (purohita) to the chief minister of Sārngadeva of Gujarat, the third Vāghel sovereign who ruled at Pattana during 1277-12972. Jayanta is quoted by Paramānanda Cakravartin and Ratnakaņtha (q.v.), and the latter states that his own commentary was based on: the Jayanti. Our Jayanta Bhatta must be distinguished from Bhatta Jayanta or Jayantaka, father of Abhinanda the author of the Kādambarī-kathā-sāra (ed. Kāvyamālā 11, 1888), who is. an earlier author quoted by Abhinavagupta (p. 142) and who lived probably in the 9th century. Extract in Bhandarkar Rep. 1883-84, App. 326.

1 Peterson i, pp. 25f, 74. 2 Bhandarkar Rep. 1883-84 pp. 17-18 ; Peterson ii, pp. 17, 20 ..

COMMENTATORS ON MAMMAȚA 159

Someśvara His commentary is called Kāvyādarśa1 (also Samketa). He describes himself as son of Bhatta Devaka of Bharad- vāja-gotra. Jhalakikara thinks that he was a native of Kanauj from his decided partiality for that country. But his reference to the Pratyabhijña School of Kashmir might indicate that he was Kashmirian. Peterson2 and following him Aufrecht3, identify him with Someśvara, author of Kīrti-kaumudī and Surathotsava, and place him in the first half of the 13th century. But this is doubtful, because this Someśvara is known as son of Kumara. R. C. Parikh would assign the commentary to a period between 1150 and 1160 A.D. Our Someśvara cites Bhāmaha, Rudrața, Mukula, Bhațța Nāyaka, Bhatta Tauta, Kuntaka (quoted pp. 135, 152, 302), Vakrokti- jīvita-kāra (p. 36), Ācārya Bhartrmitra (p. 16), Candrikā-kāra (p. 55) and Yāyāvarīya (p. 224). He is cited in his turn by a very late commentator Kamalākara4 Ed. R. C. Parikh in 2 vols. (with the text), Rajasthan Pracya Vidya Pratisthan, Jodhpur 1959.

Vācaspati Miśra Nothing is known of him or his commentary, but he is cited by Candīdāsa (as prācīna p. 131), by Viśvanātha on Mammata, and by Bhīmasena. He is to be distinguished from Vācaspati Miśra, author of the Bhāmatī, who is probably older than Mammata ; for in the list he gives of his own works at the end of the Bhamati, he does not refer to any commentary on Mammata. Sivaprasad Bhattacharya (JOI Baroda, iii pp. 359-63) states that Vācaspati Miśra belonged to Mithila and lived near about 1200 A.D. But our Väcaspati

1 The two entries Kāvyaprakāśa-țīkā and Kāvyādarsa in Aufrecht i 737b should be one, as both refer to this comm. 2 v, pp. lxxxiv. 3 i, 102a, 737b. 4 In the Bhau Daji collection (see Cat. of BRAS p. 45) a MS. of Someśvara's comm. states that it was copied from another MS dated in Samvat 1283. Hence the comm. appears to be older than 1227 A.D.

160 SANSKRIT POETICS

"is probably not the Maithili legist who wrote Acāra-cintāmani, Vivāda-cintāmani and other works (see Aufrecht i. 559-60).

Śridhara With the title Sāmdhi-vigrahika, cited by Candīdāsa (pp. 29, 59, 62, 117), and by Viśvanātha on Mammața. Śrīdhara's commentary is called Viveka. A MS the Viveka was copied in Mithila in 1405 A. D. (Sastri, Cat. ASB MSS vi, p. cclxxi). Śrīdhara's date would be about first quarter of the 13th century A D. As in the colophon to this MS the author is described as Tarkācārya Thakkura, he probably belonged to Mithila. Edition. By Sivaprasad Bhattacharya, Part i, ch. i-iv. :Sanskrit College, Calcutta 1959.

Candīdāsa His commentary, called Dīpikā, was written at the instance of his friend Laksmana Bhatta. The India Office MS of his work is written in Bengali characters, and he is cited mostly by Oriya, Maithili and Benares writers (e.g. ·Govinda in his °Pradīpa pp. 24, 36, 202, 274, Narasimha Țhakkura, Kamalākara, Vaidyanātha in his Udāharaņa- candrikā, Nāgojī Bhațța in his Prabhā, and Viśveśvara in his Alamkāra-kaustubha pp. 125, 166). He is not identical1 with Candīdäsa, the younger brother of the grand-father of Viśvanatha, author of the Sahitya-darpana. He appears to have flourished before or about 1300 A.D. He is also cited by Viśvanātha, son of Trimaladeva (q. v.), in a Kashmirian MS dated 1602 A.D. Candīdāsa mentions a Dhvani-siddhanta-grantha by himself. He also quotes a work

1 See H. P. Sastri, Cat. ASB MSS, vi, p. cclxvi, for some curious information of Candīdasa, author of Dipika. He belonged to Bengal (born in the Mukha-kula). The family lived at Ketugrama, four miles west of Uddhäranapura on the Ganges. According to Sastri, the period of.Candīdasa's literary activity was in the middle of the 15th century or earlier .- The other Candīdasa belonged to Orissa.

COMMENTATORS ON MAMMAȚA 161

called Sāhitya-hrdaya-darpana, which may be Bhatta Nāyaka's lost Hrdaya-darpaņa. Edition. By Sivaprasad Bhattacharya, Sarasvati Bhavan Texts, Benares 1933. Extract in IOC iii, 1141/491 (p. 320).

Viśvanātha Author of the commentary °Darpana. He is identical with Viśvanātha (q. v.), author of the Sāhitya-darpaņa, which is referred to in this commentary as his own. First half of the 14th century. Extract in Jhalakikara's introd.

Bhatta Gopāla Known as Lauhitya Bhatta Gopāla Sūri. The name of his commentary is Sāhitya cūdāmani, which is cited several times in Kāma-dhenu on Vāmana (ed. Benares, pp. 4, 8, 33). If he is the same as Gopāla Bhatta cited by Kumārasvāmin (p. 93), he should be earlier than the 15th century. K. P. Trivedi however, thinks that this Gopāla Bhatta of Kumāra- svāmin is the same as wrote a comm. or Rasa-mañjari. He will be identical, thus, with Gopāla Bhatta, son of Harivamśa Bhatta Dravida, who wrote commentaries on Rudra's Srngāra- tilaka (p. 95 above) and Bhānudatta's Rasa-mañjarī (q. v.) Edition. By R. Harihara Sastri and K. Sambasiva Sastri, 2 vols, Trivandrum Skt. Series 1926, 1930.

Bhāskara Wrote Sāhitya-dīpikā commentary. He is cited by Śrī- vatsalāñchana, Govinda Țhakkura (p. 21), Ravi (Peterson iii, p. 20), Narasimha Țhakkura, Bhīmasena, and Ratnakaņțha (Peterson ii, p. 17). Narasimha calls him Lāța Bhāskara Miśra. He is earlier than the end of the 15th century, being cited by Govinda (Kāvya-pradīpa pp. 25, 204, 308, 329). The commentary is also called Kāvyālamkāra-rahasya-nibandha. Extract in Mitra 1681. 11

162 SANSKRIT POETICS

Paramānanda Cakravartin His commentary is entitled Vistārika. He refers to Miśra, Dīpikākrt (Jayanta Bhatta?) and Viśvanātha ; and he must be later than Vidyānātha, whose Pratāparudrīya is cited by him. He is himself cited by Kamalākara, Narasimha Țhakkura, Vaidyanātha (Udāharaņa-can°), Nāgojī Bhatța, Ānanda and Ratnakantha. The earliest citation is probably by Prabhā- kara Bhatța in his Rasa-pradīpa (p. 20) in 1583 A.D. Probab- ly a Naiyāyika of Bengal. He mentions his guru Īśāna Nyāyā- cärya and appears to refer in a punning verse1 to the Tattva- cintāmaņi of Gangesopādhyāya. Jhalakikara thinks that the cakravarti-lakșana, found in the fourteen gādādharī laka- nas, was formulated by him. Paramānanda, from his citations, cannot be earlier than the second half of the 14th century; and he probably flourished before the 16th century2, at the end of which Gadadhara flourished. He must be distinguished from Śrīvidya-Cakravartin, apparently a South Indian writer, who commented on Ruyyaka (q. v.) as well as on Mammata, and who is also cited under the common designation of Cakra- vartin. Paramānanda also wrote a commentary on the Naișadha (IOC vii p. 1438). Extract in Peterson ii pp. 108-9. H. P. Sastri, Cat. ASB: MSS, vi, no. 4831/2492.

Śrīvidyā-Cakravartin His commentary, entitled Sampradāya-prakāśinī refers to a commentary by himself on Ruyyaka. See under Ruyyaka for futher information on this commentator. Edition. Trivandrum Skt. Series 1926, 1930, along with. the comm. of Bhatta Gopala mentioned above.

1 andhā doşāndhakāreşu ke vā na syur vipaścitah/ nāham tu drsti- vikalo dhṛtaś cintāmani sadā. 2 H. P. Sastri (Catalogue ASB, vi, p. cclxix) states Paramānanda flourished before Kamalakara Bhatta (beginning of the 17th century) who cites from his work, as we have noted above.

COMMENTATORS ON MAMMAȚA 163

Govinda Țhakkura His well-known commentary is called °Pradipa1. Govinda also wrote an °Udāharaņa-dīpikā, apparently the same as ° Śloka-dīpikā 2 in Stein (pp. xxviii 60, 269), cited by Nāgojī Bhatta. This is supplementary to the larger exegetical work, being a commentary on the illustrative verses of the text. The 'Pradipa has been commented on by Vaidyanatha Tatsat (°Prabhā and Udāharaņa-candrikā) and Nāgojī Bhațța (°Ud- dyota). Govinda was a native of Mithila, born in the family of Ravikara, eldest son of Keśava and Sonodevī, elder brother of poet Śrī-harsa who is not, however, as Peterson supposes, the author of the Naisadha2. In addition to this information about himself, Govinda tells us that he learnt kāvya and sāhitya from his elder step-brother Rucikara. His exact date is not known, but Govinda refers to Viśvanātha as arvācīna, quoting the latter's criticism of Mammata's definition of poetry, as well as the latter's own definition of the same, without actually naming him or his Sāhitya-darpana. Govinda, therefore, is probably later than the middle of the 14th century. On the other hand, he is earlier than the last quarter of the 16th century, being quoted in Prabhākara's Rasa-pradīpa which was composed in 1583 A.D. Narasimha Țhakkura, who flourished later, but not much later, than 1612 A.D. (having himself quoted Kamalākara), is supposed, on the authority of the family genealogy, to be fifth in descent from

1 The full name of the commentary is Kāvyaprakāśa-pradīpa, simplified generally as Kāvya-pradīpa; so Peterson's speculation on the name (i, 27) is idle trifling. 2 The second verse of this work refers to his Kavya-pradīpa. 3 His brother's verse is cited in ch. x (p. 355) as mad-bhratuh śrīharşasya, but the Naisadha is cited by name in the same chapter (p. 351) with iti naisadha-darśanāt. He laments, in one of the conclud- ing verses, the death of this brother Śrī-harsa, in which however he does not mention him, as he could have done, as the poet of the. Naişadha.

164 SANSKRIT POETICS

Govinda, This will roughly place Govinda towards the end of the 15th century1. Edition: (1) In Pandit vols. x-xiii, 1888-89, by Rama Sastri Bhagavatacharya. (2) With comm. of Vaidyanatha, called ° Prabhā, in Kāvyamālā 24, NSP, Bombay 1891, 1912 (our references are to the ed. of 1912). (3) With °Uddyota in Anandasrama Series 1911. (4) With °Uddyota (ch. i, ii, vii, x) by Chandorkar, Poona 1889.

Jayarāma Nyāya-pañcānana

His commentary is called ° Tilaka or Jayarāmī. The com- mentary called Rahasya-dīpikā by Jayarāma, entered in some catalogues, appears to be an alternative name. He seems to be identical with the author of the Nyāya-siddhānta-māla, the Padārtha-maņimāla, and of commentaries on the Nyāya- kusumāñjali and on the Tattvacintāmaņi-dīdhiti, which works indicate that he was a Naiyāyika. He is described as pupil of Ramacandra (or Rāmabhadra) Bhattācārya Sārvabhauma and guru of Janārdana Vyāsa. He is cited by Śrī- vatsalāñchana and Bhīmasena, but the only writer who appears to quote him extensively is Viśveśvara (as Nyāyapañcānana) in his Alam. kaus. pp. 11, 23, 106, 127, 161, 162, 172, 263, 327. Jayarāma was certainly later than Raghunātha Śiromaņi . (beginning of the 16th century), on whose Tattva-cintāmani- dīdhiti he commented, but earlier than the beginning of the 18th century, the date of Bhimasena. A more precise dating is possible because the date of Jayarāma's Nyāya-siddhānta- māla is given as Samvat 1750 (=1694 A.D.). He is said to have been patronised by Rājā Rāmakrsņa of Krishnagar (Bengal). See S. C. Vidyabhusan, Indian Logic, Calcutta 1921, pp.477f. Extract in Peterson ii, p. 107 and Mitra 1447.

1 See introd. to N. S. P. ed of the Pradipa; also the Pandit xiii, p. 74f.

COMMENTATORS ON MAMMATA 165

Śrīvatsalāñchana1 Bhatțācārya and Subuddhi Miśra Śrīvatsa's commentary is called Sāra bodhinī. It is men- tioned by Hall2 and attributed to "Maheśvara, otherwise called Śrīvatsalāñchana." Maheśvara or Māheśvara appears to be another name of Subuddhi Miśra who, Aufrecht notes,3 wrote a commentary on Vāmana called Sāhitya-sarvasva ; but Subuddhi also appears to be cited as a commentator on Mammața by Narasimha Țhakkura, Vaidyanātha (Udāharaņa- can°), Bhimasena and Ratnakantha. The two are probably different persons, as Bhimasena and Ratnakantha cite separately both Śrīvatsa and Subuddhi Miśra. Śrīvatsa is also the author of an independent work called Kāvya-parīksa4, which deals in five Ulläsas with the general characteristics of poetry and follows in the main the teachings of Mammata. This work may not be identical with Tattva-parīksā (or more fully Sabdārtha-Tattva-parīksā) by Subuddhi Miśra (perhaps the name of his commentary on Mammata, cited by Ratna- kantha and entered5 by Kielhorn in Central Prov. Cat. p. 100).

1 Also called Śrīvatsa-śarman, Śrīvatsa-varman or simply Vatsa- varman. 2 introd. Vāsavadattā p. 54. 3 Aufrecht's description (ABod 208a ; IOC, iii, 1130/566, p 321) of Subuddhi-miśra as Subuddhimiśra-maheśvara, as well as Hall's statement, makes one think that the term is not maheśvara but māheśvara, which is often, as in the cases of Abhinavagupta and Vidyā- dhara, applied as an appellation of a Saiva writer. This is thus a surname of both writers, which might have led to their doubtful identification. 4 Aufrecht i. 778b, ii. 19b ; IOC, iii, p, 342 (MS dated 1550 A.D.). The five chapters of this work correspond in the following way to those of Mammata. (i) Śabdārtha-nirņaya=M 1-3 (ii) Kāvya-bheda= M 4-5 (iii) dosa-nirņaya=M 7 (iv) Guņa-nirūpaņa=M 8-9 (v) Alamkāra= M 10. With a few exceptions it gives the Kārikās as well as the illustra- tions of Mammata with appropriate observations on them. It is thus in effect a commentary on portions of Mammata's work. It has been printed by the Mithila Institute, Darbhanga 1956. 5 See Peterson ii, p. 17 where both Subuddhi's comm. and Tattva- parīkșā are mentioned.

166 SANSKRIT POETICS

It is mentioned as a Comm. on Kav. pr. in H. P. Sastri Cat. ASB MSS, vi, no. 4839/3515, pp. 417-18. Two other works, Kāvyāmrta1 and Rāmodaya-nāțaka, are also ascribed to Śrīvatsa. A work called Siddhānta-ratna-mālā (a refutation of the Dvaita view of Vedänta) is noticed in Mad- ras Cat. Trm I, B, 362, and is said to have been composed by Śrīvatsalāñchana Sarman, son of Vişņudhvajācārya. As Śrīvatsa cites Vidyānātha, he cannot be earlier than the 14th century ; on the other hand, he is earlier than the 17th century, having been quoted by Kamalākara (1612 A.D.) and Jagan- nātha (p. 39). It appears also that the Sāra-bodhinī in many places expands or condenses Paramananda's Vistārikā. The colophon to the BORI MS of Sāra-bodhinī (no. 107, Cat. xii p. 115) informs us that Srīvatsalanchana Bhattācarya's father was Śrīvişņu Bhattācārya Cakravartin. MSS. Extract in Madras Catalcgue xxii, 12827; also BORI MS Cat. xii, no. 54, pp. 56-57 (extract) of Kāvya-parīkşā.

Paņditarāja This commentator, cited by Ratnakantha, is probably identical with Raghunandana Rāya, disciple of the legist Maheśa Țhakkura (See Jha's transl. of Kāv. prak, introd. p. ix). He should not be confounded with Jagannātha Pandita- rāja. Stein pp. 60, 269, extract given MS no. 1164 (Aufrecht i. 19a). Stein's MS of this work goes up to Ullasa ii only and quotes no authorities except the Miśras and Pratyabhijñākāra. Jha's MS appears to have been copied in 1637 A.D.

Ravi and Ratnapāņi Ravi is the author of the Madhumati commentary, the last verse of which tells us that he had a beloved daughter named Madhumatī, after whom the commentary was baptised, He also informs us that he was son of Gauri and Manodhara,

1 Autrecht 1 103a, ii 20a.

COMMENTATORS ON MAMMAȚA 167

alias Ratnapāni, and grandson of Acyuta, who was a minister of Sivasimha or Sivasiddha of Mithila (about the middle of the 15th century, 10C iv p. 875f). Ratnapāni or Manodhara also wrote a commentary or Mammata, called Kāvya-darpana1, which is cited by his son and on which the latter's commentary itself seems to have been based. The father and the son are cited by Bhīmasena, while Kamalākara and Narasimha cite the Madhumatīkāra. Extract in Peterson iii, p 332f; Madras Cat. xxii, nos. 12822-23.

Maheśvara With the title Nyayālamkara. His commentary is called · Ādarśa, or · Bhāvārtha-cintāmaņi. He is a Bengal writer, who also composed a commentary on the Dāyabhāga. As he is cited by Vaidyanatha he should be placed before the middle of the 17th century, and it is probable that he flouri- shed about the commencement of that century. Edition, By Jivananda Vidyasagar, Calcutta 1876 ; also ed. Calcutta Skt. Series 1936.

Kamalākara Bhațta He is better known as a legist, and wrote a large number of works on Smrti and Mīmāmsa. He was a Mahratta Brahman of Benares, son of Rāmakrsņa Bhatta and Umā, younger brother of Dinakara Bhatta, and grandson of Nārāyana Bhatta and great-grandson of Rāmeśvara Bhatta2, Ananta Bhatta, who wrote the Rama-kalpadruma at the request of Gariba-

1 See Peterson iii, App. p. 332 where extract is given (esp. śl. 5), A MS of this work, called Kāvya-darpana or Kāvyaprakāsa-darpaņa is noticed in Mitra 3169, and the author's name given as Manodhara. 2 For Kamalakara's place in the Bhatta family of Benares, see the pedigree given in V. N. Mandlik's ed. of Vyavahāra-mayūkha p. lxxvi. See also Bhandarkar, Rep. 1883-84, pp. 50-1. He is also the author of numerous works on Mīmāmsa and Smrti, some of which he mentions at the end of his comm. (see Aufrecht i. 80). It is said that his descendants still live in Benares.

168 SANSKRIT POETICS

dāsa, minister of Rājā Rājasimha, was his son. Kamalākara's date is known from the fact that he dates his well known Nirnaya-sindhu in Samvat 1668=1612 A.D. He also wrote a poem called Rāma-kautuka in 4 cantos. Edition. By Papa Sastri, Benares 1866. This comm. is described with extract in IOC iii, no. 1143/361, p. 327.

Rājānaka Ānanda His commentary is called °Nidarsana or Śitikantha-vibo- dhana 1. Hall (Vāsavadattā p. 16) is mistaken in attributing this work to Sitikantha and taking it as dedicated to Ānanda. The colophon, as well as the first verse2, of this commentary accounts probably for Hall's mistake ; but the author himself explains that the commentary is so named from the fact that an attempt is made in it to interpret Mammata's text as having, besides its ālamkārika meaning, a mystical sense relating to Sitikaņtha or Siva. Mammata himself might have been a Kashmirian Saiva; but this leaves no doubt that Ananda was one, although it is doubtful whether the text lends itself to such an interpretation. The date is given in the colophon. as 1665 A.D., although Hall thinks that this is the date in which the MS of the work was copied. Stein remarks: "Ananda, who composed his commentary in 1665 A.D. is still well remembered in the tradition of Kashmirian Pandits as the contemporary and friend of Rājānaka Ratnakantha3",

1 The colophon, as quoted by Bühler (Kashmir Rep. p. 69 fn), says : iti śrīmad-rājānakānvaya-tilakena rājānakānandakena viracitam kāvya- prakāśa-nidarsanam. But elsewhere in the Jammu MS of Stein, it says: iti śrī-kāvya-darśane śitikantha-vibodhane kāvyoddeśa-darśanam prathamam, col. to ch. i. It seems that the real name of the comm. is °Nidarśana as Peterson thinks, Sitikantha.vibodhana being an alternative or descriptive name arising from the second meaning relating to Sitikantha or Siva which the commentary finds in the text. 2 praņamya śāradām kāvya-prakāśo bodha-siddhaye/ padārtha- vivṛti-d vārā śitikanthasya darśyate. Jhalakikara reads: sva-śişyebhyah pradarśyate, but remarks: atra śitikaņthasya darśyate iti pātho vivaraņa- kārair angīkrtah, 3 Jammu Cat p. xxvii fn. The date is given in Kali era 4766.

COMMENTATORS ON MAMMAȚA 169

one of whose known dates is 1648 A.D. Ananda, therefore, may be assigned to the second and third quarters of the 17th century. Ananda also appears to have written a commentary on the Naişadha. Extract in Peterson (i, 74); also in Stein, Jammu Catalogue p. xxvii. Rājānaka Ratnakaņtka His commentary is called Sāra-samuccaya which, as its title indicates and the author confesses, was composed by a compilation from "the principal commentaries of Jayanta and others." He cites, therefore, some of the well-known commentaries before his time, 1 among which we find the Šāhitya-dīpikā (of Bhāskara), Sāra-bodhinī (of Śrīvatsa), the commentaries of Subuddhi Miśra and Paņditarāja, the Vistārikā (of Paramānanda), the Pradīpa (of Govinda), and two other works Tattva-parīksā and Rasa-ratna-dīpikā. This is the same Ratnakantha as copied the codex archetypus of the Raja-tarangini, mentioned by Stein (introd. p. vii f) and also transcribed MSS of the °Samketa of Ruyyaka in 1648, of Rayamukuta's commentary on Amara in 1655, and of Trilocanadāsa's Kātantra-pañjikā in 1673 A.D. He is identical with Ratnakantha, who was the son of Samkarakaņtha and grandson of Ananta-kavi of the Dhaumyayana family and who wrote a Stuti-kusumāñjali-țīkā (called Sisya-hitā) in 1681 A.D., and a Yudhişthira-vijaya-kāvya-țīkā in 1672 A.D. (Aufrecht i. 489b ; Stein, loc. cit). These dates range from 1648 to 1681 A.D., during which apparently the literary activity of Ratnakantha falls. Extract in Peterson, Report ii, p. 129 (also ii, 16f), which gives a list of authors quoted ; BORI MS no. 113 (Cat. xii, p. 121). Narasimha Țhakkura His commentary is called Narasimha-maņīşā. He belonged to the same family as Govinda Thakkura and was fifth in

1 The list of authors cited by him is given in Peterson, Report ii p. 17f.

170 SANSKRIT POETICS

descent from him. The latest writer he cites seems to be the Madhumatī-kāra (Ravi) and Kamalākara, and he is cited in his turn by Bhīmasena with the title nyāya-vidya-vagisa. Between 1620-1700 A.D. MSS. Aufrecht i. 101b, ii. 19b.

Vaidyanātha Tatsat He wrote two commentaries : (1) the ° Prabha on the °Pradipa of Govinda and (2) the Udāharana-candrika on the illustra- tive verses of the Kāvya-prakasa. The date of the latter work is given in the concluding verse as Samvat 1740=1684 A D.1 He also wrote Alamkāra-candrīkā commentary on Appayya's Kavalayānanda (q. v.). He is different from Vaidyanātha, the Maithili grammarian, son of Mahādeva and Veņī and pupil of Nāgojī Bhatta; for our Vaidya- nātha is known as son of Rāmacandra (or Rāmabuddha) Bhatta and grandson of Vitthala Bhatta of the Tatsat family, and is referred to by Nagojī himself. Our Vaidya- nātha cites Candīdāsa, Subuddhi Miśra, the Dīpikākrt (Govinda's Udāharaņa-dīpikā), Cakravartin and Maheśa, and is cited by Bhimasena. He is probably not the same as Vaidyanātha Pāyagunda who wrote commentaries on the Candrāloka of Jayadeva and the Paribhāşendu-śekhara of Nāgojī (ed. Anandasrama, Poona 1913). Edition. Prabhā, ed. with Pradīpa by Durgaprasad and K. P. Parab, NSP, Bombay 1891, 1912 (our references are to the ed. of 1912). Udāharaņa-candrikā. Extract in Peterson Report ii, p. 108, in SCC vii, 54, in IOC iii, 1151/943b.

Bhīmasena Dīkşita His commentary is called Sudha-sagara or Sudhodadhi2. It is dated in Samvat 1779=1723 A.D.3 He was a Kanauj

1 The date is given in the IOC MSS Cat. iii, p. 322, no. 1151. 2 The form Sukhodadhi given in Peterson's extract (i, p. 94) should be Sudhodadhi. 3 Peterson Report ì, p. 94.

COMMENTATORS ON MAMMAȚA 171

Brahman who describes himself as son of Sivananda and grandson of Muralidhara; his genealogy being given thus: Gangādāsa-Vīreśvara- Muralīdhara- Śivānanda- Bhīma- sena. He is also the author of two independent works, called Alamkāra-sāroddhāra and Kuvalayānanda-khandana1, the latter apparently directed against Appayya's work of the same name, and both referred to in his commentary on Mammata. The last work was composed at Jodhpur while Ajitasimha (1680-1725 A D.) was still reigning. Bhimasena also wrote a commentary on the Ratnavali2. He cites a large number of commentators, such as Candīdāsa, Bhāskara, Acyuta, Ratna- pāņi, Ravi, Jayarāma Pañcānana, Vācasapati Miśra, Cakra- vartin, Ruci Miśra, Murāri Miśra, Pakşadhara Upādhyāya, Devanātha Tarkapañcānana, Śrīvatsalñchana, Govinda and Narasiņha Țhakkura, Maheśa or Maheśvara, and Vaidyanātha. Edition. By Narayan Sastri Khiste in Chowkhamba Sansk. Ser. Benares 1927. Extract in Peterson i, p. 94 and in Jhala- kikara's ed. of Kāv. prak.

Baladeva Vidyābhūşaņa Known simply as Vidyābhūșana. His work, dignified with the name of "a vrtti on Bharata's sūtra" (so he calls Mammața's Kārikās3), is named the Sāhitya-kaumudī, on which he himself writes a tippana, called Krsnānandinī. It has the same arrangement and subject-matter as those of the Kāvya-prakāśa, but it adds an eleventh chapter on some extra Alamkāras of Sabda and Artha. A work called Kāvya- kustubha in 9 Prabhās is noticed and attributed to one Vidyābhūșaņa (in Stein pp. 59, 268) who appears to be a

1 The work is also called Alamkāra-sāra-sthiti, a MS of which is noticed in Mitra 4084 (Aufrecht ii. 23a). A very incorrect and at places illegible MS of this work exists in the BORI (Cat. xii, no. 156, pp.179-80, which gives a list of its citations). See also H. P. Sastri, Cat. ASB MSS, vi, no. 4895/3147, p. 456. 2 Aufrecht i. 492. 3 See above pp. 152f.

172 SANSKRIT POETICS

Vaisnava and probably the same as our author (see pref. to Kāvya-mālā ed. of the Sāhitya-kaumudī and Aufrecht i 101a, ii. 19b, 193b, iii. 22b) ; for this work see below under Minor Writers. Baladeva was a pupil of Rādhādāmodara-dāsa (concluding verse of Sah. k. and its commentary) and Gopāladāsa (alias Rasikānanda, commentary on śl. 1), and the Guru of Uddhava-dāsa. He was a Vaisnava and follower of Caitanya, and wrote various Vaisnava works. Though belonging to Orissa he was a champion of the later phase of Bengal Vaisnavism, and attempted a rapproachement between Madhvaism and Caitanyaism; see on this S K. De, Vaisnava Faith and Movement in Bengal, Calcutta 1942, pp. 11-12. Apart from his commentary mentioned above, his most notable works are Govinda-bhāsya on the Vedānta- sūtra and Prameya-ratnāvalī. He is said to have been a contemporary of Jayasimha of Jaypur, who flourished in the beginning of the 18th century. Aufrecht notes that his. commentary on the Utkalika-vallarī was written in 1765 A.D. A pun in the first verse of his Sāhitya-kaumudī refers, as he himself explains in the tippana, to Gajapati Pratāparudra of Utkala or Orissa. Edition. With Krsnānandini, ed. Sivadatta and K. P. Parab, Nir. Sag. Press, Bombay 1897. The Bharata-sūtra-vrtti in Peterson ii, 10 is the same work as this.

Nāgojī or Nāgeśa Bhatta He wrote two commentaries called Laghu- and Brhat- Uddyota on Pradīpa. Also author of a °Udāharaņa-dīpikā or °pradipa on the illustrations in the text (Stein, pp. xxvii, 268). He belongs to the first quarter of the 18 h century. He also wrote a commentary on Jagannatha (q.v.), from whom he was removed by two generations, and other works. See below under Jagannātha. Edition. °Uddyota, ed. with 'Pradipa in Anandasrama Series 1911 ; ed. with text and °Pradipa (ch. i, ii, vii and x by Chandorkar, Poona 1889, 1915.

COMMENTATORS ON MAMMAȚA 173

°Udāharaņa-dīpikā, Aufrecht ii. 19b (extract in Stein pp. 268-69, called °Pradīpa).

( 2 ) We give below the names of the some of the less known commentators on Mammata (alphabetically) : (1) Kalādhara, Wrote Kārikāvali, which appears to be a synopsis of the Kārikās. KBod 501. (2) Kalyāņa Upādhyāya. Name of comm. unknown, mentioned by Jha in his introd. to his trans. of Kāv. prak. p. ix. (3) Krsna Dvivedin. Comm. Madhura-rasa. Aufrecht i. 101b. (4) Krşņa Śarman. Rasa-prakāśa. HPS iii no. 58 (ex- tract ; ends with the 5th chapter) ; Cat. ASB MSS, vi, no. 4842/6581, pp. 419-20 (a fragment of only 20 leaves, containing the first and second chapters). (5) Krsnamitrācārya, son of Rāmanātha and grandson of Devīdatta: a Naiyāyika, for whose works see Aufrecht i. 121b. Țīkā. Aufrecht i. 101b. (6) Gadādhara Cakravartin Bhațțācārya. °Țīkā. Mitra 1527 ; SCC vii 13. Gadādhara is well known for his commentary on Raghunātha Siromaņi's Tattva- cintāmani-dīdhiti, one of the standard works on Navya Nyāya of Bengal. He was a pupil of Harirāma Tarkālamkāra and flourished at the end of the 16th and beginning of the 17th century. (7) Guņaratna Gaņi. Comm. Sāra-dīpikā (BORI Cat. MSS xii, p. 112). The MS is dated Samvat 1890. (8) Gokulanatha Upadhyaya, the Maithili Smarta. °Țīkā, mentioned in Jha op. cit. p. ix. See chapter on Minor Writers below. (9) Gopīnātha. Comm. Sumano-manoharā. Aufrecht i. 101b. He also wrote a comm. on Viśvanātha's Sāhitya-darpaņa (q.v.). End of the 17th century. (10) Jagadīśa Tarka-pañcānana Bhatțācārya. Comm.

174 SANSKRIT POETICS

° Rahasya-prakāśa. Aufrecht i. 101b (Mitra 1651). MS written by his pupil in Saka 1579=1657 A.D. He is different from Jagadīśa Tarkālaņkāra, the famous Naiyāyika who was a pupil of Bhavā- nanda and (Rämabhadra) Sārvabhauma of Nava- dvipa (Bengal). (11) Janārdana Vibudha, pupil of Ananta. Comm. ° Śloka dīpikā. Aufrecht i. 101b, ii. 19b (Stein 61, incomplete). He also wrote commentaries on the Raghu and Vrtta-ratnakara. He should be distinguished from the better known Janārdana. Vyāsa, son of Bābuji Vyasa and grandson of Vithala Vyāsa and pupil of Jayarāma Nyāya- pañcānana (see above p. 164). (12) Tiruvenkata, son of Cinnatimma and grandson of Tirumala-guru. A South Indian writer who quotes. Bhatta Gopala's commentary. Madras Trm. Cat. A 318. (13) Devanātha Tarkapancānana. Comm. Kāvya- kaumudī written in Samvat 1717 (=1661). BORI MSS Cat. xii, p. 81. Described as. son of Govinda. Cited by Kamalākara and Bhīma- sena. Bharata Mallika on Bhatti x. 73 quotes. one Devanātha. Our Devanātha appears to be a logician of Bengal who upheld the views of Mam- mata against the adverse criticism of Viśvanātha. See Madras Trm II, C, 1570 for extracts, and II, A, 819. See Mitra 1447 where mention is made of an Ekaşaştyalamkāra-prakāśa which quotes Deva- nātha and Jayarāma as its sources. (14) Narasimha Sūri, son of Timmaji Mantrin and. grandson of Rangaprabhu. Comm. Rju-vrtti on Kārikās only. Aufrecht ii. 19b ; Madras Trm B. 381. (15) Nāgarāja Keśava. Comm. Pada-vrtti. Aufrecht i .- 101b.

COMMENTATORS ON MAMMAȚA 175

(16) Nārāyaņa Dīkşita, son of Ranganātha Dīkșita anď brother of Bālakrsna. Ranganātha's commentary on the Vikramorvasi was finished in 1656 A. D. Hence the date of our author is the end of the 17th. century. °Țīkā. Aufrecht i. 101b (see also 292a : AFI p. 155) (17) Bhanucandra. °Țīkā. Aufrecht i. 101b. Also wrote a commentary on the Daśa-kumāra. (18) Bhavadeva, son of Krsnadeva of Mithila and pupil of Bhavadeva Thakkura. Comm. Līlā. Aufrecht ii. 20a ; Madras Cat. 12824-25 (extract). Also wrote a commentary on the Vedānta-sūtra (IOC 1428). According to the final verse in this commentary the author lived in the reign of Shah Jahan and. composed his commentary at Patna in Saka 1571 =1649 A. D, (19) Madhumatiganeśa. Comm. Kāvya-darpana, Au- frecht i. 102a. (20) Yajnesvara Yajvan. Comm. Vyākhyā. Madras Cat. 12821 (extract). (21) Raghudeva. Comm. Kārikārtha-prakāśikā. Au- frecht ii. 20a (up to the end nearly of Ullasa ii) (22) Ratneśvara, Name of Comm. unknown, but referred to by himself in his comm. on Bhoja (cf ABod 209a). (23) Raghava. Wrote an Avacūri Țippana, mentioned in Jhalakikara p. 36. (24) Rājānanda. Comm, without a name. Madras Cat. 12820 (extract) ; cf Aufrecht ii. 20a. (25) Rāmacandra. Wrote a Kāvyaprakāśa-sāra, which is apparently a summary exposition of the sub- stance of the text. Aufrecht i. 102b. (26) Rāmanatha Vidyavācaspati, a Bengal commen- tator who wrote the comm. Rahasya-prakāśa. Aufrecht i 102a. His commentary on Bhavadeva's.

176 SANSKRIT POETICS

Samskāra-paddhati was composed in 1623 A.D. (see Aufrecht i. 516a). (27) Rāmakrsņa. Comm. Bhāvārtha or Kavi-nandinī (or °nandikā). Aufrecht i. 102a, ii. 20a: also ii. 16b. (28) Vijayānanda. Wrote a °Țīkā. Deccan Coll. Cat. p. 44. The age of the MS is given as 1683 A.D. (29) Vidyāsāgara, apparently the title of some commen- tator. Cited by Śrīvatsalāñchana. One Vidyāsāgara wrote a Kalā-dīpikā-țīkā on Bhatti, and is cited by Bharata Mallika (on x. 73), and by Rāmanātha on Amara-kośa. S. P. Bhattacharya (introd. to ed. Śrīdhara's comm. p. xxx) is inclined to take this commentator on Mammata as Pundarīka Vidyāsāgara who flourished in the early decades of the 15th century. He is said to have written comm. also on Dandin and Vāmana. (30) Venkațācala Sūri, Comm. Subodhinī. Aufrecht i. 102a. H. P. Sastri, Cat. ASB MSS, vi. no. 4837/, 8736, p. 415. (31) Śivanārāyaņa Dāsa Sarasvatīkaņțhābharaņa, son of Durgādāsa. Comm. °Dīpikā. Weber i, no. 819; Aufrecht i. 102a. He wrote at the beginning of the 17th century. For his other works, see

(32) Aufrecht i. 649b. Sivarāma Tripāthin. Comm. Vişamapadī. Kielhorn, Central Prov. Cat. p. 107. See below under Minor Writers for the author. (33) Siddhicandra Gaņi, Kāvyaprakāśa-khandana in ten Ullāsas, edited by Rasiklal C. Parikh from the Bharatiya Vidya Bhavan, Bombay 1953. The writer is a notable Jaina monk and author (b. 1587-88 A.D.) in the time of Akbar and Jahangir and a contemporary of Jagannatha. His method, as he himself tells us, is anuvāda-pūrvaka khandana, i.e. first exposition and then criticism,

COMMENTATORS ON MAMMAȚA 177

although all his criticisms are not justified. Siddhi- candra appears also to have written a Bṛhatī Țīkā on the Kāvya-prakāśa. While Hemacandra would accept Mammata's work as the standard, Siddhi- candra was perhaps one of the 'Navyas', like Jagannātha, who indicates an attempt to set up a new school of poetic theories. For information about the author and his works see Introd. to ed. mentioned. - MS entitled Kāvyaprakāśa-khaņdana or Kāvyāmrta-tarangiņī noticed by Mitra 2674 goes up to the 7th Ullasa; it appears to be a different work. Besides these, there are numerous commentaries, either anonymous or with the name of the author missing, some which are entered in Aufrecht i. 101b, 778b, ii. 20a, 193b. This illustrates the saying of Maheśvara, one of the commentators: kāvya-prakāśasya kṛtā grhe grhe| fīkā tathāpy eșa tathaiva durgamaḥ !

12

CHAPTER VII

FROM RUYYAKA TO VIDYĀNĀTHA

RUYYAKA

(1 ) Ruyyaka, who also bore the name of Rucaka1 and had the Kashmirian title of Rājānaka prefixed to his name, was son of Rājanaka Tilaka2 who, Jayaratha informs us (pp. 115, 124, 205), wrote a commentary or critique on Udbhata called Udbhața-viveka or Udbhata-vicāra. The Alamkara-sarvasva, by which Ruyyaka is chiefly known, consists of two parts, viz. Sūtra and Vrtti ; but the question has been raised whether the authorship of the two parts should be attributed to the same writer. The Nirnay Sagar Press edition of the text, published under the above title, admits Ruyyaka's authorship of both the Sutra and the Vrtti, a view which is accepted by his earliest commentator

1 rājānaka-rucakāparanāmno'lamkārasarvasva-kṛtah (krtiḥ), col. to Pischel's ed. of Sahrdaya-līla; cf ed. of the same in Kāvyamālā Gucchaka v (1908), reading of MS kha. This name is given in some of the MSS of his larger work (e. g. col. to the N.S.P. and Trivan- drum eds. of the text ; in Mitra ix. p. 117), as well as by Kumārasvāmin (pp. 393, 396, 425, 448), Appayya Dīkșita (Citr. mīm. p. 72), Rāghava- bhatta on Šakuntalā (pp. 161, 179, 193), and Śrīvidyā-Cakravartin, one of Ruyyaka's commentators (Madras Cat. xii, p. 8609). Bühler (Kashmir Rep. no. 247, p. xvi) and Peterson (ii p. 13f) also found this name in connexion with Ruyyaka's Samketa comm. on Mammata. That the form Ruyyaka is the more authentic is testified to by his pupil Mankhaka (Śrīkantha-c. xxv. 30 : see below). 2 rājānaka-tilakātmaja° col. to Pischel's ed. of the Sahrdaya-līlā. Ruyyaka's father was thus, like himself, a writer on Poetics and a follower of Udbhata. See above p. 76-77. Tilaka is mentioned and quoted (one verse) by Someśvara in his comm. on Mammata (ed. Parikh, p. 295, on Ullāsa x. 106).

RUYYAKA 179

Jayaratha, who refers to the author as granthakrt with respect to both the Sutra and Vrtti portions.1 The invocatory verse to the Vrtti in this edition reads in the second line nijālamkāra-sūtrāņām vrttyā tātparyam ucyate so that the Sutra is avowedly referred to as his own by the Vrttikara.2 Some doubt, however, has been thrown on this view by the discovery of a different reading of this line in some of the South Indian MSS, which introduce a grave variant in the phrase nijālamkāra-sūtrāņām by changing it into gurvalamkāra-sūtrāņām. There is also, at the end of the Vrtti in these MSS an additional verse3, which gives the name of the Vrtti (and not of the text which is called Alamkāra-sūtra) as Alamkāra-sarvasva, and the name of the author of the Vrtti as Mankhuka or Mankhaka, who is described as a sāmdhivigrahika to the king of Kashmir. This forms the reading of the text (based on three MSS), which is published in the Trivandrum Series, as well as of MSS noticed by Burnell,4 Winternitz5 and in the Madras Catalogue6. This view is supported by Samudrabandha, a South Indian commentator, who flourished at the end of 13th century and whose text is printed in the Trivandrum edition ; for the Vrtti, the subject of his commentary, is known to him by the name of Alamkāra sarvasva and its author as Mankhaka, while the original work of Ruyyaka is called Alamkāra-sūtra.

1 pp. 19, 20, 55, 57, 67, 72, 83, 87 etc. 2 Same reading in ABod 210a, where Ruppaka is a mistake or a variant (Bühler op. cit. p. 68) for Ruyyaka; Mitra ix p. 117 .- All the Jammu MSS have nijālamkāra°. 3 iti mankhuko vitene kāśmīra-kşitipa-sāmdhivigrahikaḥ/ sukavi- mukhālamkāram tad idam alamkāra-sarvasvam / / 4 Tanjore Cat. p. 54a. 5 Cat. of South Ind. MSS in the RAS, p. 208; cf Jacob in JRAS, 1897, p. 283f. 6 xii, pp. 8606-7. The question is discussed at some length in Harichand Sastri, op. cit. p. 105f.

180 SANSKRIT POETICS

We are thus confronted with two distinct traditions, embodied in these two different views and prevailing in the North and South of India respectively, with regard to the authorship of the Vrtti, there being no dispute as to Ruyyaka's authorship of the Sutra-text itself. So far as we can judge, the North Indian tradition, obtaining in Kashmir, to which place both Ruyyaka and Mankhaka belonged, seems to be the authentic one ; for the South Indian tradition is not uniform in this respect and does not always distinguish between the author of the Sutra and the Vrtti respectively. Thus, Appayya Dīksita, a noted South Indian writer, refers1 to Ruyyaka or Rucaka as the author of the work as a whole, which is called the Alamkāra-sarvasva, attributing to him both the Sutra and the Vrtti; and Appayya is in agreement, in this respect, with Mallinātha2, Kumārasvāmin8, and Jagannätha4. On the other hand, the testimony of Jayaratha, himself a Kashmirian, cannot be very well superseded by what the much later commentator Samudrabandha says in conformity to a tradition which itself is not unanimous. It is also significant that while Ruyyaka (and even Jayaratha) is quoted and discussed extensively by later writers on Poetics, Mankhaka is not cited as a writer on the subject except once by Appayya in his Citra-mīmāmsā (p. 10).

( 2 ) This tradition of Mankhaka's collaboration with Ruyyaka would not perhaps have arisen, had not Mankhaka, as the

1 Citr. mīm. p. 14, 15, 54, 72, 84, 90, 94, 98 ; Kuvalay. p. 41, 89, 92, 96, 184. 2 Taralā pp. 21, 186, 187, 232, 237, 249, 261, 262, 266, 331, 332. 3 Ratnāpaņa pp. 393 (=Alam. sar. vṛtti p. 58), 425 (=ibid, p. 133), 448 (=ibid, p. 144) ; p. 341 (=ibid, sūtra p. 20) p. 452=(ibid, p. 156). 4 Rasa-gangādhara has numerous references, but see pp. 163 and and 200, where both the Sūtra and Vrtti are quoted under the citation Alamkāra-sarvasva. See also pp. 251, 342-43, 352, 482. Also Rāghava- bhatta on Šakuntalā p. 161 (=Alam. sarv. vṛtti p. 64), p. 179 (=ibid, p. 75), p. 193 (=ibid, p. 127).

RUYYAKA 181

tradition says, been in fact a pupil of Ruyyaka.1 Rājānaka Mańkhaka or Mankhuka, son of Viśvāvarta and grand- son of Manmatha, is well known as the Kashmirian author of Srikantha-carita (ed. Durgaprasad and K. P. Parab, NSP, Bombay 1887) which was written, according to Bühler,2 between 1135 and 1145 A. D, Mankhaka's brother Alamkāra (or Lankaka, xxv. 15, 37f) was a minister (v. 62, xxv. 43, 61) under Sussala and Jayasimha of Kashmir (1129-1150 A.D.), and another brother Srngāra held a high office (brhat-tantrapati) under Sussala, whom he assisted in his war against Harsadeva. Mankhaka tells us how after composing his poem he submit- ted it, at the house of his brother Alamkāra, to an assembly of learned scholars and officials, among whom he describes Ruyyaka as his own preceptor (xxv. 30, 135). This also ex- plains how Ruyyaka's own work contains five verses from Mankhaka's poem3 cited as illustrations ; for it is not unlikely that the Guru should in this way quote his worthy disciple. As the latest date of Mankhaka's poem is given as 1145 A.D., we may presume that Ruyyaka's Alamkāra-sarvasva, which quotes it, was composed a little later. Again, Māņikyacandra's Samketa (on Mammata) which was composed in 1159-60 A.D. quotes the Alamkara-sarvasva. We can, therefore, fix Ruyyaka's literary career in the second and third quarters of the 12th century4.

1 This fact, as well as what Jayaratha says about corruptions and additions to the text of the Alamkara-sarvasva, would perhaps explain how the tradition of Mankhaka's collaboration began. 2 op. cit. p. 50f; extract App. pp. cix f. See also Raja-tarang. viii. 3354. 3 See Jacob in JRAS, 1897, p. 283 for these verses (ii. 49, iv. 79, v. 23, vi. 16, x. 10). 4 Jacob (op. cit. p. 283) points out that Ruyyaka (p. 93) quotes Rāja-tarangiņī iv. 441 (asamāpta-jigīsasya), which work was not com- pleted till about 1150 A. D. in the reign of Jayasimha. This verse, however, occurs an an anonymous quotation in Abhinava's Comm. on Bharata (ch. vi, vol. i, p. 305). It should be borne in mind that Jayaratha

182 SANSKRIT POETICS

In this work, composed in the Sūtra-vrtti style Ruyyuka concerns himself only with poetic figures. After dealing with Punar-uktavadābhāsa, Anuprāsa (Cheka-, Vrtti- and Lāța-), Yamaka and Citra he goes on to discuss 75 figures of Artha beginning with Upama. He gives two altogether new figures Vikalpa (p. 159) and Vicitra (p. 133-34). His work is mostly drawn upon by later writers like Viśvanātha, Vidyānātha and Appayya Dīkşita. Ruyyaka quotes the Kāvya-prakāśa in many places (p. 107 on Paryāyokta; p, 102=KP iv, p. 128 ; p. 183 definition of Bhavika), and the definitions of Citra, Kāvya- linga, Vyājokti, Uttara, Milita and Samādhi are given as they occur in Mammata's work.

( 3 ) The works of Ruyyaka are numerous, of which three only have been printed : (1) Kāvyaprakāśa-samketa, a commentary on Mammața's text, referred to as Ruyyaka's by Jayaratha p. 102, and by Ratnakantha (Peterson ii, pp. 17, 19 as Brhat-samketa). For ed. see above p. 156. (2) Alamkāra-mañjarī, referred to by himself at p. 15. Not mentioned by Jayaratha as Ruyyaka's. P. V. Kane doubts if it is a work by Ruyyaka. (3) Sāhitya-mīmāmsa, referred to by himself at p. 61, also by Jayaratha p. 126. It is cited without the author's name in Vidyānātha p. 11 (Cf ABod 210a). Burnell mentions an anonymous metrical Sāhitya-mīmāmsā (p. 58a), with prose Vrtti in eight Prakaranas. This is apparently the Sāhitya- mīmamsa published in the Trivandrum Skt. Series in 1934.1 The MS sources contain large gaps and lacunae, and the name of the author is not given either at the beginning or at the

frequently complains of unauthorised additions and corruption of the text itself (pp. 50. 67, 107, 124, 126 etc), and discusses readings (pp. 21, 37, 49, 172 etc.). Jayaratha himself quotes twelve verses (p. 194) from the Raja-taranginī, which describe Lalitāditya of Kashmir. 1 See above p. 144.

RUYYAKA 183

end. A résumé of this work has been given by P. V. Kane (HSP, pp. 269-72), who believes it to be a work of Ruyyaka. It is noteworthy, however, that it speaks not of Vyañjanā, but of Tätparya-vrtti as leading to the realisation of Rasa, a view which fundamentally differs from that of Ruyyaka who declares distinctly (p. 13): asti tāvad vyangya-nistho vyāpāraḥ. It shows some influence of the Srngāra-prakāśa of Bhoja. V. Raghavan (pp. 99-100) doubts if it is the work of Ruyyaka. (4) Alamkārānusāriņī, cited by Jayaratha as Ruyyaka's (pp. 36, 57, 58 and 60). This work is supposed by Peterson1, and following him by Aufrecht2 and Jacobi3, to be a commen- tary on Jahlaņa's Somapāla-vilāsa. As Jayaratha's citations indicate, it discussed some of the poetic figures in that com- position, but there is nothing in these citations to show that it was in fact a commentary on Jahlana's Somapāla-vilāsa. The poet Jahlana is described in Mankhaka's Srīkantha-carita (xxv. 75) ; and Somapāla, king of Rājapurī (near Kashmir), whose life Jahlana appears to have recorded, is known from the Rāja-taranginī (viii, 621f) to have made war against Kashmirian Sussala. It is clear that this poet, who must have flourished in the first half of the 12th century, should be distinguished from Bhagadatta Jahlanadeva, the compiler of the Sukti-muktavali, who lived in the second half of the 13th century, but to whom Aufrecht (i. 203a) wrongly attributes the Somapāla-vilāsa itself. (5) A commentary on Mahimabhatta's Vyakti-viveka, referred to by Jayaratha as Vyaktiviveka-vicāra (p. 13). This has been identified with the anonymous commentary published with Mahimabhatta's work in the Trivandrum Sanskrit Series ; for, although the author's name does not appear in the publish- ed text of that commentary, which is called simply °V yāk hyāna,

1 Report ii, p. 17; introd. to Subhāșo. p. 106; Actes du 6me Cangrés p. 364. It is so described by Ratnakantha in his comm. on Stuti-kusumāñjali. 2 Cat. Cat. i. 32b. 3 ZDMG lxii, 291 : cf Harichand Sastri, op. cit. pp. 105-106.

184 SANSKRIT POETICS

the anonymous commentator refers in it to his other works, the Sahitya-mīmāmsā (p. 32) and Harsacarita-vārttika (p. 44, 50), which we know to be two works of Ruyyaka referred to by himself in his Alamkāra-sarvasva and mentioned by Jaya- ratha. He refers in this work also to the Candrika (on the Dhvanyāloka), Kāvya-kautuka (p. 13), Hrdaya-darpaņa (pp. 1, 13), and Kuntaka's Vakrokti-jīvita (pp. 16, 32, 36, 44). (6) Nātaka-mīmāmsā referred to as his own in the above commentary on Mahimabhatta (p. 32). (7) Harşacarita-vārttika referred to as his own in Alam- kāra-sarvasva p. 61 and in Vyaktiviveka-vyākhyāna p. 44, 50. (8) Sahrdaya-līlā, edited in the Kāvyamālā Gucchaka 5, ås well as by Pischel (Kiel 1886, along with Rudra's Srngāra- tilaka). The work consists of four Ullekhas: (i) Guna, describing the ten excellences (Rūpa, Varna. Prabhã etc.) of a woman (ii) Alamkāra, speaking of ornaments of gold, pearls etc, unguents, flowers worn by a woman (iii) Jīvita, dealing with youth which is the essence of womanly charm, and (iv) Parikara, treating of the parapharnelia of beauty. (9) Alamkāra-vārttika, cited as Ruyyaka's by Jayaratha p. 71. (10) Śrīkaņtha-stava cited as madīya in Alamkāra-sarvasva p. 19. ( 4 )

THE COMMENT ATORS ON RUYYAKA

Alaka (or Alața?) This commentator is referred to by Ratnakantha in his own commentary on Mammata1, the name given being Rājā- naka Alaka. We have seen that the identification of Mam- mata's continuator Alaka, Alata or Allata with this commen- tator is open to doubt2, and nothing more can be said on the subject until his commentary is recovered.

1 Peterson, Report ii, p. 17 f. See above p. 169. 2 See above p. 149f.

COMMENTATORS ON RUYYAKA 185

Jayaratha His commentary is called Alamkāra-vimarsinī. He informs us that his father's name was Srngara, and his patron was king Rājarāja. From his commentary °Viveka on the Tantrā- loka of Abhinavagupta, we learn that the full name of his father was Srngararatha, who had two sons named Jayaratha and Jayadratha1. He also tells us that he was a pupil of Sankhadhara and Siva. Jayaratha's pedigree is given at some length in his °Viveka, and we learn that his great-grand- father's brother was Sivaratha2, who was a minister of king Ucchala of Kashmir (1101-1111 A.D.). As four generations intervene between the minister of Ucchala and Jayaratha, Bühler thinks that the latter lived in the beginning of the 13th century. Jacobi supports this conclusion and attempts to arrive at greater precision by identifying Jayaratha's (and his father's) patron king Rājarāja with Rājadeva who is mention- ed by Jonarāja in 2 Rāja-tarangiņī v. 79-91, and who lived in 1203-1226 A.D.3 'The citations in Jayaratha are numerous, and include the names of Bhāmaha, Dandin, Udbhata, Vāmana, Rudrața, the Dhvanikāra (=Ānandavardhana), the Vakroktijīvita-kāra, Abhinavagupta, the Vyaktiviveka-kāra, Bhoja and Mammața.

1 The Kashmirian MSS of the Alamkāra-vimarsinī examined by Bühler vacillate between these two names as the name of our author; and Peterson's MS of Ratnakantha's Sāra-samuccaya reads Jayadratha (ii p. 17), which is accepted by Aufrecht, although the published texts of the °Vimarsini, as well as the °Viveka, have Jayaratha, which Bühler himself considers (op. cit. p. 68) to be the correct name of the author of the two commentaries, the other being that of his brother .- Jaya- dratha wrote a Kavya in 32 cantos, named Hara-carita-cintāmaņi (ed. NSP, Bombay 1897) 2 v. 22 ; see the extract given in Bühler op. cit. App. pp. cli f. This Sivaratha is mentioned in Raja-tarang° viii. 111. 3 Jacob (JRAS, 1897, p. 283) came to the conclusion that Jayaratha must be placed later than the end of 12th century, for he quotes (p. 64) from the Prthviraja-vijaya, a poem dealing with Prthviraja of Delhi who fell in 1193 A. D. (cf Bühler, op. cit. p. 62).

186 SANSKRIT POETICS

Besides citing his own author's other works and Rājānaka Tilaka's work on Udbhata, Jayaratha also mentions some other works on Alamkāra unknown to us, viz. Alamkāra-sūtra (p. 150), Alamkāra-bhāşya (pp. 35, 46, 83, 138, 173) and Alam- kāra-sāra (pp. 88, 97. 171, 172, 184), as well as an Alamkāra- vārttika (p. 71) ascribed to Ruyyaka. The Alamkāra-bhāşya is also cited by Jagannātha (pp. 239, 365) ; while a work bearing the name Alamkāra-sara is mentioned in Kielhorn's list1 and in Peterson iii, App. p. 393 ; but this is probably a later work attributed to Bālakrsņa Pāyagunda who also wrote a commentary on Appayya's Citra-mīmāmsā (q. v.) and was thus later than the end of the 16th century2. The ·Vimarśinī is, in its turn, cited extensively by Jagannātha (pp, 325, 327, 352, 380, 387, 414, 418), chiefly because Appayya, against whom Jagannatha's attacks are primarily directed, follows Ruyyaka and Jayaratha pretty slavishly, Jayaratha appears also to have written another work on Poetics called Alamkārodāharana,3 in the concluding verse of which he refers to his 'Vimarsini. It appears from the résumé of the work given in Mitra 2442 that it was chiefly meant to adduce illustrations to Ruyyaka's text, which the limited scope of his °Vimarsini did not allow him to do properly.

Samudrabandha Samudrabandha flourished, as he himself tells us, in the time of Ravivarman, alias Samgrāmadhīra, king of Kolamba {Quilon) in the Kerala country (Malabar), and there are nu- merous illustrative verses in the commentary itself, which sing the praise of this king4. This king was born in 1266-67 A.D.

1 Kielhorn's List 18 ; see also Aufrecht i. 32b. 2 But see Pischel in GgA, 1885, p 765 ; contra Jacobi in ZDMG xlii 293. 3 See Jammu Cat. no. 806, p. 59. 4 e. g. pp. 48, 58, 76, 133, 149 (kolambādhipati) etc. For Ravivarman and Samudrabandha, see K. Kunjunni Raja, op. cit. pp. 211-13.

COMMENTATORS ON RUYYAKA 187

and crowned himself as king of Malabar on the banks of the Vegavatī in 1312-13 A.D .; so that Samudrabandha may be taken to have flourished towards the end of the 13th and the beginning of the 14th century. The citations from earlier authors are not so numerous in Samudrabandha, but he shows his familiarity with Bhāmaha, Udbhata, Vāmana, Rudrața, the Dhvanikāra and Ānandavar- dhana, Bhatta Nāyaka, the Vakroktijīvita-kāra, Mahimabhațta, Bhoja and Mammata. He also cites Udbhata's vrtti, pre- sumably on Bhāmaha's Kāvyālamkāra. He refers to the explanation of other commentators of the Alamkāra-sarvasva (pp. 55, 96, 145, 239) and discusses readings (p. 57).

Śrīvidyā-Cakravartin A commentary, called Samjīvanī or Alamkāra-samīīvanī on Ruyyaka is cited by Mallinātha1 and Kumārasvāmin2. Appayya3 and Viśveśvara4. They apparently refer to the author of this commentary by citing Cakravartin; for the two verses5, cited by them, are attributed to this commentator by Kumārasvāmin, whose other references show that Cakravartin wrote a °Samjivanī commentary on the Alamkāra-sarvasva. This Cakravartin, who is to be distinguished from Paramā- nanda Cakravartin, appears to be identical with Śrīvidyā- Cakravartin, two MSS of whose commentary on Ruyyaka, called Samjīvanī, are noticed in the Madras Catalogue6. This commentary refers to and is referred to by the same

1 pp. 31, 57, 221, 237, 324. 2 pp. 54 (Cakravartin as commentator on Alam. sarvasva), 319, 377, 383 (°Samjivanī as a comm. on Alam. sarv.), 387, 393. 398, 435, 449-50, 465. 3 Citr. mīm. pp. 7, 74. 4 Alam. kaust. p. 11. 5 as cited above = Ratnāpaņa p. 378. 6 xit, no. 12799-12800, pp. 8609-10. Jacobi wrongly conjectured Alaka to be the author of this ° Samjivanī (ZDMG lxii, p. 292).

188 SANSKRIT POETICS

author's Sampradāya-prakāsinī Brhatī Țīkā on Mammața1, and both the commentaries are mentioned together at the close of the former work thus: kāvyaprakāśe'lamkārasarvasve ca vipaścitām/ atyādaro jagaty asmin, vyākhyātam ubhayam tataḥ/ /, which indicates the popularity of the works of Mammata and Ruyyaka in later times. Śrīvidyā-Cakravartin calls his author Rucaka. He refers to Viśvanātha as sāmdhivigrahika. He is evidently a South Indian writer, belonging to the Saiva Sampradāya or Saiva sect2. The reference by Mallinātha should place him chronologically earlier than the end of the 14th century. He appears to have flourished in the court of Vīra Vallāla III (Hoysala) in the beginning of the 14th century (V. Raghavan in ABORI xiv, 1933, p. 256). A Rasa- mīmāmsa3 is also ascribed to Vidyā-Cakravartin, as well as a Bharata-samgraha on Dramaturgy and Rasa (ABORI, xiv, 1933, p. 257).

BIBLIOGRAPHY

Alamkāra-sarvasva Editions. (1) ed. Durgaprasad and K. P. Parab, Nirnay Sag. Press, Bombay 1893 (with Jayaratha's °Vimarśinī). (2) ed. T. Ganapati Sastri, Trivandrum Sansk. Series 1915 (with comm. of Samudrabandha); 2nd ed. 1926. Translated into German by H. Jacobi in ZDMG lxii, pp. 289-336, 411-58, 597-628. Our references are to the N.S.P, ed., unless otherwise specified.

1 ibid, no. 12826-28, p. 8627 ; Burnell 55a. Ed. Trivandrum Skt. Series 1926 ; see above p. I62. He also wrote a Laghu-Țīkā prior to his writing of the Brhati Tika in which this comm. is referred to. Only one of these commentaries has been published in the Trivandrum ed. 2 The Sampradāya-mata or Sāmpradāyikas are quoted by Prabhā- kara Bhatța in his Rasa-pradīpa (before 1583 A. D.) at pp. 11, 13, 32. 3 See V. Raghavan in ABORI, xvi, 1934-35, p. 140. It is mentioned by Vidyā-Cakravartin himself ir his comm. on Ruyyaka (rasa-mīmām- sāyam vistarah), but it is not clear whether it is a separate work.

HEMACANDRA 189

Commentaries. (1) Alamkāra-vimarśinī of Jayaratha, ed. with text in Nir. Sag. Press, as above, 1893. On MSS of Alamkārodāharaņa see Aufrecht i. 32a, 773a, ii. 6b ; W Bod 1157. (2) °Vrtti of Samudrabandha, ed. with text in Trivandrum Sank. Series 1915, as above. (3) *Samjīvanī of Śrīvidyā-Cakravartin. Madras Cat. xxii, 12799-12800 ; Aufrecht i. 32b. Not yet published,

Sahrdaya-līlā Edition. (1) ed. Pischel (with Rudra's Śrngāra-tilaka), Kiel 1886. (2) ed. Kāvyamālā Gucchaka v, 1908. Comm, on the Vyakti-viveka Edition. With the text of that work, ed. T. Ganapati Sastri (anonymous but attributed to Ruyyaka), Trivandrum Sansk. Series, 1909.

HEMACANDRA AND THE VAGBHATAS ( 1 ) The versatile and volumious Jaina writer Hemacandra directed his many-sided activity to the field of Sanskrit Poetics as well, and wrote a Kāvyānuśāsana with its Vrtti named Alamkāra-cūdāmaņi and a glass called Viveka, basing it chiefly on Mammata's work, but appropriating his materials from various sources1. While as a textbook it hardly super-

1 He appropriates, for instance, long passages, without acknowledg- ment, from Rājaśekhara, Abhinavagupta, the Vakroktijīvita-kāra, Mammata and others. A passage in Abhinava-bhāratī on Bharata ch. iv is copied (Hemacandra pp. 57-66) almost literally with only a general acknowledgment at conclusion : iti śrīmān abhinavaguptācāryah, etan- matam evāsmābhir upajīvitam veditavyam (p. 66). In the Vrtti, again, at p. 83, his remarks on the sthāyi-bhāva (pp. 83-84) are copied from the same source. His extensive appropriation from Rājasekhara has been already alluded to; see p. 118 above. At p. 316 he calls himself a follower of Bharata's views (bharata-matānusārī). His dependence on earlier works is so close as to amount at times to almost slavish imitation or plagiarism.

190 SANSKRIT POETICS

sedes the Kāvya-prakāsa, it is, like most of Hemacandra's other productions, more or less an industrious compilation, displaying its author's encyclopaedic erudition, but hardly constituting an original contribution to the subject. About Hemacandra and his time, we know perhaps more than we do with regard to other writers on Poetics, and the biographical and other details will be found collected together in Bühler's erudite little pamphlet on this author1. Hema- candra was born at Dhunduka or Dhandhukā (in Ahmedabad) on the full-moon night of the month of Kärttika in the Samvat year 1145=1088 A.D. of humble Bania parents, named Caciga and Pahini. He was originally named Cangadeva. He was initiated as a Jaina monk in Samvat 1150=1093 A.D., taking the name of Somānanda. He was a pupil of Devacandra of Vajra-śākhā, author of the Sthānaka-vrtti and the Santinātha-carita. He became a sūri or ācārya in Samvat 1166=1109 A.D. changing his name, again, into Hemacandra. He spent the greater part of his life, as the acknowledged head of the Jaina community at Anahilla-pattana, under the patronage of Jayasimha Siddharāja (1094-1143 A. D.) and his successor Kumārapāla of Gujarat (1143-1172 A.D), dying shortly before the latter is Samvat 1229=1172 A. D. at the ripe old age of 84 years. He wrote most of his works at the request of his patrons, of whom he converted Kumārapāla into Jainism in Samvat 1216=1160 A.D. Hemacandra wrote voluminous works on many branches of Sanskrit learning, such as grammar (Siddha-hemacandra, Śabdānuśāsana, Lingānuśāsana, Dhātu-pārāyaņa and Uņādi- sūtra), prosody (Chandonuśāsana), lexicon (Abhidhāna-cintā- maņi, Anekārtha-samgraha, Nighaņtu-śeşa, and Deśī-nāma- māla), besides works on Jaina Sastra. His stupendous learn- ing justifies his sobriquet Kalikāla-Sarvajña. His Kāvyānuśā-

1 Ueber das Leben des Jaina Mönches Hemacandra, Wien 1889 ; trs. into English by Manilal Patel in the Singhi Jaina Series 1936. See alșo Jacobi in Ency. of Religion and Ethics, vi, 591.

THE VĀGBHAȚAS 191

sana in eight Adhyāyas has the merit of comprehending all topis of Poetics, including a brief reference to Dramaturgy. In spite of occasional differences Hemacandra borrows freely from Bharata, Anandavardhana, Abhinavagupta, Mammața and Rājasekhara. Its want of any striking originality perhaps stood on the way of its being accepted as an authoritative work. It exercised little influence on later writers and is scarcely ever quoted1. It is written in the form of Sutra and Vrtti.2 ( 2 ) There are two Vagbhatas in Sanskrit Poetics who must be distinguished from each other, viz. Vägbhata, author of the Vāgbhațālamkāra (here cited as Vāgbhața I) and Vāgbhața, author of the Kāvyānuśāsana and its Vrtti, Alamkāra-tilaka (here cited as Vägbhata II). Eggeling3 falls into the error of confounding the two and assigning both the works to the same author. From the Vāgbhatālamkāra iv. 148, we learn that the Jaina name in the Prakrit form of the author is Bahada and that he was son of Soma4. From the Kāvyānuśā- sana and its commentary5, on the other hand, we learn that its author was son of Nemikumara and Maha-(mahī ?- ) devī or Vasundhara; while his native town, calied Rahadapura from the shrine of a deity of that name, is mentioned6, as well as described in a verse by the author himself".

1 Except, as P. V. Kane notes (HSP, p. 278), by Ratnapana pp. 46, 75, 224, 233, 259, 299. 2 For [summary of the topics of Hemacandra's Kāvyānusāsana see below vol. ii, ch. vii (6). 3 IOC iii, pp. 330-1. 4 So also in Jinavardhana, Simhadeva and Ksemahamsa-gaņi's comms. on this verse. 5 p. 1 vrtti, and the concluding verse. 6 p. 1 vrtti. 7 IOC iii, p. 332. In the Nir. Sag. Press edition of the work, this verse is also given at p. 10, but the words asmabhir uktam, preceding it in the India Office MS, are wanting.

192 SANSKRIT POETICS

Vāgbhata II also appears to cite Vāgbhața I as one of his authorities1. Both the Vagbhatas, however, quote from the poet Vagbhata, author of the Nemi-nirvāna, Vāgbhata II citing the poem by name frequently for the purpose of illustrating the characteristics of a good poem (e g. p. 16)2. Vägbhata I may or may not be identical with the poet of the Nemi-nirvāna; but Vāgbhata Il should be distinguished from both3. We must also distinguish the medical writer Vāgbhața, son of Simhagupta. Vägbhata I seems to have been contemporaneous with Hemacandra, and lived under Cālukya Jayasimha Siddha- rāja of Anahilla-pattana4, who flourished from 1094 to 1143 A. D. We have references to this king and his capital in iv. 45, 76, 81, 85 and 132, and he is described as son of king Karnadeva. Both Jinavardhana Sūri and Simhadeva Gaņi in their commentaries explain that the prince referred to is Jayasimha, son of Karņadeva, of Aņahilla-pāțaka. It also appears from what Simhadeva Gani on iv. 148 says that Vagbhata was probably a mahāmātya of the said prince, a statement which is supported by the description given of our author in Prabhācandra Sūri's Prabhāvaka-carita5 (p. 205),

1 ii, p. 31: iti dandi-vāmana-vāgbhațādi-praņītā daša kāvya-guņāh, vayam tu mādhuryaujah-prasāda-lakşanāms trīneva guņān manyāmahe. 2 The verses quoted in Vāgbhatālamkāra from the Nemi-nirvāna are given by Jacob, op. cit. p. 309. 3 Winternitz thinks (Geschichte der Ind. Lit. ii, p. 338 fn 1; iii, p. 22 fn 1, also iii, p. 642) that Vägbhata I is the same as the poet of the -Nemi-nirvāņa .- Jahlaņa ascribes the verse anālocya premnah to one Vagbhata, but it does not occur in any of these Vagbhatas. It occurs, however, in Amaru 80. It is cited anonymously by Vallabhadeva 1170; while in the Sadukti-karnāmrta it is attributed to Rājaśekhara, and in Kavīndra-vacana 372 to the poetess Vikațanitambā. 4 And not Jayasimha of Kashmir, as Harichand (p. 49) erroneous- ly gives it. 5 Second half of the 13th century, see Buhler's Hemacandra note 1 ; also Vāgbhațālam° (ed. Kāvyamālā 1916) p. 1-2 fn.

THE VĀGBHAȚAS 193

from which we also learn that Vägbhata was living in 1123 A. D. and also in 1157 A.D. Vāgbhața's literary activity, therefore, may be assigned roughly to the first half of the 12th century. The Vāgbhatālamkāra, consisting of five Paricchedas, covers in 260 verses most of the topics of Poetics, but excludes Dramaturgy. Although it claims a large number of commen- taries, it is a small compilation of no superior merit. It speaks of ten Gunas instead of three of Mammata and Hemacandra, and only two Rītis, namely Vaidarbha and Gaudīya. Weber's Berlin MS no. 1718 adds a sixth chapter ; see also Burnell, Cat. Tanjore MSS, p. 576. Vagbhata II appears to be a later writer. His reference to Vägbhata I and considerable borrowing from Hemacandra give us one limit to his date. The other terminus1 is unknown; for the Jaina authors (excepting Hemacandra) are rarely quoted by later writers on the subject. He may have been earlier than Deveśvara, whose borrowings, however, are not conclusive enough for any chronological inference. Vägbhata II himself cites two of his own works, viz. Rșabhadeva-carita (p. 15, called a mahākāvya) and Chandonuśāsana (p. 20) ; but of these nothing is known. In two illustrative verses there are references to two princes called Mūlarāja (p. 45) and Vibhākara (p. 44). This Vibhākara is unknown, but Mūlarāja appears to be the same as the founder of the Calukya dynasty at Anahilla-pattana (=Anhilvād) in Gujarat2. A MS of Kāvyānuśāsana (Eggeling, Ind. Office Cat. no. 1157) is dated in Samvat 1515 (=1458-59 A.D.). Vagbhata II probably flourished in the 14th century. Like Hemacandra's work of the same name, the Kāvyānu-

I Harichand Sastri (op. cit. p. 49) places him in the 13th century, but he does not state the grounds of his opinion. 2 Peterson notes (iii, App. p. 124) a reference in the puspikā of a MS of Hemacandra's Trişasti-śalāka-purușa to one Nemikumāra, who flourished in Samvat 1295, and he queries whether this Nemikumara was our Vägbhata's father (iv, p. lxxi). 13

194 SANSKRIT POETICS

śāsana of Vāgbhata II is written in the form of Sūtra and a running commentary; but it is a much smaller work of five Adhyāyas. It covers most topics of Poetics but there is no treatment of dramaturgy. It speaks, however, of three Gunas and three Rītis after Mammata. The name and definition of poetic figures in these Jaina writers differ in some cases from those of orthodox authors. They do not exceed 40 in number, but Vāgbhata II gives nearly 70 poetic figures.

(3 ) No commentaries on Hemacandra and Vāgbhata II is known, but the Vāgbhatālamkāra1 of Vāgbhața I appears to have been fortunate in this respect. Of the commentators on this work, whose names are noted below, Jinavardhana Sūri and Simhadeva Gani are better known, and their commen- taries have been published. Jinavardhana was pupil of Jinarāja Sūri and was a priest of Kharatara-gaccha from about 1405 to 1419 A.D.2. In some catalogues (e.g. Mitra 2814), his name is given as Ādinātha.

BIBLIOGRAPHY

Hemacandra Edition. (i) ed. Sivadatta and K. P. Parab, Nir. Sag. Press, Bombay 1901, the text with Alamkāra cūdāmaņi and Viveka. (ii) with Alamkāra-cūdāmaņi and Viveka, and an anonymous Tippana, ed. R. C. Parikh and R. B. Athavale, in 2 vols., Mahavira Jaina Vidyalaya, Bombay 1938.

Vāgbhața I Editions. Vāgbhațālamkāra (1) ed. by A. Borooah, Calcutta

1 The published text of this work contains five chapters, which is also the number in the Bodleian, Stein, Madras and India Office MSS ; but Weber's MS (no. 1718) adds a sixth chapter, which appears to deal with the figure yamaka. 2 Klatt in IA xi p. 249; Bhandarkar, Rep. 1882-3, p. 25; IOC iii, no. 1156 and 2656a.

THE VĀGBHAȚAS 195

1883. (2) ed. Sivadatta and K. P. Parab (with Simhadeva Gaņi's comm.), Nir. Sag. Press, Bombay 1895, 1915 (our references are to the ed. of 1915). (3) by Jivananda Vidyasagar, 3rd ed. Calcutta 1903, (4) by Murtidhara, Venkatesvara Press, Bombay. (5) with an old gloss, by Ksemaraja Srikrishnadasa, Bombay 1894. (6) Granthamālā iii, 1889-90 (with Jinavardhana's comm.). Commentaries. (1) by Jinavardhana Sūri, who was a priest of Kharatara-gaccha from about 1405 to 1419. Ed. with the text in Granthamālā iii, as noted above. Ādinātha in Mitra 2814 (Aufrecht i. 559a) is the same as Jinavardhana. A MS copied in Samvat 1610 =1553-54 A.D. (Cat. MSS BORI xii, p. 323). (2) By Simhadeva Gaņi, ed. NSP, Bombay, as noted. In ALeip MS no. 824, p. 269, the commentary is called Curnī. But this name is not found in Jammu MS no. 1231, p. 274. (3) By Samayasundara, pupil of Sakalacandra, who was a pupil of Jinacandra. His comm. was composed in Ahmedabad for Harirāma in 1636 A. D. See Peterson iv, p. cxxvi. Also wrote a comm. on Raghu. (4) By Rājahamsa Upādhyāya, pupil of Jinatilaka Sūri who was a pupil of Jinaprabha Suri of Kharatara-gaccha. The MS noticed by Bhandarkar (Rep. 1883-84, pp. 156, 279) was copied in Samvat 1486=1430 A.D. See P. K. Gode's note in Calcutta Orient. Journal ii, pp. 312-14, in which he gives 2nd half of 14th century (between 1350 and 1400 A. D.) as the probable date of this commentary. (5) Samāsānvaya Țippaņa by Kșemahamsa Gaņi. Extract in Stein p. 274. (6) Vivarana of Ganeśa, son of Anantabhatta and disciple of Bhāskara. Aufrecht i. 559a, 794a ; IOC iii, no. 1155/702b, p. 330. A MS copied in 1713 A.D. (7) Avacuri. Author's name unknown. Aufrecht ii. 132a, iii. 118b. (8) Jñāna-pramodikā of Vācanācārya Jñānapramoda-gaņi,

196 SANSKRIT POETICS

composed in Samvat 1681 (=1624-25). See P. K. Gode, Studies in Ind. Literary Hist. i, p. 76.

Vāgbhața II Edition. Kāvyānuśāsana, by Sivadatta and K. P. Parab. NSP, Bombay 1894, 1915 with Alamkāra-tilaka.

JAY ADEV A

( 1 ) Jayadeva, author of the popular text-book Candrāloka, is otherwise known as Pīyūşavarsa (i. 2)1. He himself gives us the names of his parents as Mahādeva and Sumitra (i. 16). The name Jayadeva, however, is borne by our author in common with many other Sanskrit writers. Of the fifteen or more different persons, mentioned by Aufrecht, as bearing the same name, it seems likely that our author is identical with the poet who wrote the well-known drama called the Prasanna-räghava; for in the prologue to that drama there are two verses (i. 14-15) which inform us that the dramatist was also son of Mahadeva of the Kaundinya-gotra and Sumitrā, a coincidence of names which does not seem to be accidental. Aufrecht, however, identifies2 our author with Jayadeva who composed the well-known lyric named Gita-govinda ; but apart from all arguments derived from the style and poetic genius of the two writers, which possess few kindred excellences, the fact that the author of the lyric, in one of his

1 Also in a verse given at the end in some MSS, e.g. Peterson ii, p. 109, Madras Cat. xxii, p. 8656 : pīyūşavarsa-prabhavam candrālokam manoharam etc. Also the verse jayanti yājñika-śrīman-mahādevānga- janmanaḥ/ sūkti-pīyūşa-varşasya jayadeva-kaver girah, commented on in the Saradāgama and the Rākāgama comms. These verses are wanting in the Calcutta ed. The Rākāgama comm. of Gāgābhatta expressly states: Jayadevasyaiva pīyūşavarșa iti nāmāntaram. 2 ZDMG xxvii, p. 30.

JAYADEVA 19'

concluding verses,1 tells us that he was son of Bhojadeva and Rāmādevī (or Vāmādevī or Rādhādevī, according to other readings) stands seriously against the proposed identi- fication. The identity of Jayadeva with the logician Pakșadhara, also called Jayadeva, is equally doubtful, and Aufrecht mentions the two names separately. The name Pakșadhara, no doubt, was a mere title given to the logician from the circumstance of his having been able to maintain by subtle reasoning whatever side of a question he undertook to defend ; but the argument for his identity with our Jaya- deva, relied on by Hall2, that Jayadeva in his drama refers (i. 18) to his knowledge of pramāņa, befitting a logician, is hardly convincing and sufficient3.

( 2 ) The date of Jayadeva yet remains unsettled. There is hardly any doubt, however, that he should be placed earlier than Keśava Miśra, who cites (p. 47) the verse kadalī kadalī from the Prasanna-rāghava (i. 37). As Keśava flourished in the middle of the 16th century, we may safely assign Jayadeva to a period earlier than that. This conclusion is supported by the fact that the Saradagama commentary on the Candrāloka, was composed by Pradyotana Bhatta in 1583 A.D.4, under the patronage of a Bundella prince, named Virabhadra, of the Vaghela dynasty, who himself wrote a commentary on Vāt-

1 xii, p. 171, ed. N. S. P. 1917. It is not commented upon by Kumbha in his Rasika-priyā comm., but Samkara, in his Rasa-mañjarī comm. says: adhunā pitr-mātr-nāma nibadhnan prāthayate sajjanān (ed. N. S. P. loc. cit). The same in the colophon in Bühler's MSS (Kashmir Rep. p. 46), where read Rāmādevī for Rāmadeva. 2 Introd. to Sāmkhya-pravacana-bhāsya (Bibl. Ind. Calcutta 1956), pp. 62-63. Keith (Indian Logic p. 33f) appears to accept the identi- fication. 3 Jayadeva. author of a manual on Erotics, called Rati-mañjarī in 60 verses (ed. in Haeberlin and by Pavolini in Giornale della Soc. Asiat. Italiana, 1904 pp. 371f) is probably a different and later writer. 4 AFI no. 467 (51) p. 158 ; ALeip no. 820, p. 268.

198 SANSKRIT POETICS

syāyana (called Kandarpa-cūdāmaņi) in Samvat 1633=1577 A,D.1 We may push this limit to the date of Jayadeva's work back to the beginning of the 14th century, because some verses from the Prasanna-rāghava (i. 19 and 33)2 are quoted in Sārangadhara-paddhati (164 and 3520), compiled in 1363 A.D .; while Singabhūpāla, whose date has been fixed at 1330 A.D., cites the drama itself in his Rasārņava-sudhākara (pp. 258, 277). This gives us one terminus to the date of Jayadeva in the first quarter of the 14th century3. The other terminus is given by the inference that Jayadeva is later than Ruyyaka; for in his Candrāloka he directly adopts some of the original definitions of poetic figures given for the first time by Ruyyaka. The figure Vikalpa, for in- stance, which (as both Ruyyaka himself and Jayaratha inform us)4 was invented and defined for the first time by Ruyyaka,

1 Peterson ii, pp, 66, 132; iv, p. cxvi. Ed. Rama Chandra Sastri, Lahore 1926. 2 Other varses quoted are ii. 22 (=3557), vii. 59 (=3626), vii. 60 (=3631). 3 Paranjpe and Panse in their edition (Poona 1894) of the drama Prasanna-raghava (p. xiii f) seek to identify Jayadeva with the logician Paksadhara Jayadeva and assign him to a period between 1500 and 1577 A.D. So also Peterson in introd. to Subhāș° p. 37f. Cf also Eggeling IOC iii, pp. 332f. Winternitz (Geschichte der Ind. Lit. iii, p. 26. fn 3) thinks that Jayadeva could not have written long before Appayya. But all these scholars appear to have overlooked this quotation in Sarangadhara. No chronological conclusion is inferable from Jaya- deva's mention of the poet Cora; for Bühler's identification of this poet with Bihlana is not free from doubt (see Solf, Die Kashmir Recension der Pañcāsikā, Kiel 1886, p. xxì f ; also see on the question S. K. De, Hist. of Skt. Lit., Calcutta 1947, pp. 368-69. Nor should stress be laid on the fact that verses from the Prasanna-rāghava occur in the Mahanataka; for the date of the latter, as well as its proper text, cannot be taken to have been satisfactorily settled (see Lévi ii, p. 48 ; Sten Konow, Ind. Drama pp. 88-9). Jayadeva himself, as a rhetorician, is quoted by very late writers like Appayya, Kesava and Bhīmasena. 4 Cf Jacobi in ZDMG Ixii, p. 600, note 1. Ruyyaka says expressly

JAYADEVA 199

is literally copied by Jayadeva (v. 112). We cannot, therefore, place Jayadeva, who upholds the views peculiar to Ruyyaka as well as Mammata, earlier than the second half of the 12th century. Jayadeva, therefore, should be assigned to the period bet- ween the last quarter of the 12th century and the first quarter of the 14th, a closer approximation than which is not possible at present ; but perhaps we may tentatively place him in the frst half of the 13th century.

( 3 ) The Candraloka is a general treatise on Poetics in ten chapters (called mayūkhas) and about 350 verses, written in the Anustubh metre. The Calcutta edition of the text, published in 18741, enumerates the following divisions: (1) Vāgvicāra (śl. 16). (2) Doșa-nirūpaņa (śl. 44ž). (3) Lakșaņa- nirūpaņa (śl. 11). (4) Guņa-nirūpaņa, given as ten in number (śl. 12). (5) Alamkāra-nirūpaņa, consisting of Šabdālaņkāras (śl. 10), Alamkārānukramaņikā (śl. 16) and Arthālaņkāras (śl. 174). (6) Rasādi-nirüpaņa (śl. 24), incidentally dealing with three Rītis and five Vrttis. (7) Dhvani-nirūpana (śl. 18). (8) Guņībhūta-vyangya (śl. 10). (9) Lakşaņā-nīrūpaņa (śt. 15). (10) Abhidhā-nirūpaņa (śl. 4). This arrangement is substantia- lly followed in the Leipzig MS 819 (which contains only five mayūkhas) and correspond closely with the arrangement men- tioned by Gangadhara in his commentary (p. 9) on Appayya's Kuvalayānanda, where the chapters are given thus:, 1. Sabda- mayūkha. 2. Doșa-mayūkha. 3. Laksaņa-mayūkha. 4. Guņa-

with regard to this figure : pūrvair akrta-viveko'tra darśita ity avaganta- vyam, upon which Jayaratha remarks: anenāsya granthakrd-upajña- tvam eva darśitam (p. 159). Also the figure Vicitra (Ruyyaka p. 133= Jayadeva v. 82). 1 The Calcutta edition (by Jivananda) of 1906 substantially keeps to this arrangement and numbering of verses in the different chapters. The work contains about 300 verses, but the numbering differs to some extent in the different editions. The author gives his own illustrations.

200 SANSKRIT POETICS

mayūka. 5. Alaņkāra-mayūkha. 6. Rasa-mayūkha. 7. Dhvani- mayūkha. 8. Guņībhūtavyangya-mayūkha. 9. Lakșaņā-mayūkha and 10. Tatśakti (=Abhidhā)-mayūkha1. It will be seen from this that the section on Arthālamkāra in chapter v is the most considerable part of the work, which appears to have become, to the exclusion of the rest of the work, a popular manual of poetic figures. It was specially adapted for this purpose by Appayya Dīksita's Kuvalayānanda, which bodily incorporates the Karikas of this section (with only slight modification), himself only writing the running prose commentary and adding a few supplementary figures. This work of Appayya's, therefore, may be regarded, in a sense, as a commentary on the Arthālamkāra-chapter of the Candrāloka. Appayya himself indicates his indebtedness in one of the prefatory verses2 by saying that the definition- stanzas of the Candraloka are borrowed in his own work, but there are a few modifications and-additions3 of his own. He also explains in the concluding verse how his work came to be called Kuvalayānanda (lit. 'delight of lotuses') from the Candrāloka (lit. 'the sight or light of the moon'): candrāloko vijayatām, śaradāgama-sambhavaḥ/ hrdyaḥ kuvalayānando yat-prasādād abhūd ayam, which. apart from the obvious pun involved, praises the Candraloka, the cause of its commentary called Saradāgama, from the contact of both of which the charming Kuvalayā- nanda originated. This Saradāgama commentary obviously

1 The text as commented upon by Pradyotana Bhatta, Gāgābhatța and Vaidyanatha (Madras Cat. xii, 12876-78) contains ten mayūkhas. MSS of the complete text noticed also in Mitra ii p. 177, v p. 103, ix p 184 : Peterson ii 109. 2 yeşām candrāloke drśyante laksya-lakşaņa-ślokāh/ prāyas ta eva, teşām itareşām tvabhinavā viracyante. 3 The differences of reading in the Kärikas are noted in Halasynatha Sastri's ed. of Kuvalayānanda (with the Rasika-rañjanī of Gangādhara), Kumbhakonum 1892.

JAYADEVA 201

refers to the commentary of the same name on the Candrāloka, composed by Pradyotana Bhatta in 1583 A. D.1. But on account of the wholesale appropriation of this chapter of Jayadeva's work, the title Candrāloka appears to have been frequently applied to the Arthalamkāra-section of the work2 alone, as well as to Appayya's Kuvalayānanda3 itself. Thus, the India Office MS 2656, Weber 1721 and Madras MSS 12871-74 constitute in reality the Arthālamkāra- section of the Candrāloka, embodied in the Kuvalayānanda, and not the whole text, but they are entitled Candrāloka. Appayya's work does not end with the hundred or 108 poetic figures4 dealt with by Jayadeva, but it adds a supplementary chapter on a few additional figures. In some texts of the

1 Vaidyanātha, apparently ignorant of the existence of the Saradā- gama commentary, interprets (ed. N. S. P. 1917 p. 188) the phrase śaradāgama-sambhavah as referring to some previous original of the Candräloka itself. An instance of similar ignorance on the part of the commentator is given by the story of Asädhara in his comm. on the Kuvalay° (p. 86) that Appayya composed the Candraloka itself at the request of the king of Venkatagiri. and later on wrote his Kuvalay° on its basis. Gangādhara, a more reliable commentator on Appayya's work (who tells us that Appayya was the Guru of a brother of his grand- father) interprets the phrase correctly as : atra candrāloka-nāmā granthaḥ śaradāgama-nāmnā țīkā-granthena sambhava utpattiḥ (p. 283). The supposition (SgS ii, pp. 68-9) that Appayya's utilisation of Jayadeva's work was resented by the latter, who is said to have made a veiled reference to this fact in the prologue to the Prasanna-raghava (where the stage-manager alludes to the stealing of his name) is disproved by the fact that Appayya lived long after Jayadeva. 2 Cf. Gangādhara on Kuvalay° p. 9: candrāloko'rthālamkārātmaka eva, na tvanya iti keşāmcid bhramaḥ. 3 Thus, Regnaud (Rhétorique Sanskrite p.375) speaks of the Candra- loka as being composed of 151 ślokas, dealing with the definition and illustration of poetic figures, which description applies to the Kuvalayº. 4 This is not the largest number of poetic figures enumerated and defined in works on Alamkāra. Mammata defines 61, Ruyyaka 75 Arthālaņkāras ; but Sobhākaramitra gives 109, Appayya Dīkșita 115 Alamkāras, which go on multiplying !

202 SANSKRIT POETICS

Candraloka this appears to have been erroneously included. A considerable confusion is also noticeable in the different MSS of Jayadeva's and Appayya's works as to the arrange- ment of the three opening verses, as well as with regard to the total number of Ślokas contained in the Arthālamkāra-section. The verse paraspara-tapah-sampat° occurs in most accepted texts of this section of the Candrāloka, but it is not intelligible why Jayadeva should add this benedictory verse in a chapter, which occurs in the middle of the book. Gangadhara pointedly remarks that this verse is not Jayadeva's but was composed by Appayya himself as prefatory to his own work1.

( 4 )

THE COMMENT ATORS ON JAY ADEV A Of the commentators on the Candraloka, mention has already been made of Pradyotana Bhatta (alias Padmanābha Miśra) and his commentary, called Candrāloka prakāśa Saradāgama. He is described as son of Miśra Bala- bhadra, and his patron's name is given as Vīrabhadra (or °rudra)-deva, son of Ramacandra and grandson of Vīrabhānu, king of Ayodhya, of the Vaghela (Vandella)2 family. His commentary is dated in 1583 A. D .; while his patron lived in the second half of the 16th century, as we find Virabhadra's commentary (called Kandarpa-cūdāmaņi) on Vātsyāyana is dated in 1577 A.D. Virabhadra is said to have murdered Abul Fazl at the instigation of Prince Selim His Court-pandit

1 Gangadhara op. cit. p. 9: "tatha paraspara-tapahsampat" iti candrāloka-nāndī-śloka ity api bhrama eva; pañcama-mayūkhe śadhā- lamkārān nirūpya "upamā yatra sādrsya" ityādinā arthālamkāra-prastāve nāndyā evābhavat. The same remark applies apparently to the second verse alamkārşu bālānām and to v. 174 which alludes to “Venkața- prabhu", for they appear to be Appayya's additions. Cf IOC iii, pp, 333-34 for a discussion of this point. 2 The Madras MS reads vandella, but the Florentine MS (AFI p. 158) has vāghela.

COMMENTATORS ON JAYADEVA 203

Mitra Miśra wrote the Viramitrodaya, in which he mentions his patron's name. There is another commentary called Ramā1 written by Vaidyanātha Pāyagunda, who is probably not identical with Vaidyanatha Tatsat, the commentator on Govinda's Kāvya- pradīpa and Appayya's Kuvalayānanda, although the two writers are taken as identical in most catalogues. The colophon to their commentaries distinctly make out their respective family-names as Pāyagunda and Tatsat; while in one of the introductory verses of the Ramā our Vaidyanātha distinctly calls himself Payagunda which is a well-known Mahārāstra surname; but he does not give his own genealogy. He appears to have written a commentary called Gada on Nāgojī's Paribhāşendu-śekhara; he must, therefore, be later than the beginning of the 18th century. There is another less known commentary, called Rākāgama or °Sudhā, composed by Gāgābhatta, alias Viśveśvara, son of Dinakara (or Divākara) Bhatta, who was a Mīmāmsaka. Viśveśvara, who also wrote a number of Mīmāmsā and Smrti works (Aufrecht i. 587b), was a great-great-grandson of Rāmeśvara, nephew of the well-known Mīmāmsaka Kamalā- kara Bhatta, whose date is the first quarter of the 17th century1. Viśveśvara, therefore, is a comparatively modern writer who probably flourished in the beginning of the

1 The name of his commentary is often given, through a confusion, as Harilocana-candrikā (Aufrecht i. 182a), which itself appears as a mistaken name for the Alamkāra-candrikā comm. of Vaidyanātha Tatsat on Kuvalayānanda ; the mistake arising from the word harilo- cana-candrikā occurring in the benedictory verse to the latter com- mentary, as well as from this confusion between the commentators on Jayadeva and Appayya respectively. The benedictory verse runs thus : anucintya mahālakşmīm hari-locana-candrikām / kurve kuvalayānanda- sad-alamkāracāndrikām. See under Appayya Dīksita for the commen- tary. To Vaidyanātha Pāyagunde, however, is ascribed a Laghu Kuva- łayānanda (BORI MS Cat. xii, no. 287, pp, 342-43). 2 see above p. 167. The genealogy is given thus: Rāmeśvara ->Nārāyaņa->Rāmakrsņa->Dinakara-Viśveśvara.

204 SANSKRIT POETICS

18th century, and should not be confused with Viśveśvara, author of the Alamkara-kaustubha (q. v.). Two other little known commentaries are mentioned below.

BIBLIOGRAPHY Editions. Printed very often. (1) By Vyāpāradarpaņa Press, Madras 1857 in Telugu characters. (2) by Jivananda Vidya- sagar, Calcutta 1874, 1896, 1906, 1917. (3) by Subrah- manya, Vizagapatam 1898. (4) by Venkatacharya Sastri, in Grantha characters, Palghat 1912. (5) by V. L. Pansikar, NSP, Bombay nd ed. 1907 (contains also Kuvalayānanda with the Candrika comm. of Vaidyanatha), 1912, 1917, (6) With the Candrāloka-prakāśa Saradāgama of Praya- tana Bhatta alias Padmanābha Miśra, ed. Narayan Sastri Khiste, Chowkhamba Skt. Series, Benares 1929. (7) With Ramā of Vaidyanātha Pāyaguņde, ed. Govinda Sastri, Benares 1883 ; also ed. Mahadeva Gangadhara Bakre, Gajarati Printing Press, Bombay 1923. (8) With Kuva- layānanda and Vaidyanātha's comm. ed. Govinda Sāstri, Venkatesvara Press, Bombay 1911. (9) With Rākāgama comm. of Gāgābhatta, ed. Chowkhamba Skt. Ser. Benares 1938. (10) By Suryabalirama Chaube, Benares 1895 (with Candraloka-nigudhartha-dīpika). The ed. with Budha-rañjani comm. published from Madras (1863) is really a commentary on the Arthālamkāra-section incor- porated in the Kuvalay° and not upon the whole text. Our references are to the Calcutta ed. 1917, which contains the whole text. On commentaries on Kuvalayānanda see under Appayya Dīksita below. MSS. Madras Cat. xii, 12860 (which contains the verses of the Candraloka with Kuvalay°), 12871-73. Most of the MS mentioned in other catalogues (see Aufrecht), however, contain the Arthalamkära-section and not the whole text, see above p. 201-2. The Alamkāra-śataka of Jayadeva in Oppert ii, 2763 is a descriptive name perhaps of this section

VIDYĀDHARA 205

of the Candrāloka. The Alamkāra-samgraha in Mitra 1612 is in reality this Arthālamkāra-section. Commentaries. (1) Candrāloka-prakāśa Śaradāgama by Pra- dyotana Bhatta. Ed. as above. See Madras Cat. xxii, 12878 for a description. (2) Rākāgama or Sudhā by Viśveśvara alias Gāgābhatța. Ed. as above. Gāgābhatta is known to have officiated at the coronation of Sivaji in 1674 A.D. His Samaya-naya was composed for king Sambhāji in 1680-81 A.D. (P. K. Gode in Proc. Ind. Hist. Congress, 1939, pp. 1166-71). He belonged to the famous Maratha Bhatta family of Benares. His father Dinakara was author of Dinakaroddyota. (3) Ramā by Vaidyanātha Pāyaguņda. Ed. as above. Madras Cat. xii 12876. (4) Comm. by Vājacandra. Aufrecht i. 182a. (5) °Dīpikā. Name of author unknown. Aufrecht i. 182a. (6) Sārada- śarvarī by Virūpāksa. Hultzsch 1617; Tanjore Cat. ix 5221. VIDY ADHARA ( 1 ) The date of Vidyadhara, author of the Ekāvalī1 has been fixed with sufficient approximation by K. P. Trivedi and R. G. Bhandarkar2. The latest writer quoted and mentioned by Vidyādhara is Ruyyaka (p. 150); and this gives us one 1 Aufrecht (i. 75) mentions three different works called Ekāvalī, which appear to be the same work. The first and the third are undoubt- edly identical and refer to our Ekāvali; but the second is described by Burnell 54a (cf Oppert ii. 3605) as composed by Mahāmāheśvara ·Kavi. This, however, appears to be a title of Vidyädhara himself, and is apparently the source of the confusion of our Vidyadhara with `Abhinavagupta who also bore the same title (see Weber ii, no. 1723). The colophon in the Madras MS (Madras Cat. xii, p. 8611) reads: iti śrimato mahāmāheśvarasya kaver vidyādharasya kṛtāvekāvalī-nāmny alamkāra- sastre etc. The first verse quoted in Burnell is the same as found in all the texts of our Ekavali The commentary Tarala noticed by Weber (loc. cit.) is apparently the same as Tarala of Mallinatha. The Keli- rahasya on Erotics is ascribed to Vidyadhara by Aufrecht, but the colophon gives the author's name as Vaidya Vidyādhara. 2 Introd to the text in B. S. S. ed. and Bhandarkar Rep, 1887-91, p. lxvi f.

206 SANSKRIT POETICS

terminus to his date at the middle of the 12th century. This conclusion is apparently supported by Vidyādhara's mention (p. 19) of Śrīharșa, author of the Naisadha, who lived very probably in the 12th century1 ; but Vidyādhara's allusion in the same context to the poet Harihara2, who is said by him to have obtained amazing wealth from a prince Arjuna (presumably the ruler of Mlava of that name), puts this terminus a little lower at the first quarter of the 13th century. The Ekāvalī, in its turn, is quoted by Śingabhūpāla3, whose date is fixed at 1330 A.D .; while Mallinätha, at the end of the 14th century, commented upon it. The internal evidence of the text, therefore, assigns it to a period between the first quarter of the 13th and the first quarter of the 14th century. This approximation has been considerably narrowed down to the end of the 13th and the beginning of the 14th century by the identification of king Narasimha of Kalinga, panegyrised in the illustrative verses of the work4, with either of the two Narasimhas of Kalinga, whose dates fall between 1282 and 1327. The patron of our author is described as. one who crushed the pride of Hammira (pp. 176, 177, 257, 260), who is probably the famous Cauhan prince, the hero of

1 See Bühler in JBRAS x p. 31f, xi. p. 279f; K. T. Telang in IA ii, p. 71, iii 81f ; Bühler, Rep. 1874-75, p. 8. 2 See Trivedi's note at p. 348. 3 Rasārņava-sudhākara p. 107=Ekāv. i. 2. Cf Śgš i, p. 7f. This verse occurs, however, as the third praśasti-śloka in the printed text of Bihlaņa's Karņa-sundarī (ed, Kāvyamālā 7, 1895, p. 56) .- Singabhūpāla refers to Vidyadhara and his Ekavalī expressly in the following terms: utkalādhipateḥ śrngāra-rasābhimānino narasimha-devasya cittam anuvartamānena vidyādhareņa kavinā bādham abhyantarīkrto'si, evam khalu samarthitam ekāvalyām anena (ed. Triv. Skt. Ser. p. 206). K. P. Trivedi (Introd. p. xxiii) comes to the conclusion that Vidyadhara was patronised by Keśarī-Narasimha (1282-1307 A. D.) or by Pratāpa-Nara- siņha (1307-1327). 4 As the author himself says (Śl. 7): karomi narasimhasya cāțu- ślokan udāharan. In this respect the work resembles Pratāpa-rudra- yaśobhūşaņa of Vidyānātha, Raghunātha-bhūpālīya of Krsņa Yajvan and Alamkāra-mañjūşā of Devaśamkara.

VIDYĀDHARA 207

Nayacandra Sūri's poem1, who began his reign about 1283 A.D. and attempted a conquest of Sourthern countries. All this makes it probable that the Ekāvali was composed towards the end of the 13th and the beginning of the 14th century. Vidyādhara appears to have written a work on Erotics entitled Kali-rahasya2.

( 2 )

Mallinātha The date of Mallinatha, author of the Tarala commentary on the Ekavali, has been fixed at the end of the 14th cen- tury by Bhandarkar and Trivedi3. He must have written his commentary after a certain time had elapsed from the com- position of the original text ; for from śl. 6 it appears that the Ekāvalī was not studied for some time because it had no commentaries. He is identical with Kolacala Mallinātha Sūri (Pedda Bhatta) who is the well-known scholiast and commentator on the five standard Mahākāvyas of Kālidāsa, Bhāravi, Bhatti, Srīharsa and Māgha, in some of which he quotes from the Ekāvalī itself. The Ekāvalī, consisting of Kārikā and Vrtti in eight Un- mesas, utilises the works of Mammata and Ruyyaka in its treament of poetic figures in the last two chapters (vii-viii). After a general discussion of the definition of Kävya in ch. i it deals in ch. ii with the three Vrttis, namely, Abhidhā, Laksaņā and Vyañjanā. Ch, iii and iv are devoted to Dhvani, and ch. v-vi deal with three Gunas, three Rītis, and the Dosas. The illustrative verses are all composed by Vidyādhara himself and consist of panegyrics of the author's patron Nara- siņha of Utkala.

1 See ed. Kirtane v. 56. also p. 27 ; Bhandarkar op. cit. p. lxvii f. 2 Aufrecht i. 537 b. 3 Bhandarkar, Rep. 1887-91, p. lxix ; Trivedi, introd. to Bhatți pp. xxiv-xxviii, introd. to Ekāvalī p. xxvii f; Pāțhak, introd. to Megha-dūta pp. 11-12 ; Nandargikar, introd. to Raghu, pp. 1-6, esp. pp. 5-6.

208 SANSKRIT POETICS

BIBLIOGRAPHY Edition. ed. K. P. Trivedi in the Bombay Sansk. Series 63, 1903, with the Tarala of Mallinatha and introd, and notes. There is another comm. by Prabhakara (b. 1564 A,D.), son of Madhavabhatta and grandson of Rāmeśvara Bhatta.

VIDYĀNĀTHA

( 1 ) The latest writer that Vidyanatha cites is Ruyyaka (pp, 291, 334), whose lost work Sāhitya-mīmāmsā is also apparently referred to at p. 11. Vidyānātha, in his turn, is quoted ex- tensively but anonymously for definitions of poetic figures by Mallinatha in the latter's many commentaries on the different Kāvyas1. This gives us the same broad limits to his date as to that of Vidyädhara ; and other considerations make it probable that he was contemporaneous with the latter. The Pratāpa- rudra-yaśo-bhūşaņa of Vidyānātha was written, like the Ekāvalī, with the obvious object of panegyrising the king whose name it bears on its title. All the illustrative verses in the work eulogise the same king (also called Virarudra or Rudra), who is described as the son of Mahadeva and Mun- mudi or Mummadambā (pp. 12, 13, 16, 17, 133) ; and a short drama, named Pratāparudra-kalyāna2 after him, is introduced in the third chapter to illustrate the characteristics of a drama, discussed in the work itself. He is described as a Kākatīya king3 whose capital was Ekasilā-nagara in the Trlinga or Andhra country, and who is said to have vanquished, among other kings, the princes of the Yadava family. All these and

1 For the quotations, see Trivedi's introd. to the text p. ix. 2 Separately entered by Aufrecht i. 349a and published in the Granthamālā vol. i. 3 so called, as the Ratnāpana explains (p. 10, also Ratnaśāņa p. 485) from the goddess Käkati he worshipped.

VIDYĀNĀTHA 209

other details have led K. P. Trivedi to identify Vidyanatha's patron with Pratāparudra, the seventh Kākatīya king of Ekaśilā or Warangal, whose inscriptions date between 1298 and 1317 A.D.1, and who is placed by Sewell between 1295 and 1323 A.D., and by Sesagiri Sastri between 1268 and 1319 A.D.2 The Yadava king referred to, therefore, seems to be Ramacandra, sixth ruler of the Yadavas of Devagiri, whose dates are 1271 to 1309 A.D.3 We may, therefore, assign Vidyā- nātha approximately to the end of the 13th and beginning of the 14th century. It has been suggested that the author's real name was Agastya Paņdita, and Vidyānātha was his title. Vidyānātha's work, like the Ekāvalī, consists of Kārikā and Vrtti with illustrative verses in praise of the author's patron. In nine Prakaranas it deals respectively with the topics of Nāyaka, Kāvya, Nāțaka, Rasa, Doşa, Guņa, Śadālaņkāra, Arthlāmkāra and Miśrālamkāra. In the third Prakaraņa, as we have already noted, it illustrates the requirements of a Nāțāka by a model drama. Its treatment is based mainly on Mammața, Ruyyaka, Bharata and Dhanañjaya, but it is more comprehensive than the Ekāvalī inasmuch as it includes Dramaturgy.

( 2 ) Kumārasvāmin Vidyānātha's commentator Kumārasvāmin describes him- self as the son of Kolacala Mallinatha4, the well-known

1 Eggeling (1OC iii, p. 338) gives the dates 1268 and 1319. 2 See Trivedi, introd. pp. xvi-xxii. The correct dates appear to be 1298 and 1323 A. D. 3 Bhandarkar, Early Hist. p. 92. 4 Nārāyaņa, who describes himself as a descendant of Kumāra- svamin, gives the genealogy of his ancestors in his comm. on Campū- īāmāyana (Madras Catalogue xxi, Kāvya p. 8212) thus: Mallinātha- Kapardin-Mallinātha Peddubhatta-Kumārasvāmin. He speaks of Peddubhatta as a Mahāmahopādhyāya, a commentator on Naișadha and as having been bathed in gold by Sarvajña (Singabhūpāla?). 14

210 SANSKRIT POETICS

commentator and author of Tarala on the Ekāvalī. He may, therefore, be placed in the beginning of the 15th century. The title of his commentary Ratnāpana (wrongly called Ratnārpaņa by Eggeling op. cit. p. 338b, following Burnell 36b) signifies, as he himself explains, a market-place where are sold jewels of poetic sentiments, collected together by Vidyānātha, after they have been fashioned on the grindstone furnished by the merits of the hero. The quotations in the Ratnāpana are numerous and in- clude, besides other well-known names, the Śrngāra-prakaśa of Bhoja, the Ekāvalī, the Sāhitya darpaņa (p. 245), Cakra- vartin and his Samjivanī commentary on Ruyyaka, Sińga- bhūpāla and his Rasārnava-sudhakara, the author's own father Mallinātha and brother Peddayārya, Bhatta Gopāla and Narahari Sūri. There are numerous references to a work on Rasa, called Bhāva-prakāśa, which is now known to be a work of Sāradātanaya (q.v.). Mention is also made of Vasantarājīya Nāțya-śāstra, its author Vasantarāja being apparently king Kumāragiri (q.v.) of the same name, who was. a patron of Kātayavema. A Kavikalpadruma-kāra is also cited at p. 170, but this is a work on grammar (dhātu-pātha). by Vopadeva. We know nothing of the Nātaka-prakāśa cited. at p. 113. On Alamkāra-sudhānidhi cited on p. 44, see below under Appayya Dīksita who also quotes the same work. The Rasa-nirūpaņa may be by Narahari Sūri, and the Sāhitya- cintamani is probably the work of the same name composed ly Vīranārāyaņa (q.v.).1 There is another incomplete commentary, called Ratna- sana, included in the Bombay edition of the text. From the colophon of a MS of this work (Madras Trm, II, C, 1923), it appears to have been composed by Tirumalācārya, son of Rāmānujācārya of Šukavața family and disciple of Vātsya Rāmānujācārya. He is said to have lived in Rāmatīrtha near Koțipallī in the Godāvarī district.

1 For these authors, see chapter on Minor Writers below.

VIDYĀNĀTHA 211

BIBLIOGRAPHY Editions. (1) by K P. Trivedi in Bombay Sanks. Series 65, 1909, with Kumārasvāmin's Ratnāpaņa, and Ratnaśāņa, notes and introd. Our references are to this edition. (2) lithographed ed. in Puthi form, Poona 1849. (3) by Sarasvati- Tiruvenkada Acharya and Vangipuram Ramakrisnam Acharya, with Ratnapana, in Telugu characters, Madras 1868, 1869, 1871, 1888. (4) by S. Chandrasekhara Sastrigal, with comm. Ratnāpaņa. Balamanorama Press, Madras 1914.

CHAPTER VIII

FROM VIŚVANĀTHA TO JAGANNĀTHA

VIŚV ANĀTHA

(1) Viśvanātha never cites Ruyyaka and Mammața by name ; but, like Vidyādhara and Vidyānātha, he draws very con- siderably upon the works of botb. He adopts, for instance, the definitions of the figures upameyopamā and bhrāntimat directly from Ruyyaka, and admits the two figures vikalpa and vicitra which, both Ruyyaka and Jayaratha inform us, were inventions of Ruyyaka's.1 It is quite possible, as P.V. Kane holds, that the censorious glancing on Mammata's text, reproved by Viśvanātha (ad ii. 14, p. 57), refers in particular to Ruyyaka's Samketa commentary where the latter criticises Mammata on the particular point under discussion. But a surer indication of Viśvanātha's acquaintance with Ruyyaka's works is given by his quotation of the verse bhujanga-kundalī- vyaktaº at p. 445 (ad x. 2), which Ruyyaka himself cites (p. 19) as his own from the Śrikantha-stava. Viśvanātha quotes two other writers who, in all probability, belong to this century, namely, Jayadeva, author of the Gita-govinda,2 and

1 For other instances, where Viśvanātha is following or criticising Ruyyaka, see P. V. Kane's ed. of the text in the introd. and notes. 2 The verse hrdi vişa-latā quoted by Viśvanātha at p 506 (ad x. 39) occurs in the Gita-govinda, ed. N. S. P. iii. 11, p. 58. It is also ascribed to Jayadeva by Sārnagadhara (no. 3460) and Vallabhadeva (no. 1314). Jayadeva is quoted in the Sadukti-karņāmrta of Śrīdhara and therefore must be placed before 1206 A.D. Bühler and Peterson assign (Kashmir Rep. p. 64 and Subhas° p. 38) 1116 A.D. as the date of Jayadeva, while Haraprasad Sastri gives the date 1175 A. D. (Notices, 2nd. Ser. i, p. xxxviii). Jayadeva, however, is said to have been cited by Cāndkavi, who wrote his epic on Prthviraja of Delhi towards the end of the 12th century (but see WZKM vii, p. 189; JBRAS xi, p. 283). Viśvanātha

VIŠVANĀTHA 213

Srīharșa, author of the Naișadha.1 Viśvanatha also quotes a verse kadalī kadalī (ad iv. 3) from Prasanna-rāghava (i. 37) of Jayadeva. Again, the Rāja-tarangiņī iv. 441 is quoted in our text at p. 529, under x. 57a (possibly indirectly through Ruyyaka p. 93); but this work of Kahlana's was not com- pleted till the middle of the 12th century. All this will roughly fix one terminus to the date of Viśvanātha, who cannot thus be placed earlier than the end of the 12th or beginning of the 13th century. The other more or less terminal date is given by the date of a MS of the Sahitya-darpana, discovered by Stein at Jammu, which was written in Samvat 1440=1384 A.D.2 This certainly negatives the date (viz. the middle of the 15th century) assigned by Weber3, Eggeling4 and Harichand Sastri5, the last of whom makes the unfortunate mistake of identify- ing Candīdāsa, referred to as a relative by Viśvanātha, with Candīdasa, the Bengali poet of the 15th century. It may be noted that Kumarasvamin, at the beginning of the 15th century, names and quotes (pp. 245, 248) the Sāhitya-darpaņa (iii. 146a, 147 and 150). All this raises the most likely presumption that Viśva- nätha should be assigned to a period ranging roughly from 1200 to 1350 A. D. This approximation can be considerably narrowed down if we can draw any chronological inference from a verse in the Sāhitya-darpana (ad iv. 14, p. 232) which refers to a Muhammadan king named Allāvadīna6. This

also refers to Lataka-melaka (p. 176, ad iii. 212) of Śankhadhara, which also belongs to this century. 1 P. 526, ad x. 54 (hanūmadādyai°)=Naișadha ix. 122b p. 520, ad. x. 50 (dhanyāsi vaidarbhi)=ibid iii. 116. For the date of Śrīharsa see S. K. De, Hist. of Sansk. Lit. pp. 325-26. 2 Jammu Cat. p. 64, no. 349. 3 Hist. of Sansk. Lit. p. 231 (Eng. trans. 1904). 4 IOC iii, p. 337. 5 op. cit. p. 115. 6 samdhau sarvasva-haraņam vigrahe prāna-nigrahaḥ / al(I)avadīna- nrpatau na samdhir na ca vigrahah.

214 SANSKRIT POETICS

Allāvadīna or Alāvadīna1 may probably be Sultan2 Ala-ud- din Khalji, whose army invaded the Deccan and seized Warangal. Even if we suppose that the verse in question was composed in the life-time of that Sultan, who died in 1316 A. D., the Sāhitya-darpana may be presumed to have been composed at a date not earlier than 1300 A.D. At any rate, if this historical deduction is permissible, we may assign Viśvanātha to a period between 1300 and 1350 A.D., or roughly in the first half of the 14th century3. ( 2 ) Viśvanātha describes himself as the son of Mahākavi Candraśekhara (p. 583, concluding verses) who appears, like his son, to have been a poet and scholar4, as well as a high official5 in the court of some king, probably king of Kalinga. Nārāyana, who appears to have written also on some topics of Poetics, is either his grandfather or great-great- grandfather ; for in his commentary on the Kāvya-prakāśa, Viśvanātha speaks of Nārāyaņa as asmat-pitāmaha, while in his Sāhitya-darpaņa (p. 73, ad iii. 4a), the same person is called asmat-vrddhapitāmaha. Caņdīdāsa, who appears to be different from the Bengal author of the Dīpika commentary on Mammata, is also quoted.6 He should not be confused with Visvanatha's relative.

1 We find both these forms of the name in two inscriptions, see JASB xliii, p. 108 and Bhavanagar inscription 114=Pracina-lekha-mālā ii. 28. In Harsakīrti's Dātu-pātha this king is referred to as Allāvadi (Bhandarkar Rep. 1882-83, p. 43). 2 The sanskritised form of this word suratrāna occurs at p. 509 (ad x. 42). 3 Cf Kane op. cit. introd .; M. Cakravarti in JASB lxxii (1903), p. 146, N. S. ii, 1906, p. 157f ; Keith in JRAS, 1911, pp. 848f ; Sten Konow, Ind. Drama, p. 3. Prabhākara in his Rasa-pradīpa (1583 A.D.) quotes Sāhitya-darpaņa at pp. 18, 20, 35. 4 His verses are cited at pp. 58, 116, 170, 174, while his works, called Pușpamālā and Bhāșārņava, are referred to at pp. 263 and 316 respectively. 5 Both are described as sāmdhivigrahika-mahāpātra. 6 Viśvanātha cites one Purușottama (p. 440, ad ix. 4a). A work

VIŚVANĀTHA 215

Viśvanātha appears to have written a number of works, besides his well-known Sāhitya-darpana ; for in it he himself refers to his own productions, namely : (1) Rāghava-vilāsa-kāvya (ad vi. 325a, p. 355). (2) Kuvalayāśva-carita in Prakrit (ad vi. 326, p. 356). (3) Prabhāvatī-pariņaya (ad vi. 182b, p. 320), also referred to in his commentary of Mammata ch, vii. (4) Praśasti-ratnāvalī in 16 languages, a karambhaka (ad vi. 337b, p. 358). (5) Candrakalā (ad vi. 183a and 184. p. 320-1), a nāțikā. He also wrote a commentary called Kāvyaprakāśa-darpaņa on Mammata's work ; but this was probably composed after he had written his larger independent work on Poetics ; for in it he himself refers, while commenting on laksanā (ch. ii), to the latter work1. In the Shitya-darpana itself he draws very considerably upon Mammata; and although at the beginning of this work, he quotes and criticises at some length Mammata's definition of poetry, he distinctly reproves all irreverent criticism of this venerable writer, who is declared to be his own upajīvya (ad ii. 14 p. 57). In this commentary Viśvanātha refers to a Narasimha-kāvya by himself.2 It is not clear on what grounds Weber and Eggeling3 state that the Sahitya-darpana was composed "on the banks of the Brahmaputra", i. e. in Eastern Bengal. It appears on the contrary that Viśvanatha was probably a native of Kalinga, which we may take at this date to have been co-extensive roughly with Orissa and Ganjam. In his commentary on called Kavitāvatāra is attributed to one Purusottama in Burnell 54a .- On Viśvanātha's genealogy in relation to Nārāyaņa, Caņdīdāsa and Candrasekhara see Sivaprasad Bhattacharya Viśvanātha Kavirāja and his references in JOI, Baroda, iii (1954) pp. 35f. 1 eşām ca șodasānām lakşaņā-bhedānām iha darśitāny udāharaņāni mama sāhitya-darpaņe'vagantavyāni. Also on figure anumāna (ch. x): tad uktam matkrte sāhitya-darpaņe. 2 Anantadāsa in his comm. on Sāhitya-d. quotes a verse on p. 9 with the words: yatha mama tāta-pādānām vijaya-narasimhe. 3 Cf also Macdonell, Sansk. Lit. p. 434; SCC vii, no. 53, p. 33.

216 SANSKRIT POETICS

Mammata, he explains certain expression with Oriya equival- ents1 ; and speaking of his ancestor Nārāyana, he refers to king Narasimha-deva of Kalinga (presumably Narasimha II. about 1279-1306), at whose court Nārāyana vanquished one Dharmadatta2, who is also referred to in the Sāhitya-darpana® at pp. 73, 79. It is probably in praise of one of the Nara- simhas of Kalinga that Viśvanātha's lost poem Narasimha- vijaya was written.

( 3 ) Though not a work of much originality, the Sahitya-darpana gives in ten chapters a comprehensive treatment of all topics of Poetics, including Dramaturgy. The distribution of topics in the different chapters is as follows: (i) Definition of poetry, (ii) Three Vrttis of word and sense, (iii) Rasa, (iv) Dhvani and Guņībhūta-vyangya, (v) Establishment of Vyañjanā-vrtti, (vi) Dramaturgy, (vii) Doșa, (viii) Guņa. (three in number), (ix) Rītis enumerated as four, Vaidarbhī, Gaudī, Pāñcālī and Lāțī, (x) Alamkāras. The treatment of Dramaturgy is based mostly on Daśa-rūpaka. The commentaries on Viśvanātha are not so numerous or important as to deserve any special enumeration. Of the five commentaries mentioned below, that of Rāmacarana Tarka- vāgiśa, dated in Śaka 1622=1700 A.D., has been frequently printed with the text.

1 "vaiparītyam rucim kuru" iti pāthaḥ, atra cinku-padam kāśmīrādi- bhāșāyām aślīlārtha-bodhakam, utkalādi-bhāşāyām dhṛta-vāņdaka- drava iti, on Mammața v, p. 238 (ed. Jhalakikara). 2 Cited also in the Rasa-pradipa of Prabhākara, son of Bhatta Mādhava (Weber i. 823), in which the Sahitya-darpana is also quoted. Prabhakara's work was composed in 1583 A. D. For Dharmadatta see Sivaprasad Bhattacharya in the article cited above, p. 360-62. 3 yad āhuh śrī-kalinga-bhūmaņdalākhandala-mahārājādhirāja-śrī- narasimha-sabhāyām dharmadattam sthagayantah sakala-sahrdayagostḥī- garistha-kavi-panditāsmat-pitāmaha-śrīman-nārāyanadāsa-pādāh, etc.

VIŠVANĀTHA 217

BIBLIOGRAPHY Editions. Published frequently, of which ed. (1) by Nathu- rama, Education Press, Calcutta 1828, and (2) by E. Röer in Bibl. Indica, Calcutta 1851, are notable. These editions do not contain the comm. of Rāmacarana. Also eds. with Rāmacarana's comm. Vivrti (i) by Chandicharan Smritibhushan, Calcutta B. S. 1318. (2) ed. Durgaprasad Dviveda, Nir. Sag. Press, Bombay 1902, 1915, 1922. (3) by P. V Kane (ch. i, ii, x) with intro. and notes, Bombay first ed. 1910, 2nd. 1923 (with a Hist. of Skt. Poetics), 3rd ed. 1951 (revised and enlarged, but the commentary is not given). (4) by Karunakar Kavyatirtha, with comms. Vijña-priyā of Maheśvara and Locana of Anantadāsa, Lahore 1938 .- Translated into English by J. R. Ballantyne and P. D. Mitra, Bibl. Indica 1875. Our references are to the N. S. P. ed. of 1915 by Durgaprasad Dviveda, unless otherwise specified. Commentaries. (1) °Locana by Anantadāsa, son of Viśva- nātha, a MS of which is dated 1636 A.D. Aufrecht ii. 171a. An incomplete MS (no. 262, p. 65) in Jammu Cat. Ed. as. noted above. The commentator is described as son of Viśvanātha himself. (2) Țippaņa by Mathurānātha Śukla, a voluminous writer, under whose name Aufrecht makes no less than 64 entries. Apparently the same person as Mathurānātha Sukla, a native of Pāțalīputra in Mālava, who wrote at Benares in 1783 A. D. the Jyotih-siddhānta-sāra by order of prince Dalacandra (but see Aufrecht i. 422-23). One Mathurānātha was also author of a comm. on Kuvalayā- nanda, and may have been the same person. Aufrecht i. 715b. (3) Vivrti by Rāmacaraņa Tarkavāgīśa, a native of West- ern Bengal. He was a Chattopadhyay Brahman ; his home was at Rāyavāti in Burdwan district. He dates his com- mentary in 1700 A. D. Frequently printed with the text in Bengal editions. Also in N. S. P. ed. 1915, as noted above.

218 SANSKRIT POETICS

(4) °Prabhā by Gopīnātha. Madras Trm C 712. Gopī- natha is also the author of the Sumanomanohara comm. on Mammata. See above p. 173. He is probably identical with Gopīnātha Kavirāja who composed, among other works, a commentary on the Raghu-vamsa in 1677 A.D. (see Aufrecht i. 163b). (5) Vijña-priyā by Maheśvara Bhatta, Ed, as noted above. This Maheśvara appears to be the same as Maheśvara Nyāyālamkāra who commented also on the Kāvya- prakāśa. Middle of the 17th century. See above p. 167.

KEŚAVA MIŚRA AND ŚAUDDHODANI

( 1 ) Keśava himself tells us that he composed his Alamkāra- sekhara at the request of a ruling chief named Mānikyacandra, son of Dharmacandra and grandson of Ramacandra, who is said to have ruled near Dilhi (Dhilli) and defeated the king of Kābila (Kabul?). Eggeling1 is obviously wrong in identifying him with Manikyacandra of Tirabhukti or Tirhut; while Bühler2 did not go further than suggesting that this prince was not a Kashmirian but ruled or lived in Delhi just before the Muhammadan conquest. The patron of our author, however, appears to be Mānikyacandra of Kot-kangra, whose genealogy corresponds to that given by Keśava and whose date of ac- cession, according to Cunningham3, is 1563 A.D. The literary activity of Keśava may, therefore, be fixed in the third quarter of the 16th century.

( 2 ) The Kārikā-portion of the Alamkāra-śekhara, called Sūtra, is declared to have been based on, if not actually taken from, some lost work of an authority who is cited as bhagavān (or

1 IOC no. 1197. 2 Kashmir Rep. p. 69. 3 Arch. Survey v. 152f, at p. 160. (cf. JASB, 1907, p. 212).

KEŚAVA MIŚRA AND ŚAUDDHODANI 219

maharsi p. 50) Śauddhodani1, Keśava himself apparently as- suming the modest rôle of a commentator or interpreter in the running prose Vrtti. The name Sauddhodani, apparently Buddhistic, is otherwise unknown in Alamkāra-literature2. Whatever may be the original source of his work, Keśava shows himself conversant with the work of most of his pre- decessors, and quotes, among more recent writers, Rājaśekhara (pp. 32, 67), Bhoja (p. 7), Mahimabhatta, Mammața, the Vāg- bhațālamkāra, Deveśvara and Jayadeva author of Candrāloka. He also quotes one Śrīpāda (pp. 4, 5, 6, 23, 27, 32, 72, 81), who may be his master Sauddhodani himself designated by this honorific term, as well as the author of a Kavi-kalpalatā who is described as a follower of this Srīpada3. This Kavi- kalpalatā-kāra, however, is neither Deveśvara nor Arisimha and Amaracandra, whose works also bear a similar title. The passage cited by Keśava in this connexion (pp. 48-9, venyah sarpāsi-bhrngālyo) gives a list of more or less conventional words useful for the purpose of conveying a simile or metaphor. A comparison of an almost similar passage in Deveśvara (p. 157f), who copies it directly from Arisimha and Amaracandra (pp. 135f), will show enough verbal dis- crepancy to indicate that neither of these sources constitutes the original from which Keśava quotes. A similar discrepancy is also noticeable in another passage of Keśava's (ratnāni ya- tra tatrādrau pp. 55-6), which at first sight will seem to have been borrowed from Deveśvara (p. 36f) who, however, copies it almost literally from Arisimha and Amaracandra

1 Mentioned in exalted terms as: alamkāra-vidyā-sūtrakāro bhaga- vāñ chauddhodaniḥ parama-kāruņikaḥ (p. 2). The Alamkāra-sūtra of Sauddhodani is mentioned at pp. 2, 20. 2 This Sauddhodani should not be identified with the Sauddhodani mentioned in the mangala verse (where it apparently stands for the name of Buddha) of the Vidagdha-mukha-mandana of Dharmadāsa Sūri. 3 śrīpāda-matānusārī kavikalpalatā-kāraḥ p. 48, ed. Nir. Sag. Press. Frequently quoted, pp. 4, 5, 23, 27, 32, 72, 83 etc.

220 SANSKRIT POETICS

(p. 30f). At the same time, Keśava betrays otherwise an ac- quaintance with Deveśvara's text, from which he reproduces at least one long passage anonymously (nrpe kīrti-pratāpājñā p. 57f=Deveśvara p. 26f), which Deveśvara himself pro- bably adapted from Arisimha and Amaracandra (p. 27f) ; but it is curious that Kesava copies here the text of Deveśvara with its variations, rather than the original text of Arisimha Amara on this point. Keśava cites one Śrīharsa (p. 71) who may or may not be the same person mentioned by Prabhākara Bhatta (q. v.) as Śrīharşa Miśra, or Harşa (Śrīharsa) who wrote a Vārttika on the Nātya-sāstra. The opinions of a writer called Govar- dhana are frequently cited by Keśava (pp. 17, 29, 37, 43, 49). There is also a reference to Jayadeva pandita-kavi (p. 17) in the court of an Utkala king. If this person is identical with the poet Jayadeva, who is said to have lived under Laksmana-sena of Bengal and who also calls himself Jayadeva pandita-kavi in his Gīta-govinda (xii, p. 171)1, then it is likely that Govardhana, who is quoted immediately before this reference to Jayadeva, may be the poet of that name, who was Jayadeva's contemporary referred to in the beginning of the Gīta-govinda. Keśava, who is described in the colophon as a Nyāyā- cārya, tells us that he had already composed seven abstruse treatises on the subject before he undertook the composition of his Alamkāra-śekhara. Two of these are apparently those which are mentioned in the text as his own under the citations Alamkāra-sarvasva (p. 9) and Vākya- ratna (p. 12) or Kāvya-ratna (p. 72). A Kāvya-ratna is mentioned in Oppert ii. 6237. The Alamkāra-śekhara written in the form of Kārikā anď Vrtti, consists of eight chapters (called Ratnas) and 22 sections (called Marīcis) with topics distributed as follow: i. Definition

1 The verse unmilan-madhu-gandha° of the Gita-govinda (ed. N. S. P. p. 29) is quoted anonymously by Keśava at p. 6, as an instance of the Gaudī Rīti.

APPAYYA DĪKȘITA 221

of Kāvya, etc. ii. Three Rītis (Vaidarbhī, Gaudī and Māgadhī), Ukti, Mudrā with their varieties. iii. Three Vrttis (Abhidhā etc), iv-vi. Eight Dosas of Pada, twelve of Väkya and eight of Artha. vii-viii. Five Guņas of Sabda (Samkşiptatva, Udāttatva, Prasāda, Ukti and Samādhi), four Guņas of Artha (Bhāvikatva, Suśabdatva, Paryāyokti and Sudhar- mitā). ix. Cases when Doșas become Gunas. x-xii. Eight Alamkāras of Sabda and fourteen Alamkāras of Artha. Some of the names and definitions are different from those of orthodox writers. xiii-xvii. Devoted mostly to Kavi-śīkşā topies-poetic convention, mode of describing different objects etc. xviii-xix. Certain verbal tricks, Samasyā-pūrana etc. xx. Nine Rasas, topics of Nāyaka-nāyikā, Bhāvas etc. xxi-xxii. Rasa-doșas ; and letters favourable to each Rasa. It will be seen that although Keśava Miśra accepts Dhvani and Rasa and the general pattern of orthodox Poetics, he appears yet to follow a different tradition, especially in the treatment of Guņa, Dosa and Alamkāra. But the difference is not material; for as noted above, he draws largely upon most of his well-known predecessors.

BIBLIOGRAPHY

Editions. (1) ed. Sivadatta and K. P. Parab, Nir. Sag. Press, Bombay 1895. (2) ed. Anantaram Sastri Vetal, Chowkhamba Skt. Ser. Benares 1927. (3) ed. Ganesha Sarma, Benares 1886. Our references are to Nir. Sag. Press ed.

APPAYYA DIKȘITA

( 1 )

Appayya Dīksita himself furnishes us with a clue to his date. He tells us at the end of his Kuvalayānanda that it was composed at the instance of a South Indian prince

222 SANSKRIT POETICS

named Venkata1. Aufrecht2, and following him Eggeling3, identify this patron of Appayya with Venkata of Vijayanagara (about 1535 A. D,), while Hultzsch4 shows that he was Venkata I of Pennakoņda, whose inscriptions range from Śaka 1508 to 1535 (=1586 to 1613 A. D.).5 On the other hand, in the colophon to his Sivādityamaņi-dīpikā (Hultzsch 1056), Appayya mentions as his patron a prince Cinna Bomma, son of Cinnavīra and father of Lingama Nayaka. The inscriptions of this chief of Velur (Vellore in the North Arcot district) are dated in Śaka 1471 and 1488 (=1549 and 1566 A.D.).6 In the last verse of the Kuvalayānanda reference is made to Pradyotana Bhatta's commentary Saradāgama (on the Candraloka) which in dated 1583 A.D. The extreme limits, therefore, of Appayya's literary activity are 1549 and 1613 A. D. We may thus assign him to the third and fourth quarters of the 16th century ; and as he was alive in the time of Venkata I, he may have lived into the beginning of the 17th century7. This date is confirmed by the fact that we

1 Cf also śl 168 (ed. NSP 1913) which, though occurring also in the text of Jayadeva's Candraloka, is probably one of Appayya's additions. 2 Cat. Bod. 213a. But in his Cat. Cat. i. 22a and ii. 5a, he assigns the dates, viz. end of the 15th and end of the 16th century respectively. Regnaud's conjecture (Rhétorique Sansk. p. 375) that Appayya flourish- ed in reign of Krsnaraja of Vijayanagara in 1520 A.D. is not correct. 3 IOC iii, p. 335. 4 Rep. of South Ind. Sansk. MSS ii, p. xiii and EI iv. 271 (cf JASB 1907, p. 211). 5 South Ind. Inscrip. i, p. 69f and p. 84. Also see H. D. Velankar in Cat. JBRAS., i. no. 141. 6 IA xiii p. 155 and EI iii p. 238 Table. 7 He is said to have lived to the ripe old age of 73 (see introd. to Halasyanatha's ed. of Kuvalayananda p. 15). The usually accepted date is 1552-1624 or 1554-1626 A.D. But the date 1520-1593 is argued in JOR, Madras, 1928, pp. 225-237 and 1929, pp. 140-160. See also the Madras Univ. ed. (1929) of Sivādvaita-nirņaya (introd.) and Vanivilas Press ed. of Yādavābhyudaya vol. ii->(introd.), p. ivf. where the date argued is between 1552 and 1624. Venkata, author of the Viśvagunādarsa tells us that he hailed from Kāñci (or Conjeevaram). That Appayya is

APPAYYA DIKŞITA 223

find Appayya cited by Kamalakara Bhatta in the first quarter of the 17th century and attacked by Jagannatha about the same time.

( 2 ) We find the author himself using the forms Appa or Apya of his name in his Kuvalayānanda, but it is variously spelt as Appaya and Appayya. A champion of Southern Saivism he was a versatile and prolific writer, and tradition ascribes to him more than one hundred works,1 of which Aufrecht mentions nearly seventy. A Tamil Brahman of Bharadvaja gotra, he was the fifth son of Rangarāja (or Rangarājādhvarin) and had a brother called Apya or Ācchan. Appayya is notable in Sanskrit Poetics for his three works, viz. the Kuvalayānanda, the Citra-mīmāmsā and the Vrtti- vārttika. Of these, the last seems to have been his earliest work, after which comes the Citra-mīmamsa which is referred to in his Kuvalayānanda. None of these works displays much originality ; and we have seen that his Kavalayānanda was directly based on Jayadeva's Candrāloka, up to the section on the figure hetu.2 To the "one hundred" Alamkāras of Jayadeva Appayya, however, adds fifteen,3 and this perhaps constitutes

later than the 14th century is shown by the fact that he cites the Ekāvali, Pratāparudra-yaśobhūsaņa and Samjīvanī comm. of Jayaratha. 1 So states Nīlakaņtha Dīkşita in his Nīlakaņțha-vijaya i. 44. The question is complicated by the fact that no less than four Appayya Dīksitas belonged to the family in three generations. See V. Raghavan in Proceedings of A-I.O.C, Tirupati 1941, pp. 176-80. In the New Cata- logus Catalogorum (ed. V. Raghavan), Madras 1949, pp. 197-200, there are no less than 58 entries after careful sifting. This Catalogue may be consulted for Appayyas II, III, and IV also. 2 See above p. 200. 3 In the text of the Kuvalayānanda-kārikā with Āsādhara's com- mentary, which is translated by Schmidt and published by the N. S. P. 1906, the fourth chapter dealing with sabdālamkāras is an interpolation, or rather mistaken incorporation into the text of Appayya of a chapter from Cirañjīva Bhattācārya's Kāvya-vilāsa (IOC iii, pp. 340-44). as the

224 SANSKRIT POETICS

the largest number of such figures mentioned in any Alamkāra work, and forms the climax in the process of multiplying the poetic figures with endless minute differentiation. The Citra- mīmāmsā is a more independent work; but it was probably left incomplete. In most of the MSS,1 as well as in the prin- ted texts,2 it goes up to the atiśayokti-prakaraņa and breaks off with the curious verse: apy ardha-citramīmāmsā na mude kasya māmsalā/ anūrur iva gharmāmsor ardhendur iva dhūrjateh// which, if authentic, implies that the work was designedly left incomplete. But in some MSS there is an additional verse, which gives a list of the figures to be dealt with (pratipādyā- lamkāra-sūcī),3 which ends with the mention of utpreksā, and omits atisayokti which ought to come after it. The Candrikā commentary of Vaidyanatha supports this tradition with the remark: utprekşā-granthānantaram citra-mīmāmsā na kvāpi drśyate ; but the commentary of Dharānanda, son of Rāma- bala, includes and comments on the section on atiśayokti coming thereafter. Appayya's own references to the Citra- mīmāmsā in Kuvalayānanda (pp. 78, 86, 133) relate to the treatment of the figures of ślesa, prastutānkura and arthāntara- nyasa, which are wanting in the present-day text. The printed text of the Citramīmāmsā-khandana of Jagannātha goes only as far as apahnuti. Appayya's third work, the Vrtti-varttika,

colophon at the end of that chapter itself shows. It is well known that the Kuvalayānanda deals only with Arthālamkāras. 1 e.g. IOC iii p. 336, ends with atiśayokti at fol. 72a ; Madras Trm A 1104 : Śgś ii, p. 82. 2 Ed. V. L. Panshikar, Kāvyamālā 38, NSP, 1907. In the text pub- lished in the Pandit xiii, the work ends with utpreksā, and the atisayokti is wanting. 3 upamā sahopameyopamayāthānanvayaḥ smaranam/ rūpaka- pariņati-samsaya - bhrāntimad - ullekha-nihnavotprekşāḥ/ / This versc occurs at the end of the text printed in the Pandit and in the India Office MS referred to. The MS kha used in the Kāvyamālā ed. (see p. 101 fn) ends with utpreksa. In Madras Cat. xxii, MS no. 12879 ends with atiśayokti, but nos. 12880-81 end with utprekşā.

APPAYYA DĪKȘITA 225

which is a short dissertation, after a work called Kāvya- sarani, on the three functions of word and its sense, is also incomplete as it stands ; for it consists only of two chapters on the two functions abhidh and lakşanā, and the third chapter which should deal with the third function vyañjana is wanting. Appaya appears to have written another work, called Lakşaņa-ratnāvalī on the Lakșaņas of Rūpaka.1 Appayya Dīksita, second son of Āccān Dīkșita who was a brother of our Appayya, wrote an Alamkāra-tilaka.

( 3 ) Appayya's works appear to have started some controver- sies in his time. Thus Jagannätha, who flourished immedia- tely after him, not only attacked Appayya in his Rasa-gangā- dhara and stigmatised him as a slavish imitator of Ruyyaka and Jayaratha, but also wrote his Citramīmāmsā-khandana to demolish Appayya's work of that name. Bhimasena, in his commentary on Mammata, also refers to a Kuvalayānanda- khandana written by himself as an attack on Appayya's other work ; and we find Atirätrayajvan, a younger brother of Nīlakaņțha Dīkșita and descendant of Appayya's, taking up the cudgel to defend the fair fame of his ancestor in his Citramīmāmsā-doşa-dhikkāra.2 Among more recent writers and works cited by Appayya, we find the names of the Sāhityacintāmaņi-kāra, Ratnākara, Alamkāra-sudhānidhi3 (Vrtti-vārttika p. 19) and Kāvya-saraņi,

1 See T. R. Cintamani in JOR, Madras, iv, 1930, pp. 242-44 (text of a newly discovered fragment). 2 The authorship of this work is uncertain. Oppert 4802 ascribes it to Cinna Appayya, younger brother of Nīlakantha Dīksita; but Hultzsch (ii, p. 126, no. 1281, up to Apahnuti-prakaraņa) ascribes it to Cinna Appayya's last brother Atiratrayajvan. See New Cat. Cat. i, p. 200. 3 This is probably the work of the same name attributed to Sāyana, younger brother of Mädhava and elder brother of Bhoganātha. But the illustrative verses, which are in praise of Sāyana, appear to have been composed by Bhoganātha. As ministers of Harihara I (1336-55 A.D.) and Bukka (1355-77 A.D.), Sāyaņa belonged to the 14th century. He is better known as a commentator on Vedic works. An anthology, 15

226 SANSKRIT POETICS

the last work (of which nothing is known) being avowedly the: model or source of his Vrtti-varttika. The Sāhitya-cintāmaņi, also cited by Kumarasvamin, is probably the work of the same name by Vīranārāyaņa (q. v., about 1400 A.D.). The Alamkara-sudhanidhi is apparently the same work as cited by Kumārasvāmin at p. 44. If the Ratnakara quoted in the Vrtti-vārttika p. 20 be the same as Ratnākara cited extensive- ly by Jagannätha in his two works, then it refers to the Alamkāra-ratnākara of Sobhākaramitra, and should be distin- guished from Rasa-ratnakara cited by Mallinātha on Megha- dūta. A Kāvyāloka is cited by Appayya in his Citra-mīmāmsā (pp. 27, 53)

( 4 )

THE COMMENTATORS ON APPAYYA The popularity of the Kuvalayananda as a convenient manual is indicated by the many commentaries on it, the more important ones of which have been published. The Dīpikā of the poet Āsādhara, son of Rāmajī and disciple of Dharanīdhara, has been edited as well as translated. The Alamkāra-sudhā and Satpadānanda of Nāgeśa or Nāgojī Bhatta have not yet found an editor, but the Alamkāra- candrikā of Vaidyanātha Tatsat, son of Rämacandra (or Rāma Bhatta) and grandson of Vitthala Bhatta, has been printed several times in Madras and elsewhere. The more reliable commentary of Gangādharādhvarin or Gangādhara Vājapeyin, son of Devasimha-sumati of Vādhūla-gotra and pupil of Viśvarūpa Yati of Benares, probably preserves the text and the Appayya-traditions better, inasmuch as the commentator tells us that Appayya was the teacher of a brother of his grandfather, and he himself takes great pains to settle the readings of his text. Other less known commen- taries are mentioned below.

called Subhāșita-sudhānidhi is ascribed to him (Proc. A-I.O.C. Baroda 1935, pp. 121-24.

APPAYYA DĪKȘITA 227

The Citra-mīmāmsa has been commented upon by Dharānanda, son of Rämabala of Vasistha-gotra and grandson of Thäkura, who had, besides the author's father, two other sons named Pūraņadāsa and Devadāsa, The commentator was disciple of Paramananda and was born in Bharatapura. He wrote also a commentary on Mrcchakatika (Madras Cat. xii. 12625). No commentary on the Vrtti-vārttika is known.

BIBLIOGRAPHY

Kuvalayānanda Editions. The work as a popular text-book has been printed frequently at Poona, Madras, Bombay, Calcutta and Benares in Devanāgarī as well as in Grantha, Telugu and Bengali characters, with or without commentaries. The earliest edition appears to be that printed in Puthi form from Pathsala Press, Poona 1842 (2nd ed. 1845). It is not necessery to enumerate here all the editions, but the following publications in Devanāgarī are noteworthy. Without commentary: (1) by P. R. Subrahmanya Sarman with Eng. trs. and notes, Banerjee Press, Calcutta 1903. With the Candrika commentary of Vaidyanātha Tatsat (2) the Poona ed. mentioned above. (3) ed. Jivananda Vidyasagar, Satya Press, Calcutta 1847 etc. (4) ed. Satyavrata Samasrami in Pratna-karma-nandinī, Satya Press, Calcutta 1874. (5) ed. Kashinath Vasudev Khandekar, Jagadisvara Press, Bombay 1884. (6) printed in oblong Puthi form, Kashi Samskrita Press, Benares 1879, (7) ed. Vasudev L. Panshikar, Nir. Sag. Press, Bombay 1907 (2nd ed.), 1913 etc. (8) ed. Govinda Sastri, Venkatesvara Press, Bombay 1911. (9) the Madras eds. are mostly in Grantha (1870, 1881) or Telugu characters (1870, 1895). With the Rasika-rañjanī comm. of Gangādhara Vājapeyin. (10) ed. R. Halasyanath Sastri, Kumbhakonam 1892. With the Alamkara-dīpikā comm. of Āśādhara, (11) ed. Vasudev L. Panshikar, Nirnay Sag. Press, Bombay 1909;

228 SANSKRIT POETICS

(12) the same, with text trs. into German by R. Schmidt, Berlin 1907, along with Rāmadeva Cirañjīva's Kāvya- vilāsa. See also Bibliography under Candrāloka above p. 204f. (Our references are to the NSP ed. of 1913 by V. L. Panshikar, which also contains the Candrikā comm.). Commentaries. (1) Alamkāra-candrikā of Vaidyanātha Tatsat. Ed. as above. Published many times with the text. MSS: SCC vii, 1, 29; Madras Cat. xxii, no. 12862-67 ; IOC iii, no. 270-72, p. 33. For a list of editions see BORI MSS Cat. xii, pp. 182-83. Mitra in Bik. Cat. no 607, p. 213 gives a wrong impression of the work. Vaidyanātha also wrote a commentary on Mammata's Kav. Prak. (q.v.) which in dated in 1684 A.D. See above p. 170. (2) Alamkāra-dīpikā by Āśādhara. Printed by NSP and translated, as above. Asādhara comments only on the Kārikas, but he is not aware of Jayadeva's Candrāloka, Āśādhara himself appears to have added what is called an Uddista-prakaraņa of about 21 Kārikās with pertinent comm. See BORI MS Cat. xii, no. 153, p. 174. Āśādhara also wrote Kovidānanda and Triveņikā (see under Minor Writers below). He should not be confused with Asadhara who wrote a commentary on Rudrata ; see above p. 93. (3) Rasika-rañjanī by Gangādharādhvarin or Gangā- dhara Vajapeyin. Printed from Kumbhak'onam as noted above. MSS: Aufrecht i. 113a (the attribution to Appayya himself is wrong, as corrected later), ii. 22b ; Madras Cat. xxii, 12868-70; also see Tanjore Cat. ix, no. 5205, pp. 4024-27. This commentator describes Appayya as asmat- pitāmaha-sahodara-deśikendra ; but according to tradition the commentator lived under the Tanjore prince Sāhajī (1684-1711 A.D.). Originally he was a native of Tiravā- langadu in Chingleput district. Also wrote some comms. on philosophical works. (4) Alamkāra-sudhā by Nāgojī Bhatța. First quarter of the 18th century. Nagoji is said to have written also

APPAYYA DĪKȘITA 229

another commentary on Kuvalay° called Satpadānanda, or in full Vişamapada-vyākhyāna Șatpadānanda ; see Jammu Cat. nos. 1190 and 1191; SCC viii. 28. As its name indicates, the latter commentary probably deals only with difficult words and passages. The two commentaries are often confused. In this last commentary, Nägojī refers to his °Marma-prakāśikā on Jagannātha's work. Extracts of both in Stein pp. 270-271. (5) Kāvya-mañjarī by Nyāyavāgīśa Bhatțācārya. Aufrecht i. 113a. Is he identical with Rāmacandra Nyāya- vāgīśa, son of Vidyānidhi (q.v.) and author of Kāvya- candrīkā? (6) Comm. by Mathuranatha. Aufrecht i. 113a. See above p. 217, bibliography under Viśvanātha. (7) °Țippaņa by Kuravirāma referred to in the intro- ductory verse of his comm. on the Viśvaguņādarśa; Hultzsch i, extr. p. 57, no. 21. For the author who also wrote on Dramaturgy see above p. 127 (under Dhanañjaya). As Venktādhvarin, author of the Viśvaguņādarśa, is known to be a grandson of Appayya's, Kuravirāma, who com- mented on this poem, could not have been earlier than the middle of the 17th century. (8) Laghvalamkāra-candrikā by Devīdatta. SCB 830. (9) Budha-rañjanī by Vengala Sūri, The colophon to some MSS describes him as Śrī-rāmabhūpāla-sabhābhūşaņa. Ed. in Telugu characters, Bharati Nilaya Press, Madras 1882. Also included in the Palghat ed. of Candrāloka, see . under Candrāloka, p. 204. This is really a comm. on the Arthālamkāra-section of Candrāloka, which is co- extensive with the text of Appayya's Kuvalayānanda. (10) An anon. comm. in BORI MSS Cat. xii, no. 155, p. 177.

Citra-mīmāmsā Editions. (1) ed. Rama Sastri Tailanga in the Pandit xiii, 1891, (2) With Citramīmāmsā-khaņdana, ed. Sivadatta, and

230 SANSKRIT POETICS

V. L. Panshikar. Nir. Sag. Press, Bombay 1893, 1907 (our references are to the 2nd ed. of 1907). Commentaries. (1) Sudhā by Dharānanda, son of Rāmabala of Vasistha-gotra. Comments up to Atiśayokti. Madras Cat. xii, 12884-86 (extract). Dharānanda also wrote com- mentaries on the Anargha-rāghava (Madras Cat. xxi, Kāvya, no. 12444, p. 8355) and on Mrcchakatika (ibid, no. 1265, p. 8475). The last-named comm. was composed in 1814 A.D. In it he gives his genealogy and an account of him- self, from which we learn that he was son of Rāmabala of Bharatapura, grandson of Țhākura and disciple of Para- mānanda. (2) Gūdhārtha-prakāśikā by Bālakrsņa Pāyaguņda. Aufrecht ii. 38b. He should be distinguished from Bala- krsņa Bhatța, author of Alamkāra-sāra. See chapter on Minor Writers below. (3) Citrāloka. SCB 106.

Vrtti-vārttika Editions. (1) ed. Rama Sastri Tailanga in the Pandit xii, 1890. (2) ed. Sivadatta and K. P. Parab, Nir. Sag. Press, Bombay 1893. Our references are to the N.S.P. 2nd ed. of 1910.

Lakşana-ratnāvalī Ed. T. R. Chintamani in JOR Madras, iv, 1930, pp. 242-44 (a fragment). An incomplete Grantha MS entitled Lakşaņa-ratnāvalī-vyākhyā without the name of the author is noticed in Tanjore Cat. xi, no. 5295, p. 4079 ; but it is probably a different work which appears to deal with Dramaturgy.

JAGANNĀTHA In his Bhāminī-vilāsa Jagannātha tells us1 that he passed his youth under the patronage of the emperor of Delhi, from

1 dillīvallabha-pāņipallava-tale nītam navīnam vayah, ed. Grantha- māla vol. iv, śl 32. This verse is wanting in the N.S.P. ed. 1894. The

JAGANNĀTHA 231

whom, we are told elsewhere, he received the title of Pandita- raja,1 This emperor seems to have been Shah Jahan (1628- 1658). He also seems to have lived under the protection of Nawab Asaf Khan (d. 1641), brother of Nur Jahan and a nobleman in the court of Shah Jahan, in whose praise he wrote his Asapha-vilāsa and who is also referred to in verses quoted in his Rasa-gangādhara (p. 166 sudhīva vaņī; 457 yuktam tu yāte, referring to Asaf's death). In the latter work, there is also a reference in a verse (p. 521) to Nuradīna which is apparently the Sanskritised form of one of the names of Jahangir (1605-1627), Shah Jahan's father. Shah Jahan came to the throne in 1628 A. D., and was thrown into prison in 1658 A. D. In his Jagad-ābharana Jagannātha eulogises king Jagatsimha of Udaipur (1628-1654) and in his Prāņābharaņa king Prāņanārāyaņa of Kāmarūpa (1633-1666) ; but they are essentally identical works which have been utilised, with certain change of names and addition of verses, to eulogise two patrons. It seems, therefore, Jagannätha enjoy- ed the patronage of four rulers, Jahangir, Shah Jahan, Jagat- simha and Prāņanārāyaņa at different periods of life. His literary activity, therefore, lay in the second and third quarters of the 17th century ; and it extended roughly from about 1620 to 1660 A. D. Nageśa Bhatta commented upon his Granthamāla text is published with the comm. of Mahādeva Dīkșita who claims to be a grandson of Jagannätha himself. The phrases dilli-narapati and dillisvara occur also in Rasa-gangadhara and other works (see Aryendra Sarma, Panditarāja-kāvya-samgraha, Osmania Univ. 1958, p. vii for references). 1 See citation from Asapha-vilāsa given in introd. to Kāvyamālā ed. of Rasa-gangādhara p. 2 fn .; also Nāgeśa on Rasa-gang° p. 3. The text of the Asapha-vilasa appears to contain lacuna and ends abruptly. The poet Pandita-raja, cited in the anthology Padyāmrta- tarangiņī of Haribhaskara, whose commentary on the Vrtta-ratnākara was composed in 1676 A.D. (Bhandarkar, Rep. 1877-91, p. Ixii and Rep. 1883-84. p. 60) is probably our author. There is also a commentator on Mammața named Paņditarāja (q.v.) who is a different author. Aufrecht (ii. 40a), making a confusion between the two, attributes the Kāvya- prakāša-țīkā to our Jagannātha.

232 SANSKRIT POETICS

Rasa-gangadhara in the beginning of 18th century, while Jagannätha himself attacks Appayya Dīksita who lived in the third quarter of the 16th century.1

( 2 ) Jagannatha describes himself as son of Peru2 (or Perama3) Bhatta and Laksmī. He was, like Appayya, a South Indian writer, being a native of Tailanga4 (Telugu country),. and belonged to the Venginādu, Vegināți or Vegināda family of Brahmans5. His father was reputed for his learning, having been, as Jagannätha himself tells us,6 taught the Vedānta by Jñānendra-bhiksu, the Nyāya-vaiśeşika by Mahendra Paņdita, the Pūrva-mīmāmsā by Khaņdadeva and the Mahābhāşya by Seșa Vīreśvara. Our author learnt these subjects from his father, and also from one of his father's teachers, Sesa Vīreśvara. Very little is known of his personal history, although curious tales about his fond- ness for a Muhammadan woman, named Lavangi, and his. death by plunging into the sacred river Ganges have gathered round his name.7 Jagannätha was also the author of several

1 On Jagannatha see V. A. Ramaswami Sastri, Jagannātha Pandita (reprinted from Journal of Annamalai Univ. iii-iv) and Aryendra Sarma in the work cited above. Ramaswami Sastri gives Jagannātha's. date as 1590-1665. 2 Rasa-gang° i. 3. 3 concluding verse to his Prāņābharaņa. 4 Prāņābharaņa śl. 52. 5 colophon to Bhāminī-vilāsa. 6 Rasa-gang° i. 2. 7 P. K. Gode, in Studies in Ind. Lit. History ii, 1954, pp. 452-59,. finds the earliest mention of the tradition ina MS dated 1843 A.D. But Acyuta Rāya, whose Sāhitya-sāra is dated in 1831 A.D., discusses in his commentary on the Bhāminī-vilāsa (ed. NSP, Bombay 1933) the- autobiographical significance of the lyrics in that work and of verses like yavanī navanīta-komalāngī often attributed to Jagannātha. See also L. R. Vaidya in the introd. to his ed. of the Bhaminī-vilasa. For a full discussion of the Lavangi episode see V. A. Ramaswami Sastri, op. cit., pp. 19-21 ; P. K. Gode in Bhāratīya Vidyā iv, 1942, 57-62 and in Rāja- sthāna Bhāratī (Bikaner) ii, 1948, pp. 45-49. Aryendra Sarma assumes the Lavangi verses (given on p. 190, nos. 582-88) to be genuine.

JAGANNĀTHA 233

poetical works,1 besides writing the Rasa-gangadhara, the latest yet not the least important work on Poetics, and the Citramīmāmsā-khandana. He also wrote a grammatical work, directed against Bhattojī Diksita's famous commentary Mano- ramā, and called it Manoramā-kuca-mardana. Both the rhetorical works of Jagannätha have been obtain- ed incomplete. The printed editions of the text of the Rasa- gangādhara go up to the treatment of uttarālamkāra and break off with an incomplete verse ; and so do most MSS noticed in the various reports and catalogues. Nāgesa or Nāgojī Bhatța's commentary also ends with the same section. In conformity to a pun in the word gangadhara in the title, the work was apparently planned to consist of five heads (ānana) or chapters, of which we have got only one complete and another incomplete chapter2. The topics covered by the

1 Some of these have been published by the N.S.P. For a list, see introd. to Jagannatha's Rasa-gang° (N. S. P. ed.) ; Aufrecht i. 196b ; Kāvyamāla Gucchaka i, p. 79; and Aryendra Sarma in the work cited. These are: (1) Amrta-laharī (Km. Gucchaka ii) (2) Āsapha-vilāsa, praise of Asaf Khan (in Aryendra Sarma, op. cit.) (3) Karuņā-laharī (Km. Gucchaka ii) (4) Gangā-laharī or Pīyūșa-laharī (ed. N. S. P. Bombay 1930) (5) Jagadābharaņa, praise of Jagatsimha of Udaipur (6) Prāņābharaņa, praise of Prāņanārāyaņa of Kāmarūpa (Km Gucchaka ii; 51 verses in different metres) (7) Bhāminī-vilāsa (in four Samullāsas on Anyokti, Śṛngāra, Karuņa and Sānti, ed. NSP 1894) (8) Manoramā-kuca-mardana, directed against Bhattojī Dīksita's Manoramā (9) Yamunāvarņana-campū (quoted in Rasa-g. pp. 19, 128) (10) Lakșmī- laharī (Km. Gucchaka ii) (11) Sudhā-laharī (Km. Gucchaka i). Of these nos. 3, 6 and 11 are quoted in the Rasa-ganga°, e.g. śl 60=p. 36 ; śl 4= p. 56 ; śl 1=p. 20 ; as also Gangā-laharī p. 243 (samṛddham saubhāg- yam), 491 (samutpattih); Bhāminī-vilāsa p. 402 (digante śrūyante), 403 (pura-sarasi). The Panca-laharyah (five laharīs) are mentioned at p. 109. 2 It cannot be determined whether the work was completed; but it was certainly composed before Jagannatha wrote his Citramīmāmsā- khandana, the second verse of which refers to the Rasa-gangadhara. In Citramīmāmsā-khandana, however, a reference is made to Udā- haraņāļaņkāra-prakaraņa of Rasa-g. (višeșas tu udāharaņālaņkāra- prakaraņe rasa-gangādharād avaseyaḥ, p. 12), but this Prakaraņa is not found in the extant text of Rasa-g.

234 SANSKRIT POETICS

first Anana are : definition of Kāvya, its four varieties Uttamottama, Uttama, Madhyama and Adhama ; Rasa and Bhavas ; Gunas, whether three or ten. In the second Anana we have divisions of Dhvani, with a discussion of Abhidh and Lakșaņā, after which comes treatment of Upamā and other poetic figures, enumerated as 70 ; but it is incomplete. The work is written in the form of Sutra and Vrtti. The Citramīmāmsā-khandana, directed against Appayya's work of that name, also goes as far as the apahnuti-section and does not deal with utpreksā and atiśaya which are found in some MSS of the Citra-mīmamsa. There is a reference to a matter to be dealt with in the Nidarsanālamkāra-prakaraņa {p. 101: adhikam tu nidarśanālamkāra-prakaraņe cintayisyate) which Jagannatha obviously contemplated writing.

( 3 ) Of comparatively recent writers, Jagannātha, besides citing Mammata, Ruyyaka and Jayaratha extensively, refers to and quotes Vidyadhara (p. 254), Vidyānātha (p. 162), Viśvanātha (and the Sāhitya-darpaņa, p. 7) and Appayya, and refers very often to the navyāh (pp. 25, 149, 240, 313, 429, 478). The scathing criticism which he levels against Appayya as a slavish imitator of Ruyyaka and Jayaratha was, no doubt, prompted by the zeal of eclipsing the fame of another South Indian writer in the same field. Jagannatha also cites Śrīvatsalāñchana (p. 39), apparently the commen- tator on Mammata, an unknown Alamkāra-bhāşyakāra (pp. 239, 365 ; also referred to by Jayaratha), and Ratnākara (pp. 202, 207, 209, 211, 221, 225, 281, 313, 480, 492 etc.) which last name is also cited by Appayya. Jagannātha also refers to a work called Alamkāra-ratnākara (pp. 163, 165). An anonymous Alamkāra-ratnākara is mentioned in Burnell 54a ; but Bühler1 describes a work of that name by Sobhakara- -mitra, son of Trayīśvaramitra. Peterson informs us2 that

1 Kashmir Rep. App. ii no. 228, p. cxxviii. 2 Rep.ip. 12. Bühler (Report 1877) mentions a small work called

JAGANNĀTHA 235

the Kashmirian poet Yaśaskara extracted some sūtras on Alamkāra1 from a work called Alamkāra-ratnākara by Sobhäkaramitra, and illustrated them in his Devi-stotra by composing verses in praise of Devi, as the opening words of the latter work themselves show2. The work of Yaśaskara in Stein is for this reason called Alamkarodāharaņa- .samnibaddha Devi-stotra.3 The Ratnākara of Jagannātha undoubtedly refers to this Alamkāra-ratnākara of Sobhākara- mitra ; for the citation from Ratnākara at p. 202=sūtra 11 (as given in Peterson i p. 78).4 Jayaratha criticises (pp. 41, 52) the Kashmirian Sobhakara who deviates from Ruyyaka. Jagannātha says (p. 281) that Appayya Dīksita follows Alamkāra-ratnākara. ( 4 ) Nāgojī Bhatța The commentator on the Rasa-gangādhara is Nāgeśa or Nāgojī Bhatta, whose name we have already mentioned as a commentator on Mammata, Govinda Thakkura, Bhānudatta

Dhvani-gatha-panjika which contains explanations of Prakrit verses, apparently of the Dhvanyāloka; but there is no evidence (except the epithet Kaśmīrakācārya in the colophon) to show that he was the same as the Kashmirian Ratnakara, author of the poem Hara-vijaya. BORI MS no. 182, Cat. xii, p 207. 1 These are given in Peterson, op. cit. App. pp. 77-81. 2 ratnākarābhyantarato grhītvālamkāra-sūtrāņi yathākrameņa/ bandīva devyā girirāja-putryāh karomi samsan śruti-gocarāņi. The commentary on this verse explains: śrī-trayīśvara-mitrātmaja-śrī-śobhā- karamitra-viracite'lamkāra-ratnākare'lamkāra-sūtrāņi. The colophon to Bühler's MS of the Alamkāra-ratnākara reads trayīśvara-mantra- putrasya as a description of Sobhakaramitra, in which the word mantra is obviously a mistake for mitra. Stein's Jammu MS 58 reads Sobhā- karamitra as the name of the author (cf. also W Bod 1162). 3 The original is also called Alamkāra-ratnodāharana and the author Sobhakareśvara. See also Mitra 1822; Hultzsch's Eine Samm- lung ind. Handschriften 170. 4 The work has been recovered and edited by C. R. Devadhar, Poona 1942. See under Minor Writers below. Jagannātha refers to Alamkāra-ratnākara in as many as eleven different places (see C. R. Devadhar in Proc. A-I. O.C, Lucknow 1955, pp. 60-65).

236 SANSKRIT POETICS

and Appayya. He was a Mahratta Brahman (with the surname Kāla or Kāle), son of Siva Bhatta and Satī. He lived in Benares and was patronised by Rāmasimha of Śrngavera-pura (near Allahabad). He is one of the latter-day grammarians, who composed a number of works and commentaries on grammar, poetics and philosophy. He was a pupil of Hari Dīkşita, who was the son of of Vīreśvara Dīksita and great- grandson of Bhattoji Diksita, the well-known author of the Siddhānta-kaumudī. Bhatțojī is known as a pupil of Seșa Krsņa,1 whose son Śesa Vīreśvara was, as we have noted, a teacher of Jagannātha himself. Nāgojī was, thus, separated from Jagannatha roughly by two generations, and flourished in the beginning of the 18th century. The India office MS of his commentary on Bhānudatta's Rasa-mañjarī is dated in Māgha Samvat 1769=Feb. 1713 A.D.2 Nāgojī was the teacher of Vaidyanātha, the Maithili grammarian, and of Gangārāma, the great-grandfather of Maņirāma (1802 A.D.).3 Nagojī wrote the following commentaries on various works on Poetics: (1) Guru-marma-prakāśikā on Jagannātha's Rasa-gangadhara. (2) Brhat and Laghu Uddyota on Govinda's Pradīpa on Mammata. (3) Udāharaņa-dīpikā or °Pradīpa on Mammața. (4) Alamkāra-sudhā and Vişamapada-vyākhyāna Şatpadānanda on Appayya's Kuvalayānanda. (5) Prakāśa on

1 author of the Pada-candrikā and Prakriyā-prakāśa, and son and pupil of Sesa Narasimha or Nrsimha. For the Seșa family of Benares see IA, 1912, p. 245f .- Nāgojī's relation to Jagannātha is illustrated thus: -Śesa Vīreśvara-Perubhatța-Jagannātha (son) (pupil) (pupil of Śea Vīreśvara Śeșa Śeșa Krşna- and son of Perubhațța Nrsimha -> (son and pupil) Bhattojī Vīreśvara Hari Nāgojī Dīkșita Dīksita - Dīkșita -> Bhațța (pupil) (son and (son and (pupil) pupil) pupil) 2 IOC iii, p. 355 ; cf. Belvalkar's Systems of Skt. Grammar, p. 49. 3 Maņirāma wrote a comm. on Jagannātha's Bhāminī-vilāsa in 1802 A.D. See IOC vii, p. 1526.

JAGANNĀTHA 237

Bhanudatta's Rasa-mañjarī. (6) A commentary on Bhānu- datta's Rasa-tarangiņī. The relative chronology of some works of Nāgoji is fixed by P. K. Gode (Oriental Thought i, no. 2, 1955, pp. 45-52) between circa 1670 and 1750 A.D.

BIBLIOGRAPHY

Rasa-gangādhara Editions. (1) Ed. Durgaprasad and K. P. Parab (with Nagojī's comm.), NSP, Bombay 1888, 1913, 3rd ed. 1916, 6th ed. 1947 (2) Ed. Gangadhar Sastri (with Nāgoji's comm.), Benares Skt. Series 1885-1903. Our references are to the NSP ed. of 1916. Commentaries. (1) Guru-marma-prakāšīkā by Nāgeśa or Nāgojī Bhatta. Printed in NSP and Benares ed. as noted above. (2) Vişamapadī. Anonymous. Aufrecht i. 494b.

Citramīmāmsā-khaņdana Edition. Ed. Sivadatta and K. P. Parab, with Citra-mīmāmsā, NSP, Bombay 1893, 1907.

CHAPTER IX

LATER WRITERS ON RASA AND KAVI-SIKȘĀ

WRITERS ON RASA

( 1 ) ŚĀRADĀTANAYA Śāradātanaya, author of the Bhāva-prakāśa, °prakāśikā or °prakāśana, appears to have been a popular writer on the Rasas and Bhavas, who is quoted extensively by Kumāra- svāmin,1 and in the Kāma-dhenu on Vāmana,2 as well as by commentators like Rāghavabhatta, Ranganātha3 and Vāsu- deva.4 He is described as son of Bhatta Gopāla and grandson of Krsna, and great-grandson of Laksmana of Kāśyapa-gotra who is said to have been a resident of Mātara- pūjya village in the Merūttara country of the Āryāvarta, and to have written a commentary on the Vedas, called the Veda- bhūșana. Our author was so named, because he was con- sidered to have been born by the grace of Sāradā, worshipped in Benares. He learnt Nātya-śāstra from Divākara, and says that he is following Abhinavaguptācārya; but his work in some parts is in reality a condensed epitome chiefly of Bhoja's Śrngāra-prakāśa, which it cites and practically summarises. This fact will place āradātanaya chronologi- cally later than Bhoja ; and the citation of the Bhāva-prakāsa by Śinga-bhūpāla (pp. 20, 139, 169, 202 etc.) will fix its other terminus at 1330 A.D. We may, therefore, assign our author roughly to the period between 1100 and 1300 A.D. Śāradātanaya cites Agastya (p. 2), Kohala, Mātrgupta Subandhu and Āñjaneya (p. 251) as authorities on the dramatic 1 pp. 12, 15, 44, 68, 102, 106, 118, 121. 127, 129, 139, 143, 145, 219, 223 etc. 2 e g. on I. 3. 30. 3 on Vikramo°, ed. N. S. P. 1885, p. 10. 4 on Karpūra-mañj° ed. N.S.P. 1900, pp. 5, 7, etc.

ŚIŃGA-BHŪPĀLA 239

art. The work consists of ten adhikaras as follows: 1 and 2. Bhāva. 3. Avāntara-bhāvabheda-svarūpa. 4. Śṛňgārālambana- nāyakādi-svarūpa. 5. Nāyaka-bhedāvastha-rasa-bhāva-vikāra. 6. Śabdārtha-sambandha-bheda-prakāra. 7. Nāțyetivrttādi- lakşaņa. 8. Daśa-rūpaka-laksaņa. 9. Nrtyabheda-svarūpa- lakşaņa. 10. Nāțya-prayoga-bheda-prakāra. The Kavi-kalpalata cited in this work (pp. 131, 175) is different from works of the same name of Arisimha and Deveśvara ; for our author believes that the Kāvya-prakāsa borrowed from it. For quotations in Sāradātanaya's work see Adyar Library Bulletin xix, 1-2, pp. 47-51. Ed. Yadugiri Yatiraj of Melkote and K. S. Ramaswami Sastri in Gaekwad's Orient. Series 1930.

( 2)

ŚINGA-BHŪPĀLA Śinga-bhūpāla, whose name is also given as Śinga-dharaņīśa Śinga-rāja or Singa-mahīpati has been identified by Sesagiri Sastri with Singama Nāyadu, rājā of Venkatagiri who ruled, on the authority of a biographical sketch of the Rājās of Venkatagiri, about 1330 A.D.1 This date is probable, as Mallinātha (on Kumāra i. 25, iti bhūpālaḥ=Rāsārņava i. 181), as well as Kumārasvāmin, quotes our author. The intro- ductory verses of the Rasārnava-sudhākara, attributed to this South Indian prince, show that he belonged to the Recarla dynasty, and ruled over the country lying between the Vin- dhyas and Śrīśaila, of which Rājācalam (Rācakoņda) was the hereditary capital. He was the son of Ananta (or Anapota)

1 For detaills see ŚgS i, pp. 7-11 ; also introd. to Trivandrum ed. of the Rasārnava°. M. T. Narasimha Aiyangar (ed. Subhāșita-nīvī,. Vanivilasa Press, Srirangam 1908) states that Singama Nāyadu, our author, was a contemporary of Praudha Devarāja of Vijayanagar (1422-1477 A.D.) ; but P. R. Bhandarkar in his Note on Śinga-bhūpāla (in Proc. First Orient. Conf. Poona, ii, 1916, p. 425) doubts the correct- ness of this date. A. N. Krishna Aiyangar would assign the period bet- ween 1340 and 1360 A. D. (Proc. A-I.O.C., Mysore 1937, pp. 264-73).

240 SANSKRIT POETICS

and Annamämba, while his grandfather and great-grandfather were respectively named Śinga-prabhu (or Śińgama Nāyaka) and Yacama Nayaka. Our author was, like Hemacandra, called sarvajña on account of his great knowledge, and was a patron of letters1. The Rasarnava° appears to be mainly based on previous works like Bhoja's Śrngāra-prakāśa (Bhoja cited 57, 69, 149, 168, 190) and Śāradātanaya's Bhāva-prakāśana (cited pp. 139, 169, 202), although it draws directly on Bharata, Rudrabhatta (cited pp. 29, 30, 87), the Daśa-rūpaka and other authors and works on Rasa and Dramaturgy. The author Singa-bhūpāla is in the third person throughout the work. We have a large number of dramas actually quoted or cited for illustration, among which may be noted: Prabodha-candrodaya (pp. 265, 291), Anargha-rāghava (72, 83, 261, 266, 274), Prasanna- rāghava (pp. 258, 277), Dhanañjaya-vijaya-vyāyoga2 (p. 287), Abhirāma-rāghava 3 (anapotanāyakīya pp. 119, 243, 265, 273, 275), Mādhavī-vīthikā (p. 290), Māyākurangikā-īhāmrga (p. 298), Padmāvatī (p. 263, 266), Kāma-datta 4 (p. 285), Rāmānanda (pp. 248, 255, 269), Karuņākandala-anka (pp. 163, 197, 198, 286), Vīrabhadra-vijrmbhana-dima (pp. 272, 274, 276, 278, 298), Maheśvarānanda (p. 275), Ānandakośa-prahasana pp. 40, 41, 278, 291, 297), Śrngāra-mañjarī-bhāņa 5 (p. 288),

1 Śgš loc. cit. 9 .- For Viśveśvara Kavicandra's Camatkāra-candrikā panegyrising Singa-bhūpāla who is called here sarvajna, see under Minor Writers below. 2 By Kāñcana, son of Nārāyana, in Aufrecht i. 266b (ed. Kāvya- mālā 54, 1895). 3 A drama of this name was written by Manika in Nepal in 1390 A.D. (Lévi 268). 4 This little known work is also alluded to in the Padmaprābhrtaka- bhāna (ed. Madras 1922), ascribed to Sūdraka. According to the editors of this Bhana (Pref. p. iv), the Kāma-datta was a prakarana composed by sudraka himself. This Bhana is quoted anonymously by Hemacandra (p. 198,1. 12). 5 A Śṛňgāra-mañjarī-bhāņa by Gopālarāya, son of Jakkula Venk- tendra and Vīramāmbā, is mentioned in Hultzsch i, extract p. 77 (no. 385), p. x.

BHÂNUDATTA 241

Payodhimathana-samavakāra (p. 290), Kandarpa-sarvasva by himself (p. 151), and Vīrānanda (pp. 159, 160). The Rasārnava in three Vilasas comprehends, with profuse illustrations, practically all the topics of Dramaturgy and Rasa. The topics, in brief, are as follow: Vilāsa I. Nātya- lakșaņa and Rasa-lakșaņa ; qualities and classification of the Nāyaka; his assistants in love affairs : the qualities and classification of the Nayika; incidental treatment of three Rītis (Gaudī, Vaidarbhī and Pāñcālī), four dramatic Vrttis (Sāttvatī, Kauśikī, Ārabhațī and Bhāratī), Pravrttis and Sāttvika Bhāvas. II. Detailed treatment of 33 vyabhicāri- and 8 sthāyi-bhāvas ; aspects of Rati ; other Rasas besides Śrn- gāra ; opposition and commixture (Samkara) of Rasas; Rasābhāsa. III. Different kinds of Rūpaka and its subject- matter ; five Artha-prakrtis ; Patākā-sthānaka ; five Avasthās; five Samdhis with all their Angas treated in detail ; Bhūșaņas ; Nāțaka as the principal Rūpaka; other types of Rūpaka; languages to be employed ; naming of different characters. The printed text in the Trivandrum Skt. Series covers more than 300 pages. śińga-bhūpāla appears to have written a Nātaka-pari- bhāşā1, which topic is also treated briefly at the end of his Rasārņava. Singa-bhūpāla also wrote a work entitled Samgīta-sudhākara (ed. Kalivara Vedantavagisa and S. P. Ghosh, New Arya Press, Calcutta 1879) which appears to be a commentary on the Samgīta-ratnākara of Sārngadeva. Editions. (1) by Sarasvatisesa Sastri, Venkatagiri 1895. (2) by T. Ganapati Sastri in the Trivandrum Sansk. Series, 1916.

( 3 ) BHĀNUDATTA Bhanudatta2 is notable for his two popular works on the subject of Nāyaka-nāyikā and the Rasas, called the Rasa- 1 It is a small work of 289 verses. See India Office Cat. vol. ii (Keith and Thomas), no. 5248, p. 346. 2 The form Bhanukara of this name is given by Sesa Cintāmaņi's 16

242 SANSKRIT POETICS

mañjarī and the Rasa-tarangiņī. In the latter work, the author excuses himself1 from giving further details about a certain point because, he says, they are already given in his Rasa-mañjarī which was, therefore, the earlier composition. Besides some verses from Bharata and Rudra's Śrngāra-tilaka and a verse from the Dhvanyāloka2, Bhānudatta cites a work called the Rasa-ratna-dipika3, which may or may not be the same work as quoted by Ratnakantha in his commentary on Mammata4. In the Rasa-tarangiņī, Bhānudatta refers to Pūrvācāryāḥ, Pūrva-granthakāra-sammati and Prācīna- sammati. These citations, however, give us no clue to his exact date. A lyrical poem, called the Gita-gaurīsa or Gīta-gaurīpati in ten cantos, published in the Granthamāla 1887-88, also professes to have been composed by Bhanudatta, who is probably identical with our author5. As indicated by their respective writings, both the authors are Saivas, and both possess poetical pretensions. Our Bhanudatta gives the name of his father as Ganeśvara6, Ganapatinātha7 or ·Parimala, Gopāla's °Vikāsa and Rangaśāyin's °Āmoda commentaries. The title miśra is also appended sometimes to the name. 1 bahavo bhedāś ca rasa-mañjaryām višeşato darsitāh, iha punar vistara-bhiyā na pradarśyanta iti, ed. Granthamālā, p. 35 ; ed. Regnaud, p. 57, 1. 32. 2 The verse anaucityad rte in Dhva° p. 145 is cited with the remark tatra pracīna-granthakrtah, implying that Bhanu was much younger than Anandavardhana, in whose Vrtti the verse occurs. 3 In Rasa-tarang° ed. Granthamālā ad i. 31 ; ed. Regnaud, p. 44, 1.32. 4 Peterson ii, 17. An incomplete MS of a work called by this name is entered in Bhandarkar, Rep. 1884-87, no. 533. where the name of the author is given as Allaraja, which occurs as Mallaraja in Bühler Rep. 1874-75, no. 19, p. 16. See ch. on Minor Writers below under Allarāja. 5 Aufrecht apparently distinguished the two authors at first by separate entries under their name (i. 405), but later on (i. 793) he states. that the author of the lyric "is most likely identical with the writer of the Rasa-tarangini." (MS described in IOC vii, pp. 1443-45). 6 Rasa-mañj°. 168. 7 Rasa-tarang°.ed. Regnaud p. 66, col .; also Weber 824.

BHĀNUDATTA 243

Gananātha1, while the author of the poem describes himself as the son of Ganapati or Gananatha2. There are also a few verses in the two works of our Bhanudatta which occur in the poem. Thus the mangala-verse of the Rasa-mañjarī (ātmīyam caraņam)=Gīta-gaurīśa ii, p. 90; akaroḥ kimu netra° in Rasa-mañjarī 51=Gīta-gaurīśa ii, p. 14 ; prāņeśasya prabhavati in Rasa-tarangiņī iv, p. 40 ed. Granthamālā= Gīta-gaurīśa ii, p. 77. As the Gīta-gaurīśa is not a mere compilation in which we may expect verses from other writers, the presence of verses in it from the two works of Bhanudatta can be reasonably explained by a presumption of common authorship of the three works. The Gita-gaurisa appears to have been modelled on Jayadeva's Gīta-govinda which, like Kālidāsa's Megha-dūta, gave birth to numerous imitations in later times.3 Even a cursory examination of these two works will show the close connexion between them ; and it is remarkable that not only the general scheme, but even the verses in some chapters of Bhanudatta's poem correspond in metre with those of Jayadeva's. Thus: Jayadeva pralaya-payodhi-jale dhrtavān usi vedam vihita-vahitra-caritram akhedam keśava dhrta-mīna-śarīra, jaya jagadīśa hare and Bhānudatta bhramsi jagati sakale pratilavam aviśeşam 1 In MSS of Rasa-tarang° noticed in ALeip 835, Weber 1726. 2 kavi-gaņanātha-sutasya kaver iti vacanam tri-jagati dhanyam, iì, p. 50; kṛta-hara-vinayo gaņapati-tanayo nigadati hita-kāraņam ii, p. 58. 3 e.g. the Gīta-gangādhara by Kālyaņa, the Gīta-girīsa by Rāma, the Gita-digambara by Vamsamaņi (HPS i. 18), the Gīta-rāghava by Prabhākara, son of Bhūdhara, dated Samvat 1674 (Bhandarkar Rep. 1882-33 p. 9). A Gīta-rāghava by Hariśamkara is mentioned in HPS ii, no. 53. See also Rāma-gīta-govinda (IOC vii, p. 1480) characterised by Eggeling as "a weak imitation of Jayadeva's Gita- govinda" but attributed to Jayadeva. Eggeling quotes from Garcin de Tassy about such passing off of imitations of Gita-govinda for that of Jayadeva. Cf. Pischel. Die Hofdichter des Laksmanasena p. 23-

244 SANSKRIT POETICS

śamayitum iva jana-khedam aśeşam purahara kṛta māruta-veśa, jaya bhuvanādhipate. Again, Jayadeva nibhrta-nikuñja-grham gatayā niśi rahasi nilīya vasantam cakita-vilokita-sakala-diśā rati-rabhasa-rasena hasantam sakhi he keśi-mathanam udāram ramaya mayā saha madana-manoratha-bhavitayā savikāram and Bhānudatta abhinava-yauvana-bhūşitayā dara-taralita-locana- tāram kimcid-udañcita-vihasitayā calad-avirala-pulaka- vikāram sakhi he śamkaram udita-vilāsam saha samgamaya mayā natayā rati-kautuka-darśita- hāsam. These two extracts, taken at random, will show how close the imitation is. We may presume reasonably from this that Bhanudatta's work was written some time after Jayadeva's lyric had achieved sufficient literary reputation to be thus imitated. Whether we place Jayadeva in the first or second half of the 12th century, Bhänudatta cannot be put earlier than that century, and this conclusion gives us one terminus to his date. The other terminus is furnished by the date of one of the commentaries on the Rasa-mañjari, called the . Rasamañjarī-vikāsa (or -vilāsa) by Gopāla (alias Vopadeva), son of Nrsimha, which is expressly dated in 1572 A. D.1 In the Sarngadhara-paddhati, which was

1 Stein, Jammu Cat. p. 63, corrected at p. 421, also p. 273. As to the date of this commentary and the era used, see below Bibliography .- Kumārasvāmin, at the beginning of the 15th century, quotes (p. 280) the authority of a work called Rasa-mañjari to show that viraha-

BHĀNUDATTA 245

compiled about 1363 A.D., several verses are cited under the names of Bhanu-pandita and Vaidya Bhānu-pandita (790, 973, 1032, 1271, 3328, 3685), none of which, however, can be traced in the known works of our author. Jahlana, whose anthology (ed. GOS, Baroda 1938) was compiled about the middle of the 13th century, also quotes and ascribes to Bhanu-pandita and Vaidya Bhānu-pandita as many as 36 verses, which are similarly untraceable, but three of which are found under the same citation in the Paddhati (790=p 68, 973=p. 107, and 3328=p. 183). It may, however, be presumed that the author of the Rasa-mañjari was not unknown at this period, and that in the anthologies the epithets vaidya and pandita were annexed to an earlier or later Bhanu in order to distinguish him from our author, whose namesake he was1. A closer approximation of Bhanudatta's date is possible if any inference is permissible from the mention of Nijāma- dharaņīpāla in the Rasa-mañjari. It is suggested2 that the Nizam referred to as ruler of Davagiri by Ananta Pandita's commentary is Ahmad Nizam Shah, who obtained possession of Daulatabad (Devagiri) sometime between 1499 and 1507 A.D. and founded the Nizam Shahi dynasty of Dekkan. P. V. Kane brings in fresh evidence by pointing out (HSP, p. 296-

vipralambha is also termed pranaya-māna. It is not clear whether the reference is to Bhänudatta, for the dictum cannot be traced in his Rasa- mañjarī. 1 In the Sabhyālamkarana of Govindaji (Bhandarkar Rep. 1887-91, p. lxiii) we have selections from the poems of Bhanukara and Bhanu- pandita, by which obviously a distinction, is meant between the two poets .- Har Datta Sarma (ABORI, xvii, 1936, pp. 243-58), relying chiefly on the ascription of a large number of Bhanudatta's verses to poet Bhänukara by some late anthologies, suggests their identity, which is very doubtful. See on this question G. V. Devasthali in NIA, vii, 1944, pp. 111-17 ; P. K. Gode in Ind. Culture iii. pp. 751-56 and S. K. De, Some Problems, p. 147. 2 S. K. De in Some Problems, p. 144-45. But Ramnath Jha (Journal of Patna Univ. iii, no. 1-2) thinks that the Nizam is the second ruler of the dynasty, and Krsna (so interpreted also by Ananta Pandita) is Krşna Deva Rāya of Vijayanagar (1509-1530).

246 SANSKRIT POETICS

97. that Bhanudatta married a sister of the Smrti-writer Misaru Miśra, author of Vivādacandra, who flourished in the middle of the 15th century. It would not be unreasonable, therefore, to assign Bhanudatta to a period between 1450 and 1500 A.D. In the last verse of the Rasa-mañjari, Bhanudatta's native country is given as Videha (videhabhūh)1 or Mithila, which agrees with Burnell's description of Bhanudatta as a native of Mithila. As a Maithili writer, it is not surpising that he was acquainted with the Gaudiya Jayadeva's well-known lyric, and tried to emulate it with a similar work on Siva and Gaurī. In another work called Kumāra-bhārgavīya2 attributed to Bhānu- datta, the author is called the son of Ganapati or Gananatha (obviously the same as our author), and his pedigree is given thus: Ratneśvara-> Sureśvara (author of a Śārīraka-bhāşya- vārttika)-> Viśvanātha-> Ravinātha-> Bhavanātha-> Mahā- deva-> Gaņapati-> Bhānudatta. Ganapati appears to have been a poet whose verses are quoted by Bhänudatta himself in his Rasa-tarangiņī. A poet Ganapati is praised, in a verse ascribed to one Rājaśekhara in Jahlaņa's anthology (p. 45), as the author of a work called Mahamoda. An Alamkāra- tilaka in five chapters and a Śrngāra-dīpikā are also attributed to our Bhnudatta. The Srngāra-dīpikā is not available, but the Alamkāra- tilaka has been published3. It is written in mixed prose and verse and cover the usual topics of Sanskrit Poetics. The first Pariccheda speaks of Kavya, of which Rasa is declared

1 In Madras ed. of 1872, as well as in the MSS noticed by Aufrecht (Bod. Cat. 213b) and Bhandarkar (Rep. 1883-84 p. 12), the reading vidarbhabhüh is a mislection; for the author represents the river Ganges as flowing through his country, which is true of Videha and not of Vidarbha. Cf. Weber ii no. 1726. The title miśra, often appended to Bhanudatta's name, would indicate that he was a Maithili Brahman, and that he was probably not a Vaidya. 2 The work is in 12 Ucchvasas in mixed prose and verse. See IOC vii p. 1540, where the genealogical verses are quoted in full. 3 By G. V. Devasthali in JBRAS (New Series), xxiii-xxv (1947-49).

BHĀNUDATTA 247

to be the soul, and deals with its three varieties, Uttama, Madhyama and Adhama. It then goes on to six Rītis and four Vrttis (Kaisiki etc.). The second and third Paricchedas are devoted respectively to Dosa and Guna, in the treatment of which the work closely follows Bhoja. The fourth and fifth Paricchedas deal with Alamkāras of Sabda and Artha respectively, enumerated as 77. A work of the author's called Citra-candrikā is cited. The Rasa-mañjari is a much smaller work which deals generally, with illustrations, the topics of Nāyaka-Nāyikā, their adjuncts and excellences, the two varieties of Śrngāra, and ten stages of Vipralambha-Srngara. The Rasa tarangiņī, divided into eight Tarangas, deals more elaborately with Rasas, with a detailed treatment of Srngāra. The chapter- arrangement of topics is as follows: (i) Sthāyi-bhāva (ii) Vibhāva (iii) Anubhāva (iv) Sāttvika-bhāva (v) Vyabhicāri- bhāva (vi) Śrngāra Rasa (vii) Other Rasas (viii) Three kinds of Drsti in relation to Sthāyi-bhāva, etc. The commentaries on the two works of Bhnudatta, as detailed below, are numerous. Of these, the Naukā on Rasa- tarangiņī by Gangārāma Jadi, and the Vyangyārtha-kaumudī and °Prakāśa on Rasa-mañjarī by Ananta-paņdita and Nāgojī Bhatta respectively have been published. Bhanudatta also appears to have compiled an anthology called Rasa-parijata (printed by Matilal Banarsidas, Lahore 1939). BIBLIOGRAPHY Rasa-mañjarī Editions. (1) In Kāvya-samgraha of Jivānanda Vidyasagar, 2nd. ed. Calcutta 1886. (2) ed. Rama Sastri Tailanga, with Vyangyārtha-kaumudī of Ananta Paņdita and °Prakāśa of Nāgojī Bhațța, Benares Skt. Series, 1904. (3) ed. Venkatarama Sastri in Telugu characters, Madras 1909. (4) In Grantha- ratna-mālā vol. 1. Our references are to the Benares ed. Commentaries. (1) Vyangyārtha-kaumudī by Ananta Paņdita, son of Tryambaka Pandita (Timaji) and grandson of

248 SANSKRIT POETICS

Bālo Paņdita, and great-grandson of Nīlakaņțha Paņdita. His native place was Punyastambha (Puntambem in Ahmednagar) on the Godävarī. The comm. was written at Benares in Samvat 1692=1636 A. D. at the request of Candrabhanu, son of Virasenadeva and grandson of Madhukara. The date of composition is specified in the concluding verse which is given in the India Office MS (Eggeling, IOC iii, p. 356), but omitted in the printed text. Ed. as above. Ananta also wrote a Mudrārāksasa- pūrvapīthikā (Mitra 1654), and a commentary on Govar- dhana-saptaśati in 1645 A.D. (ed. N.S.P. Bombay, 1886). (2) Prakāśa by Nāgojī or Nāgeśa Bhatta, for whom see under Jagannatha. Ed. as above. IOC, iii, no. 1222/, 2602 ; BORI Cat. xii, nos. 223-25, p. 258f. (3) °Parimala by Śeşa Cintāmaņi, son of Seşa Nrsimha and younger brother of Sesa Krsna. Extract in Mitra 3115, vol. ix p. 194, and Bhandarkar, Rep. 1883-84, p. 365. The MS in Mitra appears to have been copied in Samvat 1609 (=1552-53 A.D.); but the MS in Bhandarkar bears no date. There are six MSS of this comm. in BORI Cat. xii, nos. 217-222 ; see also IOC iii, nos. 1226-27, p. 357. Cintāmani also wrote several other works including one on Prosody called Chanda-prakāśa (Aufrecht 189a). (4) °Vikāsa (or °Vilāsa) by Gopāla Ācārya, alias Vopadeva of Kaundinya-gotra, son of Nrsimha and grandson of Gopāla of Jabala-grāma in the Mahārāștra country, and pupil of Menganatha. He calls his author Bhanukara. The comm. is dated in Samvat 1484=1428 A. D. See Jammu Cat. no. 1221 (extract). Sridhara Bhandarkar however, points out (Rep. of Second Tour 1904-06, p. 36) that the date is 1494 and not 1484 as given by Stein (extr. p. 273), but he thinks that it is Saka era, in which case the date of the commentary will be 1572 A. D. See P. K. Gode in ABORI, xvi (1934-35) pp. 145-47. (5) Rasika-rañjanī by Gopāla Bhatta, son of Harivamśa

BHĀNUDATTA 249

Bhatta Dravida and grandson of Nrsimha Bhatta. He also wrote a comm. on Rudra's Srng. til. There is also a commentator of the same name on Mammata. They may all be identical. See above pp. 95, 161. MSS: Aufrecht i. 495b, ii. 116a, iii. 106a ; BORI MSS Cat. xii, nos. 226-30 (extracts). The commentary by Harivamśa Bhatța in Oudh Cat. xi. 10 is probably a mistake for this comm. of his son. For Gopala's other works see Aufrecht i. 161. (6) Samañjasā or Vyangyārtha-kaumudī by Viśveśvara, son of Laksmīdhara, for whom see the chapter on Minor Writers below. MSS: Aufrecht i. 495b, ii. 116a, iii. 106a. (7) Āmoda by Rangaśāyin alias Gurujālaśāyin or Gurujāla Rangaśāyin, who is said to have studied under Mahādeśika of Vādhūla-gotra and under his own paternal uncle Anantācārya. He is described as son of Dharmācārya of Cilukamari family. He was a Śrīvaișņava. V. Raghavan points out that Gurujāla is a village in Palnad Taluq in Guntur district ; the author's real name was thus Rangasāyin. Chilakamari is in the East Godavari district ; this epithet attached to the names of his father and grandfather shows that this must have been their original home. Rangasāyin frequently criticises Parimala which, as noted above, cannot be later than 1553 A. D. He quotes also Kuvalayānanda of Appayya, as well as Praudha-manoramā of Bhattojī Dīksita, whose literary activity belongs to the last quarter of the 16th century. Rangasāyin, therefore, may be assigned to the 1st half of the 17th century A.D. Besides the Amoda commentary he wrote a Śrngāra-laharī which he himself quotes. Madras Cat. xxi1, 12941-42, (extracts). (8) Vyangyārtha-dīpikā by Ānanda Śarman, son of Tryambaka. Aufrecht i. 495a. ii. 116a. (9) Bhānu-bhāva-prakāśinī by Mahādeva. Tanjore Cat ix, no. 5284, p. 4070. (10) Rasika-rañjana by Vrajarāja Dīksita, son of Kamaraja and father of Jīvaraja. North Western Pov. Cat.

250 SANSKRIT POETICS

1877-86, ii. 120. See below under Jīvarāja's comm. on Rasa-tarangiņī. Burnell wrongly enters Rasika-rañjana as a commentary by "the author of the text." It appears from Madras Cat. xx, Kāvya, p. 8008 that Vrajarāja wrote an independent poem, entitled Rasika-rañjana, in three Stabakas, describing feminine attractions and charms. Are the two works identical? (11) Rasamañjarī-sthūla-tātparyārtha. IOC iii, 1230/ 543, p. 358.

Rasa-tarangiņī Editions. (1) with Naukā of Gangārāma Jadi, in Puthi form, Kashi Samskrita Press, Benares 1886 (2) in Grantha- ratna-mālā vol. i. 1887-88 (3) ed. Regnaud in his Rhétorique Sanskrite, Paris 1884 (text in Roman transliteration). Commentaries. (1) °Naukā by Gangārāma Jadi or Jadin. Madras Cat. xxii 12930 (extract)-31. This commentary is dated in Samvat 1799=1742-43 A.D. (P. K. Gode in ABORI, xiii, p. 186). Gangārāma is also the author of an independent work called Rasa-mīmāmsa (ed. with the author's Chāyā, Kashi Samskrita Press, Benares 1885), in which he refers to his Naukā. He was son of Nārāyana and pupil of Nīlakantha, and also wrote two works on logic, one of which is Tarkāmrta-caşaka; in it he gives the names of his father and preceptor (see Aufrecht i. 140). This is a comm. on the Tarkāmrta of Jagadīśa. On Gangārāma Jadi see G. V. Devasthali in Journal of Univ. Bombay, xi, pt. 2, 1942, pp. 84-88. (2) Rasika-rañjanī by Veņīdatta Tarkavāgīśa Bhattācārya, son of Vireśvara and grandson of Laksmaņa. From a corrupt verse giving the date Eggeling (IOC, no. 1216) surmises that the commentary was dated is 1553 A.D .; but G. V. Devasthali (IA v, 1942, p. 195f) shows that this early date cannot be accepted. In his opinion the work . belongs to the beginning of the 18th century, about 1708 A. D. The author belonged to the Ahicchatra-dhara Kula,

BHĀNUDATTA 251

and his genealogy is thus given: Mahīdhara (a Māntrika of Kasīpati and author of mantra-mahodadhi)->Kalyāņa-> Lakşmaņa->Vīreśvara->Veņīdatta. Veņīdatta also wrote an independent work on Poetics called Alamkāra-candro- daya. See under Minor Writers below. MSS: Aufrecht i. 494b, ii. 115b, 220a, iii. 106a ; IOC iii, no 1216/1703a p. 354 (extract) ; Madras Cat. xxii, no. 12932. (3) °Setu or °Setu-bandha by Jīvarāja, son of Vrajarāja Dīkșita (see bibliography above under Rasa-mañjurī). MSS: Aufrecht i. 494b, ii. 220a, iii. 106a. The Rasa-setu in Aufre- cht i. 494b is probably the same commentary. He was the great-grandson of Sāmarāja Dīkșita (q. v. under Minor Writers), who lived in the latter part of the 17th century. He speaks of Gāngārāma's comm. Naukā with contempt (introd. verse 9). Extract in Ulwar Cat. no. 226. (4) Rasodadhi by Gaņeśa. MS dated 1698 A.D. Bühler, Cat. Gujarat, Kathiawad etc. 3. 54. (5) Rasodadhi by Mahādeva. Kielhorn, Central Prov. Cat. 104. (6) Sāhitya-sudhā or Kāvya-sudhā by Nemiśāha, son of Bhīmaśāha, described as mahārājādhirāja. Aufrecht i. 494b, iii. 106a. See Cat. BORI M S xii, pp. 234-35. P. K. Gode (Cal. Orient. Journal. i, pp. 217-20) would identify the author with Nemi Shah II of Jawhar line of chiefs in Bombay Presidency-about 1650 A D. (7) Nūtana-tarī by Bhagavadbhațța. Aufrecht i. 494b. (8) Comm. by Ayodhyāprasāda. Aufrecht i, 494b. The author also commented on the Vrtta-ratnakara. (9) Comm. by Dinakara. Aufrecht ii. 115b. Possibly this is the comm. ascribed to Nemisaha, mentioned above, who was Dinakara's patron, as the nominal author. Alamkāra-tilaka Ed. G. V. Devasthali in JBRAS, New Series, xxiii pp. 57-82, xxiv-xxv pp. 92-120 (1947-49). The name of the author is given as Bhanukara in Burnell 54a and Bhau Daji Catalogue. The work is in five paricchedas (Peterson vi, App. p. 29).

252 SANSKRIT POETICS

Śrngāra-dīpikā Aufrecht i. 661a (=Oudh Cat. iii. 12).

( 4 ) Following upon these, we have innumerable works of a similar nature which take Rasa, especially Sṛňgāra, as its principal theme, and which were composed apparently with the purpose of guiding the poet with rules and illustrations in the composition of erotic pieces so popular and profuse in Sanskrit literature. The most important of these works and their authors will be mentioned in the chapter on Minor Writers below. Some Vaisnava authors, like Rūpa Gosvāmin, however, attempt to bring Vaisnava ideas to bear upon the general theme of poetic or dramatic Rasa. We shall, for conve- nience of treatment, take them in a group here.

RŪPA AND JĪVA GOSV ĀMIN Rūpa Gosvāmin, son of Kumāra and grandson of Mukun- da, is the author of numerous Vaisnava works, including the Bhakti-rasāmta-sindhu and Ujjvala-nīlamani, which give an analysis and exposition of Bhakti-rasa on the analogy of the orthodox Rasa, especially the Śrngāra or Erotic. He is well known as a contemporary of Caitanya, the Vaisnava reformer of Bengal, and must have, therefore, flourished towards the end of the 15th and the first half of the 16th century. This date is confirmed by the dates which some of his works bear. His Dāna-keli-kaumudī and Vidagdha-mādhava were composed respectively in 1495 and 1533 A.D.1; while his Lalita-mādhava, Bhakti-rasāmrta and Utkalikā-vallarī are dated respectively in 1537, 1541 and 1550 A.D. The most flourishing period of Rupa's literary activity thus falls between 1533 and 1550 A.D., but it must have begun as early as 1495 A.D.

1 For information about these authors see S. K. De, Vaisnava Faith and Movement in Bengal, Calcutta 1942.

RŪPA AND JIVA GOSVĀMIN 253

The commentary Lacana-rocanī on Ujjvala-nīlamaņi was composed by his nephew Jīva Gosvāmin, son of his younger brother Vallabha (alias Anupama).1 His Mādhava-mahotsava was composed in 1555 A.D. and his Gopāla-campū bears the dates 1589 and 1592 A.D. Rūpa Gosvāmin also appears to have written a treatise on Dramaturgy called Nātaka-candrikā, which is quoted in the commentary on the Vidagdha-madhava and in the Vaisnavatoşinī on the Bhāgavata. At the commencement of this work, Rūpa Gosvamin states that in composing it he consulted the Bharata-śāstra and Rasa-sudhākara (Rasārņava-sudhākara of Singa-bhūpāla), and rejected generally (i. 2) the treatment of the Sāhitya-darpana as- being opposed to Bharata's views. The topics dealt with in eight sections are: (1) general characteristics of a drama, (2) the hero (nāyaka), (3) the divisions of a Rūpaka (nāndī etc), (4) elements in the action (samdhi, patākā etc) and their classification, (5) arthopaksepaka and its divisions (vis- kambhaka etc), (6) division of acts and scenes, (7) distribution of dialects (bhāșā-vidhāna), (8) styles of dramatic composition (vrtti) and their employment according to the Rasa intended. The work is not a small one, and the illustrations taken mostly from Vaisnava works are fairly minute and numerous. In his Ujjvala-nīlamaņi Rūpa gives illustrative verses from most of his own poetical and dramatic productions, such as Uddhava-dūta, Vidagdha-mādhava, Dānakeli-kaumudī and other works2. A Rasāmrta-śeșa is also attributed to Rūpa.

1 On the question of date see S. K. De, Padyāvalī (ed. Dacca Univ. 1934) pp. li-liii. A list of Rūpa's works will be found on pp. xlix-1. 2 For a list of his works see also S. K. De, Vaisnava Faith and Movement pp. 113-118 ; also analysis of Bhakti-rasāmrta and Ujjvala- nīlamani at pp. 126-167. This work may also be consulted for the novel application of the orthodox Rasa-theory to the doctrine of Bhakti and for an exposition of the erotic mysticism in these Vaisnava works, which expresses religious longings in the language and imagery of earthly passion. The question is, therefore, not discussed here.

254 SANSKRIT POETICS

VIŚV ANĀTHA CAKRAV ARTIN Viśvanatha Cakravartin, who wrote the commentary, called Ananda-candrikā or Ujjvala-nilamaņī-kiraņa, lived at the end of the 17th and beginning of the 18th century; for he composed a commentary called Sārārtha-darsinī on the Bhāgavata in Saka 1626=1704 A.D., while his Ananda- candrikā is also dated in Śaka 1618=1696 A.D. He also wrote, besides several Vaisnava works, a commentary on the Alamkāra-kaustubha of Kavikarnapūra, mentioned below. He mentions a work called Rasa-tarangiņī by Nārayāņa Bhatta in his commentary on Ujjvala-nīlamani (ed. NSP, p. 25).

KAVIKARN APŪRA Paramānanda-dāsa Sena Kavikarņapūra described as the youngest son of Sivananda Sena and pupil of Śrīnātha, wrote a work called Alamkāra-kaustubha. The author was a well- known Vaisnava of Bengal belonging to the Vaidya family, and wrote several Sanskrit Vaisnava works, including a metrical life of Caitanya (Caitanya-caritāmrta ed. Radharaman Press, Murshidabad 1884) and a drama on the same theme (Caitanya- candrodaya, ed. Bibl. Ind. 1854). His drama Caitanya- candrodaya was composed in Saka 1494 or 1501=1572 or 1579 A.D.1, and his Gaurānga-ganoddeśa-dīpikā was composed in 1576 A.D.2 His father Sivänanda was an elder disciple of Caitanyadeva, who used to organise and lead the annual

1 The date is given in the verse: śāke caturdasa-śate ravi-vāji-yukte/, gauro harir dharani-mandala āvirāsīt| tasmin catur-navati-bhāji tadīya- Tīlā-/grantho'yam āvirabhavat katamasya vaktrāt, which tells us that Gaurahari or Caitanya was born in Saka 1407, and that the drama, which deals with him, was written in Saka 1494 or 1501=1572 or 1579 A.D. See discussion about the date in Vaisnava Faith mentioned above. Aufrecht is incorrect in stating that it was composed in 1543 A.D. (But see Sten Konow Ind. Drama p. 93, section 104). For the author and his works see S. K. De, Padyāvali (ed. Dacca Univ. 1934), pp. 188-90 and Vaisnava Faith and Movement pp. 32-34. 2 HPS ii p. 50, as well as ALeip 721, reads śāke vasu-graha-mite which gives 1576 or 1577 A.D. but IOC no. 2510 reads sāke rasārasa- mite which would give 1540 A.D.

KAVICANDRA 255

exodus of Caitanya's followers to Puri from Bengal. Mitra in the introduction to his edition of the drama (p. vi) says that Kavikarņapūra was born in 1524 A.D. at Kāñcanapallī (Kāñcdāpādā) in Nadiya a few years before Caitanya's death. The Alamkāra-kaustubha is composed in ten kiraņas, as follow: 1 Kāvya-laksaņa, 2 Śabdārtha, 3 Dhvani, 4 Guņībhūta-vyangya, 5 Rasa-bhāva-tadbheda, 6 Guņa, 7 Śabdālaņkāra, 8 Arthālamkāra, 9 Rīti, 10 Doșa1. It is thus a more comprehensive work than Rūpa Gosvämin's and the Vaișnavite proclivities are not so prominent ; but most of the illustrative verses are in praise of Krsna and the subject-matter follows the Kāvya-prakāśa in treatment. The commentaries on this work are noticed below.

KAVICANDRA Kavicandra describes himself2 as a physician and as son of Kavikarņapūra and Kauśalyā, grandson of Vidyāviśārada, and father of Kavibhūşana and Kavivallabha belonging to the Datta family of Dīrghānka-grāma. He wrote a Cikitsā- ratnāvali3 in Śaka 1583 (=1661 A.D.) where the same personal history is given. The date precludes his identification with the poet of the same name quoted in the anthology called Padyavali4. He wrote, among other works, a Kāvya-candrikā in 15 prakāśas dealing with (1) kāvya-lakşaņa, (2) śabda-śakti, (3) rasa, (4) bhāva, (5) rasa-bheda, (6) rasābhāsa, (7) kāvya-bheda, (8) pramāņa-

1 For a detailed résumé of its contents see Mitra 1662. 2 IOC, iii, pp. 344-45; ABod pp. 211-12. 3 IOC, v, pp. 958-59. Aufrecht (ii. 166) mentions the date with a query, but the date appears to be correct. 4 See ed. S. K. De, Dacca Univ. 1934, nos. 162, 166, 188, 189, 190-191. See also IOC vii, p. 1534, at p. 1535. The verses quoted from Kavicandra's work in ABod 212a gives the same account of the author, and the colophon says: iti dirghanka-grāma-nivāsi-dattakulodbhava- vaidya-śrī-kavicandra-viracitāyām etc. It is, therefore, not possible that Paramānanda Sena Kavikarņapūra, author of the Caitanya-candrodaya, is identical with Kavikarnapūra, father of Kavicandra.

256 SANSKRIT POETICS

nirūpaņa, (9) rīti, (10) guņa, (11) śabdālaņkāra, (12) arthā- laņkāra, (13) doșa, (14) kavitopāya, and (15) nāțya1. He quotes, besides older authors, the Kavi-kalpalatā, Sāhitya- darpaņa, Rāmacandra-campū, Ratnāvalī-kāvya, Sānti-candrikā, Stavāvalī and an author called Purusottama, as well as two of his own works called Sāra-laharī and Dhātu-candrikā. His date would be latter half of the 17th century.

BIBLIOGRAPHY Ujjvala-nīlamaņi Editions. (1) with Locana-rocanī of Jīva Gosvāmin, Murshidabad 1889, 1917 in Bengali characters (2) with the same and with the comm. Ananda-candrikā of Viśvanātha Cakravartin, ed. Kedarnath and V. L. Panshikar, Nir. Sag. Press, Bombay 1913 (our references are to this edition). Commentaries. (1) Locana-rocanī by Jīva Gosvāmin, the name often confused with that of Sanātana Gosvāmin in the Catalogues. Published with the text, as noted above. (2) Ananda-candrikā (also °Kirana which is a summary) by Viśvanātha Cakravartin. Aufrecht i. 62a. Published in NSP ed. of the text, as noted above. The comm. Kirana-leśa in Mitra 580 (also in SCC vii. 5; Kathvate Rep. 1891-95, 318) is probably this commentary. (3) Agama-candrika and Ātma-prabodhikā. Aufrecht i. 62a. (4) An anonymous °Țīkā in SCC vii. 3. Nāțaka-candrikā Edition. Rasavihari Samkhyatirtha, with a Bengali trans. (in Bengali characters), Kashimbazar 1907. MSS: Aufrecht i. 284b, ii. 61b, 207b. Extract in Ulwar Cat. 1061 and Mitra 3160 ; Madras Cat. xxii. 12900.

Rasāmṛta-śeșa Aufrecht ii. 220b. Also called Bhakti-rasāmrta-śeşa. Ed. 1 Aufrecht's Bodleian MS contains eight prakāsas named after the first eight topics given here; but our enumeration follows the India Office MS. Cf. Regnaud p. 377.

WRITERS ON KAVI-SIKȘĀ 257

Haridas Das, Haribole Kutir, Navadvipa 1941 in Bengali characters.

Alamkāra-kaustubha Editions. With the commentary of Viśvanātha Cakravartin, Murshidabad 1899, in Bengali characters. Also ed. Sivaprasad Bhattacharya, with an anonymous comm. and a gloss, Varendra Research Society, Rajshahi, vol. i (ch. i-v), 1923 in Devanāgarī characters. Commentaries. (1) Sāra-bodhinī by Viśvanātha Cakra- vartin. Aufrecht iii. 7b. Published with the text, as noted above. (2) Țippaņī by Sārvabhauma, pupil of Cakravartin, Dacca Univ. MS nos. 2363, 2394, 3471. (3) ° Dīdhiti prakāśikā by Vrndāvanacandra Tarkālamkāra Cakravartin, son of Radhacarana Kavīndra Cakravartin. IOC iii, 1195/240, p. 344. (4) °Țīkā by Lokanātha Cakravartin. Aufrecht i. 31b. This is printed anonymously in the Varendra Research Soc. ed. mentioned above as an 'old commentary'.

Kāvya-candrikā MSS: Aufrecht i. 101a; KBod 499. The Bodleian MS appears to contain 16 Prakāśas.

WRITERS ON KAVI-SIKȘĀ ARISIMHA, AMARACANDRA AND DEVEŚVARA

( 1 ) The Śvetāmbara Jainas, Arisimha and Amaracandra, are notable in Sanskrit Poetics for a work on the composition of verses called the Kavitā-rahasya or Kāvya-kalpalatā and its commentary entitled Kavisikşā-vrtti. This work was com- posed in part by Arisimha and completed by Amaracandra1,

1 kimcic ca tad-racitam ātmakṛtam ca kimcit/ vyākhyāsyate tvarita- kāvya-krte'tra sūtram, says the Vrtti. 17

258 SANSKRIT POETICS

who also wrote the commentary1. Arisimha, described as son of Lāvanyasimha or Lavanasimha, wrote also a poem called Sukrta-samkīrtana (ed. Bhavnagar 1917) in honour of his patron Vastupāla (d. 1242 A.D.), the Jaina minister of the Dholkā Rāņā Vīradhavala ; and he also lived in the time of Vīradhavala's son Vīsaladeva2 (1243-66 A,D.). Amaracandra appears to be a more prolific writer, being also the author of the Jinendra-carita (otherwise called Padmānanda-kāvya)3, the Bāla-bhārata4 and a grammatical work called Syādiśabda- samuccaya (Aufrecht i. 180).5 Rājaśekhara Sūri, the Jaina author of the Prabandha-kośa (p. 61, ed. Singhi Jaina Granthamālā), also attributes a Sūktāvali and a Kalā-kalāpa. In the Vrtti to the Kāvya-kalpalatā itself, three other works by Amaracandra are cited, viz. the Chandoratnāvalī, a work on Prosody (p. 6), Kāvya-kalpalatā-parimala6 (pp. 19, 63), probably an epitome or continuation of the work of that name, and an independent work on Poetics called Alamkāra- prabodha (p. 117). Amaracandra was a pupil of Jinadatta Sūri of the Väyada-gaccha, who is identified with the author of the Viveka-vilāsa and who lived about the first half of the 13th century8. Amaracandra appears to have been a pupil or a 1 The colophon calls the Vrtti kavi-sikşā-vrtti. Bühler thinks, from i. 1 and 2, that the original Kārika-verses of Arisimha were called Kavitā-rahasya, while Amaracandra's Vrtti was entitled Kāvya-kalpalatā. 2 See the question discussed in detail in Bühler, Das Sukrtasamkir- tana des Arisimha, Wien 1889, pp. 5f, 38 ; also IA, vi 210-12. Amara- candra is said to have added a postscript or Sargānta-śloka to the Sukrta-samkīrtana. It is a Kāvya in 11 cantos (ed. Jaina Ātmananda Sabhā, Bhavnagar 1917). See IA, xxxi, pp. 477-95. 3 Ed. H. R. Kapadia, GOS, Baroda 1932. 4 Ed. in Kāvyamālā 45, Nir. Sag. Press, Bombay 1894 ; also ed. in the Pandit iv-vi, (1869-72). 5 Ed. Chandraprabha Press, Benares 1915. 6 Cf. IOC 848, which has a reference to Kāvya-kalpalatā-mañjarī. 7 See the concluding verse of the Bala-bhārata and the colophon to the Kāvya-kalpalatā. 8 See Peterson i, p. 58-59 ; but Peterson's conjecture about Jinadatta

WRITERS ON KAVI-SIKȘĀ 259

fellow-student of Arisimha1 and lived, according to the account given in Jaina Rājasekhara's work2, in time of Vīradhavala and his minister Vastupāla, as well as in that of Vīsaladeva, before the latter's accession to the throne of Anhilvād, i.e. about the second quarter or the middle of the 13th century. The Kāvya kalpalatā consists of four Pratānas, namely, I. Chandah-siddhi, II. Śabda-siddhi, III. Śleşa-siddhi, and IV. Artha-siddhi. For a full account see Aufrecht, Bodleian Cat. no. 497 and IOC iii, no 1183/848, pp. 340-41.

( 2 ) The date of Deveśvara, author of the Kavi-kalpalatā, can be approximately settled from the internal evidence supplied by the work itself ; for it closely follows, in its treatment of the subject and general arrangement, Arisimha and Amara- candra's work ; and it is not difficult to show that he also copies wholesale from the text of his predecessors. He borrows literally most of the rules and definitions, and even repeats the illustrative stanzas. Thus, Deveśvara pp. 157-60 (veņyāḥ sarpāsi°) and pp. 36-7 (ratnādi yatra)=Arisiņha pp. 135-37 and pp. 30-1 ; the definition of adbhuta-vidhi in Arisimha p. 93=Deveśvara p. 130. This copying is not sporadic but systematic, and is found practically throughout the work, so that it is highly probable that Deveśvara had the text of the Kāvya-kalpalatā before him, when he composed

is not right; see also Merutunga's Prabandha-cintāmaņi, p. 258 ; Peterson iv, pp. viii, xxxvi and App. 115; Bhandarkar Rep. 1883-84, pp. 6, 156; Bühler op. cit. pp. 25, 48. Jinadatta Sūri's Viveka-vilāsa is dated 1220 A.D. 1 Rājaśekhara Sūri says that Amara was a pupil of Arisimha. Cf. Sridhar Bhandarkar, Rep. 1904-6. pp. 23-24; Bühler op. cit. pp. 5-6 .: contra Bhandarkar .- One of the sketches of Rājasekhara's Prabandha- kośa deals with Amaracandra. 2 His work is dated in 1348 A.D. See introd. to Amara's Bāla- bhārata, ed. Kāvyamālā. Rājaśekhara Sūri was a pupil of Tilaka Sūri of Koțika Gaņa (Peterson iv, p. cv).

260 SANSKRIT POETICS

his own work on the same subject. This gives us one terminus to his date at the middle of the 13th century. The other terminus is given by the fact that a verse of Deveśvara's (nāga-viśeșe śeşe p. 155) is quoted under his name in the Paddhati of Sārngadhara (545, deveśvarasya). As this anthology was compiled about 1363 A.D., we get the middle of the 14th century as the other terminus. Allowing half a century to elapse between Deveśvara and Arisimha, on the one hand, and a similar period of time between Deveśvara and the compiler of the Paddhati, on the other, we may roughly fix the beginning of the 14th century as his approximate date. Deveśvara describes himself as son of Vāgbhata, who was a mahāmātya to some prince of Mālava (?) ; and in one of the samasyā-ślokas, there is a panegyric of Hammīra- mahīmahendra, who is apparently the Cauhan prince of that name who reigned about 1283-1301 A.D.1

( 3 ) Two other Kavi-śiksā works by Jaina authors are known, namely, by Ācārya Vinayacandra (about 1250 A.D.) and by Jaymangala who wrote in the times of Jayasimha Siddharāja (1094-1143 A. D.). A large part of Rājaśekhara's Kāvya- mīmamsa is concerned with similar topics. There is another work on Kavi-śiksa by Gangādāsa, the well-known author of Chandomañjarī. It appears to be an elementary treatise on miscellaneous topics, such as Chandaḥ-kathana, Sāmānya-

1 See S.K. De in JRAS 1922, pp. 577f on the date of Deveśvara. In the footnote there, omit the words "from the author's own comment on the word." Deveśvara also refers to another work of his, called Candra-kalāpa (matkrta-candrakalāpe'mala-matibhis tad budhair jñeyam, ŚgŚ ii, p. 225; also in Bibl. Ind. ed.). The variant in the Calcutta edition (1900) reads (p. 42) instead matkrta-kavikalpalatā- parimalatas°, while an alternative reading noticed in the Bibl. Indica ed. (p. 52) is matkta-kavi-kalpalatāyām amala-matibhiḥ.

WRITERS ON KAVI-SIKȘĀ 261

śabda, Rasa, Guņa, Sabdālaņkāra, Kāvya-doșa and Samasyā- pūraņa (see IHQ xxiv, p. 315-16). The Kavi-śikşā of Jaya- mańgala and Kavi-kalpalatā of Rāghava-caitanya will be noticed below in the chapter on Minor Writers. All these manuals are directed to the practical object of Kavi-śikșā or instruction of poets in the composition of their works. They have little direct relation to Poetics proper.

BIBLIOGRAPHY Kāvya-kalpalatā and its Vrtti Kavi-śikşā Editions. (1) Rama Sastri, Benares 1886. (2) Vamana Sastri, Bombay 1891. (3) ed. Jagannath Sastri Hoshing, Chow- khamba Skt. Ser. Benares 1931 (our references are to this ed.). Commentary. °Makaranda by Šubhavijaya Gaņi, pupil of Hīravijaya Sūri of Tapā-gaccha who lived in the reign of 'Akabbar Sāhi'. Aufrecht i. 101a, iii. 22b ; KBod 497. Our author lived in the reign of Salem or Jahangir (śrīmat- salem-śāhi-rajye) and wrote the commentary in Samvat 1665=1608-9 A.D. at the request of Vijayadeva Sūri (Peterson vi, p. 25f).

Kavi-kalpalatā Editions. (1) with Comm. by Vecārāma Sārvabhauma, in the Hindu Commentator, vols. 1-3, Benares 1867-70. (2) with Comm. by Ramgopal Kaviratna 1900 (our references are to this edition). (3) together with his own comm. by Saraccandra Sastri, Bibl. Indica, Calcutta 1913. (4) in Pratna-krama-nandinī, Benares nos. 1-31. The work is divided into four Stabakas. Commentaries. (1) Bāla-bodhikā by Sūrya Kavi, also known as Sūrya-dāsa or Sūrya Sūri. His genealogy is given thus: Rāma of Pārthapura (under Rāma, king of Devagiri)-> Vişņu-> Nīlkaņțha-> Nāganātha-> Nrsimha-> Nāganātha ->Jñānarāja (author of Siddhānta-sundara)-> Sūrya (Weber i, p. 231). He was a versatile author (for his works see

262 SANSKRIT POETICS

Aufrecht i. 731b, ii. 175b). His Līlāvatī-țīkā was com- posed in 1542, while Sūrya-prakāśa on Bhāskara's Bīja- ganita is dated in 1539 A,D. He also wrote an artificial poem called Rāma-krsņa-viloma-kāvya (ed. in Haeberlin's Kāvya-saņgraha, and Kāvyamālā Gucchaka xi, p. 147f), which praises in alternate verses Rāma and Krsna, and gives the same text to be read forward or backward respectively. He also wrote a Nrsimha-campū in five Ucchvāsas (IOC vii, p. 1548 ; see also ibid, p. 1478). He belonged to the Bharadvāja-gotra and lived in Pārtha- pura near the confluence of the Godāvarī and Vidarbhā. Aufrecht i. 87a ; iii. 19a. (2) Comm. by Vecārāma. Probably the same work as printed in the Benares ed. Aufrecht ii. 16b ; the name is given as Vecārāma Sārvabhauma in Jammu MS no. 3482 (Jammu Cat. p. 59). See under Minor Writers below. (3) anonymous °Țīkā in SCC vii. 8. (4) Padārtha-dyotanikā by Mahādeva, son of Pațta- vardhana Mudgala. H. P. Sastri, Cat. ASB MSS, vi, MSS no. 4.99/10004, 4800/8999, pp. 393-94.

CHAPTER X

MINOR WRITERS ON ALAMKĀRA ( 1 ) We propose in this chapter to deal with the minor writers on Alamkära, other than the commentators as well as less known authors already mentioned in the preceding chapters, arranging them alphabetically, and giving the names of their works and such details as can be collected about them1.

1. AKABARA ŚĀHA, alias Bade Sāheb Śngāra-mañjarī (Ed. V. Raghavan, Hyderabad Arch. Dept, 1951) The author is described as son of Shaha Raja, and Guru of Sultan Abul Hasan Qutb Shah of Golconda (1672-87 A,D.), who was captured by Aurangzeb in 1687 and who died in 1704 A.D. Akbar was born about 1646 and died between 1672 and 1675 A.D. The Sanskrit work mentioned is said to be a translation of a Telugu original by Akbar Shah himself. Possibly at the instance of Akbar some Telugu scholar wrote the original Srngāra-mañjarī in Telugu, and a Sanskrit scholar translated it (V. Raghavan, introd. p. 7). The work is based on Bhanudatta's Rasa-mañjarī, and is concerned mainly with the theme of Nāyaka-Nāyika, topically dealing with Śrngāra- rasa. Total number of verses 312. 2. ACYUTA ŚARMAN or ACYUTARĀYA MOŅAKA Sāhitya-sāra and its commentary Sarasāmoda (Ed. in litho MS form, Bombay 1860; ed. W. L. Panshikar, NSP, Bombay 1906) The work is dated in Saka 1753=1831 A.D. The author describes himself as pupil of Sasti Nārāyana. He was son of 1 Doubtful names to be found in some catalogues, as well as those about which there is no reliable information, have been omitted in this list ; and care has been taken to avoid useless or trifling entries:

264 SANSKRIT POETICS

Nārāyaņa and Annapūrņā and lived at Pañcavatī near Nasik. He appears to be identical with Acyutaraya who wrote the Praņaya-prąkāśa commentary on Jagannātha's Bhāmiņī-vilāsa; for in it he refers to the Sahitya-sara as his own (ed. N. S. P. 1894, p. 1), citing the verses i. 14-15 from the latter. In his Sāhitya-sāra, again, he refers to the Bhāminī-vilāsa at p. 7. The work consists of twelve chapters called Ratnas, the metaphor being that these precious doctrines are churned by the author from the ocean of Alamkara-sastra. The chapters are accordingly named: 1. Dhanvantari-ratna (the general characteristics of kāvya), 2. Airāvata-ratna (function of śabda and artha), 3. Indirā-ratna (the vyangya and its occasions), 4. Dakşiņāvarta-kambu-ratna (divisions of dhvani, including rasa-dhvani), 5. Aśvavara-ratna (other inferior divi- sions of dhvani), 6. Vișa-ratna (dosa), 7. Guņa-ratna (guna), 8. Kaustubha-ratna (arthālamkāras), 9. Kāmadhenu-ratna (śabdālamkāras), 10. Rambhā-ratna (nāyikā), 11, Candra-ratna (nāyaka) and 12. Amrta-ratna (conclusion). Quite a recent writer, who holds some novel views which are neither historically nor theoretically sound. He is identical with Acyuta, author of the Bhāgīrathī-champū (written in the 1814 A.D.), who is described by Aufrecht (i. 770b) as son of Narayana. This work is divided into seven chapters called manorathas. For the author and his other works, some of which are dated, see New Cat. Cat. i, p. 59-60.

3. AJITASENĀCĀRYA or AJITASENA-DEVA YATĪŚVARA a. Alamkāra-cintāmaņi (Ed. by Padmaraja Pandit in the Kāvyāmbudhi 1893-94 ; see Ind. Office Printed Bks., 1938, p. 72) b. Šrňgāra-mañjarī (ŠgŚ ii, pp. 83, 231 extract; Madras Cat. xxii 12956-57) The author was a Digambara Jaina priest of Cāmuņda- rāya, minister of the Ganga king Räcamalla, and flourished in

MINOR WRITERS ON ALAŅKĀRA 265

the latter part of the 10th century. He was a teacher of Nägavarman, a Kanarese poet, who lived under the protection of Rakkasa Ganga, younger brother of Räcamalla. His works were written in the Sāntīśvara temple at Bangavādipura. He wrote the Alamkāra-cintāmaņi in five chapters. Ajitasena also wrote Cintāmaņi-prakāśikā on Yakşavarman's Cintāmaņi, which is itself a commentary on Sākațāyana's Šabdānuśāsana. See Rice p. 308. The Ajitasena who was the author of the Kālidāsa apocrypha Śruta-bodha on Metrics is probably a different person. There is another Ajitasena of the Senagana who wrote the Śrngāra-mañjarī at the instance of a Jaina Alūpa prince of the lunar race1, named Rāya or Kāmirāya, son of Vitthala-devī, for his instruction. It consists of three chapters and 128 stanzas, dealing with (1) pada-dosa (viz. alakşaņa, śruti-kațu, vyāghātārtha, anarthaka, aprasiddha, neyārtha, grāmya, and asammata), which ends with a discussion of the Vrttis, (2) the ten guņas of Vāmana and (3) arthālamkāras (viz. upamā, rūpaka, jāti, bhrāntimat, hetu, samśaya. prativastūpamā, ākșepa, drstānta and tulyayogitā). On the two Ajitasenas see New Cat. Cat. i, p. 69 which, however, distinguishes all the known Ajitasenas. The entry of Alamkāra-cintāmaņi as a work of Śāntarāja is some MS Catalogue is a mistake, for Śāntarāja was the scribe and not the author of the work.

4. AŅURATNAMAŅDANA or RATNAMAŅDANA GAŅI a. Jalpa-kalpalatā (Weber 1722, long extract given ii, pp. 278-80) This Jaina author was a pupil of Ratnaśekhara Sūri2 of Tapa-gaccha, who died in Samvat 1517=1460-61 A.D.3 1 Among the Alūpa princes matriarchy prevailed. Bangavādi was the capital perhaps of a branch of the Ālūpas. 2 For Ratnaśekhara, see Bhandarkar Rep. 1883-84, pp. 156-7: Peterson iv, p. cii f ; IA xi, p. 256. 3 Ratnaśekhara Sūri composed his Kriyā-ratna-samuccaya (ed. Jaina Yośovijaya Granthamālā Series) in Samvat 1466=1410 A.D.

266 SANSKRIT POETICS

Anuratna, therefore, belonged roughly to the middle of the 15th century. The work, in three stabakas, gives practical instruction on poetical composition (kavi-śikşā). b. Mugdha-medhākara Alamkāra-vrtti (Peterson vi, p. xv, extract given at p. 31 ; also BORI Cat. xii, p. 222-23) A manual on poetic figures and kindred topics.

5. ANANTA Sāhitya-kalpavallī (Madras Trm. Cat. no. 5483) The author was of Tirumala family and Sathamarsana- gotra and son of Tocamamba. He was a protégé of the Orissa king Gajapati Purusottama-deva. The work is fully called Gajapati-Puruşottamadeva-Sāhityakalpavallī

6. ANANTĀRYA or ANANTĀCĀRYA (Anantālvan) Kavi-samaya-kallola (Madras Cat. xxii 12808, extract) This recent South Indian writer of the Sesācārya family was son of Singarācārya. His family was resident at Yāda- vagiri or Melkote in Mysore. The work cites from Dharma- sūri, Narasiņha's Nañjarāja-yaśo-bhūșaņa, and Pratāpa- rudrīya. He flourished in the court of Krsnaraja Wodeyar III. His dates are 1822-62 A.D. He refers to a Kāvya by himself, called Krsnarāja-yaśo-diņdima. He was a Viśistādvaitin and wrote a large number of Vädas which have been published in Vedanta-vādavalī, Bangalore 1898 etc. (See New Cat. Cat. i, p. 143).

7. AMRTĀNANDA YOGIN Alamkāra-samgraha (ed. Calcutta 1887 with Engl, trans. ; ed. also in the Adyar Library Series, and by the Sri Venkatesvara Oriental Institute, Tirupati. MS in Madras Cat. xxii, 12794, extract) The work consists of five chapters dealing with (1) varna- gaņa, (2) śabdārtha, (3) rasa-bhāva, (4) nāyaka-bheda, and (5)

MINOR WRITERS ON ALAMKĀRA 267

alamkāra. The author says that he wrote this work at the request of Manvasamudra, son of Bhakti-bhūmipati, who was a devotee of Siva. It appears from the introductory verse that the author also intended to deal with guna, dosa and daśa-rūpaka. He is not identical with the Tāntrika Amrtānandanātha (pupil of Puņyananda) who is said to have corrected (Weber p. 361) the Tantra-sāra of Krsņānanda ; but in the introd. to his Yoginī-hrdaya-dīpikā (ed. Sarasvati Bhavan Text, no. 7) it is said that this Amrtananda could not have done so because Krsnānanda was much later in date.

8. ARUŅAGIRI KAVI Godavarma yaśo-bhūşaņa (Printed in Journal of Travancore Univ. MS Library, vol. i) The work is on Arthālamkāras only. The author is des- cribed as belonging to the Kaundinya-gotra, son of Seşādri and pupil of Venkatādri. He was patronised by Godavarma, king of Vadakkumkūr. Circa 1550-1650 A.D.

9. ALLARĀJA or MALLARĀJA

Rasa-ratna-pradīpikā (Ed. R. N. Dandekar, Bharatiya Vidya Series 8. Bombay 1945 ; the work called is Rasa-ratna-pradīpikā) A work called Rasa-ratna-dīpikā is cited by Bhānudatta in his Rasa-tarangiņī (=Allarāja v. 57), and by Ratnakaņtha on Mammața (Peterson ii, p. 17). See above p. 242, fn 4. The work is called Rasa-ratna-pradīpa by Allarāja in Bhandarkar Rep. 1884-87, no. 533. The author was son of Hammira, the Cauhan king of Ranathambor who conquered Konkana and whose regnal period was 1283-1301 A.D.1 The work is a manual of six chapters (called Paricchedas) which deals in prose and verse exclusively with the topic of Rasa and Bhāva.

1 The question of Allarãja's date is discussed by the editor of the text. In honour of this Hammīra Nayacandra Sūri wrote his Hammīra- Mahākāvya in 1486 (ed. N. J. Kirtane, Bombay 1879).

268 SANSKRIT POETICS

It borrows verbatim extensively from Bharata and the Daśa- rūpaka. There is hardly any originality either in subject- matter or treatment.

10. ĀŚĀDHARA Kovidānanda (Aufrecht ii. 25a) Triveņika (ed. Batuknath Sarma and J. S. Hoshing, Sarasvati Bhavana Texts, Benares 1925). We have already mentioned above (p. 228) Āsādhara as the author of the Alamkāra-dīpikā commentary on Kuvalayā- nanda. He was son of Rāmaji and disciple of Dharaņīdhara, and should be distinguished from the much earlier Jaina Āādhara, son of Sallakșaņa and commentator on Rudrața (see p. 93). The Kovidānanda, with its commentary called Kādambinī, is concerned (as the author himself states) with Śabda-vyāpāra-nirņaya. His (Šabda-) Triveņikā 'having three streams' also deals with the same theme of the three Vrttis, namely, Abhidhā, Lakşaņā and Vyañjanā and refers frequently to his Kovidānanda. As Āśādhara comments on Appayya's work and quotes Bhattojī's Siddhānta-kaumudī he must be later than the first half of the 17th century. A MS of his Alamkāra-dīpikā is dated in Śaka 1775 (=1850 A.D.), while a MS of his Kovidānanda appears to have been copied in Śaka 1783 (=1861 A.D.)1 In all probability Āśādhara flourished in the middle and latter half of the 18th century.

11. INDRAJIT Rasika-priyā in 16 Pravāhas (Peterson vi, no. 379) A MS of this work in BORI MS Cat. xii, p. 293 is dated Samvat 1729 (=1672-73 A.D.). In two of the BORI MSS the author is called Mahārāja-kumāra in the colophon. This is not a Sanskrit but an old Hindi work. The author also wrote a Bāla-bodha commentary on the Vairāgya-śataka

1 R. G. Bhandarkar, List of Skt. MSS, pt. i, Bombay 1893, p. 68.

MINOR WRITERS ON ALAMKÄRA 269

(Aufrecht iii. 13b). Bühler mentions (ZDMG, xliii. p. 543) a work on Alamkāra called Rāmacandra-candrikā by Indrajila (sic), the date of which is given as 1712 A.D.

12. KACCHAPEŚVARA DIKȘITA

Rāmacandra- yaśo-bhūșaņa (Madras Cat. xxii, 12950, extract) The author was son of Väsudeva and grandson of Kālahastīśvara who was a native of Brahmadeśa, a village in the North Arcot district. He wrote also a commentary on the Bhagavata. His grandfather had two other sons, named Nārāyana and Krsna, his father being the second son. The work consists of three chapters dealing with the Rasas, viz. (1) Śrngāra, (2) the other 8 Rasas, (3) Bhāva-nirūpaņa. The illustrations are in praise of the valour of Bommarāja (probably of the Karvetnagar zamindary in North Arcot).

13. KANDĀLAYĀRYA

Alamkāra-śirobhūşaņa (Hultzsch i no, 371, extract at p. 75; Madras Trm A 168) The author was son of Rāmānujārya of the Kausika-gotra and grandson of Kesavārya of the Rāyalūri family. He tells us that he lived in the court of Venkata-bhūpati, son of Soma-bhūpati (and Giryambā), who was son of Nallareddi of Mustipallī (also called Pākanādu) family and of Mițilla- gotra. He resided in Nadigadda country which lies between the Tungabhadra and Krsna. But see New Cat. Cat. i, p. 297a. It seems that he probably lived at the court of either Venkata I or II of the third Vijayanagar dynasty1, and was thus a contemporary of Appayya Dīksita. The work is in ten Ullāsas as follow: (i) Upodghāta (ii) Kāvya-lakșaņa (iii) Dhvani-prakaraņa (iv) Rasa-prakaraņa (v) Doșa-praka-

1 Hultzsch ibid, p. viii. But see New Cat. Cat. i, p. 297a.

270 SANSKRIT POETICS

raņa (vi) Guņa-prakaraņa (vii-ix) Kāvyaviśeşa-prakaraņa (x) Nāyaka-prakaraņa.

14. KALYĀŅA-SUBRAHMAŅYA SŪRI

Alamkāra-kaustubha with commentary (Śgś ii, pp. 80, 220 extract ; Madras Cat. xxii, 12790) The author was son of Subrahmanya and grandson of Gopāla of the Perūru or Perur family. The work invokes and sings the praise throughout of Padmanäbha, the god of the temple of Anantaśayana (Travancore), and of the Vañji- pāla (Bāla-) Rāma Varma Kulaśekhara, king of Travancore (1758-98). It deals with Arthālamkāras only, gives the characteristics of those figures that are dealt with in the Candraloka, and illustrates them with examples composed by the author himself in praise of his patron and his deity.1

15. KĀĪ or KASHĪKARA LAKSMAŅA KAVI Alamkāra-grantha (Burnell 54a). This work was probably written at the end of the 17th or beginning of the 18th century, as the examples are all in praise of the Tanjore prince Sāhajī (1684-1711 A.D.). It is also called Sāharājīya. See Tanjore Descriptive Cat. ix, nos. 5304-05. Also see V. Raghavan's ed. of Sāhendra-vilāsa, Tanjore Sarasvati Mahal Series (p. 23).

16. KĀSĪŚVARA MIŚRA Rasa-mīmāmsā This work is mentioned and cited by Viśveśvara Kavi- candra (q.v.) in his Camatkāra-candrikā, ch. v: tathā coktam

1 On this author see K. Kunjunni Raja, Contribution of Kerala to Skt. Lit. (Madras 1958), p. 175. Bāla Rāma Varman appears also to be eulogised in the illustrative verses of Bālarāma-varma-yasobhūșana of Sadāśiva Dīkșita which includes (as some Yaśobhușaņa works do) a play in five Acts, called Vasu-lakşmī-kalyāna, of which the king is the hero. See Cat. Trivandrum Palace Lib. vi, p. 2354.

MINOR WRITERS ON ALAMKĀRA 271

samad-ācāryaiḥ kāśīśvara-miśraiḥ rasa-mīmāmsāyām. As Viśveśvara's teacher his date would be about 1300 A.D. See V. Raghavan in ABORI, xvi (1934-35), p. 139-40.

17. KUMĀRAGIRI

Vasanta-rājīya Nātya-śāstra The author is quoted by Katayavema as his patron1. The work is cited by Kumārasvāmin as vasanta-rājīya p. 178, by Mallinātha on Śiśu ii, 8 and by Sarvānanda on Amara-kośa2. These citations make it clear that this metrical work on Dramaturgy can not be later than the 14th century. This Nātya-śāstra, probably an Āndhra work, is also mentioned in a commentary on the Southern recension of the Sakuntalā by Kāțayavema who is described as a minister of king Vasanta- rāja Kumāragiri. This commentary proposes to follow the exposition of Vasantarāja's Nātya-śāstra (IOC vii, p. 157-76). The author's genealogy is given in a MS of the work3, which states that Vasantarāja Kumāragiri was son of Anapota and grandson of Vema Reddi. Kāțayavema was son of Kāțaya- bhūpati by his wife Woddāmba, who was the daughter of Vema Reddi. The Reddi king Kumāragiri ruled in the Telugu country in the second half of the 14th century4. As his work is lost, we know nothing about its scope and extent, but later citations show that it dealt chiefly with Dramaturgy and incidentally with Rasa.

18. KUMBHA or KUMBHAKARŅA, Śrīrājādhirāja Rasa-ratna-kośa (Aufrecht i. 495b) A MS of this work in Devanāgari characters in the Paris

1 Burnell 173a. 2 ŚgS ii, p. 30. 3 In Madras Trm I A 295 (6), but the stanzas are missing in Burnell and IOC MSS. 4 For a discussion of genealogy and time of this Reddi king (2nd half of the 14th century) see Introd, to Vanivilasa Press ed. (1906) of Pārvafi-pariņaya; N. Venkata Rao, Vasantarājīya in Pathak Comm. Vol., Poona 1934, pp. 401f.

272 SANSKRIT POETICS

Biblioth. Nationale (no. 243) is described by Regnaud p. 379. It is a treatise on Rasa and kindred topics in eleven chapters1, dealing with (1) 1-4 rasas,- (2) 5-6 nāyaka and nāyikā, (3) 7 abhinaya, (4) 8-9 anubhāva and vyabhicāri-bhāvas, (5) 10-11 rasa and bhava. The treatment and subject-matter correspond to those of the 3rd chapter of the Sāhitya-darpana and Bhanudatta's two works on Rasa. The author is probably king Kumbha of Mewad (1428-1459 A.D.) who wrote, besides some treatises on Samgīta, a commentary entitled Rasika- priyā (ed. NSP. 1917) on Jayadeva's Gita-govinda, and flourished in the first half of the 15th century.

19. KURAVIRĀMA

Daśarūpaka-paddhati See above pp. 127, 229 for information about this author.

20. KRSNA Sāhitya-tarangiņī (Aufrecht ii 171a)

21. KRSŅA DĪKȘITA or KRSNA YAJVAN Raghunātha-bhūpālīya (Aufrecht i. 446a ; Madras Trm C 656d ; Adyar II, p. 336) The work was written in honour of the author's patron, whose name it bears in its title, after the manner of the Pra- tāpa-rudra-yaśobhūșana of Vidyānatha who is referred to as Vidyāpati in the introductory part. Raghunātha, son of Acyuta, distinguished himself as one of the Nāyaka rulers of Tanjore (17th century A.D.) and patron of literature. His mistress Rudrāmba wrote the semi-historical poem, Raghunātha- bhyudaya,2 to celebrate in twelve cantos the greatness of her

1 The work appears from citations in his Rasika-priyā to be a part of the author's voluminous Samgita-raja ; see V. Raghavan in ABORI, xiv, 1933, pp. 258-62. Kumbha also wrote a comm. on the Samgita- ratnākara of Šārngadeva. 2 Ed. T. R. Chintamani, Madras Univ. 1934.

MINOR WRITERS ON ALAMKĀRA 273

lover. Krsna Dīksita's work consists of eight Vilasas dealing with (i) Nāyaka-guņa (ii) Kāvya-svarūpa (iii) Samlaksya- krama-vyangya (iv) Asamlaksya-krama-vyangya (v) Guņī- bhūta-vyańgya (vi) Śabdālamkāra (vii) Arthālamkāra and (viii) Guņa. Krsna Yajvan also appears to have written an Alamkāra-muktāvalī. A commentary called Sāhitya-sāmrājya by Sumatīndra Yati, pupil of Sudhīndra-pūjyapāda, is men- tioned in Rice 288 .- Another work of this type which praises Shahaji of Tanjore (1648-1710) is Sāharājīya of Lakşmaņa Kavi ; see above p. 270, no. 15.

22. KRSNA BHAȚȚA or JAYAKRSŅA MAUNIN

Vṛtti-dīpikā (Aufrecht i. 598a) The author was a grammarian, and the work probably dealt with the grammatico-rhetorical question of the Vrttis of words. For his other works, see Aufrecht i. 198a. He is described as son of Raghunätha Bhatta and grandson of Govardhana Bhatta.

23. KRSŅA ŚARMAN or KRSŅĀVADHŪTA

a. Mandāra-maranda-campū (ed. Sivadatta and K. P. Parab, Nir. Sag. Press Bombay 1895, with Mādhurya-rañjanī· com- mentary) b. Kāvya-lakşaņa c. Sārasvatālamkāra, Sūtra and Bhāșya The first-named work is a so-called Campu dealing in reality with Prosody, Dramaturgy and Poetics, as well as practical Kavi-iksā. The work consists of eleven chapters called bindus, dealing with (1) chandas (2) nāyaka-varņana (3) śleşa (4) yamaka and citra (5) the different bandhas (6) enigmatology (7) dramaturgy (8) nāyaka-laksaņa (9) bhāva and rasa (10) alamkāra with dhvani-nırūpaņa etc. and (11) doșa, which however includes sections on sabdärtha, the three vrttis etc., pāka, kāvya-bheda, and a section of practical hints for 18

274 SANSKRIT POETICS

descriptive poetry. The work appears to be ill-arranged and ill-digested, having no fixed theory but forming a cyclopaedic compilation from various sources, meant to serve as a com- plete handbook for the poet. The author, who is described as an inhabitant of Guhapura and a pupil of Vāsudeva Yogīśvara, gives us no clue to his date ; but his work belongs to quite recent times. He copies, for instance, many definitions and illustrations from Appayya's Kuvalayānanda, and even appro- priates the whole section on Pāka from Vidyānatha. Some of the new poetic figures, which appear to be first adduced and illustrated by Appayya, find a place in this comprehensive compilation. For the author and his work see Karnatak Univ. Journal (Humanities), 1957, pp. 127f. His date is given as 1835-1909 A.D. He also wrote a commentary Rasa-prakāśa on Mammata (see p. 173).

24. KRSŅA SUDHĪ Kāvya-kalānidhi The author was son of Sivarama and grandson of Upadestr- paņdita Nārāyana. He was a native of Uttaramerūr in Tondaimandalam on the banks of the Ceyyar near Kāñci. He wrote the Kavya-kalanidhi in 1845 A.D. under the patronage of Ravivarman, rāja of Kolattanād. The work is in ten sections in which the illustrations are all in praise of the poet's patron.1

25. KRSŅA SŪRI Alamkāra-mīmāmsā (Madras Cat. xxii, no .: 2700) The author was son of Gopālācarya, who was son of Krsnārya of the Sāntalūri family. There is one Krsņa Sūri who wrote a commentary called Ratna-śobhākara on the Alamkāra-muktāvalī of Ramasudhī, son of Nrsimha (ed. in Telugu script, Vizagapatam 1897-98).

1 See K. Kunjunni Rāja, Contribution of Kerala to Skt. Lit. pp. 62, 244.

MINOR WRITERS ON ALAMKĀRA 275

26. KEŚAVA BHATȚA

Rasika-samjīvanī (Aufrecht i. 127b, 497b ; Br. Mus. no. 424, extract)

The work, dealing with Rasa is in three vilasas. The author, son of Harivamsa Bhatta, is claimed as a Vaișnava disciple of Vitthaleśvara, who is apparently the son of the reformer Vallabhacarya. Hence our author probably belongs to the second half of the 16th century. He is not the same as Keśava Bhatta Kāśmīrī, son of Śrīmangala and a veteran champion of the Nimbarka school1. The first Vilasa (11 śl. only) of the work is mainly introductory ; the second deals with nāyikās, and the third with māna, pranaya, rāga, śrngāra etc.

27. KOLLŪRI RĀJASEKHARA

Alamkāra-makaranda (Madras Trm 2285) The author belonged to a family of Dravidas of Perūru in Andhra. The work quotes Catmatkāra-candrikā (see below under Viśveśvara Kavicandra) and praises a chief called Rām- eśvara of Aņipiņdivamśa, son of Viśveśvara and Kāmakşī, and described as Ammanna Mahī-mahendra of Mukteśvara (near Godāvarī). This Rājaśekhara is said to have been also patronised by Peshwa Madhava Rao (1760-72 A.D.).

28. GANGĀDHARA MIŚRA

Catura-cīntāmaņi (H. P. Sāstri, Cat. ASB MSS, vi, 4934/8162, p .. 485-86) This is a work on the nine Rasas, but chiefly on Śrngāra Rasa, in eighteen Prakāsas. The author is described as son of Miśra Samdoha.

1 See S. K. De, Vaisnava Faith and Movement, 1942, p. 55 fn.

276 SANSKRIT POETICS

29. GANGĀNANDA MAITHILA

Karņa-bhūşaņa (ed. Bhavadatta and K. P. Parab, Nir. Sag. Press, Bombay 1902)

Kāvya-dākinī (ed. P. Jagannath Hoshing, Sarasvati Bhavana Texts, Benares 1924) The first is a work on the Rasas in five chapters, dealing with (1) vibhāvas (2) anubhāvas (3) vyabhicāri-bhāvas (4) sthāyi-bhāvas and (5) rasa. It was written, as the author himself says, at the command of king Srīkarņa of Bikaner (bikāneri- purī), who appears to be the same as Lūņakarņajī who ruled at Bikaner from 1505 to 1526 A.D. The Kāvya-dākinī deals in five chapters (called Drstis) with Dosas.

30. GANGĀRĀMA JAŅI or JAŅIN

Rasa-mīmāmsā (ed. with his own Chāyā, Kashi Samskrita Press, Benares 1885) It is a small work of 114 verses on the poetic sentiments. For details about the author and his commentary on Bhanudatta's work, see above p. 250. He belongs to the second quarter of the 18th century.

31. GADĀDHARA BHAȚȚA

Rasika-jīvana (Aufrecht i. 497b, ii. 116b). This work on Rasa, which bears the character of an anthology rather than a work on Poetics, is in ten prabandhas and contains 1562 verses comprising selection from no less than 122 authors. The author is the son of Gaurīpati or Gaurīśa and Umā, and grandson of Damodara. Regnaud (p. 379) gives an account of the Paris Biblioth. Nationale MS of this work. The work quotes Jagannatha's Rasa-

MINOR WRITERS ON ALAMKĀRA 277

gangadhara, and hence it must be later than the middle of the 17th century.1

32. GIRIDHARA

Kalyāņa-kallola (H. P Sastri, Cat. ASB MSS, vi, 4932/ 8312, p. 482) This is a work on the nine Rasas (including Sānta) composed under the patronage of Kalyāna Dāsa, son of Todara Malla.

33. GOKULANĀTHA MAITHILA

Rasa-mahārņava The author is the celebrated Maithili Smarta and Naiyāyika Gokulanātha, son of Pītāmbara and Umādevī of Phanadaha family in Mangraunī, who lived and wrote in Benares at the end of the 17th and beginning of the 18th century. The work is referred to by himself in his Pada- vākya-ratnākara.2 He is probably the same as wrote a commentary on Mammata, see above p. 173. His drama Amrtodaya (written about 1693 A.D.) has been published in Kāvyamālā 59, Nir, Sag. Press, Bombay 18973.

34. GAURANĀRYA

Lakşaņa-dīpikā Prabandha-dīpikā or Padārtha-dīpikā (Madras Cat. xxii, 12951, extract) The first work, which breaks off with the sixth prakāśa, deals with (1) kāvya-svarūpa (2) paribhāsā (3) kāvya-lakşaņa-

1 P. K. Gode in ABORI xii, pp. 296-99 and Hara Dutt Sarma in Jha Comm. Volume pp. 359-65. For list of authors quoted see BORI MS Cat. xii, no. 247, at p. 288-90. 2 ABod 246a. 3 See HPS i p. 17f.

278 SANSKRIT POETICS

bheda (4) kalikotkalikādi (5) udāharņa-bheda and (6) nāyikā. The other work Padartha-dipika covers common ground. The author is described as son of Ayamaprabhu and brother of Mitaraja, who was minister of Singaya Madhava of Recarla family. The work cites Sāhitya-cūdāmaņi (of Bhatta Gopala?). Another work in four paricchedas, also called Lakşaņa-dīpikā (consisting of Kārikā and Vrtti), is attributed in the same Catalogue (no. 12952, extract) to Gauranārya; but the author is here described as son of Ayyalu-mantrin, brother of amatya Potama. They may be identical. The latter work cites the Alamkāra-samgraha, Kavi-kantha-pāśa, Camatkāra-candrikā, Sāhitya-candrodaya and Sāhitya-ratnākara (of Dharma Sūri?). The Kavi-kantha- pasa, of which two MSS (without the name of the author) are noticed in Madras Cat. xii 12802-03, is a treatise (said to be based on some work of Pingala's) on the characteristics of a poet's personal appearance and qualities, on the effect of the initial words of a poem, on the auspicious day for beginning a composition and so forth.

35. GHĀSI or GHĀSĪ RĀMA PAŅDITA

a. Rasa-candra (IOC iii, 1210/295, pp. 351-53 ; extract) b. Rasa-kaumudī (Madras Cat. xii 12921, extract ; BORI MSS Cat. xxii, no. 197, p. 223) The first work was composed in 1696 A.D. The second work describes the nine Rasas. The anonymous Rasa-kaumudī in Peterson v. 414 refers apparently to this work.1 The Rasa-candra is in four chapters, dealing with: 1. nāyikā-gaņa- bheda (198 śl.), 2. nāyaka-samgha (85 śl.), 3. anubhāvādi gaņa

1 P. K. Gode (Cal. Orient. Journal iii, pp. 35-37) gives the latter half of the 18th century A.D. as the probable date of this anonymous work.

MINOR WRITERS ON ALAMKĀRA 279

(150śl.), 4. rasa-daśaka (162śl.). Is this Ghāsīrāma of Gautama- vaņśa identical with Ghāsīrāma Bhatta, father Śrīnātha who wrote the medical work Jagat-prakāśa (Stein pp. 193, 348)? He, however, appears to be the same as wrote the Padya- muktāvalī (erotic verses).

36. CANDIDĀSA Dhvani-siddhānta-grantha The work is referred to by himself in his commen tary on Mammata. See above p. 160.

37. CANDRACŪŅA Prastāva-cintāmaņi (Ulwar 1064, extract 223; Weber 826) It is a work in five Ucchväsas on the art of writing poetic descriptions. The author is described as son of Purusottama Bhatta. For citations in this work, see Weber loc. cit. It cites Candrasekara-campū-prabandha which, Regnaud thinks, is a campū by Candraśekhara, father of Viśvanātha (q. v.).

38. CIRAÑJIVA or RĀMADEVA (VĀMADEVA) CIRAÑJĪVA BHATȚĀCĀRYA a. Kāvya-vilāsa (Ed. Batuk Nath Sarma and Jagannath Sastri Hoshing, Sarasvati Bhavana Texts, Benares 1925. See IOC, iii no. 1191, pp. 343-44 for a detailed summary of contents. b. Śrngāra-taținī (Aufrecht i. 660b). The author was son of Rāghavendra (described as ācārya- śatāvadhāna) and grandson of Kāśīnātha of Rādhāpura in Gauda country. He is the author of the Vidvanmoda- tarangiņī (a campū)1 and also of Vrtta-ratnāvalī, a work on

1 Ed. Venkatesvara Press, Bombay 1912 ; also ed. Satyavrata Sama- srami in the Hindu Commentator iv, nos. 1-4, 1871 ; and ed. Kali- krishna Deb, Serampore Press 1832 (text and trs.), 2nd ed. 1834. The author's genealogy as above is given by himself in this work.

280 SANSKRIT POETICS

prosody. The India Office MS, as well as the printed text, of the Kāvya-vilāsa consists of two chapters (called Bhangī), dealing with Rasa and Alamkara respectively. We are told that the definitions in this work are taken from old standard writers, while the illustrative stanzas are the author's own. The section on sabdālamkāra from this work has been printed and inadvertantly included in the text of Appayya's Kuvalayā- nanda, published by N. S. P. (ed. Vāsudeva L. Panshikar Bombay, 1909) with Asadhara's commentary1. See above p. 223 fn 3. The illustrative verses of his Vrtta-ratnāvalī2. panegyrise Yaśovanta Simha, Nayeb-dewan of Dacca under Sujau-d-daulah of Bengal, about Saka 1653=1731 A.D. His Kāvya-vilāsa was composed in 1703 A.D. He belongs, there fore, to the last quarter of the 17th and first half of the 18th century. Cirañjiva also wrote Mādhava-campū (ed. Satyavrata Samasrami, in the Hindu Commentator iv, no. 4-7, Calcutta 1871), as well as Kalpa-latā and Siva-stotra mention- ed in his Kāvya-vilāsa.

39. JAYAMANGALA

Kavi-śikșā (Peterson i, no. 120, extract) This work is apparently cited by Ratnakantha on Stuti- kusumāñjali i. 1. The author was a Jaina who wrote at the time of Jayasimha Siddharāja (1094-1143 A.D.) and was thus a contemporary of Hemacandra3.

40. JINAVALLABHA SŪRI Praśnottara (Br. Mus. MS no. 426, extract) This is a collection of riddles and verbal puzzles. On the author (about 1110 A.D.), see Klatt p. 36 and Bhandarkar Rep. 1882-83, p. 48, where other works of his are mentioned. It is accompanied by an avacūri by Kamalamandira.

1 See pp. 97-100 of this edition. 2 HPS iii, no. 280. 3 Peterson, Detailed Report 1883, p. 68

MINOR WRITERS ON ALAMKÄRA 281

41, JIVANĀTHA Alamkāra-śekhara (Oudh iii. 12)

42. TIRUMALA or TRIMALLA BHATȚA Alamkāra-mañjarī (extract in ALeip 851) This South Indian author was son of Vallabhabhatta, and his name is also given as Trimmala or Tirmala, and some- times incorrectly as Nirmala, The work, written in Benares in 43 verses, deals only with arthālamkāras1. The Arthā- lamkāra-mañjarī entered under this author's name in Bühler's. Catalogue (1871-73) is probably this work2. The author appears to be identical with Trimalla Kavi, son of Vallabha and grandson of Singhana Bhatta, who wrote some works on medicine (see ALeip 1182-85). His date3 is fixed between 1383-1499 A D.

43. TRILOCANĀDITYA Nāțya-locana (Aufrecht i. 284b, iii. 61a) This work (without the author's name) is extensively cited, e.g. by Rāghavabhațța on Šakuntalā ed. NSP, 1886, p. 7, by Vāsudeva on Karpūra-mañjarī, by Ranganātha on Vikramor- vaśīya i. 1, by Dinakara and Cāritravardhana on Raghu. As Dinakara's date is 1385 A D.4, this work cannot be placed later than the middle or third quarter of the 14th century. A commentary called, Locana-vyākhyāñjana, by the author himself, is mentioned in Oppert 2695.

44. TRYAMBAKA Nāțaka-dīpa (Aufrecht i. 284b) Three commentaries on the work are entered by Aufrecht 1 The list of 38 figures dealt with is given at the outset; the passage in quoted in ALeip, MS no. 851, p. 273. 2 New Catalogus Catalogorum i, p. 295 enters them separately. Most of the MSS contain Arthalamkara only. 3 See Cat. R.A.S (Bombay Branch) vol.i, no. 126,p. 42. 4 Nandargikar's ed. of Raghu, 1897, Pref. p. 17.

282 SANSKRIT POETICS

loc. cit. One of the Deccan College MSS of the work contains a Prakrit commentary1.

45. DĀMODARA BHAȚȚA HARȘE Alamkāra-krama-mālā (Aufrecht i. 32a)

46. DĪNA KRSŅADĀSA

Rasa-kallola The work was written about 1480 A.D. when Gajapati Purusottama was reigning2.

47. DEVANĀTHA Rasika-prakāśa (Aufrecht i. 497b) The author is probably the same as Devanātha Tarkapañ- canana who wrote a commentary named Kāvya-kaumudī, on Mammata (see above p. 174). A Devanātha is cited by Bharatamallika (18th century) on Bhatti x. 73.

48. DEVAŚAMKARA, surnamed Purohita

Alamkāra-mañjūşā (Ed, S. L. Katre, Scindia Orient. Ser. Ujjain 1940. See Bhandarkar, Rep. 1887-91 p. lxiiif, extract) The author, a Gujarat Brahman, was son of Nahanabhai and a native of Raner (Rander near Surat), and lived at Urahpattana (probably Olpād in the same district). The work deals with poetic figures alone, and the illustrations sing the glory of the Peshwas Madhava Rao I of Poona and Nārāyaņa Rāo. and their uncle Raghunātha Rão who flourished between 1761 and 1772 A.D. The author, there-

1 Deccan Coll. Catalogue p. 417 no. 38 .- The comm. by Rāmakrņa Pandita on Nātaka-dīpa is not a comm. on this work, but (as Aufrecht points out, i. 791a) a comm. on the Nātaka-dīpa in the Pañcadasi. Correct this error in Schuyler's Bibliography p. 18 and in Harichand Sastri, p. 35, no. 361, 2 See IA i, p. 215.

MINOR WRITERS ON ALAMKĀRA 283

fore, belonged to the third and fourth quarters of the 18th century.1 The work deals almost exclusively with poetic figures, which are enumerated as 115 and classified into Arthā- laņkāras (102), Pramāņālaņkāras (103-6), Dhvanyālaņkāras (107-13) and Miśrālaņkāras (114-15). The Kārikās, as well as prose explanations, are generally based on the corresponding passages of Appayya's Kuvalayānanda. In spite of its editor's eulogy, the work does not seem to be a striking or original contribution to the subject. Devaśamkara also wrote a commentary on the Amaru-śataka (Mitra x, pp. 81-82).

49. DHARMADĀSA SŪRI

Vidagdha-mukha-mandana, with vtti (Ed. Haeberlin in Kāvyasamgraha, Calcutta 1847, p. 269f. Also in Kāvya-kalāpa (pub. Haridas Hira- chand) Bombay 1865 Printed many times in India; but ed. N.S.P., Bombay 1914, is useful. MSS: Aufrecht i. 572b, ii. 135b, 225a, iii. 121a). The work of Dharmadasa is in four paricchedas, dealing with enigmatology and Citra-kavya, and at the same time describes the feeling of separation from a lover. The conclud- ing verse in the Bombay edition of the text (wanting in Br. Mus. MS, Bendall no. 427), as well as the first verse (which invokes Sauddhodani) and the colophon to Jinaprabha's com- mentary, makes the author a Buddhist ascetic. The known dates of Jinaprabha put the limit of Dharmadāsa's date earlier than the last quarter of the 13th century. This work is also cited by name by Kumārasvāmin (p. 122=iv. 1), by Rāyamukuța on Amara2, and quoted in the Paddhati of Sarngadhara. These citations themselves would put the date of our author earlier than the 14th century. Commentaries on this work are numerous: (1) By Jinaprabha Sūri, pupil of Jinasimha Sūri (Weber

1 See ABORI xv, pp. 92-96 and xxi p. 152-54. 2 Composed 1431 A.D. ; see Bhandarkar, Rep. 1883-84, p. 63.

284 SANSKRIT POETICS

1728). For this Jaina writer, whose known dates are 1293 and 1309 A. D. see Peterson iv, p. xxxvii and Klatt's Onomasticon. His Guru Jinasimha founded the Laghu- kharatara-gaccha in 1275 A.D. The date of our commentator, therefore, will be the last quarter of the 13th and begin- ning of the 14th century. (2) Commentary by Ātmārāma. Aufrecht i, 573a. The full name of this writer appears to be Svātmārāma Yogīndra. (3) Vidvan-manoramā by Tārācandra Kāyastha. Aufrecht i. 573a, ii. 135b, iii. 121a. For his other works, see ibid i. 229a. (4) Śravaņa-bhūşaņa by Narahari Bhatta, Aufrecht i. 573a. (5) Subodhini by Trilocana. Aufrecht ii. 135b (extract in Stein p. 274). (6) Commentary by Sivacandra. Aufrecht iii 121a. Its date is 1613 A.D.1 (7) °Țīkā by Durgādāsa, son of Vāsudeva and pupil of Bhațța Devacandra. Aufrecht ii. 135b, iii. 121a ; extract in Peterson iv p. 36.

50. DHARMA SUDHĪ or DHARMA SŪRI

Sāhitya-ratnākara (ed. Tiruvenkatacharya with commentary, Madras 1871 ; ed. Nellore 1885. MSS: Aufrecht i. 716a, ii. 171a, iii. 148a; BORI MSS Cat. xxii, no. 301, pp. 366-70 (extract) ; Madras Cat. xxii, 12970-75 ; HPS ii, no. 246, extract). The author's name is given variously as Dharma-simha or Dharma-pandita. Son of Parvatanātha and Allamāmba. he belonged to a Benares family distinguished for the high proficiency of its members in philosophical studies, and his genealogy is thus given (Hultzsch i. p. 70): Tripurāri->Dhar- ma->Parvatanātha or Parvateśa->Dharma Sūri. He was also the author of two plays, called Naraka-dhvamsa or Narakā- suravijyaya1 (a vyāyoga) and Kamsa-vadha (a nātaka), and of

1 P. K. Gode in Journal of the Univ. Bombay, 1954, pp. 126-29. 1 Ed. Madras 1885 (in Telugu characters) ; Hultzch 323, Aufrecht i. 277a.

MINOR WRITERS ON ALAMKĀRA 285

some Kavyas and Stotras. The author is cited by Anantārya (q. v.) in his Kavi-samaya-kallola. A commentary on this Sāhitya-ratnākara, called Naukā, by Venkața Sūri is men- tioned in Madras Cat. xxii, 12974-75 (ed. Madhusudan Mishra, Bomra 1901). Venkata Sūri was son of Lakșmaņa Sūri and Suramāmbā, and grandson of Brahmāntara-vāņi and disciple of Venkatācārya. There is another commentary called Mandara by Mallādi Laksaņa-sūri (ed. Madras 1891 in Telugu characters). The Sāhitya-ratnākara (c. 1425 A.D.) is in ten tarangas, dealing with the conventional topics of Poetics as follow: (i) Granthārambha (ii) Vācaka-śabdārtha- vṛtti (iii) Lakşaņā-śabdārtha-vrtti (v) Guņa (vi) Šabdālamkāra (vii) Arthālamkāra (viii) Doșa (ix) Dhvani-bheda (x) Rasa. Most of the illustrative verses are in praise of Rama as a deity (śrīmat-raghu-tilaka-yaśoghanasāra-surabhita). Dharma Sūri must have been later than Vidyanatha; for in one of his verses he anonymously ridicules Vidyānātha's method of praising his patron (alamkriyāh pūrvataraiḥ praņītāḥ / prayo- gitāḥ kāścana nāyakena / kaiścit tu kukşimbharibhir nibaddhāḥ / kşodīyasā kāścana nāyakeņa). His date is roughly the first half of the 15th century.1

51. NARASIMHA

Guņa-ratnākara (Tanjore Cat. ix, no 5207, p. 4028) This work deals with a hundred poetic figures. It was written under Serfoji of Tanjore (1684-1710 A.D.)

52. NARASIMHA or NRSIMHA KAVI

Nañjarāja-yaśo-bhūşaņa (Ed. E. Krishnamacharya, Gaekwad Orient. Ser. Baroda 1930) The author, son of Sivarama-sudhī-mani and disciple of an ascetic Yogänanda, belonged to the Sanagara class of Brah- 1 See E.M.V. Raghavacharya in Proc. A-I.O.C ix. Trivandrum 1940, pp. 503-17 ; also NIA ii, 1939, pp. 428-441 for Dharma Sūri's date and works. The date of the work is given as c. 1425 A.D.

286 SANSKRIT POETICS

mans, and was patronised by Nañjarāja, whose name is borne by the title of his work and whose glory it sings in the illustra- tive verses. Nañjarāja was Sarvādhikarana (revenue minister) of Chikka Krishnaraja of Mysore from 1739 to 1759 A.D., after which came Nañjaraja's downfall terminating with his miserable death in Hyder Ali's imprisonment in 1773. Nara- simha's work must have been written during the twenty years of his patron's flourishing period. The work is divided into seven chapters called Viläsas; and it deals with all topics of Poetics including Dramaturgy. The seven Ullasas deal with (1) Nāyaka (2) Kāvya-svarūpa (3) Dhvani (4-5) Doșa-Guņa (6) Nāțya, including Candrakalā-kalyaņa (a short typical drama) (7) Alamkāra. It is obviously modelled on the Pratāparudra-yaśobhūşaņa of Vidyānātha and freely appro- priates much of its subject-matter verbatim. He inserts, after Vidyānātha, a model five-act drama called Candrakalā- kalyāna to illustrate his treatment of Dramaturgy in ch. vi. The author had the grandiose title Abhinava-Kālidāsa ; and he describes himself as the friend of Abhinava Bhavabhūti (alias Ālūra Tirumala-kavi). 53. NARASIMHĀCĀRYA or VENKAȚA NRSIMHA KAVI Alamkārendu-śekhara (Madras Cat. xxii, 12978, extract ; contains the first prakaraņa only) This South Indian author was son of Dasamācārya of Śrīsaila family and wrote a commentary on a work on music, called Śānta-vilāsa, composed by Subrahmaņya Sudhī (alias. Hariśāba Kavīndra) son of Harirāya and grandson of Krsna- rāya of Carișnusāla village. He refers to a Gīta-mañjarī by this Harisaba Kavīndra, and to a Campu of his own called Jānakī-pariņaya, and quotes Sāhitya-ratnākara of Dharma Sūri. The manual on Alamkāra mentioned above deals in five Prakaraņas with (1) nāyaka lakşaņa (2) kāvya svarūpa (3) rasa-lakşana, especially śngāra (4) doșa and guņa, and

MINOR WRITERS ON ALAMKĀRA 287

(5) alamkāra. It is based generally upon the Pratāparudrīya. Our author also wrote a commentary, also called Alamkā- rendu-śekhara, on some Kārikās on Poetics entitled Lakşaņa- mālikā1

54. NARAHARI SŪRI

Rasa-nirūpaņa This work and the author are mentioned by Kumāra- svāmin at p. 224.

55. NARENDRAPRABHA SŪRI (Maladhāri)

Alamkāra-mahodadhi (Ed. L. B. Gandhi, Gaekwad Orient. Series, Baroda 1942) The author was a pupil of Naracandra of Harsapurīya- gaccha. The work was composed at the request of Vastupāla (d. 1242 A.D.) in whose honour he wrote three Praśastis. It consists of eight chapters and deals with (i) Kāvya- phalādi (ii) Śabda-vaicitrya (iii) Dhvani including Rasa (iv) Guņībhūta-vyangya (v) Doșa (vi) Guņa (viii) Śabdālaņkāra and (viii) Arthālamkāra. The work is stated to have been composed in Samvat 1282 (=1225-26 A.D.).2

1 See Madras Cat. xxii 12955, extract ; ŚgS i, 98-99 extract, also p. 11. V. Raghavan (New Cat. Cat. i p. 300) thinks that probably "the basic text of the Lakşana-mālikā is also by Nrsimha himself." 2 The Alamkāra-candrīkā of Nārāyaņa Deva referred to by himself in his Samgīta-nārāyana (ABod 201) is not a work on Rhetoric but deals with the subject of musical Alamkaras. The author, also called Gajapati Vīranārāyana-deva, was son of Padmanābha and disciple of Purușottama Miśra. Similarly, the Kāma-samūha of Ananta, son of Maņdana and grandson of Nārāyaņa (composed in 1457 A.D.), is really an anthology of erotic verses; see P. K. Gode in JOR, Madras, xiv, pp. 74-81. The Srngārālāpa of Rāma, of which a MS is dated I556 A.D., is a similar work (see P. K. Gode in Journal of Bom. Univ. xv (N.S.), pt. 2, 1946, pp. 81-88.

288 SANSKRIT POETICS

56. NĀRĀYAŅA Kāvya-vṛtti-ratnāvalī (Tanjore Cat. ix, no 5173) The work is in nine Prakaranas. Its avowed object is to deal with Kavi-svarūpa, Kāvya-lakșaņa, Rasa-svarūpa and general principles of poetry.

57. NĀRĀYAŅA

Sabda-bheda-nirūpaņa The work deals with the three Vrttis of word (Abhidha etc). The author refers to Sāha Mahārāja whose protégé he was {=Sbahaji, king of Tanjore 1686-1710). There is another work of Laksaņa-kavi, called Sāharājīya (Tanjore Cat. ix, no. 5304), the illustrations of which eulogise this prince. See above p. 270, no. 15. Several works called Śaba-bheda-nirūpaņa are found in Tanjore Cat. ix, no 5301-3.

58. PADMASUNDARA (Akabara-śāhī) Śrngāra-darpaņa (Ed. Anup Skt. Series, Bikaner 1943) The author was a Jaina monk of Akbar's time. The work is in four Ullasas, but it is a rehashing of Rudrabhatta's Śrngāra- tilaka. The illustrative verses are addressed to the Mughal emperor Akbar. MS (Bikaner 9356) is dated 1569 A,D. See paper on the work by V. Raghavan in C. Kunhan Raja Presentation Vol.

59. PUÑJARĀJA a. Dhvani-pradīpa (Aufrecht i. 273b) b. Kāvyālamkāra-śiśu-prabodha or bodhālamkāra (Aufrecht i. 103a) Siśu-pra-

The author was son of Jīvana (or Jivananda) and Maku, of the Śrīmāla family of Malabar (mālabhāra), and his genea- logy is given thus: Sādhu-sadepāla->Kora(ā?)->Pāma(ā?) ->Govā->Yāmpaca->Jīvana. Jīvana's brother Megha and him-

MINOR WRITERS ON ALAMKĀRA 289

self were ministers of Khalaci Sāhi Gayāsa; and Jīvana had two sons Puñja and Muñja. Puñja became king, but aban- doning his kingdom to his younger brother, devoted himself to study, and wrote some works1. The colophon 2 to Puñja- rāją's Sārasvata-fīkā on Sārasvata-prakriyā says: śrīmāla-kula- śrīmālabhāraśrī-puñjaraja°, on which Bhandarkar remarks that Puñjarāja was the ornament of the Malava circle3. Aufrecht thinks that the patron of Puñja's father and uncle was Ghiyas Shah Khalji of Malava (about 1475 A.D.) and that Puñjarāja must have lived between 1475 and 1520 A. D. or at the end of the 15th and beginning of the 16th century.

60. PUNDARĪKA

Nāțaka-laksaņa (Aufrecht i. 284b ; SCB 308)

61. PUNDARĪKA (or PAUŅŅARĪKA) RĀMEŚVARA

Rasa-sindhu (Aufrecht iii. 106a) The work consists of fourteen chapters called Ratnas. It quotes Darpana (of Viśvanātha) and Rasa-tarangiņī of Bhanudatta ; hence later than 1500 A.D. For its date see P, K. Gode in Calcutta Orient. Journal ii, pp. 30-32, dating the work at about the beginning of the 15th century.

62. PURUȘOTTAMA SUDHINDRA

Kavitāvatāra (Aufrecht i. 87a) The work, in ten chapters (called Vihäras), is dedicated to one Nāgabhūpala, whom its illustrative verses panegyrise. A Puruşottama is cited by Viśvanātha in his Sāhitya-darpana, see above p. 214, fn 6.

1 Peterson Report v, pp. xliii, 166-69. 2 Peterson Report v, p. 169 ; AFI 181. 3 Rep. 1882-83, p. 12; cf also the colophon to his Siśu-prabodha quoted in op. cit p. 199. See P. K. Gode, Studies in Ind. Literary Criticism, i, pp. 68-72 on Puñjarāja's date. 19

290 SANSKRIT POETICS

63. PRAKĀŚVARȘA

Rasārņavālamkāra The text in Roman transliteration was published by V. Venkataram Sarma with an account of the work in IHQ v, 1929, pp. 173f. The Madras MS gives the work in five chapters, the first four of which deal with Dosa, Guna and Alamkāra, the last (incomplete) with Rasa. It is possible that the whole of ch. v (now lost) dealt with Ubhayālamkāras and the major part of ch. vi (now recovered incomplete) with Rasa. See the question discussed by S. K. De in IHQ v,. 770-780 and by V. Raghavan in JOR viii, 1934, pp. 267-276. The work is later than Bhoja whose works are extensively utilised. See also on this work S. P. Bhattacharya in JOI, Baroda, vii, 1957, nos. 1-2 and ix, 1959, pp. 5-16.

64. PRABHĀKARA BHAȚȚA

a. Rasa-pradīpa (Ed. Narayan Sastri Khiste, Sarasvati Bhavana Text, Benares 1925. MSS: Weber 823 ; SCC vii 42, extract in both) b. Alamkāra-rahasya, cited in his Rasa-pradīpa,. pp. 8, 9, 10, 13, 15, 20, 37, 38, 39, 40, 51. The author was son of Madhava Bhatta and grandson of Rāmeśvara Bhatta, and younger brother of Raghunātha and Viśvanatha, the last of whom he calls his 'vidya-guru'. The Rasa-pradīpa was composed in Samvat 1640=1583 A.D. at the age of nineteen (Weber loc. cit.)1. Prabhākara was thus born in 1564 A. D. His Laghu-saptaśatika-stotra, an epitome of the Devī-mahatmya was written in 1629 A.D. He belongs, therefore, to the last quarter of the 16th and first quarter of the 17th century. The Rasa-pradipa consists of three chapters (called ālokas) dealing with (1) kāvya-lakşaņa (2).

1 The date given in SCC vii, no. 42 is Samvat 1170=1114 A.D; but. this must be a mistake.

MINOR WRITERS ON ALAMKĀRA 291

rasa-viveka (3) vyañjanā-nirūpaņa. The author cites Śrīharșa Miśra, Miśra Rucinātha, Dharmadatta, Locanakāra (Abhi- navapguta), Pradīpakrt, Sāhityadarpaņa-kāra. The Dharma- datta may be the same as quoted by Viśvanātha in his Sāhitya- darpana. Ananta in his commentary on Āryā-saptaśatī quotes a Rucinatha Miśra as a writer on Poetics. For Prabhākara's other works, see Aufrecht i. 353b. On citations in Rasa- pradīpa see S. K. De in IHQ viii, 1932, p. 358.

65. BALADEVA

Śṛṅgāra-hāra

(Kielhorn, Rep. 1880-81, p. 71=BORI MS Cat. xii, no. 295, p. 351) The author is described as son of Keśava. MS is dated in Samvat 1845 (=1789-90 A.D.).

66. BALADEVA VIDYĀBHŪȘAŅA

Kāvya-kaustubha (Ed. Haridas Das, Navadvip, Bengal, 1957) This work consists of nine prabhas and deals respectively with (1) Kāvya-phalādi (2) Śabdārtha-vrtti (3) Rasa (4) Guņa (5) Rīti (6) Doşa (7) Dhvani-bheda (8) Madhyama-kāvya and (9) Sabdārthālamkāra. See above pp. 171-72 under com- mentators on Mammata.

67. BĀLAKRSŅA BHAȚȚA

Alamkāra-sāra (Aufrecht i. 32b) The work consists of ten chapters. Bālakrsna Bhatta, styled Tighara, was son of Govardhana Bhatta and belonged to the Vallabha Sampradaya. A work of this name is cited by Jayaratha (pp. 88, 97, 171, 172, 184) ; also in Bühler's Catalogue 1871-73. The Alamkāra-sāra quotes Kuvalayānanda and Citra-mīmamsa ; and the Deccan College MS of the work (no. 23 of 1881-82) appears to have been copied in Samvat

292 SANSKRIT POETICS

1758 (=1702 A.D.). We can, therefore, assign it to a period between 1625 and 1700 A.D. The ten Ullasas of the work have the following topics respectively: (i) Kāvya-prayojana-kāraņa-svarūpa (ii) Śabda- nirņaya (iii) Artha-nirņaya (iv) Dhvani-nirņaya (v) Guņī- bhūta-vyangya-nirņaya (vi) Šabdārtha-nirņaya (vii) Doșa (viii) Guņa (ix) Šabdālamkāra and (x) Arthālaņkāra.

68. BHVA MIŚRA or MIŚRA BHĀVA

Śrńgāra-sarasī (SCC vii 43, extract) The author of this treatise on amorous sentiments is des- cribed as son of Miśra Bhataka.

69. BHĀŞYKĀRĀCĀRYA (or ? BHĀSKARĀCĀRYA)

Sāhitya-kallolinī (Madras Cat. xxii, 12964, extract) The author is described as a descendant of Varadaguru of Śrīvatsa-gotra, and a resident of Bhūtapurī or Śrīperumbū- dūr. The verses are taken copiously from several well-known rhetorical works, e.g. Mammata, the Bhāva-prakāśa etc. The author states his indebtedness to the Rasārnava-sudhākara of Singa-bhūpāla ; hence he sould be placed later than the middle of the 14th century. The topics dealt with are pra- bandha-bheda, nātya-nrtta-nrtya, vastu, samdhi nāyaka- lakşaņa, rūpaka, uparūpaka and kāvya-lakşaņa.

70. BHĪMASENA DĪKȘITA

a. Alamkāra-sāroddhāra b. Alamkāra-sāra-sthiti or Kuvalayānanda-khaņ- dana, see above p. 225 under Appayya Dīkşita. He refers to both these works in his Sudhā-sāgara commen- tary on Mammata (see p. 171). Date between 1650-1725 A.D.

MINOR WRITERS ON ALAMKĀRA 293

71. BHĪMEŚVARA BHATȚA Rasa-sarvasva (Burnell 57a) The author was son of Ranga Bhatta.

72. BHŪDEVA ŚUKLA

Rasa-vilāsa (Ed. Prem Lata Sarma, Poona 1952) The author, son of Sukadeva of Jambusara in Gujarat, flourished between 1660 and 1720 A.D.1 For his other works, see Aufrecht i. 414b and introd. to above ed. p. xii. The India Office MS of the present work (no. 1209/2526b) contains only three stabakas and the beginning of a fourth. Our author is the same as Bhūdeva Sukla who wrote the drama Dharma- vijaya2 in five Acts. The Rasa-vilasa consists of seven chapters, called Stabakas. The topics dealt with are as follow, according to chapters: 1-2 Rasa, counted as nine including Sānta. 3 Bhava. 4 Guņa. 5-6 Doșa. 7 Vrtti (Abhidhā, Lakșaņā and Vyañjanā). As its editor rightly says it is a mediocre manual on Rasa and allied topics, which derives its material chiefly from Mammata and Jagannatha and shows little originality.

73. MĀNASIMHA

Sāhitya-sāra (Aufrecht i. 716a)

74. MOHANADĀSA Rasodadhi

The work is cited by himself in his commentary on the

1 P. K. Gode, however, in ABORI xiii, p. 183, thinks that the Rasa- vilāsa was composed about 1550 A.D. As the Rasa-vilāsa refers to the definition of poetry given by Rasa-gangadhara, it could not have been composed earlier than 1660 A.D. 2 Ed. Granthamālā iii, 1889; also ed. Narayan Sastri Khiste, Sarasvati Bhavana Texts, Benares 1930. See Mitra i, p. 37; Weber 1561; IOC vii, p. 1596.

294 SANSKRIT POETICS

Mahānāțaka (ABod 143a). The author was son of Kamalā- pati,

75. YAJÑANĀRĀYAŅA DĪKȘITA

Alamkāra-ratnākara (Tanjore Cat. ix, no. 5131) Sāhitya-ratnākara (ed. T. R. Chintamani, Madras 1932) The author was son of Govinda Dīksita, minister of Raghu- natha Nayaka of Tanjore who ruled between 1614 and 1633 A.D. Almost all the verses of the first work eulogise Raghu- natha. The second work is really a Kāvya in sixteen cantos, dealing with the exploits of the same prince. Both the works form a companion to the author's Raghunāthābhyudaya. K. Kunjunni Raja (Contribution of Kerala, p. 134) is not correct in identifying him with Yajñeśvara Dīkșita mentioned below.1

76. YAJÑEŚVARA DIKȘITA

Alamkāra-rāghava (Tanjore Cat, 5132-33) Alamkāra-sūryodaya (Tanjore Cat. 5140-41) The author was son of Cerukūri Kondubhatta and brother of Tirumala Yajvan.2 The first work quotes Rasārnava- sudhākara of Singa-bhūpāla and Sāhitya-cintāmaņi (apparently of Vīranārayana, q.v.) ; hence later than the 15th century (c. 1600 A.D.). The work is so called from the circumstance that the illustrative verses refer to Rama. The author may be identical with Yajñeśvara already mentioned as a commen- tator on Mammața (see p. 175). He may have been related to Lakșmīdhara (q.v.) who also came from Cerukūri.

1 So also in Tanjore ix, 1933, no. 5132. The Alamkara-ratnākara ascribed Yajñanārāyaņa (no. 5131), also eulogises Raghunātha Nāyaka, the poet's patron. Yajnesvara and Yajñanārāyana appear to be different persons. 2 See Šgś ii, p. 65.

MINOR WRITERS ON ALAMKĀRA 295

77. YAŚASVIN KAVI Sāhitya-kautūhala and its commentary Ujjvala- padā (Aufrecht i. 715b, ii. 171a) The author is described as son of Gopāla and Kāsī. The India Office MS (Cat. iii, p. 337) was copied in 1730 A.D .; it contains only the first chapter which deals with enigma- tology and Citra-kāvya. There is another Sāhitya-kutūhala of Raghunatha, a protégé of queen Dipabai of Tanjore (between 1675-1712 A.D.). also on Citra-kāvya (see Journal Bomb. Univ. x, p. 132f).

78. RATNABHŪȘAŅA

Kāvya-kaumudī (HPS ii, no. 35, extract) . This work in ten paricchedas is apparently a very modern composition by a Vaidya Pandit of East Bengal. It deals with (1) nāma. (2) lingādi. (3) dhātu-pratyaya. (4) kāvya- lakşaņa. (5) dhvani. (6) guņībhūta-vyangya. (7) guņa. (8) and (9) alamkāra. and (10) doșa, the first three chapters being devoted to grammar. The date Śaka 1781 (=1859 A.D.) may be the date of its composition (HPS ibid, preface p. viii).

79. RAGHUNĀTHA MANOHARA

Kavi-kaustubha P. K. Gode (Poona Orientalist vii, 1943, pp. 157-64) places this work between 1675 and 1700 A.D.

80. RĀGHAVA-CAITANYA

Kavi-kalpalatā (Aufrecht i. 87a) Possibly the poet of the same name cited in the Paddhati as Rāghavacaitanya Śrīcaraņa (71, 168, 877, 1557-8), which title apparently indicates that he was a well-known Vaisnava. In the colophon to the codex containing Deveśvara's Kavi- kalpalatā in SCC vii, no. 7 (cf. ABod 211b), the reading is māgha-caitanya-viracita-kavi-kalpalatāyāh etc. This may be

296 SANSKRIT POETICS

a corruption of or mistake for the name Rāghavacaitanya, whose work may have got mixed up with that of Deveśvara himself.

81. RĀJACŪŅĀMAŅI DĪKȘITA

a. Kāvya darpaņa (Ed. S. Subrahmanya Sastri, Vani Vilasa Press, Srirangam (no date). MSS: Madras Cat. xxii, 12809-814, with the commentary of Ravi-pandita)

b. Alamkāra-cūdāmaņi Mentioned in his Kāvya-darpaņa (Madras Cat, xxii. 12809) or Alamkāra-śiromaņi (Hultzsch i. extract p. 86) The author, who is a well-known and prolific South Indian author, was son of Satyamangala Ratnakheta Śrīnivāsa Dīkşita and Kāmāksī, and step-brother of Keśava Dīkșita and Śeșādri- śekhara Dīksita, He was grandson of Bhāvasvāmin and Lakşmī and great-grandson of Krsnabhatța Dīksita, and pupil of Ardhanārīśvara Dīksita who was his brother (Hultzsch ii, p. x). The date of composition of his Tantra-śikhamani, a work. on Mīmāmsā, is stated to be 1636 A.D. He was, thus, a contemporary of Nīlakaņtha Dīkșita, whose Nīlakaņtha- vijaya Campū was also composed in 1636 A.D., and belonged to the first half of the 17th century. In his poem Rukmiņī- kalyāna (in ten cantos) he states that he composed it when Raghunatha, son of Acyuta, was ruling at Tanjore ; and his two dramas Ananda-rāghava and Kamalinī-kalahamsa were staged in the court of the same prince. His genealogy and a long list of his other works are given in the concluding verses of his Kāvya-darpana1 and in his drama Ānanda- rāghava 2. The Kāvya-darpaņa in ten ullāsas covers generally all the topics of Poetics, while his other work deals specially with the poetic figures. The ten Ullasas of the Kāvya-darpana

1 Extract in Madras Cat. xxii no. 12809 and Hultzsch i, pp. 85-6. 2 Madras Cat. xiii, no. 12495. The Kāvya-darpaņa mentions 26. works of the author.

MINOR WRITERS ON ALAMKĀRA 297

deal with (i) Kāvya-svarūpa (ii) Śabdārtha (iii) Vyangārtha (iv-vi) Kāvya-bheda (vii) Doșa (viii) Guņa (ix) Śabdālaņkāra and (x) Arthālamkāra. The family to which Rājacūdamaņi belonged was known as atirātra-yājin. Śrīnivāsa Atirātrayājin in Aufrecht i. 672a is identical with Rājacūdāmaņi's father. He lived in Surasamudra in Tondīra (ie. in the region of Kāñcī). Cf Sten Konow, Ind. Drama p. 94. For a list of his other works see Hultzsch i, pp. ix-x, and introd, to the Vani- vilasa Press ed. of Kamalinī-kalahamsa. His Śamkarā- bhyudaya has also been published by Vanivilasa Press, Srirangam.

82. RĀMACANDRA and GUŅACANDRA

Nāțya dārpaņa (Ed. G. K. Srigondekar and L. B. Gandhi in 2 vols. Gaekwad Oriental Ser. Baroda vol. i 1929. Ed. based on a single MS ; Peterson v, p. 188) A work of this name, but probably not identical, is cited by Ranganātha on Vikramorvaśīya and Bharatamallika on Bhațți. The present work is in four vivekas, dealing with Dramaturgy, and mentions twelve varieties of Rūpaka and a number of Uparūpakas. The author Rämacandra was the one-eyed pupil of Jaina Hemacandra1 of whom Gunacandra was also a pupil. He thus flourished between 1100 and 1175 A.D. He also wrote two dramatic works respectivly called Raghu-vilāsa2 or Raghu-vilapa3 where he mentions four other works by himself, as well as Satya-hariścandra (ed. B. R. Arte, Nir. Sag. Press, Bombay 1898) which gives a curious Jaina version of the Hariścandra legend. Rāmacandra is said to have been the author of a hundred works (prabandha-śata-kāra) : and no less.

1 Peterson, Report iv, pp. 16-7 ; Bühler's Hemacandra p. 44. The present work was first brought to notice by Sylvain Lévi in JA, cciii, 1923. P. K. Gode (Studies, i, pp. 36-42) places the work at 1150-1170 A.D. 2 Peterson Report v. 145. 3 Bühler Kashmir Rep. p. xlix.

298 SANSKRIT POETICS

than eleven of his dramatic works are quoted in the Nātya- darpana.

83. RĀMACANDRA NYĀYAVĀGĪSA

Kāvya-candrikā or Alamkāra-candrikā (Aufrecht i. 101a, 778b), with commentary called Alamkāra- mañjūșā (ed. Comilla 1885 ; ed. Dacca 1886 with commentary of Jagabandhu Tarkavagisa ; ed. Venkatesvara Press, Bombay 1912, with commentary Alamkāra-mañjūşā by Rämacandra Sarman who may be the author himself) A Bengal writer described as son of Vidyānidhi. Is he identical with Nyāyavāgīśa Bhațțācārya, author of the Kāvya-mañjarī commentary on the Kuvalayānanda (see above p. 229)?

84. RĀMA ŚARMAN or RĀMA KAVI

Nāyikā-varņana in 42 stanzas (Madras Cat. xxii, no. 12901).

85. RĀMA SUBRAHMAŅYA

Alamkāra-śāstra-samgraha (Hultzsch 1562) or Alamkāra-šāstra-vilāsa (Madras Trm II C 1802, 1805 ; extract) The author, also called Rāmasubba, belonged to Tiruvisa- lore. He seems to be a very recent author, who appears to have also written some philosophical works noticed in the Catalogues cited above.

86. RĀMA SUDHĪ or SUDHĪŚVARA

Alamkāra-muktāvalī (Ed. with Ratna-śobhākara comm. of Krsņa Sūri in Telugu script, Vizagapatam 1897-98) The author was son of Nrsimha.

MINOR WRITERS ON ALAKĀRA 299

87. LAKŞMĪDHARA DĪKȘITA

a. Alamkāra-muktāvalī (Aufrecht i. 32a)1 b. Rasa-manjarī, cited by himself in his commen- tary on the Gita-govinda c. Bharata-śāstra-grantha (BORI MS no. 40 1916-18)2 The author was son of Yajñeśvara and Sarvāmbikā (or Ambikāmbā), grandson of Timmaya Somayājin, and brother and pupil of Kondubhatta, He belonged to the Kāśyapa Gotra and bore the surname of Dakşiņāmūrti-kimkara. His family came from Cerukūri on the Krsņā river (in Andhra country) which place Hultzsch thinks to be identical with modern Peddacerukūru near Bāpațla. He is identical with Lakşmīdhara, author of the Prakrit grammar Sadbhāşā-can- drika. He also wrote commentaries on the Anargha-rāghava, Prasanna-raghava and Gita-govinda. In the first of these commentaries it is said that after having led the life of a householder for a long time, he travelled to different countries and conquered all literary opponents, and then having renoun- ced the cares of the world he became a samnyāsin or yati with the name Rāmānanda or Rāmānandāśrama, under a Guru called Krsņāśrama. Lakșmīdhara (who is also sometimes called Lakşmaņabhatța or Lakșmaņa Sūri) was patronised by Tiru- malaraja, probably Tirumala I of the third Vijayanagar dynasty8, to whom his Śruti-rañjani commentary on Jayadeva's

1 V. Raghavan (New Catalogus Cat. i, p. 296) queries whether this work is really the Alamk. muktāvalī of Viśveśvara (q.v.), son of Lakşmī- dhara, and refers to ABORI, xviii, 1937, p. 200. 2 See P. K. Gode ABORI xv, 1953, p. 240-42. Mentions Bharatārņava and Kavi-kantha-pasa. The Kavi-kantha-pāsa (Madras Cat. xxii, no. 12802), said to be based on some work of Pingala's, gives miscellaneous information about a poet's personal appearance, qualities etc. (cf. Rāja- śekhara, Kav. Mim. ch. x) ; the name of the author is not known. See above under Gauranārya, p. 278. 3 EI iii p. 238 Table. He died in 1572 A.D.

300 SANSKRIT POETICS

Gita-govinda is sometimes attributed1. As the prince flourish- ed in the middle of the 16th century, Laksmīdhara's date would be the same2.

88. VALLABHA BHATȚA

Alamkāra-kaumudī (ed. Granthamālā ii, 1889) A short treatise of very recent times, dealing with poetic figures, the illustrations being in praise of Rāma,

89. VIȚȚHALESVARA or VIȚȚHALA DĪKȘITA

Rīti-vrtti-lakşana (Kielhorn, Central Prov. Cat. p. 104) The author, also called Agnikumara, was the second son of Vallabhacārya the famous religious reformer, and brother of Gopinätha, and father of seven sons, Giridhara, Raghu- natha and others. He was born in 1515 A.D. For his other works, see Aufrecht i. 572ab, 135a, 225a, iii. 121a. His Śrngara-rasa-mandana (ed. Mulchand Tulsidas Telivala, with a Gujarati trs. Bombay 1919) in ten Ullāsas is not a work on Śrngāra Rasa, but an erotico-religious poem on Rādhā-Krsņa modelled obviously on Jayadeva's Gita-govinda and introdu- cing songs in rhythmic rhymed metres.

90. VIDYĀRĀMA

Rasa-dīrghikā (Peterson iii. no. 336 ; for a description of this work and extracts see BORI MS Cat. xii, no. 210, p. 240. MS incomplete) Nothing is known of the author, but the work in five Sopanas was composed in Samvat 1706 =1649-50 A. D.). It names Kavi-kalpalatā as one of the sources.

1 See Hultzsch 2112 ; SgS ii, pp. 203-5. See also Sgś ii, pp. 63-5, 67; Hultzsch iii, pp. viii-ix. 2 See K. P. Trivedi, introd. to his ed. of the Sad-bhāşā-candrikā (Bombay Skt. Ser. 1916) pp. 14-17. P. K. Gode (ABORI, xv, pp. 240-42) would place him in the 3rd quarter of the 16th century.

MINOR WRITERS ON ALAMKĀRA 301

91. VIŚVANĀTHA

Sāhitya-sudhā-sindhu (Ulwar Catalogue, extract 235 ; also Jammu Cat. no. 1254)

This South Indian author, who wrote in Benares, was son of Trimala or Trimalla Deva and grandson of Ananta of Dhārasura city on the Godāvarī. Stein's Kashmirian MS1 is dated in 1602 A.D.2 He quotes at the beginning of his work from Mammata and Bhoja, and elsewhere cites Candī- dāsa (probably the same as the commentator on Mammata) and Mahimabhatta. The work is in eight tarangas. Viśvanātha also wrote a drama called Mrgānka-lekhā3, a MS of which is dated Samvat 1664 (=1608 A.D.).

92. VIŚVANĀTHA NYĀYA-(or SIDDHĀNTA-) PAÑCĀNANA

Alamkāra-parişkāra This work is mentioned under Viśvanātha Nyāya-pañcā- nana in S. C. Vidyabhushana's Indian Logic p. 479 (also p. 392). The author was son of Vidyānivāsa Bhattācārya and a brother of Rudra Vacaspati. He composed his well known Vaiśeșika treatise Bhāșā-pariccheda in 1634 A.D. and also wrote a Pingala-prakāśikā. He was a native of Navadvipa (Bengal) and an adherent of the Navya Nyaya school of Raghunātha Siromaņi. See H. P. Sastri in JASB vi, 1910, p. 313.

1 Jammu Cat. p. xxix. 2 Stein speaks of a MS "transcribed from an autograph copy of the author. In the colophon referring to this original copy, which is added by another hand at the end of the Jammu MS, the date samvat 1659 (=A.D. 1602) can be made out with difficulty." 3 Sten Konow, Ind. Drama p. 113. The work has been published in the Sarasvati Bhavana Text Series, Benares.

302 SANSKRIT POETICS

93. VIŚVEŚVARA KAVICANDRA

Camatkāra-candrikā (IOC vii, p. 1507; Madras Trm Cat. 1916-19, 1918-19, R 2679) The author, a protégé of Singa-bhūpāla (1330 A. D.) wrote this work in eight Vilasas or chapters on principles of rhetoric, the illustrative verses being in praise of the author's patron (siņhabhūpāla-kīrti-sudhā-sāra-śitalā). He gives seven elements. of Camatkära in poetry, and the names of the chapters will sufficiently illustrate its scope. They are as follow: (i) Varna, Pada and Pada-doșas (ii) Vākya and Vākya-doșas (iii) Artha and Artha-dosas; varieties of composition (iv) Guņas; Rīti, Vrtti, Pāka and Śayyā (v) Rasa (vi) Šabdālaņkāras (vii) Arthālamkāras and (viii) Ubhayālamkāras. The work is note. worthy as one of the few Alamkära-treatises which generally follow Bhoja ; but the author does not do so in respect of the treatment of Rasa, eight of which he accepts, dismissing Sānta. Four Rītis are recognised, and called Asamāsā, Madhyama- samāsā, Atidīrgha-samāsā and Miśrā. He anticipates later writers in describing Rasa as Lokottarāhlāda, Anubhavaika- vedya and Vigalita-vedyāntara. It is perhaps the first work which makes an approach through Camatkra, on the basis of which poetry is classified into three groups: Camat- kāri (Śabda-citra), °kāritara (Artha-citra and Guņībhūta- vyangya) and kāritama (Vyangya-pradhāna). For detailed information and estimate of the work see V. Raghavan, ABORI, xvi (1934-35), pp. 131-39.

94. VIŚVEŚVARA BHATȚA

a. Alamkāra-kaustubha (with his own gloss, ed. Sivadatta and K.P. Parab, Nir. Sag. Press, Bombay 1898) b. Alamkāra-muktāvalī (ed. Visnuprasad Bhandari, Chowkhamba Skt. Ser. Benares 1927)

MINOR WRITERS ON ALAKĀRA 303

c. Alamkāra-(kula)-pradīpa (ed. Visnuprasad Bhandari, Chowkhamba Skt. Series, Benares 1923) d. Kavīndra-karņābharaņa (ed. in Kāvyamālā Gucchaka viii, 1891) e. Rasa-candrikā (ed. Visnuprasad Bhandari, Chow- khamba Skt. Series, Benares 1926) The author was son of Laksmidhara. He was born in Almoda ; hence he is called Parvatīya. He flourished in the first half of the 18th century and died about its middle1 at the age of 34. In the first work, the author refers to two dramas called Śrngāra-mañjarī (sațțaka)2 p. 347 and Rukmiņī-pari- naya pp. 381, 387 by himself ; and the latest writers that he quotes appear to be Appayya Dīksita and Jagannatha (both cited extensively). He cites also Mallinātha (p. 69) as a commentator on Dandin (see above p. 71), Candīdāsa (pp. 125, 166), Maheśvara (p. 49, 111) who is probably the com- mentator on Mammața, cited as Nyāyālaņkāra (p. 82), as well. as a work (p. 157) called Kāvya-dākinī 3. The Nyāyapañcā- nana, so extensively (eleven times) quoted, is probably Jayarāma Nyāyapañcānana (q.v.), another commentator on Mammata. He gives the name of his elder brother as Umā- pati (p. 357). In this work he deals with 61 poetic figures. The second work of our author, as he himself says, was written as an easier and briefer manual for beginners, after his first more extensive work. The third work Alamkāra- pradīpa deals entirely with poetic figures, enumerated as 119 and defined with illustrations. The fourth work deals in four chapters with enigmatology and Citra-kāvya (58 varieties). The fifth work deals with the different classes of heroes and heroines, and their characteristics. Viśveśvara was a scholiast of considerable activity and wrote a commentary called Vyangyārtha-kaumudī or Samañjasārthā on Bhānudatta's

1 See Kāvyamālā, Gucchaka viii, pp. 51-52 fn. 2 Also quoted in his Rasa-candrikā p. 90. 3 See above p. 276.

304 SANSKRIT POETICS

Rasa-mañjarī (see above p. 249). For his other works, see Aufrecht ii. 139b. The Kāvyamālā editors (Gucchaka viii, p. 52) mention two other works Kāvya-tilaka and Kāvya-ratna by Viśveśvara.

95. VIȘŅUDĀSA

a. Šiśu-prabodha Alamkāra (AFI 469) b. Kavi-kautuka cited by himself in chapter vii of the above work The author was son of Madhava, The Florentine MS, referred to above, contains only chapters 6 and 7, which deal with artha-guna and śabdālamkāra respectively.

96. VĪRA NĀRĀYAŅA

Sāhitya-cintāmaņi with a commentary (Madras Cat. xxii, 12265-68, extract) According to the colophon, Vīra Nārāyana is the author ; but in the work itself he is in the vocative case and praised, as in the Prataparudra. Vāmana Bhatta Bāna1 is the real author of the work which bears the name of his patron. The alleged author seems to be the same as the Reddi prince Vema of Kondavīdu (end of the 14th and beginning of the 15th century), the hero of the prose Vemabhūpāla-carita or Vīranārāyaņa-carita of Vāmana (or Abhinava) Bhatta Bāņa (ed. R. V. Krishnamachariar, Srivani-vilasa Press 1910). There is also a reference to Pedakomati Vema-bhūpala who is the same person. A commentary called Śrngāra-dīpikā on the Amaru-śataka is attributed to Vīranārāyaņa (Aufrecht ii. 141b) or Vema-bhūpāla (ibid i, 609b). The Sāhitya-cintāmaņi (also called °cūdāmaņi) consists of seven chapters2 which deal

1 For Vāmana Bhatța Bāņa, see introd to the Vani Vilas ed. of his drama Pārvatī-pariņaya. 2 In the Tanjore Catalogue, ix, no. 5308, p. 4100 the work is described as having thirteen chapters.

MINOR WRITERS ON ALAMKĀRA 305

with (1) dhvani (2) śabdārtha (3) dhvani-bheda (4) guņībhūta- vyangya (5) dosa (6) guņa and (7) alamkāra. Possibly this is the work cited under the name Sāhitya-cintāmaņī by Kumāra- svāmin (p. 97) and in the Vrtti-vārttika (p. 4).

97. VĪRESVARA PANDITA (BHAȚȚĀCĀRYA) surnamed Śrīvara Rasa-ratnāvalī (IOC iii, 1233/12576, p. 359) This Vireśvara is the son of Laksmana and father of Venīdatta, the last of whom, besides being the author of the Alamkāra-candrodaya, wrote a commentary on Bhānu's Rasa- tarangiņī (see above p. 250). The present work quotes Rudrabhatta's Śrngāra-tilaka, and is limited chiefly to Šrngāra- rasa and treatment of the Nāyikās.

98. VECĀRĀMA NYĀYĀLAMKĀRA

Kāvya-ratnākara The author was a Bengal writer, son of Rājārāma. He mentions this work in his Ananda-tarangini, which describes an itinerary from Chandernagar to Benares (Mitra 305). He also wrote a work on Jyotisa, and is probably identical with Vecārāma, who wrote a commentary on Deveśvara's Kavi- kalpalatā (see above p, 262).

99. VENKAPAYYA PRADHĀNA

Alamkāra-maņi-darpaņa (Rice 280) The author is known as Pradhāni Venkayāmātya of Mysore, ca. 1763-80 A.D.

100. VEŃKAȚA NĀRĀYAŅA DĪKȘITA

Śrngāra-sāra (Madras Cat. xxii, 12958-9, extract) Fbe author, son of Kameśvara Vaidika of the Godavarti family and Laksmi, refers in this work to his larger Srngāra- 20

306 SANSKRIT POETICS

sāravalī for fuller treatment. The present work consists of six ullāsas dealing with (1) kāvya-svarūpa (2) nāyaka-nāyikā- lakşaņa-vibhāga (3) nāyakādi-sahāya-nirūpaņa (4) rasa-bhāva- svarūpa (5) caturvidha-śṛngāra (6) daśarūpaka-svarūpa. The author is said to have composed works in eight languages.

101. VEŃKAȚĀCĀRYA (also called Kirīți Venkațācārya) surnamed Tarkālaņkāra Vāgīśvara Alamkāra-kaustubha

(New Catalogus Catalogorum i, p. 292-93) This writer, son of Annayārya Dīksita of Surapuram and of the Tirumala Bukkapattanam Śrīśaila family, should be distinguished from the poet Venktācārya (author of the Viśvagunadarsa) who was son of Raghunatha and grandson of Appayya. Our author was patronised by Venkata, son of Pāmi Nāyaka (died in 1802 A.D.). See Journal of Andhra Hist. Res. Society xiii, i, pp. 17 and 20-22.

102. VEŅĪDATTA SARMAN, TARKAVĀGĪSA BHATȚĀ- CĀRYA, surnamed Śrīvara

Alamkāra-candrodaya (IOC iii, 1198/235) This author, son of Vireśvara Śrīvara, also wrote a commentary on Bhānudatta's Rasa-tarangiņī (q.v.). His genealogy is given thus: Mahīdhara (a māntrika of Kāśīpati)->Kalyāņa->Laksmaņa->Vīreśvara. He had the surname Śrīvara and belonged to nāgacchatra-dhara-dvijottama- kula. The Alamkāra-candrodaya is in six ullāsas dealing with (1) kāvya-svarūpa (2) kāvya-vibhāga (3) doșa (4) guņa (5) alaņkāra and (6) upamā.

103. ŚANKHA, ŚANKHADHARA or SANKHACŪŅA (sometimes called ŚAMKARA), surnamed Kavirāja Kavi-karpațī or Kavi-karpațika-racanā

MINOR WRITERS ON ALAMKĀRA 307

(Jammu Cat. no. 1135 (p, 267) extract; BORI MS Cat. xii, nos. 42-46 ; extracts. Printed at Durbhanga 1892) The word kavi-karpați means "the ragged cloth of a poet," and the work is a strange effort at supplying a profuse stock of expressions which may be of use in poetic compositions for ideas of frequent occurrence. Various ways of expressing one and the same thought are indicated to suit various metres. The author, who also wrote the Latakamelaka-prahasana (ed. Durgaprasad and K. P. Parab, Nir. Sag. Press, Bombay 1889), was court-poet of mahāmāndalikādhirāja Govinda- nrpati, king of Kānyakubja, and wrote in the first half of the 12th century (about 1113-1143 A.D.). His verses are quoted in the anthologies of Sārngadhara (nos. 155, 3632) and Jahlaņa, and in the Sāhitya-darpana (ad iii. 219 p. 176, guror girah pañca, anonymously). The verse cited under Kārpațika in Kşemendra's Aucitya-vicāra (under śl. 15) is attributed to Mātrgupta by Kahlaņa (iii. 181) and in the Subhāşitāvali1 (3181).

104. ŚAMBHUNĀTHA

Alamkāra-lakșaņa

(Peterson v. 407 ; BORI Cat. xii, no. 19, p. 18)

105. ŚĀTAKARŅI

Cited as a writer on Dramaturgy (on Sūtra-dhära) by Śamkara in his commentary on Sakuntalā 2, and by Sāgara Nandin in his Nātaka-laksaņa-ratna-kośa (on Sūtradhāra).

1 See Peterson's paper on Aucitya-vicāra, 1885, p. 21. There is a Kavi-karpațikā of Vādīndra noticed in Tanjore Cat. vi, no. 3753-56 (pp. 2711-14) ; one of these MSS belonged king Serfoji who acquired it during his pilgrimage to Benares. 2 ABod 135a. Mention is also made of an anonymous work Kavi-kaņtha-hāra.

308 SANSKRII POETICS

106. ŚIVARĀMA TRIPĀȚHIN

a. Rasa-ratna-hāra and its commentary Lakşmī-vihāra (ed. Kāvyamālā Gucchaka 6, 1890, pp. 118-140; Jammu Cat. p. 273 (extract) b. Alamkāra-samudgaka, cited at the end of his Rāvaņa- puravadha, where he gives a list of his own 34 works. Stein p. 292. For the author, son of Krsnarama and grandson of Triloka- candra, and brother of Govindarāma, Mukundarāma and Keśavarāma, see JAOS xxiv 57-63. He appears to be a comparatively recent writer, quoting Paribhāşendu-śekhara, which alone will suffice to place him in the beginning of the 18th century. He is identical with the commentator on the Vāsavadattā (see Fitzedward Hall, Bibl. Ind. ed. 1859) ; for in this commentary he refers to his Rasa-ratna-hāra pp. 4, 9, 193, 206, 207. The present work, in 100 stanzas, deals with the characteristics of rasa and nāyaka-nāyikā, and quotes exten- sively Bhānudatta and Dasa-rūpaka. For his other works, see Aufrecht i. 652b, ii. 155b ; also Stein, Jammu Cat. p. 292. He also appears to have written a commentary (Vişama-padī) on Mammata (see above, p. 176) and a work on metrics called Kāvya-lakşmī-prakāśa or °vihāra, as well as a commentary on the Siddhānta-kaumudī, called Vidyā-vilāsa1,

107. ŚOBHĀKARAMITRA

Alamkāra-ratnākara (Ed. C. R. Devadhar, Poona 1942) The author belonging to Kashmir was son of Trayīśvara- mitra. The work is written in the form of Sütra (107 in num- ber), Vrtti and illustrations. The poet Yasaskara of Kashmir extracted the Sutras from this work and wrote his Devi-stotra for the purpose of illustrating them (Peterson i, pp. 77-78 ; extract p. 81). Ratnakantha (q.v.) appears to have commented

1 See P. K. Gode, Studies in Ind. Lit. Hist. ii, p. 237-41.

MINOR WRITERS ON ALAMKĀRA 309

upon both the Sutras and the Stotra. The date of Sobhākara is not known, but from the definitions and number of Alamkāras given, he appears to be a comparatively recent writer, considerably later than Ruyyaka whom he criticises. But as he is quoted by Jagannātha (p. 202=sūtra 11)1 and by Appayya (Vrtti-vārttika, p. 20), he is earlier than the end of the 16th century. It appears, however, that Jayaratha in his Vimarśinī commentary defends Ruyyaka against Sobhā- kara's attacks. Coming after Ruyyaka and preceding Jayaratha Sobhäkara probably belonged to the end of the 12th or begin- ning of the 13th century. The work deals entirely with poetic figures, the number of which is 109.

108. ŚRĪKAŅȚHA

Rasa-kaumudī (Aufrecht i. 494a=BORI MS no. 303 of 1880-81 ; Cat. xii, no. 347, p. 463f. Also H. P. Sastri, Cat. ASB MSS vi, no. 4931/8383, p. 481 ; the MS was copied in Samvat 1652=1596 A.D. The work combines Sahitya and Samgīta in ten chapters divided into two Khandas, Purva and Uttara. It was composed in 1575 A.D. The author was patronised by Śatrughna or Śatruśalya Jāma (Jam. Sattarsal) of Navanagar (1569 to 1608 A.D.).2

109. ŚRĪKARA MĪŚRA

Alamkāra-tilaka (Aufrecht i. 32a)

110. ŚRĪNIVĀSA DIKȘITA

a. Alamkāra-kaustubha (Aufrecht i. 31b) b. Kāvya-darpaņa (Rice 282)

1 See above p. 235 (under Jagannātha). 2 P.K. Gode in ABORI xii, 1931, p. 202-4 ; also xiv, 1933, p. 329 see MSS Cat. BORI, xii, pp. 463-66.

310 SANSKRIT POETICS

c. Kāvya-sāra-samgraha (Aufrecht i. 102b ; SCC vii 19) d. Sāhitya-sūkșma-saraņi (Rice 244) This author may be identical with Ratnakheta Śrīnivāsa, father of Rājacūdāmaņi Dīksita (q.v.). If this were so, then the Kävya-darpana above is the work of his son bearing the same title (see above p. 296), mistakenly entered here in most catalogues. As the first verse of the third work shows, it is three parts: (1) kāvya-lakșaņa-samgraha (2) varņa-sam- graha and (2) subhașita-samgraha. It quotes the Kāvya- prakāśa. About 1800 A.D.

111. SĀGARA NANDIN

Nāțaka-lakşaņa-ratna-kośa (Ed. Myles Dillon, Oxford Univ. Press 1937) The work is published from Devanāgarī transcript of a unique MS discovered by Sylvain Lévi in Nepal.1 As its name signifies, the work brings together a number of views of different notable writers on important dramaturgic topics. Its date2 is uncertain ; but as it cites from Rājaśekhara's Kāvya- mīmāmsā (p. 44, vilāsa-vinyāsa-kramo) it cannot be earlier than the first quarter of the 10th century. On the other hand, it is known to Rāyamukuța (1431 A.D.), Viśvanātha (between 1300 and 1350) and Bahurūpa Miśra (later than 1250 A.D.). The topics dealt with are: 1. Rūpaka and its ten varieties. 2-5. Five Avasthās, dialects to be employed, five Artha-prakrtis. 6-10. Five Upakşepakas, five Samdhis, 21 Pradeśas of Samdhi, four Patakasthana, Vrttis and their division. 11. Excellences of the Nayaka. 12-13. Thirty-six Nātya-lakșaņas, ten Guņas,

1 S. Lévi in JA, xciii, 1923, p. 210f. 2 For a discussion of date see P. K. Gode in ABORI, xix, 1938, pp. 280-88 (Studies, i, pp. 48-56) ; M. Ramkrishna Kavi in NIA ii, p. 412-19. For textual study see V. Raghavan in Journal of the Univ. of Gauhati iii, 1952, pp. 17-33 and Annals of Orient. Research, Madras Univ., xvi, 1958-59.

MINOR WRITERS ON ALAMKĀRA 311

thirty-four Natyalamkaras. 14-16. Rasas and Bhavas. 17. Types of Nayika and their excellences. 18. Minor forms of Rupaka. The work is important not only for its collection of various views on these topics, but also for its citation of a large number of dramatic and dramaturgic works.1

112. SĀMARĀJA DIKȘITA

Śṛngārāmrta-laharī (ed. Kāvyamāla Gucchaka xiv. MSS: Jammu Cat. no. 1243 ; Madras Cat. xxii, 12961) The author, also called Syāmarāja, was son of Narahari Bindupurandara, and wrote also Tripura-sundarī-mānasa- pūjana-stotra (ed. Kāvyamālā Gucchaka ix) and other poems. He lived in Mathura at the latter part of the 17th century. His son Kamarāja, whose Śrngāra-kalikā-kāvya is published in Kāvyamālā Gucchaka xiv, as well as a Prahasana named Dhūrta-nartaka2 was also a poet; while his grandson Vrajarāja and his great-grandson Jīvaraja wrote commentaries on the Rasa-mañjarī and Rasa-tarangiņī of Bhānudatta respectively (q.v.). The present work deals with Rasa, especially Śrngāra, after Bhanudatta. Our author wrote his drama Śridāma-carita in 1681 A.D. for the Bundela-prince Anandarāya. His son Kāmarāja also appears to have written a Kāvyendu-prakāśa in 15 Ulläsas (Kalās), which is apparently the same work as entered anonymously in Bhandarkar, Rep. 1887-91, no. 601 and in BORI MSS Cat. xii, no. 142, pp. 158-60, which see for information about this work. Sāmarāja, who wrote Rati-kallolinī in 1719 A.D. but who does not give his parentage, is probably a different person.3

1 For an index of authors and titles, see ed. as above pp. 145-47 and P. K. Gode as cited above p. 485 fn. As authors on Nātya-śāstra are mentioned Aśmakutta (lines 83, 437, 2766, 2775), Cārāyaņa (1. 392 ; also mentioned in Vatsyāyana's Kāma-sūtra i. 1.12 ; i. 5. 22), and Bādara. 2 Wilson ii. 407; Keith, Sanskrit Drama, pp. 262-63. 3 See P. K. Gode in ABORI x, pp. 158-59.

312 SANSKRIT POETICS

113. SĀYAŅA

Alamkāra-sudhānidhi Cited by Appayya Dīkşıta and Kumārasvāmin ; see above p. 225 fn 3.

114. SUKHADEVA MIŚRA

Śrngāra-latā (Aufrecht i. 661a) A bhāşā-work, called Rasārņarva, by Sukhadeva is men- tioned in Peterson iv, no. 770 (App. p. 29).

115. SUKHALĀLA

Alamkāra-mañjarī (AFl 213) The author, pupil of Gangeśa Miśra and his son Hari- prasāda (q.v.), professes to follow the Kārikās of Jayadeva. Aufrecht thinks that he must have flourished about 1740 A.D. The work begins with upamā and takes up rūpaka, pariņāma, smtimat, bhrāntimat, samdeha, utpreksā, where the MS breaks off. A Kāvya called Śrngāra-mālā, composed in Samvat 1801=1745 A.D. by Sukhalāla, son of Bābūrāya Miśra, is entered in Stein 75 and Ulwar Cat. no. 1083 (extract 230).

116. SUDHĀKARA PUŅDARĪKA YĀJIN Śrngāra-sārodadhi (Aufrecht iii. 137b)

117. SUDHĪNDRA YOGIN or YATI

Alamkāra-nikaşa (Madras Cat. xxii, 12976, extract) Alamkāra mañjarī with commentary, Madhu-dhāra by Sumatīndra (Tanjore Cat. ix, 5129-30) The Alamkāra-nikaşa is a short work on Arthālamkāra. The colophon gives the author's name as above; but in the work it is said that the author, following the views of ancient and

MINOR WRITERS ON ALAMKĀRA 313

modern authorities on the subject, deals with and illustrates the Arthālamkāras by means of examples eulogising the virtues of Sudhīndra Yati himself who is made out to be a follower of the Mädhva sect. He is probably the same as Sudhīndra Yati, disciple and successor of Vijayīndra Yati (d. 1623 A. D.). The Alamkāra-nikarşa by Sudhendra in Oppert 4797 is probably this work. The Alamkāra-mañjarī appears to be separate work by Sudhindra Yati in which the illustrative verses are in praise of the teacher Vijayīndra. Most of the MSS contain the Sabdālamkāras only. There is a commentary on this work called Madhu-dhāra by Sumatīndra, a successor of Sudhindra. We are told that Sudhindra was living in the Tanjore district in the 17th century. A drama called Subhadrā-pariņaya is attributed to Sudhīndra Yati in Madras Cat. xxi no. 12729 and a drama called Subhadrā- dhanañjaya to Vijayīndra Yati in ibid no. 12728.

118. SUNDARA MIŚRA AUJĀGARI Nātya-pradīpa (Aufrecht i. 284b, 791a) The work is dated in 1613 A.D. It is cited by Rāghavabhatta on Śakuntalā (ed. N.S.P. 1886, p. 6). This work repeats verbatim a large portion of the Daśa-rūpaka (see pref. to Hall's ed.). In the work itself the Sāhitya-darpana is referred to. The author is the same as Sundara Miśra who wrote the Abhirāmamaņi-nāțaka in seven acts in 1599 A. D. which is mentioned as his own (ABod 137b-138a ; Kielhorn Central Prov. p. 68 ; Wilson ii p. 395). See IOC iii, pp. 347-48, no. 1199/1148d (extract).

119. SOMANĀRYA Nāțya cūdāmaņi (Madras Cat. xxii 12998, with a Telugu commentary) This is a very recent work on dancing and music1. The

1 See V. Raghavan on Later Samgīta Literature in the Journal of Madras Music Acad. iv.

314 SANSKRIT POETICS

author is described as one famous for astāvadhana (attention to eight things at a time).

120. HARIDĀSA

Prastāva-ratnākara (Weber 827; Aufrecht i 360a, ii 212a iii 77a) The author was son of Purusottama of the Karana family, and the work is a metrical compilation dealing with kūta, samasyā and enigmatic composition in general, as well as with miscellaneous subjects including Nīti. Jyotisa etc. It was compiled in 1557 A.D.

121. HARIPRASĀDA MĀTHURA

a. Kāvyārtha-gumpha (Aufrecht ii. 20b; BORI MS Cat. xii, no. 131, p. 145, the MS is dated Samvat 1775) b. Kāvyāloka (Aufrechti. 103a, extract in Petersoniii p. 356-7) The second work in seven prakāśas is dated in Samvat 1734=1728 A.D. A MS of the first work bears the date 1775 which is possibly the date also of its composition. Hariprasada is also author of a work on ācāra (māsādi-nirūpaņa), see Peter- son iv, p. cxxxvii. A Kāvyāloka is cited by Appayya in his Citra-mīmāmsa ; but this must have been an earlier work. The Kāvyāloka cited by Kumārasvāmin p. 73 refers the Dhvanyā- loka (p. 221), and is not, as supposed by Harichand Sastri (p. 27, no. 234) a reference to the present work. The author was son of Māthura Miśra Gangeśa (see above under Sukhalāl).

122. HARIHARA

a. Śrngāra-bheda-pradīpa (Burnell 59a) b. Bindvalamkāra, cited in Ekāvalī p. 242 (on Dīpaka) One Harihara is referred to in the Ekāvalī p. 19 as having received amazing wealth from king Arjuna, who is supposed

MINOR WRITERS ON ALAMKĀRA 315

by Bhandarkar and Trivedi (see above p. 206) to be identical with Arjunavarman of Mälava, whose earliest and latest known dates are 1211 and 1216 A.D. If this Harihara be our author, then his date will be the first quarter of the 13th century. The Śrngara-bheda° deals, among other topics, with the ten Avasthas of Vipralambha Śrngāra, as we know from a passage which is quoted by Viśveśvara in his Rasa-candrikā (p. 55).

123. HALADHARA RATHA

Kāvya-tattva-vicāra (H. P. Sastri's Report, 1895-1900, p. 16).

( 2 )

ANONYMOUS WORKS

We give below a list of some minor works on Alamkāra, of which the names of the authors are unkown or uncertain:

1. Alamkāra-kārika. Aufrecht i. 31b. 2. Alamkāra-kaumudī-vyākhyā. Madras Cat. xxii, 12784. Neither the name of the author of the original treatise nor that of the commentary is given. It treats of poetic figures. 3. Alamkāra-candrikā. Rice 284 (Aufrecht i. 32). 4. Alamkāra darpana in Prakrit, consisting of 134 Ślokas. devoted to the treatment of poetic figures. Monatsber. Berl. Akad. 1874, 282. 5. Alamkāra-prakaraņa. Śgš i, no. 52. 6. Alamkāra-prakāśikā. Madras Cat. xxii, 12791. It deals with poetic figures and quotes from the Kāvya- prakāśa. 7. Alamkāra-mayūkha. Oppert 1754 (Aufrecht i. 32). 8. Alamkāra-vādārtha. Śabda-bheda discussion, starting. with the Sahitya-darpana. HPS i, 12. 9. Alamkāra-samgraha. Madras Cat. xxii, 12795. It enumerates and classifies the various poetic figures. 10. Alamkāra-sarvasva. Madras Cat. xxii, 12798 (MS. incomplete). The author says that his teacher composed a work on Alamkāra in praise of king Gopāladeva. It is a treatise on the general topics of Alamkāra, but the MS is incomplete, breaking off with the Guņa-prakarana. It appears to be a recast of Pratäparudrīya. For the uncertain name of the author see V. Raghavan, New Cat. Cat. i, 2976 ; Number of Rasas. p. 50, also note in Addendum. 11. Alamkārānukramaņikā. Oppert 5489 (Aufrecht i. 32b).

MINOR WRITERS ON ALAMKĀRA 317

12. Alamkāreśvara, cited by Sivarāma on Vāsavadattā p. 4. 13. Kavi-kantha-pāśa. See above p. 278. 299 fn. Madras Cat. xxii, 12802-03. 14. Kavi-kalpalatikā. Burnell 54a. 15. Kavi-saraņa-dīpikā. A work on poetic composition by one Ratneśvara. H. P. Sastri, Cat. ASB MSS. vi. no. 4915A/8069, pp. 471-73. 16. Kāvya-kalāpa. Aufrecht i. 100b. 17. Kāvya-kaustubha. Oppert ii. 3616 (Aufrecht i. 101a). 18. Kāvya-dīpikā, Oppert 541, 636 ; Madras Cat. xxii. 12815. A compilation for beginners. Probably the same as Kānticandra's Kāvya-dīpikā (see below). 19. Kāvya-pariccheda, Oppert ii. 8727. 20. Kāvya-ratna. Oppert ii. 6237. See above p. 220. 21. Kāvya-lakșaņa. Madras Cat. xxii, 12829. It is based on the Kārikās of Kāvya-prakāśa, but adds a section on dramaturgy (cf. Oppert i. 1793 and ii. 6238). 22. Kāvya-lakşaņa-vicāra. Madras Cat. xxii, 12979. A comparatively modern work on the general topics of Alam- kāra, citing the Citra-mīmāmsa and the Rasa-gangadhara. 23. Kāvyāmrta-tarangiņī. See above p. 177. 24. Kāvyopadeśa, cited by Hemādri on Raghu (Aufrecht i. 103a). 25. Daśarūpaka-vivaraņa. Madras Cat. xxii, 12892, This work is not a commentary on the Daśa-rūpaka, but a short treatise, mostly in the nature of a compilation, explain- ing the characteristics of dramatic composition. It may have formed the Nataka-section of some comprehensive work on Poetics. It refers to Daa-rūpaka by name. 26. Nāțaka-ratna-kośa, cited by Rāyamukuța and Bhānujī. ABod 182b. This may be the Nātaka-lakşaņa-ratnakośa of Sāgara Nandin, see above p. 310. 27. Nāțakāvatāra, mentioned by Mohanadāsa (q.v.). ABod 142a.

318 SANSKRIT POETICS

28. Nātya-darpana, cited by Ranganātha on Vikramor- vašīya (ed. N.S.P. 1914, p. 7) and Bharatamallika on Bhatti xiv. 3. See p. 297, no. 82 above. 29. Nātya-sarvasva-dīpikā. BORI MS no. 41 of 1916-18 (Cat. xii, no. 344, p. 453). Purports to be a comm. on the so-called Adibharata in 5 Skandhas, 32 Adhyayas and 221 Prakaranas, of which a fragment is found in a Mysore MS. named Adibharata. For an account of this MS and the work see S. K. De. The Problem of Bharata and Adi-Bharata in Some Problems of Skt. Poetics, Calcutta 1959, pp. 156-76. 30. Rasa-kalikā, cited by Vāsudeva on Karpūra-mañjarī (Aufrecht i 494a). For a Rasa-kalikā by Rudrabhatța see V. Raghavan Number of Rasas, p. 53f. This work is found in two MSS in the Govt. Orient. Library, Madras (nos. R. 2241 and 3274). It is identical with the work cited by Vasudeva, for all the six verses quoted by him are found in it. 31. Rasa-kaumudī. Peterson v, no. 414. P. K. Gode (Cal. Oriental Journal iii, pp. 35-37) gives the latter half of the 18th century A. D. as the probable date of this anony- mous work. 32. Rasa-gandha. Rice 286 (Aufrecht i. 494b). 33. Rasa-gandhāra. Aufrecht i. 494b (may be a mistake for Rasa-gangādhara of Jagannātha), 34. Rasa-ratnākara, cited by Mallinātha on Kirāta ix. 71 and on Meghadūta (ed. Nandargikar, 1894, pp. 64, 67. 85, 91). Aufrecht, i. 496a (commentary by Hrdayarāma Miśra). 35. Rasa-ratna-kośa, anonymous, is mentioned in Aufrecht i. 495b ; but it may be Kumbha's work of the same name (see p. 271-72). 36. Rasa-bindu and Rasāmrta-sindhu, Kathvate no. 703. and 707; BORI MS Cat xii, no. 212, pp. 245-46. 37. Rasa-viveka. Madras. Trm C 589. (Cf Oppert 5144). 38. Rasa-samuccaya. Aufrecht i. 496b. 39. Rasa-sāgara, cited by Mallinātha on Śiśu xv. 89. 40. Rasa-sudhākara, cited by Mallinātha on Raghu vi. 12g-

MINOR WRITERS ON ALAMKĀRA 319

It is Rasārņava-sudhākara of Singa-bhūpāla which Mallinātha cites in his com. on Kumara. For this work see above p. 239. 41. Rasākara, cited by Mallinātha on Megha-dūta (ed. ibid, p. 87, 97). 42. Rasika-sarvasva, cited by Nārāyaņa on Gīta-govinda v. 2; also by Rucipati in his comm. on Anargha-rāghava: (NSP ed.) p. 13. 43. Rahasya, probably an abbreviation of some more definite title, cited by Mallinātha on Kirāta iii. 60, xiv. 40, on. Śiśu xiii. 10. 44. Śrngāra-kaustubha. Rice 288 (Aufrecht i. 660b). 45. Śrngāra-candrodaya, cited in Prastāva-cintāmaņi. Weber i, p. 229. 46. Śrngāra-tarangiņī. Oppert 2465; Rice 288 (= Aufrecht i. 660b). 47. Śrngara-pavana. Oppert 5766 (Aufrecht i. 661a). 48. Śrngāra-mañjarī. Aufrecht i. 661a. 49. Śrngara-vidhi. Oppert 5680 (Aufrecht i. 661a). 50. Śrngāra-ratnākara. Aufrecht ii. 158a. It is not always clear, from the citations or descriptions. in the catalogues, whether some of the works on Rasa and Śrngāra noted here are really works on Poetics or partake of the nature of erotic Kävya. But care has been taken to ex- clude the latter wherever possible.

Among recent publications (in Sanskrit) on Poetics, pro- duced late in the 19th century, may be mentioned : (1) Alamkāra-sūtra by Candrakānta Tarkālaņkāra, a Bengal Pandit who lived within living memory. (Publ. Cal- cutta 1899). (2) Yaśovanta-yaśo-bhūşaņa by Pandit Rāmakarņa in praise of a native prince of Rajaputana of that name. (Publ. Godhapur 1897). (3) Alamkāra-maņi-hāra by Śrīkrsņa Brahmacārin, pub- lished in Mysore Govt. Oriental Series in 4 vols, nos. 51, 85, 68, 72. The author is called Krsnabrahmatantra Parakāla --

320 SANSKRIT POETICS svāmin, a recent pontiff of the Parakāla Srīvaisnava Math in Mysore. (4) Kāvya-dīpikā of Kānticandra Mukhopādhyāya Vidyā- ratna (ed, Calcutta 1870, 1886, with a comm. by Jivananda Vidyasagar 1919 ; ed. Haridatta Sastri, Lahore 1939, with Skt. and Hindi commentary). It is a compilation for begin- ners from Mammata and other authors by a modern writer belonging to the 19th century. (5) Alamkāra-sāra-mañjarī with Skt. text and Hindi comm. by Narayan Sastri Khiste, ed. Narahari Sastri Thatte, Chowkhamba Skt. Series 1933.

CONCLUSION

(1 ) An attempt has been made in the foregoing pages not only to indicate the diversity as well as immensity of Sanskrit Alamkāra literature, but also to settle its relative chronology as a workable basis for an historical treatment. If we leave aside its unknown beginnings and Bharata, the historic period of its growth covers broadly a thousand years from 800 to 1800 A.D. It is marked by a speculative activity, surprising alike for its magnitude and its minuteness. This activity in its early stage centres in Kashmir, to which place belong most of the famous and original writers on Poetics. We do not indeed know the place of origin of the two earliest writers, Bharata and Bhämaha, but immediately after them we find Vāmana, Udbhața, Rudrața, Mukula, Ānandavardhana, Lollața, Bhatta Nāyaka, Abhinavagupta, Kșemendra, Kuntaka, Mammata and Ruyyaka flourishing in Kashmir. The only important exception is found in Dandin who was probably a South Indian writer. Coming to later times we find the study extending itself to Central India, Gujarat, the Dekkan and Bengal. In South India, no doubt, this study was kept alive by a succession of brilliant, if not very original, writers ; but these contributions of later times, though greater in bulk and sometimes superior in a certain acuteness, never supersede the volume of original work done in Kashmir, which may be fit- tingly regarded as the home-land, if not the birthplace, of the Alamkara-sāstra. The writers of Central India, Gujarat, the Dekkan and Bengal only carry on the tradition, as well as acknowledge the authority, of the Kashmirian originators of the discipline. ( 2 ) Although our history covers a period of more than a thousand years, it is yet marked by several well-defined stages. 21

322 SANSKRIT POETICS

With the date of Anandavardhana, we arrive for the first time at a distinct landmark in its chronology as well as its. history ; and we may take it as the central point from which we may proceed backward and forward, although the system of Anandavardhana itself was raised to almost exclusive re- cognition by the classical work of Mammata. The mutual relation of this system to the other systems flourishing before and after Änandavardhana furnishes the best and safest cri- terion for the orientation of the divergent streams of thoughts and tendencies, which gather together in one clear, dominant and finally authoritative doctrine in Mammata. Indeed, one- of the obvious objects of Anandavardhana's work was not only to fix the new principle of Dhvani in poetry, but also to work up and rationalise into a synthetic and comprehensive system the already accumulated ideas, elaborated by previous. thinkers but flowing through different channels in the res- pective systems of Bhamaha, Vamana and the post-Bharata dramaturgic Rasa-writers; while Mammata gathered the- results up and uttered them in the convenient and concise: form of a systematic text-book.

( 3 ) Although in Bhamaha's Kāvyālamkāra, the earliest known work on Poetics, we meet for the first time with a more or less systematic scheme of Poetics, there is enough evidence to show that it must have been preceded by a period, covering perhaps several centuries, of unknown beginnings. All that we know of this period consists of glimpses of rhetorical speculations, such as we find in Bharata, in the recorded opinions of (or stray references to) pre-Bhämaha writers like Medhāvin, or in such treatises on Alamkāra as was presumab- ly utilised by the Kāvya-poets in general and by Bhatti in particular. This period begins with the enumeration and definition of only four poetic figures, ten Gunas and ten Doșas, but ends with the elaborate characterisation of thirty-eight independent figures in Bhatti. But what is important to note

CONCLUSION 323

in this period is Bharata's more or less elaborate exposition of Dramaturgy, and incidentally of Rasa, which element, how- ever, is considered not in relation to Poetry and Poetics, but in connexion with Drama and Dramaturgy. This is followed by a comparatively brief but important period of extraordinary fertility and creative genius, beginning with Bhämaha and ending with Anandavardhana, in which we find most of the fundamental problems of Sanskrit Poetics discussed and settled in their general outlines. We have, on the one hand, Bhämaha, Udbhata and Rudrata, devoting them- selves to the consideration of those decorative devices of poetic expression which are known as Alamkāras (poetic figures), and confining themselves chiefly to an external art or theory of adornment, from which the discipline itself takes its name and its original tradition. Dandin and Vämana, on the other hand, emphasise in poetry the objective beauty of representa- tion realised by means of what they call Märga or Rīti (roughly 'diction') and its constituent excellences, the ten Gunas. Both these systems, which emphasise respectively the elements of Alamkära and Rīti in poetry, content them- selves with the working out of the outward forms of expression, the advantages of which were considered sufficient for poetry. They point out the faults to be avoided and the excellences to be attained, and describe the poetical embellish- ments which should enhance its beauty, insomuch so that the whole discipline came to receive the significant designa- tion of Alamkära-āstra or the Science of Poetical Embellish- ment. Side by side with these early writers, however, we have the commentators on Bharata (like Lollata, Śankuka and others) who were bringing into prominence the aesthetic importance of Rasa, the consideration of the moods, senti- ments and feelings, which we find reacting upon and influen- cing even the theorists of rival persuasion (e.g. Dandin, Udbhata, Vāmana and Rudrata) who betray themselves more and more alive to the significance of this element in poetry.

324 SANSKRIT POETICS

But the discussion of Rasa appears to have been, so far, confined chiefly to the sphere of the dramatic art, and its bearings on poetry were not fully realised until the Dhvanikāra and Anandavardhana had come into the field. These new theorists, headed by Anandavardhana, maintain that no system of Poetics, like no system of Dramaturgy, can entirely ignore the moods, feelings and sentiments as essential factors in poetry, and must therefore find an important place for Rasa in its scheme. What was thus already established in the drama was taken over and applied to poetry, profoundly modifying, as it did, the entire conception of the Kävya. The Rasa came to be considered as the "essence" (ātman) of poetry ; and in order to harmonise it in poetic theory, the new school evolved a theory of "suggestion" (dhvani) as the means of its expression. Not satisfied, however, with working up the concept of Rasa into their system, the new theorists devoted themselves to the examination of the already accumulated ideas of Alamkāra and Rīti (with its constituent Guna and Dosa), with a view to correlate them to the new idea of Dhvani (and Rasa), and thus by synthesis evolve a comprehensive theory of Poetics. The interval between Anandavardhana and Mammata was taken up in settling precisely the details of the new system, which was raised to almost exclusive recognition by the final text-book of Mammata. Its success was so complete that the new concept of Dhvani was unquestionably accepted by most later writers, and the systems which emerged after Mammata could no longer be strictly regarded as entirely independent systems.

( 4 ) But a new theory, however systematic or comprehensive, is never accepted without some opposition. Ānanda- vardhana's system, no doubt, absorbed and overshadowed in course of time all the earlier systems ; but in the interval between Änandavardhana and Mammata, while it was still

CONCLUSION 325

striving for supremacy, we find a few vigorous but short-lived reactionary movements which refused to accept Ananda- vardhana's new interpretation. Thus, we have Kuntaka who strove to make Bhamaha's concept of Vakrokti elaborate and comprehensive enough to include the new ideas; Bhatta Näyaka who raised his voice on behalf of the Rasa-systems against their acceptance ; and Mahimabhatta who attempted to settle the new concept of Dhvani with the technical process of logical inference. All these writers, however, do not deny the newly established doctrine of Dhvani, but they try to explain it in terms of already recognised ideas. In spite of these nonconformist schools, however, whose feeble opposi- tion languished for want of support even in the time of Mammata, the system of Poetics, as finally outlined by Anandavardhana and worked out in detail by Mammata and his followers, was established without question in almost all writings from the 12th century downwards. Here and there we have some surviving exponents of some old tradition, like the Vagbhatas or the followers of Bhoja, as well as specialised departments which stood apart like the group of Kavi-śiksā- writers or the erotic Rasa-writers; but in the main, the creative days of the science were over, and no new theory forthcoming, the system of Anandavardhana, as represented by Mammata, reigned supreme, even influencing, to an obvious extent, the writers who would pretend to stand apart.

( 5 ) These considerations, which will become clearer as we proceed in our study of details in the next volume of this work will enable us to fix the rough outlines of the history of Sanskrit Poetics and divide it, for convenience of treatment, into several periods in conformity to chronology and the stages of development through which its doctrines passed. The dim beginnings of the discipline, like the beginnings of most other departments of Indian speculation, are hidden from us, until it issues forth in the works of Bharata and

326 SANSKRIT POETICS

Bhamaha in a more or less self-conscious form. Then starts a period, ending with Anandavardhana, which may be characterised as the most creative stage in its history, a stage in which the dogmas and doctrines of the different systems were formulated and settled in their general outlines, giving us at least four different systems which emphasise respectively the theories of Rasa, of Alamkara, of Rīti and of Dhvani in poetry. To this period belong Bhāmaha, Udbhata and Rudrata, Dandin and Vämana, the commentators on Bharata (Lollata, Śankuka and others), the Visnu-dharmottara and Agni-purāņa, and lastly, the Dhvanikara and Anandavardhana. Between Anandavardhana and Mammata, we have a third definitive period which ends with the ultimate standardisation of a complete scheme of Poetics, with the Dhvani-theory in its centre, in which the divergent gleams of earlier speculations are harmonised into a focus, and which finds itself finally set forth in a well-defined and precise form in the text-book of Mammața. To this period also belong reactionary theorists, like Kuntaka and Mahimabhatta, as well as Bhoja who carries on the same tradition as that of the Agni-purāna, and Dhanañjaya who writes on Dramaturgy. The period which follows this is necessarily a scholastic period of critical elaboration, the chief work of which consists in summarising and setting forth in a systematic form (generally after Mammata) the results of the final speculations, and also in indulging in fine distinctions and hair-splitting refinements on minute questions. This stage, therefore, is marked by great scholastic acumen, if not by remarkable originality or creative genius ; but at the same time it denotes a progressive dete- rioration of the study itself. The branching-off of some specialised and practical groups of writers from the main stem is to be explained as due rather to this degenerate spirit of the times than to any real split in the domain of poetic theory or to any desire for independent thinking. It is also the age of numberless commentators, and of com- mentators on commentators, who busied themselves with the

CONCLUSION 327

hardly inspiring task of explanation, of expansion or restric- tion of the already established rules. We have also now a number of popular writers who wanted to simplify the study for general enlightenment, the lowest stage being reached when we come to the manuals and school-books of quite recent times.

( 6 ) We may, therefore, conclude here by broadly indicating the bearings of the chronological results of this volume on our enquiry in general, in the light of which (as well as in the light of what follows in the next volume) we may tentatively put forward a rough division of the different periods of our history, noting the different groups of writers comprised in them, with a view to facilitate the study of the problems which will con- front us in the next volume: I. From unknown Beginnings to Bhamaha. (Formative Stage). II. From Bhamaha to Anandavardhana. Circa middle of the 7th to the middle of the 9th century. (Creative Stage). (1) Bhāmaha, Udbhața and Rudrața (alamkāra- theory). (2) Dandin and Vāmana (rīti-theory). (3) Lollata, Sankuka, Bhatța Nāyaka and others (rasa-theory). (4) The Vişnu-dharmottara and Agni-purāņa. (5) The Dhvanikāra and Ānandavardhana (dhvani- theory). III. From Anandavardhana to Mammata. Circa middle of the 9th to the middle of the 11th century. (Definitive Stage). (1) Abhinavagupta (2) Kuntaka (3) Rudrabhatta (4) Dhanañjaya and Dhanika

328 SANSKRIT POETICS

(5) Bhoja (6) Mahimabhatta IV. From Mammata to Jagannātha. Circa middle of the 11th to the 18th century. (Scholastic Stage), (1) Mammața, Ruyyaka and Viśvanātha (including Hemacandra, Vidyādhara, Vidyānātha, Jaya- deva, Appayya and others). (2) The Vāgbhatas and Keśava Miśra. (3) The writers on Rasa, especially Sṛgāra : Śāradā- tanaya, Śinga-bhūpāla, Bhānudatta, Rūpa Gosvämin and others. . (4) The writers on Kavi-śiksā: Rājaśekhara, Kşemendra, Arisimha and Amaracandra, Deveśvara and others. (5) Jagannātha.

( 7 ) Looking at the question from another point of view, we may classify the systems of Poetics broadly into (1) Pre- dhvani (2) Dhvani and (3) post-Dhvani systems, taking Dhvani-theory as the central landmark. In the Pre-dhvani group, we include all writers (flourishing before Ananda- vardhana), mentioned in Groups I and II above, with the exception of the Dhvanikara and Anandavardhana, with whose names the Dhvani-system is associated. In the Post-dhvani systems may be comprised the followers of the Dhvani-system from Mammata to Jagannātha, together with reactionary or unorthodox authors like Kuntaka or Mahimabhatta, as well as the writers on Śrngāra and on Kavi- sikşā. On the other hand, the systems of Poetics have been grouped, on the basis of the particular theory emphasised by a particular group of writers, into (1) the Rasa School (2) the Alamkāra School (3) the Rīti School and (4). the Dhvani School. The convenience of this classification is obvious, but it is doubtful whether we may safely apply the term "school" to indicate affiliation to a particular

CONCLUSION 329

system of opinion,1 when we consider that one has to admit a great deal of mutual and (to a certain extent) inevitable contamination of the different "schools", which makes the existence of any particular school by itself almost impossible. Thus, the "Dhvani School" admits Rasa and Alamkāra as important factors of poetry, which are thus not exclusively monopolised by the so-called Alamkara and Rasa Schools. It is doubtful, again, if a Rasa School, properly so-called was at all founded by Bharata, who is taken as its original exponent, or a similar Alamkāra School by Bhāmaha. All that we can say is that Bharata and Bhamaha laid stress on the elements of Rasa and Alamkära which became in course of

1 Sovani in Bhandarkar Commemoration Volume pp. 387f. Reliance has been placed on Ruyyaka's review of previous opinions and Samudrabandha's classification. But Ruyyaka only takes the concept of dhvani or pratiyamāna artha as the starting point and considers how far it was accepted, explicitly or implicitly, by his predecessors. Samudra- bandha, commenting on this passage, speaks of five paksas or theses, including the dhvani-theory (which he calls the last paksa) with which his author identifies himself. His classification is based upon the conventional theory that poetry consists of a "special" disposition of word and its sense (visista śabda and artha). This speciality, in his opinion, may be realised by putting emphasis on their (1) dharma (inherent characteristic) (2) vyāpāra (operation) and (3) vyaňgya (suggestiveness). In the first case, the dharma may proceed from alamkāra and guņa (i.e. rīti). In the second case, the vyāpāra may consist of bhaniti-prakāra or bhogīkarana. Thus, we get five standpoints associated respectively with the names of Udbhata, Vāmana, Kuntaka, Bhatta Nayaka and Anandavardhana. This classification, though very significant, is obviously overlapping and historically incorrect. The vyañjanā, it may be objected, which is taken as one of the bases of differentiation, is admittedly as much a vyāpāra as bhaņiti postulated by Kuntaka. Besides, Kuntaka, as a matter of fact, developes Bhāmaha's idea of vakrokti as bhaniti-vaicitrya. and therefore may be properly included among those who put emphasis on alamkāra. Similarly, Bhatta Nāyaka describes bhoga as a peculiar function (dharma) or process by which Rasa is said to be realised; and therefore he is in reality an exponent of the Rasa-theory as an interpreter of Bharata's dictum on Rasa.

330 SANSKRIT POETICS

time established ideas in the realm of Poetics. As a rule, each great writer who puts forward, consciously or un- consciously, a new theory, takes over from his predecessors those ideas which have stood the test of criticism and which he can combine in a self-consistent system of his own. In this way, really valuable ideas have been generally adopted, although sometimes other ideas, perhaps of the same author, have by common consent been rejected. This is illustrated by the case of the Vakrokti-jīvitakāra, whose theory of Vakro- kti was universally rejected, although the main principle (analysis of an Alamkära) for which he was contending is accepted by Ruyyaka and others. It is not maintained here that the history of Sanskrit Poetics consists of only one stream of development, and that within it we have mere currents and counter-currents. The latter were indeed very important, but they never succeeded in forming into separate rivers; and the different channels originating independently or breaking away from the main course ultimately merge into one dominant and clear stream.

ABBREVIATIONS

ABORI=Annals of the Bhandarkar Oriental Research Insti- tute, Poona. Aufrecht=Th. Aufrecht's Catalogus Catalogorum i-iii. Leip- zig 1891-1903. ABod=Aufrecht's Catalogus Codicum Manuscriptorum Sans- criticorum Bibliothecae Bodleianae. Oxonii 1864. AFI=Aufrecht's Florentine Sanskrit MSS. Leipzig 1892. ALeip=Aufrecht's Katalog der Sanskrit-Handschriften der Universitäts-Bibliothek zu Leipzig. Leipzig 1901. Bendall=C. Bendall's Catalogue of Sanskrit MSS in the British Museum. London 1902. Bhandarkar. R. G. Bhandarkar's Reports on the search of Sanskrit MSS are quoted with reference to the particular years of operations, as indicated on the respective title- page of the Reports. Other lists by him are cited as in Aufrecht. Sridhar Bhandarkar's Reports and Catalogues are separately referred to. Bibl. Ind .= Bibliotheca Indica Series of Sanskrit publication. Bik, or Bikaner=Rajendralal Mitra's Catalogue of Sanskrit MSS in the Library of the Maharaja of Bikaner. Calcutta 1880. BORI Cat. MSS=P. K. Gode's Descriptive Catalogue of the Govt. Collections of MSS deposited at the Bhandarkar Oriental Research Institute. Volume xii deals with MSS of Alamkāra and Nāțya. BSOS or BSOAS=Bulletin of the School of Oriental (and African) Studies. Br. Mus .= British Museum. BSS=Bombay Sanskrit Series. Ben. S. S .= Benares Sanskrit Series. Burnell or Tanjore Catalogue=Burnell's Classified Index to Sanskrit MSS in the Palace at Tanjore. London 1880.

332 SANSKRIT POETICS

Also P. P. S. Sastri's revised Catalogue of the same collec- tion, esp. vol. ix which deals with Alamkāra (1930). Srirangam 1928-31. Comm .= Commentary. Deccan Coll. Cat .= Sridhar Bhandarkar's Catalogue of MSS deposited in the Deccan College. Bombay 1888. Ed .= edition or edited. EI=Epigraphia Indica. F or f=following. Fn=Footnote GgA=Göttingische gelehrte Anzeigen. GN=Nachrichten der Göttingischen Gesellschaft der Wissen- schaften. Hall Index=Hall's Contribution towards an Index to the Bibliography of the Indian Philosophical Systems. Calcutta 1859. HPS or H. P. Sastri=Haraprasad Sastri's Notices of Sanskrit MSS. Second Series. i-iv. Also his Report 1895-1900. Also his Catalogue of Sanskrit MSS in the Asiatic Society of Bengal, vol. vi. HSP=History of Sanskrit Poetics by P. V. Kane, prefixed to his ed. of the Sahitya-darpana, 3rd ed. Bombay 1951. Hultzsch=E. Hultzsch's Reports on Sanskrit MSS in Southern India. i, 1895 ; ii, 1896 ; iii, 1905. Madras 1895- 1905. 1A=Indian Antiquary. IHQ=Indian Historical Quarterly. IOC=J. Eggeling's Catalogue of Sanskrit MSS in the India Office Library. Part iii deals with works on Alamkāra. London 1891. JA=Journal Asiatique. JAOS=Journal of the American Oriental Society. JASB=Journal of the Asiatic Society of Bengal. JRAS=Journal of the Royal Asiatic Society. JBRAS=Journal of the Bombay Branch of the Royal Asiatic Society. JDL=Journal of the Department of Letters, Calcutta Univer- sity. JOI=Journal of Oriental Institue, Baroda.

ABBREVIATIONS 333

JOR=Journal of Oriental Research, Madras. Kashmir Rep .= G. Bühler's Detailed Report of a Tour in search of Sanskrit MSS in Kashmir, Rajputana and Central India (Extra no. JBRAS 1877). Bombay 1877. As regards Bühler's other Reports and Lists, the references are as in Aufrecht. Kathavate=A. V. Kathavate's Report on the Search of Sanskrit MSS in the Bombay Presidency during 1891-95. Bombay 1901. Kāvyamālā=Kāvyamālā Series published by the Nirnay Sa- gar Press, Bombay. KBod=A. B. Keith's Catalogue of Sanskrit MSS in the Bod- leian Library, Appendix to vol. i. Oxford 1909. Kielhorn, Rep. 1880-81=F. Kielhorn's Report on the search of Sanskrit MSS in the Bombay Presidency during the year 1880-81. Bombay 1881. Kielhorn, Central Prov. Cat .= Kielhorn's Classified Alpha- betical Catalogue of Sanskrit MSS in the Central Provinces. Nagpur 1874. Kielhorn's other reports and lists are cited as in Aufrecht. KM=Kāvyamālā publications in 14 Gucchakas, also referred to as Kāvyamālā in parts. Madras Cat .= A Descriptive Catalogue of Sanskrit MSS in the Govt. Oriental MSS Library, Madras, by S. Kuppu- svami Sastri. Vol. xxii (dealing with works on Alamkāra). Madras 1918. Madras Trm A, B and C=A Triennial Catalogue of MSS, collected during the Triennium 1910-11 and 1912-13 for the Govt. Oriental MSS Library, Madras, by M. Rangacarya and S. Kuppusvami Sastri, Vol i (A, B, C). Madras 1913. Also vol. iv 1928 ; vol. v 1932 etc. Mitra=Rajendralal Mitra's Notices of Sanskrit MSS. i-x. Calcutta 1871-90. NSP or N.S.P .= Nirnay Sagar Press, Bombay, and its edition of Sanskrit works. Oppert=G. Oppert's Lists of Sanskrit MSS in the Private

334 SANSKRIT POETICS

Libraries in Southern India. i, Madras 1880 ; ii, Madras 1885. Our Heritage=Journal of the Postgraduate Research Dept. - Calcutta Sanskrit College. Peterson=P. Peterson's Reports on the search of Sanskrit MSS, as follow: i Detailed Report 1882-83 ; ii 1883-84; iii 1884-86; iv 1886-92 ; v 1892-95 ; vi 1895-97. Bombay 1883-99. Rep .= Report. Rice=L. Rice's Catalogue of Sanskrit MSS in Mysore and Coorg. Bangalore 1884. Raghavan, V=The Number of Rasas and Some Concepts of the Alamkāra Sāstra. Adyar Library, Madras 1940 and 1942. Regnaud=Regnaud's Rhétorique Sanskrite. Paris 1884. Sb. der Preuss. Akad=Sitzungsberichte der Königlich Preus- sischen Akademie der Wissenschaften. Sb. der Wiener Akad .= Sitzungsberichte der Wiener Akademie der Wissenschaften. SCB=Lists of Sanskrit, Jaina and Hindi MSS deposited in the Benares Sanskrit College, comprising collections 1897-1901, 1904-05, 1909-10, 1911-12, 1912-13, 1914-15. Allahabad, Separately published 1902-15. SCC=Descriptive Catalogue of Sanskrit MSS in the Calcutta Sanskrit College by Hrisikesa Sastri and Sivacandra"Guin. Vol vii (dealing with Alamkāra works). Calcutta 1904. Sgs=Sesagiri Sastri's Reports on the search of Sanskrit and Tamil MSS. Madras, i, 1898; ii, 1899. Śl=Śloka. Sten Konow=Sten Konow's Indische Drama (in the Grundriss Series), Berlin and Leipzig 1920. Stein or Jammu Cat .= M. A. Stein's Catalogue of Sanskrit MSS in the Raghunath Temple Library of the Maharaja of Jammu and Kashmir. Bombay 1894. Ulwar=P. Peterson's Catalogue of Sanskrit MSS in the Library of the Maharaja of Ulwar. Bombay 1892.

ABBREVIATIONS 335

W Bod=Catalogue of Sanskrit MSS in the Bodleian Library vol. ii, begun by M. Winternitz and completed by Keith. Oxford 1905. Weber=A. Weber's Verzeichnis der Sanskrit und Prakrit Handschriften der Königl. Bibliothek zu Berlin. I, 1853 ; II, pt. i 1886, pt. ii 1888, pt. iii 1892. Berlin 1853-92. WRAS=Winternitz's Catalogue of South Indian Sanskrit MSS in the Royal Asiatic Society. London 1902. WZKM=Wiener Zeitschrift für die Kunde des Morgenlandes. ZDMG=Zeitschrift der Deutschen Morgenländischen Gesells- chaft. Obvious abbreviations of texts referred to (e.g. Kāv. prak .= Kāvya-prakāśa) are not given in this list; but the texts are often quoted only with the author's name, e.g. Dandin= Daņdin's Kāvyādarśa. Other Reports and Catalogues are cited as in Aufrecht.

ADDITIONS AND CORRECTIONS

P. 4. Footnote 1, line 2. Read kanīyāmsam. P. 34. Footnote 1, lines 4 and 5. Read 151 (for 161) and 160 (for 16). P. 44. Bibliography. Line 4. Read ch. 1 (for ch. i), and line 8 read du (for de). Next page, line 1 read Sanskrite (for Sanscrit). P. 55. The last footnote should be numbered 4. P. 93. Footnote, line 8. Add after the paragraph: The verse is missing in ASB MS no. 4801/5456 (H. P. Sastri, Cat. vi, p. 395). P. 140. Line 2. Add: (6) Laghu-tīkā by Harihara Miśra (H. P. Sastri, Cat. ASB MSS vi, no. 4808/4851, p. 399). P. 189. Line 21. Read gloss (for glass). P. 207. Line 7. Read Keli- (for Kali-). P. 233. Footnote 1. Lines 4 and 9. Read i (for ii). P. 259. Footnote 2. Add: Maladhāri Rājaśekhara Sūri, author of the Prabandha-kośa, was a ŚvetāmbaraJaina who became pontiff of the Harsapuriya-gaccha about 1350-54 A.D. P. 270. Paragraph 15. Read KĀŚĪ (for KASHĪ). P. 274. Add after line 15: Krsņāvadhūta also wrote a drama called Ihamrga or Sarva-vinoda in four Acts dealing with Śrńgāra, Bībhatsa, Hāsya and Vairāgya.

INDEX

OF AUTHORS AND WORKS ON ALAMKĀRA CITED IN VOL. I (References are to pages ; asterisk indicates footnote) AUTHORS

Akabara Śāha 263 Abhinava Kālidāsa 286 Agnikumāra. See Vitthaleś- Abhinava Bhatta Bāņa 304 vara Abhinava Bhavabhūti 286 Agastya Pandita, name of Amaracandra 257f Vidyānātha 209 Amrtānanda Yogin 266-67 Acyuta, cited as commentator Ayodhyāprasāda 251 on Mammata 171 Arisimha 257f Acyuta Śarman or Acyutarāya Aruņagiri Kavi 267 Modaka 263-64 Alamkārabhāsya-kāra 186, 232 Ajitasenācārya (or Ajitasena- Alaka, Rājānaka. See Allata deva Yatīśvara) 264-65 Allața (or Alața or Alaka) Aņuratnamandana (or Ratna- 149f, 184 maņdana) Gaņi 265-66 Allarāja (or Mallarāja) 267 Atirātrayajvan 225 Avantisundarī 117, 119

Atirātra-yājin, a family name Aśmakutța 311*

of Śrīnivāsa Dīkșita 297 Āccān Dīkșita 225 Ananta 266 Āñjaneya 238 Anantadāsa 215*, 217 Ātmārāma. See Svātmārāma Ananta Pandita 247, 249-50 Yogīndra Anantārya or Anantācārya Adi-bharata 20*, 318 266 Ādinātha, a name of Jinavar- Aparājita, Āparājiti, 37*, 118 dhana Sūri 194, 195 Appaya (or Appa, Apya or Ānanda, Rājānaka 168 Appayya) Dīkşita, 200f, Anandavardhana 101f ; ques- 221f ; his works 223f; 235 tion of his identity with Appayya Dīksita, son of the Dhvanikāra 102f Āccān Dīkșita 225 Ānanda Śarman 249 Abhinavagupta 32, 41, 110f Āparājiti, See under Aparājita

22

338 SANSKRIT POETICS

Āśādhara, son of Sallakșaņa Kāśī or Kāśīkara Lakșmaņa 93,268 Kavi 270 Āśādhara, son of Rāmajī 93, Kāśīśvara Miśra 270 226. 227, 228, 268 Kāśyapa 16, 23, 44, 67-68,.

Indurāja. See Pratīhārendu- 70f

rāja or Bhattendurāja Kīrtidhara 20, 24*, 32, 42-43,.

Indrajit 268 44 Kucamāra (or Kucumāra) 1, 2. Uktigarbha 1 Kuntaka 127f Utathya 1 Kubera 1 Utpaladeva 11-12, 33 Kumāragiri Vasantarāja 210,. Udbhata 46, 72f ; as commen- 271 tator on Bharata 24*, 32, Kumārasvāmin 209-10 34-35; as a general writer Kumbha or Kumbhakarņa 271 on Poetics 72f Kuravirāma 127, 229, 272 Uddyotakṛt. See Nāgojī Krśāśva 16 Upamanyu 1 Krsņa 272

Audbhatas 73, 80, 118 Krsnakiņkara Tarkavāgīśa

A upakāyana 1 Bhattācārya 71 Krsna Dīksita or Krsna Kacchapeśvara Dīkșita 269 Yajvan 272 Kandālayārya 269 Krsna Dvivedin 173 Kamalamandira 280 Krşņabrabmatantra Parakāla- Kamalākara Bhațța 167 svāmin. See Śrīkrsņa Kalādbara 173 Brahmacārin Kalyāņa Upādhyāya 173 Krşņa Bhatta or Jayakrsņa: Kalyāņa Subrahmaņya Sūri 270 Maunin 273

Kavikarņapūra. See Paramā- Krsna Mitrācārya 173 Krsņa Śarman or Krsnāvadhūta nanda Dāsa Kavicandra 255-56 173, 273

Kavikalpalatā-kāra 219, 307 Krsņa Sudhī 274

Kānticandra Mukhopādhyāya Krsņa Sūri 274 Keśava Bhațța 275 Kāmadeva 1 Kāmarāja Dīkșita 320 Keśava Miśra 218f; in relation to Deveśvara and Ari- Kārpațika 309 simha-Amaracandra 219f

INDEX 339

Kolācala Mallinātha. See Gopāla Bhatta, son of Hari- Mallinātha vamśa Bhațța 95, 161, 248- Kollūri Rājaśekhara 275 49 Kohala 21f, 24* 238 Gopāla Bhatta or Lauhitya Kşemendra 129f; works of Gopāla Bhațța 161, 210 132-33 ; question of identi- Gopīnātha 173, 218 ty with Ksemarāja 130f; Gopendra (or Govinda) Tri- makes use of Rājaśekhara purahara (or Tippa) Bhū- 118 pāla 82 Govinda Țhakkura 163 Gangādhara Miśra 275 Gauranārya 277 Gangādhara Vājapeyin or Gangādharādhvarin 226, Ghaņtaka 32 227 Ghāsi or Ghāsī Rāma Paņdita Gangānanda Maithila 276 278 Gangārāma Jadi (or Jadin) 250, 276 Cakravartin. See Paramā- Gajapati Vīra-nārāyaņadeva. nanda Cakravartin and See Nārayana-deva Śrīvidyā Cakravartin Gaņeśa 251 Caņdīdāsa 160, 214, 279, 301 Ganeśa, son of Anantabhatta Candrakānta Tarkālaņkāra 195 319 Gadādhara Cakravartin Bha- Candracūda 279 ttācārya 173 Gadādhara Bhațța 276 Candrikā-kāra 101

Gāgābhatta (alias Viśveśvara) Cārāyaņa 311* Citrāngada 1 203-4 Giridhara 277 Cinna Appayya Dīkşita 225 fn

Gunacandra 297 2

Guņaratna Gaņi 173 Cirañjīva or Rāmadeva Cirañ-

Gurujālaśāyin. See jīva Bhațțācārya 279 Ranga- śāyin Jagadīśa Tarkapañcānana Gokulanātha Maithila 173, Bhattācārya 173 277 Jagaddhara 139 Gopāla Ācārya alias Vopa- Jagannātha Paņditarāja 229f ; deva 244, 248 his patrons 231; his perso-

340 SANSKRIT POETICS

nal history and Lavańgī- Dakşiņāmūrtikiņkara. See episode 232 ; his works Lakşmīdhara Dīkșita 233-34 Daņdin 57 f ; in relation to Janārdana Vibudha 174 Vāmana 59f; in relation to Janārdana Vyāsa 164, 174 Bhämaha 62f ; commenta- Jayakrsņa Maunin. See Krsņa tors on 69-72 Bhatta Dattila (or Dantila) 22, 44 Jayadeva 196f Darpaņa-kāra. See Viśvanātha Jayanta Bhatta 158 and Hrdaya-darpaņa-kāra Jayamangala 260, 261 Dāmodara Bhatța Harșe 282 Jayaratha 185, 234, 235 Dinakara 251 Jayarāma Nyāyapañcānana Dīnakrsņadāsa 282 164, 303 Durgādāsa 284 Jinaprabha Sūri 283-84 Devanātha 54 fn 1, 171, 174, Jinavardhana Sūri 194, 195 282 Jinavallabha Sūri 280 Devapāņi (shortly called Jīva Gosvāmin 253, 256 Pãņi) 126 Jīvanātha 281 Devaśamkara Purohita 282 Jīvarāja Dīkșita 149, 251, 311 Devīdatta 229 Jñānapramoda Gaņi 195 Deveśvara 219, 220, 259f

Tandu 20, 33 Dhanañjaya 121f Taruņavācaspati 69, 70, 72 Dhanika 123f

Tārācandra Kāyastha 284 Dharānanda 229, 230

Tilaka, Rājānaka 76, 77, 178 Dharmakīrti 15 fn 2, 49 Tirumalācārya 210 Dharmadatta 216, 291

Tiruvenkata 174 Dharmadăsa Sūri 82, 283

Tauta, Bhatta 33, 102, 111 Dharmavācaspati, wrongly for Tribhuvanacandra 71 Taruņavācaspati 72

Trimalla (or Tirumala, Tirma- Dharma Sudhī or Sūri 284-85

la, also wrongly Nirmala) Dhișaņa 1

Bhatta 281 Dhūrtila 22

Trilocana 284 Dhvanikāra or Dhvanikrt 103f Trilocanāditya 281 Dhvanyācārya (=Ānanda- Triśaraņatațabhīma 71 vardhana) 109 Tryambaka 281 Nakhakuțta 23

INDEX 341

Nandikeśvara 1, 2, 19f Pakşadhara Upādhyāya, cited Nandisvāmin 69 as a commentator on Nami-sādhu 92-93 Mammața 171 Narasimha 285 Narasimha or Nrsimha Kavi Paņditarāja, alias Raghu- nandana Rāya (?) 166 285 Narasimha Țhakkura 163 169 Paņditarāja, a title of Jagan-

Narasimha Sūri, son of nātha (q. v.)

Gadādhara 70 Padmanābha Miśra, alias

Narasimha Sūri, son of Pradyotana Bhatța 202

Timmajī 174 Padmasundara 288

Narasiņhācārya or Venkața Paramānanda Cakravartin Narasimha (or Nrsimba) 162, 187 Kavi 286 Paramānanda-dāsa Kavikarņa- Narahari Bhatta 284 pūra 254 Narahari Sarasvatī-tīrtha 158 Parameśvarācārya 113 Narahari Sūri 210, 287 Pārāśara 1, 23 Narendraprabha Sūri, Mala- Pālyakīrti 119 dhāri 287 Pīyūșavarşa, a title of Jayadeva Nāgarāja Keśava 174 Puñjarāja 288-89 Nāgeśa or Nāgojī Bhațta 172, Puņdarīka Rāmeśvara 289 228, 235f, 248 Puņdarīka Vidyāsāgara 176 Nānyadeva 32, 43-44 Purușottama 214-15, 289 Nārada 23, 44 Purușottama Sudhīndra 289 Nārāy aņa 288 Pulastya 1 Nārāyaņa cited by Viśvanātha Pedakomați · Vema-bhūpāla. 214 See Vīranārāyaņa Nārāyaņa Dīkșita 175 Prakāśavarșa 290 Nāyaka. See Bhatta Nāyaka Pracetāyana 1 Nirmala. See Trimalla Bhatta Pratīhārendurāja 73f; question Nrsimha Bhatta 126 of his identity with Bhatt- Nemi-śāha 251 endurāja 74-76 Nyāyavāgīśa Bhațțācārya, title Pradīpakrt or Pradīpakāra. of some unknown writer See Govinda Thakkura 229. See Rāmacandra Pradyotana Bhatta 197, 202, Nyāyavāgīśa 204, 205, 290

342 SANSKRIT POETICS

Pradhāna Venkapayya. See Śańkuka 37 Venkapayya Pradhāna Bhatta Nayaka 38 Prabhākara, commentator on Harşa 42 Ekāvalī 208 Kīrtidhara 42 Prabhākara Bhatta 216 fn 3, Abhinavagupta 43 290 f Nānayadeva 43 Prīyātithi 32 Bharata as author of Kāvya- lakşaņa 31* Baladeva 291 Bharata as a technical term Baladeva Vidyābhūșaņa 171- 20* 72, 291 Bhanucandra 175 Bādara 311* Bhānudatta (or Bhānukara) Bālakrșņa Bhațța 230, 291-92 Miśra 241 ; his date 243- Bālakrsņa Pāyaguņda 186, 230 44 ; his native place 246 291 Bṛhat-Kāśyapa 44, 68 Bhānu-pandita 245 Bhāmaha 8, 29f, 46f; in Bhagavadbhatta 251 relation to Bhatti 51-56 Bhagīratha 71 Bhāva Miśra (or Miśra Bhāva) Bhatta Gopāla 32 292

Bhatta Gopāla. See Gopāla Bhāskara Miśra 161

Bhatta Bhāşyakārācārya or Bhāskarā- Bhatta Tauta. See Tauta cārya? 292 Bhatta Nāyaka 24*, 32, 38-42 Bhīmasena Dīkșita 170-71, Bhatța Yantra 32 292-93 Bhatta Vrddhi 32 Bhīmeśvara Bhatța 293 Bhatta Samkara 32 Bhūdeva Śukla 293 Bhatti, alamkāra-section in Bhoja 118f, 133f, 238, 240

51f Brahmadatta 67, 71

Bhattendurāja 74-76, 102, 111 Mankhaka (or Mankha) 179-81 Bhavadeva 175 Mangala 81, 118 Bharata 1, 2, 16f, 18f ; his Matańga 10, 22, 44 date 26f; commentators Mathurānātha Sukla 217, 229 on 31f Madhumatīkāra. See Ravi Mātrguptācārya 33 Madhumatigaņeśa 175 Udbhața 34 Manodhara. See Ratnapāņi Lollața 35 Mammața 145f, 212, 215, 346-7

INDEX 343

Mallarāja. See Allarāja Yajñeśvara Yajvan 175, 294 Mallinātha, Kolācala 207, 209 Yantra. See Bhatta Yantra Mallinatha, son of Jagan- Yaśasvin Kavi 294 nātha 71 Yaştika 44 Mahādeva 249 Yāmuna or Yāmuneya 71 Mahādeva (same as above?) Yāyāvara or Yāyāvarīya 116f 251 Mahimabhatța 140f ; in rela- Raghudeva 175 tion to Śańkuka as an anu- Raghunandana Raya. See mitivādin 142f Paņditarāja Maheśvara 82, 165. See Subu- Raghunātha Manohara 295 ddhi Miśra Rangaśāyin, alias Gurujāla- Maheśvara Nyāyālamkāra 167 śāyin 249 218, 303 Ratnakaņțha, Rājānaka 169 Māgha-caitanya. See Rāghava- Ratnapāņi, alias Manodhara Caitanya 166 Māņikyacandra 157 Ratnabhūșaņa 295 Mātrgupta Ācārya 32, 33-34, Ratnamandana Gaņi. See 238, 307 Aņuratnamaņdana Mānasimha 293 Ratna-śrījñāna 70, 71 Māheśvara or Mahāmāheś- Ratnākara 225, 226, 235 vara as a title of Abhinava Ratneśvara 139, 175 and Vidyadhara 165* Ravi 166 Miśra 166 Ravi Pandita 296 Mukula 73 Rāghava 175 Muni, as a title of Bharata 18, Rāghava-Caitanya 295 135 Rājacūdāmaņi Dīkșita 296-97 Murari Miśra, cited as a com- Rājan, as a title of Bhoja 135 mentator on Mammata 171 Rājaśekhara 1, 11, 115f; his Medhavin (or Medhāvirudra) Kāyya-mīmāmsā 117f 16, 50, 92, 118 Rājahaņsa Upādhyāya 195 Mohanadāsa 293 Rājānanda 175 Rāma Kavi or Rāma Sarman Yajñanārāyaņa Dīkșita 294 298 Yajnesvara or Yajñeśvara Rāmakarņa Paņdita 319 Dīkșita 294 Rāmakrsņa 176

344 SANSKRIT POETICS

Ramacandra 175 181 ; his works 182-84, Rāmacandra, pupil of Hema- 198, 212, 234 candra 297 Rūpa Gosvāmin 252f Rāmacandra Tarkavāgīśa 216, 217 Lakşmaņa Bhāskara 20

Rāmacandra Nyāyavāgīśa 229, Lakşmīdhara Dīkşita, Lakş-

298 maņa Bhatta or Laksmaņa

Rāmadeva Cirāñjīva Bhattā- Suri with the surname

cārya. See Cirañjīva Dakşiņāmūrtikiņkara 294,

Rāmanātha Vidyāvācaspati 299

175 Lakşmīnātha Bhațța 139

Rāma Paņdita, an abbrevia- Lakşmaņa Sūri, Mallādi 285

tion for Ghāsīrāma Paņdita Lāța Bhāskara Miśra. See

(q.v.) Bhāskara Miśra

Rāma Sarman or Kavi 298 Lokanātha Cakravartin 257

Rāma Sudhī or Sudhīśvara Locanakāra 32. See Abhi-

274, 298 navagupta

Rāma Subrahmanya, also Lollața 24*, 35-37

called Rāma Subbā 298 Lauhitya Bhatta Gopāla. See

Rāmānanda or Rāmānandā- Gopāla Bhațța

śrama, a name of Lakşmī- Vakroktijīvita-kāra. See Kun- dhara Dīkșita 299 taka Rāhala or Rāhula 32-33 Vatsavarman. See Śrīvatsa- Rucaka, another name for lāñchana Ruyyaka (q.v) Vararuci 16, 67 Rucinātha Miśra 291 Vallabhadeva Paramārtha- Ruci Miśra, cited as a com- cihna 83, 91-92 mentator on Mammata 171 Vallabha Bhatta 300 Rudra or Rudrabhatta 85 f Vasantarāja 210, 271. See Rudrața 82f ; commentators Kumāragiri on 91f Vākpatirāja 119, 121 Ruyyaka (Rucaka or Ruppa- Vāgbhata, son of Soma 191-93 ka), Rājānaka 144; as Vāgbhata, son of Nemiku- commentator on Mammata māra 118, 191, 193 156 ; in relation to Mań- Vācaspati 72. See Taruņa- khaka 178-180; his date vācaspati

INDEX 345

Vācaspati Miśra 159 Viśveśvara Bhatta 302 Vājacandra 205 Vişņudāsa 304 Vādijanghāla (or° ghanghala) Vīranārāyaņa (Vema-bhūpāla) 71 210, 304 Vāmana 8, 78f ; commenta- Vīreśvara Kavicandra 240 fn tors on 81f 1, 270, 302 Vāmanīyas 80, 118 Vīreśvara Paņdita Bhatțā- Vitthaleśvara (or Vitthala or cārya Śrīvara 305 Vițthaleśa) Dīkșita, alias Vrddha-Kāśyapa 44, 68 Agnikumāra 300 Vṛndāvanacandra Tarkālam- Vijayānanda 71, 176 kāra Cakravartin 257 Vidyā Cakravartin. See Śrī- Venkapayya Pradhāna 305 vidyā Cakravartin Venkata Narasimha (or Nr- Vidyādhara 205f simha) Kavi. See Narasim- Vidyānātha 208f, 285 hācārya Vidyābhūșaņa, See Baladeva Venkața Nārāyaņa Dīkșita Vidyādhūșaņa 350 Vidyārāma 300 Venkața Sūri 285 Vidyāsāgara 176 Venkațācārya (Kirītī) Tarkā- Vinayacandra 260 laņkāra Vāgīśvara 306 Virūpākșa 205 Venkațācala Sūri 176 Viśākhila 23, 44 Vengala Sūri 229 Viśvanātha (commentator on Vecārāma Nyāyālamkāra 262, Daņdin) 71 305 Viśvanātha, son of Trimala Vecārāma (Sārvabhauma) 261, Bhatta 160, 301 262 Viśvanātha, son of Candra- Veņīdatta Tarkavāgīśa Bha- śekhara 161, 163, 212f ; his ttācārya (surnamed Śrīvara) works 215 250-51,306 Viśvanātha, son of Dinakara Vema-bhūpāla. See Vira- Bhatta. See Gāgābhațța Nārāyaņa Viśvanatha Cakravartin 254, Vaidyanātha Tatsat 170, 203 257 227 Viśvanātha Nyāyapañcāna 301 Vaidyanatha Payagunda 170, Viśveśvara, son of Lakșmī- 203,205 dhara 249 Vaidya-Bhanupaņdita 245

346 SANSKRIT POETICS

Vopadeva. See Gopāla Ācārya Krsnabrahmatantra-Parakala- Vyaktiviveka-kāra. See Mahi- svāmin 320 mabhatta Śrīdhara Sāmdhivigrahika 160 Vrajarāja Dīkșita 249 Śrīnivāsa Dīkșita 309-10 Samkara. See Śankha Śrīpāda 219 Śakalīgarbha 35 Śrīvatsalāñchana (or Śrīvatsa- Śańkuka, 24, 32, 37 142-43 śarman or Śrīvatsavarman Sankha (or Sankhadhara or or simply Vatsavarman) Śankhacūda, also called 165, 234 Śamkara) 306-7 Śrīvara, a surname of Veņī- Śambhunātha 307 datta (q. v,) Śākyācārya 32 Śrīvidyā Cakravartin. 162, 187 Sāņdilya 22, 23 Śrīharşa Miśra 40*, 220. 291 Śātakarņi 23, 307 Śvetāraņya-Nārāyaņa 131 Śātātapa 44 Samayasundara 195 Śāradātanaya 137, 238, 240 Samudrabandha 186-87 Šinga-bhūpāla 238, 239f Sarasvatī-tīrtha. See Narahari Śilālin 16 Sarasvatītīrtha Sivacandra 284 Sahadeva 82 Šivanārāyaņa Dāsa Sarasvatī- Sahasrākșa 1 kaņțhābharaņa 175 Sahrdaya, a suggested name Śivarāma Tripāțhin 175, 308 of the Dhvanikāra 105-6 Śubhavijaya Gaņi 261 Sāgara Nandin 307, 310 Śeșa 1 Sāmarāja Dīkșita 311 Śeșa Cintāmaņi 248 Sāyaņa 225 fn 3, 312 Sobhākaramitra or Śobha- Sārvabhauma 257 kareśvara 226, 235, 308-9 Sāhityacintāmaņi-kāra. See Śauddhodani 219f Vīranārāyaņa Śyāmala or Śyāmadeva 118 Sāhityadarpaņa-kāra. See 142 Viśvanātha 214 Śyāmarāja. See Sāmaraja Simhadeva Gaņi 194, 195 Śrīkara Miśra 309 Siņha-bhūpāla. See Śinga- Śrīkaņțha 309 bhūpāla Śrīkrşņa Śarman. See Krşņa Siddhicandra Gaņi 176f Śarman Sukhadeva Miśra 312 Śrīkrsņa Brahmacārin or Sukhalāla 312

INDEX 347

Sudhākara Puņdarīka Yājin 312 Abhidhā-vrtti-mātrkā 74 Sudhīndra Yati (or Yogin) Abhinaya-darpaņa 19 312-13 Abhinava-bhāratī 43, 110 Sundara Miśra Aujāgari 313 Arthālaņkāra-mañjarī 281. Subandhu, cited as a writer on See Alaņkāra-mañjarī Dramaturgy 238 Alamkāra-kārikā 316 Subuddhi Miśra 165f Alaņkāra-(kula)-pradīpa 303 Sumati 20 Alamkāra-kaumudī (1) anon. Sumatīndra Yati 273 316 (2) by Vallabhabhatta Suvarņanābha 1, 2 300 Sūrya Kavi 261-62 Alamkāra-kaustubha by Kavi- Somanārya 313 karņapūra 254, 255, 257 Someśvara 159 ·Dīdhiti-prakāśikā 257 Sphoțāyana 8 Sāra-bodhinī 257 Svātmārāma Yogīndra 284 ·Țīkā 257

Harşa 32, 42 Țippaņī 257

Harşa Miśra. See Śrīharșa Alamkāra-kaustubha (2) by Miśra Kalyāņa Subrahmaņya 270 Harikrsņa Vyāsa 140 (3) by Viśveśvara 302-3 (4) Haridāsa 314 by Venkațācārya 306 (5) Harinātha 70, 139 by Śrīnivāsa Dīkșita 309-10 Hariprasāda Māthura 314 Alamkāra-krama-mālā 282 Harihara 314 Alaņkāra-grantha 270 Haladhara Ratha 315 Alaņkāra-candrikā (1) anon. Hrdayadarpaņa-kāra. See 316 (2) by Vaidyanātha Bhatta Nāyaka Tatsat 170, 203, 227, 228 Hrdayarāma Miśra 318 (3) by Rāmacandra Nyāya- Hemacandra 118, 189-91, 194 vāgīśa 298

2. WORKS Alamkāra-candrodaya 251,306 Alamkāra-cintāmaņi by Ajita- Akabaraśāhī-Śṛngāradarpaņa sena 264, 265 288 Alaņkāra-cūdā-maņi (1) by Agni-purāņa, alamkära-section Hemacandra 189, 194 (2) in 97-100, 138 (also called °śiromaņi) by Añjana. See Kāvyāloka- Rājacūdāmaņi Dīkșita 296 locana Alaņkāra-tilaka (1) by Bhānu-

348 SANSKRIT POETICS

kara or Bhānudatta 246 (2) Krsna Yajvan 273 (3) by comm. on his own Kāvyā- Lakșmīdhara 299 (4) by nuśāsana by Vāgbhața, son Viśveśvara 302-3 of Nemikumāra 191 (3) by Alamkāra-ratnākara (1) by Appayya, son of Accan Śobhākaramitra 226, 234- Dīkșita 225 (4) by Śrīkara 35, 308f (2) by Yajñanārā- Miśra 309 yana 294, 295* (3) by Alamkāra-darpaņa in Prakrit Vecārāma 305 316 Alamkāra-rahasya 290 Alamkāra-dīpikā by Āśādhara Alamkāra-rāghava 294 227, 228, 268 Alamkāra-lakșaņa 307 Alaņkāra-nikașa (or °nikarşa) Alamkāra-vādārtha 315 312 Alamkāra-vārttika 184, 186 Alaņkāra-parişkāra 301 Alamkāra-vimarśinī 185 Alamkāra-prakaraņa 316 Alamkāra-śataka 204 Alamkāra-prakāśikā 316 Alaņkāra-śāstra-samgraha (or Alamkāra-prabodha 258 °vilāsa) 298 Alaņkāra-bhāșya 186, 234 Alamkāra-śirobhūșaņa 269 Alamkāra-makaranda 275 Alamkāra-śiromani (also called Alaņkāra-mañjarī (1) by Tri- cūdāmaņi) 296 mala Bhatta (also called Alamkāra-śekhara (1) by Arthālamkāra-mañjarī) 281 Keśava Miśra 218f, 220 (2) by Ruyyaka 182 (3) by (2) by Jīvanātha 281 Sukhalāla 312 (4) by Su- Alaņkāra-samgraha (1) anon. dhīndra Yogin 312, 313 316 (2) by Amrtānanda. Alaņkāra-manjūșā (1) by 266 (3) by Jayadeva (same Devaśamkara 282 (2) by as Candrāloka) 205 Rāmacandra Nyāyavāgīśa Alaņkāra-saņjīvanī. See (on Kāvya-candrikā) 298 Alaņkāra-sarvasva-samjī- Alaņkāra-maņi-darpaņa 305 vanī Alaņkāra-maņi-hāra 319 Alaņkāra-samudgaka 308 Alamkāra-mayūkha 316 Alaņkāra-sarvasva (J) anon. Alamkāra-mahodadhi 287 316 (2) by Keśava Miśra Alaņkāra-mīmāņsā 274 220 Alaņkāra-muktāvalī (1) by Alamkāra-sarvasva by Ruyya- Rāmasudhīśvara 274 (2) by ka 178,182; commentaries;

INDEX 349

°Vimarśinī by Jayaratha tara 280 (3) on Mammata 185 175 °Vrtti by Samudrabandha °Avaloka. See under Daśa- 186 rūpaka °Saņjīvanī by Śrīvidyā Cakravartin 187 Āgama-candrikā and Ātma-

Alaņkāra-śāstra-samgraha or prabodhikā 256 Alamkāra-śāstra-vilāsa 298 Ānanda-candrikā or Ujjvala-

Alamkāra-sāra (i) by Jayara- nīla-maņi-kiraņa 254, 256

tha 186 (ii) by Bālakrsņa °Āmoda. See under Rasa-

Bhatta 186, 230, 291-292 mañjarī

Alamkāra-sāra-mañjarī 320 Alamkāra-sāra-sthiti or Alam- Ujjvala-nīla-maņi 252f, 256f.

kära-sthiti 171*, 292 (see Āgama-candrikā and Āt-

Kuvalayānanda-khaņdana) maprabodhikā 256

Alamkāra-sāra-samgraha. See °Kiraņa. See under Ānan-

Kāvyālamkāra-samgraha of da-candrikā

Udbhața Locana-rocanī 253, 256 °Kiraņa-leśa 256 Alamkāra-sāroddhāra 171, 292 Alamkāra-sudhā 226, 228, 236 °Țīkā 256

Alamkāra-sudhā-nidhi 210, Ujjvala-padā 294 °Udāharaņa-candrikā. 225, 226, 312 See

Alamkāra-sūtra (1) quoted by under Kāvya-prakāśa

Jayaratha p. 150 (2) by ·Udāharaņa-dīpikā or °pradī-

Śauddhodani, mentioned pa. See under Kāvya-pra-

by Keśava Miśra 219 (3) kāśa

by Candrakānta Tarkālam- °Uddyota. See Kāvya-pradī-

kāra 319 pa

Alamkāra-sūryodaya 294 Udbhața-viveka or °vicāra 76,

Alaņkārānukramaņikā 316 77,178

Alamkārānusāriņī 183 ° Rju-vrtti. See under Kāvya- Alamkārendu-śekhara 286 prakāśa Alamkāreśvara 317 Alaņkārodāharaņa 186 Ekaşastyalamkāra-prakāśa 174

Avacūri (1) on Vāgbhațālam- Ekāvalī 205, 207, 209

kāra 195 (2) on Praśnot- Aucitya-vicāra-carcā 129, 131

350 SANSKRIT POETICS

Kamalākarī, title of Kamalā- Kavi-śikșā (1) by Jayamańga- kara's commentary on la 260, 280 (2) °Vrtti by Mammata, or simply Țīkā Amaracandra 257 (3) by 167 Gaňgādāsa 260 Karņa-bhūșaņa 276 Kavi-samaya-kallola 266, 285 Kalā-pariccheda 68 Kavi-saraņa-dīpikā 317 Kalyāņa-kallola 271 Kavīndra-karņābharaņa 303 Kavi-kaņtha-pāśa 278, 299*, Kadambini, comm. on Trive- 317 nikā 268 Kavi-kaņțha-hāra 307 fn 1 °Kāmadhenu. See under Kā- Kavi-kaņțhābharaņa 129, 132 vyālaņkāra-sūtra of Vāmana Kavi-karņikā 129 °Kārikārtha-prakāśikā. See Kavi-karpațī 306-7 under Kāvya-prakāśa Kāvi-kalpa-latā (1) by Deveś- °Kārikāvalī. See under Kāvya- vara 259f, 261; comm. prakāśa ºȚīkā 262; comm. by Kāvya-kalānidhi 274 Vecārāma 261, 262 ; Bāla- Kāvya-kalāpa 317 bodhikā 261-62; Padārtha- Kāvya-kalpalatā Kavitā-raha- dyotanikã 262 (2) by sya and Kavi-śiksā Vrtti Rāghava-Caitanya 260, 295 by Arisimha and Amara- (3) cited by Sāradātanaya candra 257, 259, 261 239 ° Parimala by Amaracandra Kavi-kalpa-latikā 317 258 Kavi-kautuka 304 °Mañjarī 258 fn 6 Kavi-kaustubha 295 °Makaranda 261 Kavi-gajāńkuśa 82 Kāvya-kautuka and its Viva- Kavitāvatāra 214-15 fn 6, 289 raņa 111 Kavitā-rahasya. See Kāvya- Kāvya-kaumudī (1) by Deva- kalpalatā 257, 258 fn 1 nātha on Mammata 174 Kāvi-nandikā or °nandinī or (2) by Ratnabhūșaņa Kāvyaprakāa-bhāvārtha 176 295 Kavi-priyā 80 Kāvya-kaustubha (1) anon. Kavi-rahasya 1. See Kāvya- 317 (2) by Baladeva Vidyā- mīmāmsā bhūsaņa 171, 291 Kavi-vimarśa, a name of Rāja- Kāvya-candrikā (1) by Kavi- śekhāra's Kāv. mīm. candra 255f, 257 (2) by.

INDEX 351

Rāmacandra Nyāyavāgiśa °Rju-vrtti 174 298 (also called Alamkāra- °Kārikārtha-prakāśikā 175 candrikā), Kārikāvalī 173 Kāvya-dākinī 276, 303 Kāvya-kaumudī 174 Kāvya-tattva-vicāra 315 Kāvyālamkāra-rahasya-dī- Kāvya-tattva-viveka-kaumudī pikā 161 on Daņdin's Kāvyādarśa 71 Kāvyādarśa or Samketa 159 Kāvya-tilaka 304 · Khaņdana or Kāvyāmrta- Kāvya-darpaņa (1) by Madhu- tarangiņī 176, 317 matigaņeśa on Mammața Țīkā (Kamalākarī) 173, 175f 175 (2) by Ratnapāņi ·Tilaka or °Rahasya-dīpikā. alias Manodhara on ibid or Jayarāmī 164 167 (3) by Rājacūdāmaņi °Darpaņa 161, 167, 215. See Dīkşita 296-97, 310 (4) by also Kāvya-darpaņa Śrīnivāsa Dīkșita 309-10. °Dīpikā (1) by Candīdāsa Kāvya-dīpikā (1) anon. 317 160, 214 (2) by Jayanta (2) by Kānticandra 320 Bhatta (also called Kāvya-nirņaya 125 Jayantī) 158 (3) by Kāvya-pariccheda 317 Nāgojī Bhațta 172 (4). Kāvya-parīkșā 165 by Śivanārāyaņa 176 Kavya-prakāśa 148, 154; its Narasiņha-maņīșā 169 dual authorship 148f; its °Nidarśana (also called Siti- Kārikās ascribed to Bha- kaņțha-vibodhana) 168 rata 152f ; commentaries °Pada-vrtti 174 on 156f °Pradīpa (also called Avacūri 175 Kāvyapradīpa) 163 °Ādarśa or °Bhavartha- °Pradīpa-prabhā 163, 170 cintāmaņi 167 °Pradīpa-uddyota, laghu ° Udāharaņa-candrikā 170 and brhat 163, 172, 236. ·Udāharaņa-dīpikā (also Bāla-cittānurañjanī 158 called Śloka-dīpikā) (1) ·Bhāvärtha. See Kavi-nan- by Govinda 163 (2) or dinī ·pradīpa by Nāgojī 172, °Bhavartha-cintāmani See- 236 ·Ādarsa above °Uddyota, See °Pradīpa- Madhura-rasā 173 uddyota below Rasa-prakāśa 173

352 SANSKRIT POETICS

· Rahasya-dīpikā. See Miśra 220 (2) by Viśveś- °Tilaka above vara 304 (3) anon. 317 °Rahasya-prakāśa (1) by Kāvya-rātnākara 305 Jagadīśa 174 (2) by Kāvya-lakșaņa (1) by Bharata? Rāmanātha Vidyā- 31 fn 1 (2) by Krsna vācaspati 175 Śarman 273 (3) anon. 317 Līlā 175 Kāvya-laksaņa-vicāra 317 °Viveka 160 Kāvya-vilāsa 223 fn 3, 279-80 Vișamapadī 176 Kāvya-vrtti-ratnāvalī 288 Vistārikā 162 Kāvya-saraņi 225-26 °Vyākhyā 175 Kāvya-sāra-samgraha 310 ° Śloka-dīpikā (1) by Govin- Kāvya-sudhā. See Sāhitya- da. See °Udāharaņa- sudhā dīpikā above (2) by Kāvyādarśa by Daņdin 57f ; Janardana 174 commentaries 70f ·Samketa (1) by Māņikya- Candrikā 71 candra 157 (2) by Ruy- Mārjanā 70 yaka or Rucaka 156,182 Muktāvalī 70 (3) by Someśvara 159 Rasika-rañjanī 71 Saņpradāya-prakāśinī 162, Ratnaśrī 70, 71 188 Vivrti or Kāvya-tattva- °Sāra 175 viveka-kaumudī 71 °Sāra-dīpikā 173 Vaimalya-vidhāyinī 71 Sāra-bodhinī 165 Śrutānupālinī 71 Sāra-samuccaya 169 Kāvyādarśa, comm. on Mam- Sāhitya-kaumudī 171 mata by Someśvara 159 Sudha-sagara or Sudho- Kāvyānuśāsana (1) by Vāg- dadhi 169 bhata, son of Nemikumāra Subodhini 176 191, 193-4 (2) by Hema- Sumanomanoharā 173 candra 189, 190-91, 194 Kāvya-pradīpa. See under Kāvyāmrta 166 Kāvya-prakāśa-pradīpa Kāvyāmrta-tarangiņī. See Kāvya-mañjarī 229, 298 Kāvya-prakāśa-khaņdana Kāvya-mīmāņsā 1, 2, 115f, Kāvyārtha-gumpha 314 117-18 Kāvyālaņkāra (1) by Bhāma- Kāvya-ratna (1) by Keśava ha 47f

INDEX 353

° Vivrti or Bhāmaha-viva- Kāvyāloka, cited by Appayya raņa or Bhāmahīya Udbha- 225, 314 ța-laksaņa by Udbhața Kāvyopadeśa 317 46, 73,187 °Kiraņa. See under Alam- Kāvyālaņkāra (2) by Rudrața kāra-kaustubha and Ujjva- 85f la-nīla-maņi Țīkā by Āśādhara 93 ° Kiraņa-leśa. See under Țīkā by Vallabhadeva 91 Ujjvala-nīla-maņi Țippaņa by Nami-sādhu Kuvalayānanda in relation to 92 Candrāloka 200f, 223, 226, Kāvyālāņkāra-kāmadhenu. 227 See under Kāvyālaņkāra- Alamkāra-candrikā 176, sūtra by Vāmana 227, 228 Kāvyālaņkāra Śiśu-prabodha Alamkāra-dīpikā 226 or Śiśu-prabodhālamkāra Alamkāra-sudhā 228, 229, 288 236 Kāvyālamkāra-samgraha (or °Karikā (same work) °sāra-samgraha) by Ud- Kāvya-mañjarī 229 bhata 46, 72, 73 Budha-rañjanī 229 °Laghu-vrtti by Pratīhā- Rasika-rañjanī 228 rendurāja 74 Laghvalamkāra-candrikā Kāvyālaņkāra-sūtra-vrtti by 229 Vămana 80 (the Vrtti is Șatpadānanda 229 entitled Kavi-priyā) °Khandana, also called Commentaries 82f Alamkāra-sthiti or Alam- °Kāmadhenu 81, 82 kārasāra-sthiti 171, 225, 292 Sāhitya-sarvasva 82, °Țippana 229 Kāvyāloka, also called Dhva- Kriyākalpa as the original nyāloka or Sahrdayāloka name of Poetics 7* 102f Krsņānandinī (on Sāhitya- ° Candrikā 101 kaumudī) 171 °Locana 101, 110f Kovidānanda 268 ° Añjana 113 Kohala-rahasya 22 · Vyākhyā-kaumudī 113 Kohalīya Abhinaya-śāstra 22 Kāvyāloka by Hariprasāda Kaumudī. See under Kavyā- 314 loka

23

354 SANSKRIT POETICS

Guru-marma-prakāśikā on Jayarāmī. See under Kāvya- Rasa-gangādhara 229, 236, prakāśa-tilaka 237 Jalpa-kalpa-latā 265f Gūdhārtha-prakāśikā. See Citra-mīmāmsā °Țīkā. See under Alaņkāra- Godavarma-yaśo-bhūșaņa 267 kaustubha, Ujjvala-nīla- maņi, Kavi-kalpalatā, Catura-cintāmaņi 275 Kāvyālaņkāra, Kāvya-pra- Candrāloka 196, 199f ; in re- kāśa, Daśa-rūpaka, Saras- lation to Kuvalayānanda vatī-kaņțhābharaņa and 200f Vidagdhamukha-mandana ° Prakāśa Śaradāgama °Țippaņa. See under Kuva- 197, 202, 205 layānanda, Sāhityadar- Ramā 203, 205 paņa, and Kāvyālamkāra ·Dīpikā 205 of Rudrața ° Sudhā. See Rākāgama °Țippaņī. See under Alam. Śārada-śarvarī 205 kaustubha °Candrikā. See under Kāvyā- loka (Dhvanyāloka) and Tattva-parīksā or Śabdārtha- Kāvyādarśa Tattvaparīkşā 165, 169 Camatkāra-candrikā 240 fn Tattvāloka 110 1, 270, 302 Tattvokti-kośa 144 Citra-candrikā 247 Taralā 207, 210 Citra-mīmāmsā 224f, 229 Tāla-lakşaņa 21 °Khaņdana 224, 225, 234 Tilaka. See under Jayarāmī ° Gūdhārtha-prakāśikā or Kāvyaprakāśa-tilaka 230 Triveņikā 268 ·Doşa-dhikkāra 225 Sudhā 230 Dattila 23

Citrāloka 230 Dattila-kohalīya 22 °Darpana See under Hrdaya- Chayā (on Rasa-mīmāmsā) darpaņa, Sāhitya-darpaņa 250 or Kāvyaprakāśa-darpaņa Chando-viciti 68 Daśarūpaka or Daśarūpa 122f °Avaloka 123, 125 Jayantī or Dīpikā. See under Lagu-țīkā 126 Jayanta Bhatta °Țīkā 126

INDEX 355

°Paddhati 127, 272 Natyā-darpaņa (anon.) 318 ºSāhasānkīya Țīkā 126 Nātya-locana 281 Daśarūpaka-vivaraņa 317 · Vyākhyāñjana 281 'Dīdhiti-prakāśikā. See under Nāțya-śāstra by Bharata 2; its Alamkāra-kaustubha text 19f; its date 28f; Com. °Dīpikā. See under Kāvya- on 31f prakāśa and Candrāloka Nāțya-śāstra by Vasantarāja Dușkara-citra-prakāśikā 139 Kumāragiri 210 Doşa-jitkāra, a mistake for Nātya-sarvasva-dīpikā 318 Citramīmāmsā-doșa-dhik- Nātyārņava 20 kāra (q.v.) Nāyikā-varņana 298 Doşa-dhikkāra. See under °Nidarśana. See under Kāvya- Citra-mīmāmsā prakāśa Nūtana-tarī 251 Dhvani-pradīpa 288. °Naukā. See under Rasa- Dhvani-siddhanta-grantha 160, tarańgiņī and Sahitya- 279 ratnākara Dhvanyāloka. See Kāvyāloka °Pada-vrtti. See under Kāvya- Nañja-rāja yaśo-bhūșaņa 285- prakāśa 286 °Paddhati. See under Daśa- Nata-sūtra 16, 24 rūpaka Nandi-bharata 19f °Parimala. See under Rasa- Nandi-mata 20, 44 mañjarī and Kāvya-kalpa- Narasiņha-maņīșa 169 latā Nāțaka-candrikā 253, 256 °Prakāśa. See under Candrā- Nāțaka-dīpa 281 loka and Rasa-mañjarī Nāțaka-paribhāșā 241 Pratāparudra-yaśo-bhușaņa Nāțaka-prakāśa 210 208f Nāțaka-mīmāmsā 184 Padārtha-dyotanikā 262 Nātaka-ratna-kośa 317 °Pradīpa See under Kavya- Nāțaka-lakșaņa 289 prakāśa Nāțaka-lakșaņa-ratna-kośa 310 °Prabhā. See under Kāvya- Nāțakāvatāra 317 prakāśa-pradīpa and Sāhi- Nāțya-cūdā-maņi 313 tya-darpaņa Nāțya-darpaņa 297 Praśnottara 280

356 SANSKRIT POETICS

Prastāva-cintāmaņi 279 Matanga-bharata 20 Prastāva-ratnākara 314 Madhu-dhārā 313

Bāla-cittānurañjanī 158 Madhumatī 166 Madhu-rasā or Madhura-rasā Bāla-bodhikā 261-62 173 Bindvalamkāra 314 Mandara 285 Budha-rañjanī 229 Mandāra-maranda-campū 273 Bṛhat-Kāśyapa 44, 68 Brhat-samketa. See Kāvya- Marma-prakāśikā. See Guru- marma-prakāśikā prakāśa-samketa by Mādhurya-rañjanī 273 Ruyyaka ·Mārjanā. See under Kāvyā- Brhad-uddyota. See Kāvya- darśa and Sarasvatī-kaņțhā- prakāśa-uddyota Brhad-deśī 20, 44 bharaņa Muktāvalī. See under Kāvyā-

Bhakti-rasāmrta-sindhu 252 darśa

Bharata-bhāșya or Bharata- Mugdha-medha-kara 266

vārttika 43 Yaśovanta-yaśo-bhūșaņa 319 Bharata-vyākhyāna by Mātr- gupta 34 Raghunātha-bhūpālīya 272 Bharata-śāstra-samgraha 299 Ratna-darpaņa 139 Bharata-samgraha 188 Ratna-śāņa 210 Bharata-sūtra-vrtti 172 Ratna-śobhākara 274, 298

Bharatārņava 20, 299 fn 2 Ratnaśrī on Kāvyādarśa 70t

Bhānu-bhāva-prakāśinī 249 Ratnāpaņa 210

Bhāmaha-vivaraņa or Bhāma- Ramā 203, 205

hālaņkāra. See under Kā- Rasa-kalikā 318

vyālaņkāra by Bhāmaha Rasa-kallola 282

Bhāva-prakāśa 137, 210, 238, Rasa-kaumudī (1) by Ghāsi- 240 rāma 278 (2) by Śrīkaņtha ·Bhāvārtha. See under Kāvya- 309 (3) anon. 318 prakāśa Rasa-gangādhara 233-34 ; °Bhavartha-cintāmaņi. See comm. on: under Kāvya-prakāśa Gurumarma-prakāśikā 236, 237 °Makaranda and °Mañjarī. Vişamapadī (anon.) 237 See under Kāvya-kalpalatā. Rasa-gandha 318

INDEX .357

Rasa-gāndhāra 318 Vyangyārtha-kaumudī (1) Rasa-candra 278 by Ananta 247 (2) by Rasa-candrikā 303 Viśveśvara 249, 303 Rasa-tarangiņī by Gopāla- (also called Samañjasā bhatta (on Śṛngāra-tilaka) or Samañjasārthā) 95 Vyangyārtha-dīpikā Rasa-tarangiņī by Bhānudatta 249 242f, 250f "Sthūla-tātparyārtha 250 Nūtana-tarī 251 Rasa-mahārņava 277 °Naukā 250 Rasa-mīmāņsā (1) by rīvi- °Rasika-rañjanī 250 dyā-cakravartin 188 (2) by Rasodadhi (1) by Gangārāma Jadi 250, 276 Gaņeśa and (2) by (3) by Kāśīśvara Miśra 270 Mahādeva 251 Rasa-ratna-kośa (1) anon. 318 Sāhitya-sudhā or Kāvy- (2) by Kumbha 271 sudhā 251 Rasa-ratna-pradīpikā or °pra- °Setu or ° Setu-bandha dīpa 169, 242, 267 251 Rasa-ratna-hāra 308 Rasa-dīrghikā 300 Rasa-ratnākara 226, 318 Rasa-nirūpaņa 210, 287 Rasa-ratnāvalī 305 Rasa-prakāśa 173 Rasa-vilāsa 293 Rasa-pradīpa 216 fn 2, 290 Rasa-viveka 318 Rasa-bindu 318 Rasa-samuccaya (anon.) 318 Rasa-mañjarī by Lakșmīdhara Rasa-sarvasva 293 299 Rāsa-sāgara (anon.) 318 Rasa-mañjarī by Bhānudatta Rasa-sindhu 289 242f, 247 Rasa-sudhākara 318 ° Āmoda 249 Rasākara 319 °Parimala 248 Rasāmrta-sindhu 252, 318 °Prakāśa 247, 248 Rasāmrta-śeșa 253, 256 Bhānu-bhāva-prakāśinī Rasārņava or Rasārnava- 249 sudbākara 239, 240; Rasika-rañjana 249 works cited in 240 Rasika-rañjanī 248 Rasārņavālaņkāra 290 °Vikāsa or °vilāsa 243 Rasika-jīvana 276 248 Rasika-prakāśa 282

358 SANSKRIT POETICS

Rasika-priyā 268 Laghvalaņkāra-candrikā 229 Rasika-rañjana 249 Līlā, comm. on Mammața 75 Rasika-rañjanī (1) by Gopāla- °Locana. See under Kāvyā- bhatta on Rasa-mañjarī loka and Sāhitya-darpaņa 248 (2) by Gangadhara on Locana-añjana. See under Kuvalayānanda 228 (3) by ° Añjana Veņīdatta on Rasa-tarań- Locana-rocanī 253, 256 giņī 250 (4) by Viśvanātha Locana-vyākhyā-kaumudī. on Kāvyādarśa 71 See under Kāvyāloka Rasika-samjīvanī 275 °Locana-vyākhyāñjana. See Rasika-sarvasva 319 Nāțya-locana Rasodadhi (1) by Ganeśa 251 (2) by Mahādeva 251 (3) by Vakrokti-jīvita 127f Mohanadāsa 293 Vana-tarangiņī wrongly as- Rahasya 319 cribed 95 °Rahasya-dīpikā and °Rahasya- Vasantarājīya Nāțya-śāstra prakāśa. See under Kāvya- 210, 271 prakāśa Vākya-ratna 220 Rākāgama or Candrāloka- Vāgbhațālaņkāra 191, 193, 194 sudhā 203-4 ° Avacūri 195 Rāma-candra-candrikā 269 Cūrņī 195 Rāmacandra-yaśo-bhūșaņa 269 Jñāna-pramodikā 195-96 Rīti-vṛtti-lakșaņa 300 °Vivaraņa 195 ·Samāsānvaya Țippaņa Lakşaņa-dīpikā 277 195 Lakşaņa-mālikā 287 Other comm. 195 Lakşaņa-ratnāvalī 225, 230 Vārttika, See Harşa-vārttika Lakşaņaratnāvalī-vyākhyā 230 Vikāsa or Vīlāsa. See under Lakşmī-vihāra 318 Rasa-mañjarī Laghu Udoyota See Kāvya- Vidagdha-mukha-mandana 82, prakāśa 283 Laghu Kuvalayānanda 203 fn 1 Comm. on 283-84 Laghu-Țīkā, see under Daśa- Vidvan-manoramā 284 rūpaka °Vimarśinī. See under Alam- ·Laghu-vrtti. See under Kāvyā- kāra-sarvasva lamkāra-samgraha Vivaraņa. See under Kāvyā-

INDEX 359

lamkāra of Bhāmaha, Śabda-bheda-nirūpaņa 288 Kāvya-kautuka, and Vāg- Śabda-vyāpāra-paricaya 148, bhațālaņkāra 156 · Vivrti. See under Kāvyā- Śabdārtha-tattva-parīkșā. See darśa and Sāhitya-darpaņa Tattva-parīkşā °Viveka. See under Kāvya- Śaradāgama Candrāloka-pra- prakāśa kāśa. See under Candrāloka Viveka, sub-comm. on Śārada-śarvarī 205 Kāvyānuśāsana Śāharājīya 270, 273 Vişama-pada-vyākhyāna. See Šinga-bhūpālīya Alamkāra, Şatpadānanda descriptive name for Rasā- Vișama-padī (1) anon. on rņava-sudhākara of Śińga- Rasa-gangādhara 237 (2) bhūpāla by Sivarāmą on Mammața Singabhūpāla-kīrti-sudhā-sāra- 176 śītalā, descriptive name for Vișņu-dharmottara Upapurāņa, Camatkāra-candrikā (q.v.) Alamkāra-section in 95- Sitikantha-vibodhana. See 96 under Kāvya-prakāśa Vistārikā 162 Śiśu-prabodhālaņkāra. See °Vrtti. See under Alamkāra- Kāvyālaņkāra-śiśu-prabo- sarvasva dha (by Puñjarāja) Vṛtti-dīpikā 273 Śiśu-prabodhālamkāra by Vrtti-vārttika 224, 230 Vişņudāsa 304 Vrddha-Kāśyapa 44 Śṛngāra-kaustubha 319 ° Vaimalya-vidhāyinī. See Śṛngāra-candrodaya 319 under Kāvyādarśa Śṛngāra-taținī 279 Vyakti-viveka 143 ; Vicāra Śṛngāra-tarangiņī 319 or °Vyākhyāna 141, 183 °Vyangyārtha-kaumudī Śṛngāra-tilaka 86f ; comm. on and · Vyangyārtha-dīpikā, See 95 (Rasa-tarangiņī) Śṛngāra-darpaņa (Akbara- under Rasa-mañjarī · Vyākhyā. See under Kāvya- śāhī) 288 Śṛňgāra-dīpikā 246 prakāśa and Bhāva-pra- Śṛngāra-pavana 319 kāśa Vyākhyā-kaumudī. See under Śṛgāra-prakāśa 136f, 140f, 183, 238, 240 Kāvyāloka Śṛngāra-bheda-pradīpa 314-15

360 SANSKRIT POETICS

Śṛgāra-mañjarī (1) by Sarasvatī-kaņțhābharaņa 136f, Akabara Saha 263 (2) Commentaries on 138f. by Ajitasena 264, 265 (3) Dușkara-citra-prakāsikā anon. 319 139, Śṛńgāra maņdana or Śṛňgāra- °Marjanā 139 rasa-mandana 300 Ratna-darpaņa 139 Śṛgāra-ratnākara 319 ·Țīkā by Jagaddhara 139 Śṛngāra-latā 312 Sarasvatī-hrdayālaņkāra 43* Śṛngāra-vidhi 319 Śṛňgāra-sarasī 292 Sahrdaya-toşiņī 131 Sahrdaya-līlā 184 Śṛňgāra-sāra and Śṛńgāra- Sahrdayāloka. See Kāvyā- sārāvalī 305-6 loka Sṛngāra-sārodadhi 312 Sāra-bodhinī (1) by Viśva- Śṛngāra-hāra 291 nātha on Alamkāra-kaustu- Śṛngārāmrta-laharī 311 bha 257 (2) by Śrīvatsa- Śravaņa-bhūșaņa 284 lāñchana on Mammata 165 Śrutānupālinī on Daņdin 71 Sāra-dīpikā on Mammața 173 °Śloka-dīpikā. See under Sāra-samuccaya on Mammața Kāvya-prakāśa 169 Sārasvatālamkāra 273 Satpadānanda Vișama-pada- ° Sāhasānkīya Țīkā. See under vyākhyāna 226, 229 Daśa-rūpaka Sāhitya-kalpa-vallī 265 °Samketa. See under Kāvya- Sāhitya-kallolinī 292 prakāśa Sāhitya-kutuhala 295 °Samjīvanī or °Sarvasva-samjī- Sāhitya-kautuhala 295 vanī or Alamkāra-samjī- Sāhitya-cintāmaņi and its. vanī. See under Alamkāra- Țippaņa Krsņānandinī on sarvasva Mammața 171 Sampradāya-prakāśinī Brhatī Sāhitya-cintāmaņi of Vīranā- and Laghu Tīkā 162, 188 Samañjasā or Samañjasārthā. rāyaņa 210, 225, 226, 304-5 Sāhitya-cūdā-maņi 161 See under Rasa-mañjarī Sāhitya-tarańgiņī 272 Samāsānvaya Țippaņa. See Sāhitya-darpaņa 161, 213f, 216 under Vāgbhatālaņkāra °Tippaņa 217 Sarasāmoda 263 °Prabhā 218

INDEX 361

°Locana 217 Sāhitya-hrdaya-darpaņa 161 ° Vijña-priyā 218 Sudhā 230 ° Vivrti 217 ° Sudhā. See Rākāgama Sāhitya-dīpikā 161 Sudhā-sagara or Sudhodadhi Sāhitya-mīmāņsā 182, 208 170 (on Mammata), 292 Sāhitya-ratnākara by Dharma Subodhinī (1) by Trilocana Sudhī 284, 285 284 (2) by Venkațācala °Naukā 285 Sūri 176 Mandara 285 Sumanomanohara 173 Sāhitya-ratnākara by Yajña- °Setu or °Setu-bandha. See nārāyaņa 294 Rasa-tarańgiņī Sähitya-sarvasva 82, 165 Sāhitya-sāmrājya 273 Sthūla-tātparyārtha. See under Rasa-mañjarī Sāhitya-sāra (1) by Acyuta 263-64 (2) by Mānasimha Harilocana-candrikā, a mis- 293 taken name for Ramā (q.v.) Sāhitya-sudhā or Kāvya-sudhā 203 fn 1 251 Harşa-vārttika 42 Sāhitya-sudhā-sindhu 301 Hdayangama 69, 70 Sāhitya-sūkșma-saraņi 310 Hrdaya-darpaņa 38f

N.B .- There are numerous passing references to some of these works and authors; but the figures in the index indicate the places where they are dealt with or cited substantially.

HISTORY OF SANSKRIT POETICS

SUSHIL KUMAR DE

IN TWO VOLUMES

VOL. II

CALCUTTA 1960

VOLUME II

SYSTEMS AND THEORIES

CHAPTER I

FROM BEGINNINGS TO BHĀMAHA

( 1 )

Of the unknown beginnings of Poetics as a discipline, our enquiry in the preceding volume1 has indicated that we can only make a few surmises, by implication, from the oldest surviving works on the subject, from stray references in general literature, from the elaboration of similar ideas in other disciplines, and from the fully developed Kāvya-style which would warrant the pre-existence of some doctrines of Poetics regulating its art and usage. Apart from such surmises, the sixteenth chapter of Bharata's Nātya-śāstra gives us for the first time an outline of Poetics which is probably earlier in substance, if not in date, than the earliest existing Kavya. In this chapter, one meets with a developed dogma, if not a theory, of Poetics which enumerates four poetic figures (alamkāras), ten excellences (guņas), ten defects (dosas), and thirty-six characteristics (lakşanas) of poetic composition. These apparently constituted the principal contents of the discipline as it existed at a very early period ; and this may be taken, in the absence of other data, as the first known period in the history of Sanskrit Poetics. It is proper to note in this connexion that in the Nātya- sastra, Bharata is principally concerned with Dramaturgy and allied topics2, and deals with Poetics in so far as it applies to the theme in hand. In later poetic theories, Dramaturgy is taken as a part of the discipline of Poetics, and the drama

1 See vol. i, pp. 1-17. 2 An outline of the different chapters of Bharata's Nātya-śāstra is given in Winternitz GIL iii, pp. 7f, and in Kane HSP, pp. vi-vii. 24

2 SANSKRIT POETICS

is accordingly considered to be a species of the Kāvya. But. there are reasons to believe that in older times Dramaturgy and Poetics formed separate disciplines, the former being probably the earlier in point of time, as well as in substance. We have seen3 that the existence of nata-sūtras, which were presumably works in the sūtra-style on the histrionic art, was. known even in the time of Panini; but there is no reference,. direct or indirect, to such alamkāra-sūtras, and indeed the term Alamkära itself in the technical sense was unknown in early literature. The earliest surviving works on Poetics,. on the other hand, do not include a treatment of the theme of Dramaturgy which, having been a study by itself, was. possibly excluded from the sphere of Poetics proper. Both Bhāmaha and Daņdin, no doubt, speak of nātaka as a species. of kāvya, but they refer to specialised treatises for its detailed treatment4. Vamana, the next important writer on Poetics, shows indeed an unusual partiality towards the drama (1. 3. 30-32), but even he did not think it proper to devote any special attention to it. Among later writers, it is not until we come to the time of Hemacandra, Vidyanatha and Viśvanātha, when the study was already entering upon a period of critical elaboration and summing-up of results, that we find special chapters dealing with the topic of Dramaturgy. Of these late writers, Vidyānātha and Viśvanātha explicitly refer to and summarise the Daśa-rūpaka, a recognised. work on the dramatic art; while the encyclopaedic Hemacan- dra, who professes great admiration for Bharata and his commentator Abhinavagupta, deals with the subject rather summarily, referring the reader to the standard works of Bharata and Kohala. It seems, therefore, that the school of Dramaturgy had an existence separate from the orthodox school of Poetics.

3 Vol. i, p. 16. 4 Kāvyādarśa i. 31, Bhāmahālamkāra i. 24. The word anyatra in Dandin is interpreted by the cmmentators as referring to Bharata.

FROM BEGINNINGS TO BHĀMAHA 3

It is thus not surprising that Bharata should set apart, as he does, a chapter of his work for dealing with the ornaments of Poetry, so far they apply to the drama (nātakāśraya). In his discussion of the gunas and dosas in their application to the drama, he expressly designates them as kāvya-guņas and kāvya-doșas (xvi. 92, 84) respectively ; and with reference to the alamkāras he says kāvyasyaite hyalamkārāḥ (xvi. 41)5, making it clear at the same time that he considers them only as embellishments of the dramatic speech. Bharata opens this chapter on Poetics with the discussion of what he calls the laksanas (lit. characteristics), which ap- pear to be partly formal and partly material elements of poetry6. Bharata mentions 36 of them and devotes a consider- able part of this chapter to their definition ; and the whole discipline appears to have received from them the designation of Kāvya-lakșaņa referred to in xvi. 177. From his treatment it appears that he considers Laksanas to be of greater importance

5 Ed. Kāvyamālā (N.S.P.), ed. GOS xvi. 41; but the Chowkhamba Skt. Ser. ed. xvii. 42 reads the line differently. Our references in the fol- lowing pages are throughout to the Kavyamāla ed .- Cf also verses 104, 110. Bharata uses the word kāvya many times here, as in other chapters, to signify the drama, but we must bear in mind that his conception of poetry is dramatic and justifies such employment of the term kāvya. But in this chapter he appears to imply a distinction between the kāvya and the nātaka as species of composition. 6 The part of the text dealing with Laksanas and Gunas exists in two recensions, which we shall call here A and B. Both the recensions are known to Abhinavagupta; but he follows the text of A on Laksaņas, as it had been handed down to him through his teachers (asmad- upādhyāya-paramparāgatah, p. 384). The editions of Kāvyamālā and Gaekwad's Series give this recension A, which consists of 39 verses start- ing with Upajati and proceeding with Anustubh stanzas. The B recen- sion, found in Chowkhamba Skt. Ser. ed. (which also gives A in the footnote), consists of 42 stanzas all in Anustubh. A is followed by the Daśa-rūpaka, but B is accepted by Šinga-bhūpāla and Viśvanātha. Bhoja appears to know both the recensions, but he makes out 64 Laksaņas; he is followed by Śāradātanaya. 7 In Kāvyamālā and Gaekwad's ed .; ed. Chowkhamba p. 204, st. 16.

4 SANSKRIT POETICS

than Alamkāras which are mentioned as just a few in number. It is not very clear, however, from Bharata's treatment as to what position these Laksanas should occupy in a formal scheme of Poetics; but the function of most of these is assigned in later Poetics to Alamkāras or Guņas. Daņdin mentions them summarily (ii. 366) under Alamkāras in the wider sense, along with samdhyanga and vrttyanga which belong properly to the drama, and refers to āgamāntara (interpreted by Tarunavācaspati as alluding to Bharata) for their treatment. So does Dhanañjaya (ed. N. S. P. iv. 84) ; while Viśvanātha (ed. Durgaprasad, vi. 171-211, pp. 316-332) . takes them in connexion with the drama, calling some of them nātyālamkara (dramatic embellishment), and is at the same time of opinion that although some of them are pro- . perly included under guņa, alamkāra, bhāva and samdhi, they require a particular mention inasmuch as in the drama they are to be accomplished with some care (p. 332). In later literature the Laksanas, which linger conventionally in Dramaturgy, entirely disappear from Poetics proper, Jaya- deva's Candraloka being the only later work on Poetics which deals with them. This phenomenon would probably indicate not only that the Laksanas were regarded as strictly proper to the drama, but also the conclusion that what were, in the infancy of Poetics, considered so important as to deserve sepa- rate treatment and to be differentiated from the Gunas and the Alamkāras, were with the growth of critical insight assigned to the Gunas and Alamkāras themselves to whose sphere in ultimate analysis they were thought properly to belong8. V. Raghavan has already given9 a detailed account of the 8 For instance, asīh is one of the lakșaņas of Bharata, Bhāmaha mentioning it rather doubtfully as an alamkāra with the remark: āsir api ca keşāmcid alamkāratayā matā (iii. 54). In Daņdin ii. 357 (as well as in Bhatti), it is already established as an Alamkāra. It is significant that Kuntaka finds fault with those who regard it as an Alamkāra. 9 Some Concepts of the Alamkāra-śāstra, Adyar Library 1942, pp. 1-47.

FROM BEGINNINGS TO BHĀMAHA 5

history of the concept of Laksaņa; but since the Laksaņa- paddhati perished very early, or lingered as a superfluous relic in the history of Poetics and Dramaturgy, it is not necessary for us to make more than a passing reference. Abhinavagupta, while explaining Bharata's text, mentions as many as ten different views concerning Laksana ; but it appears that Laksana, otherwise called Bhūșaņa, is generally taken (on the analogy perhaps of Sāmudrika Lakșaņa), to be an innate beautifying element belonging to the body of poetry, or rather as constituting the body itself. Although similar in function to the Alamkāra in being a Kāvya-śobhākara Dharma, it is not a separate entity, but Aprthak-siddha; that is to say, it imparts beauty to poetry by itself, and is not added, as an Alamkāra is added, for extra beauty. It is obvious that the concept of Laksana, even at its birth, had an overlapping of function with Alamkära, which in course of time swallowed it up. Even as a Nātaka-dharma, connected with dramatic Samdhyangas, it had little individuality, and the attitude of the Daśa-rūpaka in not considering it separately is significant. The main view, however, which takes Laksaņa, like Alamkāra, as a beautifying characteristic, appears to have died out with Abhinavagupta's somewhat apologetic formulation. From Lakșanas Bharata goes on to the more interesting topic of kāvyālamkāras or poetic figures. It appears from his treatment that only four such poetic figures were known or recognised in his time, viz. upamā (simile), rūpaka (metaphor), dīpaka (lit. illuminator13) and yamaka (repetition 10 It is difficult to translate some of these terms, for there are no equivalents for them in European Rhetoric, and therefore attempts at such translation are as a general rule avoided here. The poetic figure dipaka may be generally explained as a figure in which two or more objects, some relevant and some irrelevant, having the same attributes, are associated together; or in which several attributes, some relevant and some irrelevant, are predicated of the same object. It is called dipaka or "illuminator" because it is like a lamp which, when employed for illuminating one object, illuminates others.

6 SANSKRIT POETICS

of words or syllables similar in sound). The Upamā is sub- divided into four kinds, according as the object compared (upameya) or the standard of comparison (upamāna) is one or many11, Bharata expressly making use of these technical terms. From another standpoint, five varieties of Upamā are distinguished and illustrated, viz. (1) praśamsopamā (2) nindo- pamā (3) kalpitopamā (4) sadršī upamā and (5) kimcit-sadršī upama12, according as these qualifications apply to the upamāna. Bharata is apparently unaware of the finer shades of distinction (grammatical or otherwise) introduced later on fnto the treatment of Upamā by Bhāmaha, Dandin and Udbhata, or of its comprehensive definition given by Vāmana; but the very fact that the idea of comparison was even by this time analysed thus far shows a considerable amount of speculation on this point13. Bharata's first two kinds, however, are criticised by Bhāmaha (ii. 37), but accepted by Dandin without question (ii. 30-31); while the name, if not the idea, of the third kind lingers in Vāmana iv. 2. 2. Of Rūpaka and Dīpaka14 no subvarieties are mentioned, and possibly these were comparatively late inventions. Of Yamaka, on the other hand, ten subspecies are elaborately defined and illustrated15, a number exceeding even that given by Bhamaha. It would appear that in the earlier stages of Poetics, what in later authors is known as a śabdālamkāra

11 Viz. (i) ekasya ekena (ii) ekasya anekena (iii) anekasya ekena (iv) bahūnām bahubhiḥ. 12 Abhinavagupta notices the reading asadrsi. 13 This figure is certainly one of the most ancient, and the idea of it was not unknown to Yāska (see vol. i, pp. 3-6). 14 The arrangement of the text dealing with these two figures differ in the different editions. But the wordings agree with the exception of samprakīrtitam xiv. 55 (in Kāvyamālā ed.), which is obviously a misread- ing for sampradīpakam (as Abhinava's comm. shows), given correctly in other editions. 15 Most of the names of these varieties have survived in Bhatti, Dandin and other writers; but they are in most cases differently defined, See vol. i, p. 54 footnote 2.

FROM BEGINNINGS TO BHĀMAHA 7

(of which the artifices of Yamaka in particular seem to have found the greatest favour) received a more elaborate treat- ment16, although the process repeats itself in comparatively modern decadent authors who delight in such external poetic devices. The later distinction between śabdālamkāra and arthālamkāra is not referred to by Bharata17, as also by Bhamaha; but Bharata uses the word sabdābhyāsa with reference to Yamaka, which term might have suggested, as Abhinava's commentary on this point indicates, the later classification, which is implied for the first time by Dandin's treatment. After the Alamkāras, comes the treatment of ten dosas (xvi. 84f) and ten gunas (xvi. 92f), which seem to have consti- tuted the orthodox number of faults and excellences of poetic composition. We shall have occasion to deal with the doctrine of Guna and Dosa in connexion with the Rīti- theorists, who for the first time take it up seriously; but it 16 As in Bhatti, Dandin, Vāmana, Rudrața, the Agni-purāna and Bhoja among older authors. Bhamaha gives only five varieties, and Udbhata is the only old writer who altogether omits its treatment. This figure, as Bharata's elaborate treatment would show, must have been very early comprehended, e.g., in Rāmāyaņa, Sundarakāņda v 15-17, in Rudradaman inscription of the 2nd century A.D. Possibly it was favoured as a not unlikely substitute for rhyme, which is nearly absent in earlier Sanskrit and which probably originated from antyānuprāsa in later literature. But Mammata and later writers, following perhaps the dictum of Anandavardhana (on ii. 16 f) that yamaka, in order to be really poetical, requires a special effort on the part of the poet, and is in no way accessory to rasa, allude to it but dismiss it in a few words. As critical insight into the aesthetic requirements of poetry grew, the number of such figures as depended for their appeal chiefly on clever verbal arrange- ment, as well as their treatment in Poetics, naturally dwindled, although yamaka itself (as well as anuprāsa) played a much larger part in later decadent poetry. Bhoja, with an inaccuracy characteristic of later writers, speaks of tricks like muraja-bandha as having been bharata- kathita! 17 Abhinava, however, reads into Bharata such a distinction, and in his ° Locana p. 5 he says: cirantanair hi bharata-muni-prabhrtibhir yamakopame śabdārthālamkāratveneste.

8 SANSKRIT POETICS

may be pointed out here that Bharata's enumeration and' definition of individual Doșas and Guņas do not exactly correspond to those of his nearest successors. Except keep- ing to the conventional number of ten (although Bhamaha introduces an eleventh fault from the standpoint of logical correctness18 and a list of ten intrinsic poetic faults in a different context, as he also mentions onlv three poetic excellences), both Bhamaha and Dandin do not appear to have accepted implicitly this part of Bharata's teaching. The faults mentioned by Bharata (xvi. 84) are:19 i. gudhārtha=circumlocution or periphrase (paryāya- śabdābhihitam20), ii. arthāntara=digression into irrelevant matter (avarņyasya varņanam21),

18 With the exception of the eleventh fault, Dandin is not only in complete agreement with Bhamaha but really follows the latter's enu -- meration and definitions almost literally. This point will be discussed later. 19 On Bharata's Doşas see V. Raghavan, Śrngāra-prakāśa, pt. ii, pp. 229-233. In Kautilya's Artha-sastra the defects of the art of writing are: vyaghata (contradiction), punarukta (repetition), apaśabda (gram- matical incorrectness) and samplava (misarrangement of words). 20 This phrase in Bharata cannot mean "expression by means of a synonym", for it would then be difficult to differentiate this Dosa from ekartha given below. Possibly this is the fault which appears in later writers as the figure of speech known as paryāyokta (=roughly, circum- locution or periphrase as a poetic figure). It is apparently so explained by Abhinavagupta. It is possible that later rhetoricians perceived that periphrase might sometimes be an ornament of expression, and thus analysed it into a poetic figure; and this may be taken as an instance in point of the process by which poetic figures were multiplied in later speculation. 21 Abhinava takes it in the sense of description of matters which. should not be described in words (sabdenāvarņanīyam api varnitam), and rejects the sense of aprakrta-varnanam. We are, however, inclined to prefer the sense of digression rejected by Abhinava. The defect vācyāvacana, mentioned by Mahimabhatta (p. 100). would possibly be. this Dosa of Bharata. This seems to be the fault in Māgha i. 43. Abhinava regards the sva-sabda-vācyatā-doșa of Rasa and Bhāva as.

FROM BEGINNINGS TO BHĀMAHA 9

iii. artha-hina=incoherence (asambaddha), or multipli- city of meaning (aśeşārtha), iv. bhinnartha=(a) rusticity or want of refinement (asabhya or grāmya), or (b) changing the desired sense by another sense (vivakşito'nya evārtho yatrānyārthena bhidyate), V. ekārtha=tautology (ekārthasya abhidhānam), vi. abhiplutārtha=aggregation of complete lines without merging them into a complete sentence (yat padena samasyate22), vii. nyāyād apetam = defective logic (pramāņa-varjitam), viii. vişama=defective metre (vrtta-doșa), ix. visamdhi=disjunction in which the words are not well knit23, x. śabda-hina=use of ungrammatical words (aśabdasya yojanam24).

But the faults discussed by Bhämaha are (ch. iv): i. apārtha=absence of complete sense25, ii. vyartha=incongruity with the context, iii. ekārtha=tautology (Bhāmaha noting that others call

included in Bharata's arthāntara-dosa, although it is not clear whether Bharata himself regarded this sva-sabda-vācyatā to be a Doșa at all. 22 Abhinava explains: abhiplutārtham yathā-sa rājā nīti-kuśalaḥ saraḥ kumuda-śobhitam / sarva-priyā vasanta-śrīḥ grīşme mālatikā- gamah/iti ; atra pratipadam arthasya parisamāptavād abhiplutārtham, eka-vākyatvena nimajjanābhāvāt. 23 The reading anupratisthasabdam yat is obviously incorrect. Abhinava's explanation is not clear ; but he appears to read anupārūdha- śabdam. By samdhi or samdhana he appears to mean compactness, congruity or merging, i. e. where the words are well knit. The Gaekwad ed. reading anupaślista-śabdam is more apposite, but perhaps it is not original. 24 aśabda= apaśabda, Abhinavagupta. 25 Both Bhamaha and Dandin say samudāyārtha-sūnyam yat, and this incompletion of the total sense arises, they explain, frem the non- satisfaction of the natural expectancy of words in a sentence (akānksā), a point already dealt with by grammarians and Mīmāmsakas.

10 SANSKRIT POETICS

it punarukta, which well-known term is apparently unknown to Bharata), iv. sasamśaya=ambiguity, v. apakrama=violation of syntactical regularity, vi. śabda-hīna=use of words not approved by correct usage (grammatical), vii. yati-bhrasta=deviation from the rules of metrical pause, viii. bhinna-vrtta=use of long or short syllables in the wrong place in a metre, ix. visamdhi=disjunction of euphonic liaison when it is necessary, X. deśa-kāla-kalā-loka-nyāyāgama-virodhi=inconsisten- cy with regard to (a) place (b) time (c) the fine or mechanical arts (d) worldly usage (e) logic (f) āgama (=dharma-śāstra26 i. e. codes of law or jurisprudence). Bhamaha also adds another fault to these ten, viz .. the Doșa which arises from a faulty logical proposition (pratijñā), à faulty middle term (hetu), and a faulty logical illustration (drstānta); but this blemish is treated in a separate chapter (ch. v), being interesting to Bhämaha from the standpoint of the logic of poetry. In another context, while discussing the general characteristics of poetry (i. 37f), Bhämaha mentions ten other defects which a poet should avoid, viz .: i. neyärtha=farfetchedness, when the sense does not follow from the logical order of words but has to be guessed from the general intention, ii. klista=obstruction of the sense, iii. anyārtha=disappearance of the sense, iv. avācaka=inexpressiveness, when in the expressed words the sense does not appear to be openly dominant,

26 āgamo dharma-śāstrāņi, loka-sīmā ca tat-kṛtā | tad-virodhi tadā- cāra-vyatikramanato ... // says Bhāmaha himself in explanation (iv. 48).

FROM BEGINNINGS TO BHÄMAHA 11

v. gūdha-śabdābhidhāna=use of difficult expressions, vi. ayuktimat=impropriety, e.g. making clouds massen- gers in poetry, vii. śruti-dusta=expressly indecent27, viii. artha-dusta=implicitly indecent (later writers calling it aślīlatva), ix. kalpanā-dusta=defective conception, where in the alliance of two words an undesirable sense is produced, x. śruti-kasta=unmelodious or harsh in sound28. After giving these ten Dosas Bhāmaha points out (i. 54-58) how these Dosas may sometimes become Gunas. As to how these two different series of ten faults are to be distinguished, Bhämaha says nothing ; but it is conceivable from his treat- ment that the latter concerns the inner nature or essence of poetry, while the former refers to only such defects as are more or less external. A glance at these two lists of faults, given by Bharata and Bhämaha respectively, will at once shew that while some of Bhamaha's faults correspond generally to Bharata's in name or in substance, Bhämaha in his elaborate treatment is certainly more advanced than his predecessor. It is also noteworthy that Bhamaha lays down, in his discussion of the last-named fault of śruti-kasta, the general proposition (i. 54) that a particular combination or arrangement sometimes makes even defective expressions allowable; in other words, a fault sometimes is converted into an excellence. Bharata, on the other hand, regards, as we shall see, all Gunas or excellences to be mere negations of Dosas or faults. After dealing with the Dosas, Bharata speaks of the Gunas or excellences of composition which are also enumerated as

27 Bhamaha's text is obviously corrupt on this point. But the distinction between śruti-dusta and śruti-kasta of earlier writers is explained by Abhinava in his ° Locana p. 82 (on ii. 12). 28 The examples given of this fault are words like ajihladat (Bhāmaha), or adhākşīt, akșautsīt, trnedhi (Abhinava, °Locana loc. cit.).

12 SANSKRIT POETICS

ten in number29. He states summarily at the outset (xvi. 91) that the Gunas are negations of the Dosas (gunā viparyayād eşam), an opinion which is indeed extraordinary in view of the fact that later writers like Vamana (ii. 1. 1-3) rightly consider Gunas in a theory of Poetics to be positive entities, of which the Dosas are the negations known by implication. It appears, on the other hand, that Guņas like mādhurya and audārya, mentioned by Bharata in xvi. 91 and 92, are not really, as defined by himself, negations of any particular defect discussed by him. Jacobi's explanation30 is probably right that Bharata's description of the Gunas as negations of the Dosas. is in conformity with the common-sense view of the matter, for it is not difficult for one to seize upon a fault instinctively, while an excellence cannot be conceived so lightly unless its essence is comprehended by differentiating it from a more easily understood fault. The Gunas, according to Bharata (xvi. 92), are the following31 : i. ślesa=coalescence of words, connected with one another through the aggregate meaning desired by the poet, and consisting of a subtlety which in appearence is clear but in reality difficult to compre- hend32. ii. prasāda=clearness, where the unexpressed sense appears from the word used through the relation of the easily understood word and sense33. 29 Abhinava reads kāvyasya gunāh in xvi. 92 as in Gaekwad and Chowkhamba eds., and not kāvyārtha-guņāḥ as in Kāvyamālā ed. 30 In Sb. der preuss. Akad. xxiv, 1922, p. 223. 31 The alternative readings show that the text for some Gunas also existed in two recensions. Abhinavagupta attempts throughout to approximate Bharata's Gunas to those of Vamana and strains to make Gunas of Sabda and Artha out of each. Hence he is not a safe guide for this portion of the text. On the Gunas of Bharata see V. Raghavan Šṛňgāra-prakāśa, pp. 271-81. 32 Hemacandra (p. 196) and Māņikyacandra (p. 191) remark: svabhāva-spaştam vicāra-gahanam vacah ślistam iti bharataḥ. Abhinava thinks that this excellence corresponds to Vāmana's śabda-guna śleşa. 33 Hemacandra and Māņikyacandra explain: vibhakta-vācya-

FROM BEGINNINGS TO BIIĀMAHA 13

iii. samatā=evenness, which is easy to understand and in which there is no redundance of expression nor excess of cūrna-padas34. iv. samādhi=superimposition (samādhāna) of some- thing special or distinguishing in the sense35.

ācakāyogād anuktayor api sabdārthayoḥ pratipattiḥ prasāda iti bharatah. Jacobi proposes (ZDMG lxiv, p. 138 contd. fn.) to read mukhya instead of mukha in the text, and thinks that Bharata's prasāda corresponds to Dandin's samadhi. But Abhinava reads sukha and ·explains: sukhayati, na prayatnam apekşate yah śabdārthah. Perhaps by this Guna, Bharata means to imply some kind of hint (anukta artha), transparent from the words used (such as we find, e. g., in the figure mudrā in Candrāloka, ed. Jivananda, v. 139, and Kuvalayānanda, ed. N.S.P., 1917, pp. 146-7), which may correspond partly to the metaphori- cal mode of expression included by Vämana in his peculiar definition of vakrokti (iv. 3. 8.), or comprised by later writers under lakşanā or upacāra. Referring to Vāmana's definition of artha-guņa prasāda as .artha-vaimalya (iii. 2. 3), Abhinava seems to support our suggestion when he says so'rtho vaimalyāśrayo'pi vaimalyam upacārāt, thus attempting to approximate Bharata's prasāda to Vāmana's artha-guņa .of that name. 34 The cūrna-pada is defined by Bharata himself in xviii. 50b, 51 Cf. Vāmana 1. 3. 22, 24 where cūrna is the name given to a kind of prose, which contains short compounds. Commenting on this passage in Vāmana, Gopendra Tippa Bhūpāla interprets the word as: cūrna-padena upacārād vyasta-pada-samāhāro lakşyate, tena vyasta-pada-bahulam cūrņam. Vāmana himself in his Vrtti gives two characteristics, viz., adirgha samāsa and anuddhata pada, short compounds and soft vocables. Referring to Vāmana's śabda-guņa of the same name in iii. 1. 12, and trying to approximate it to Bharata's samata, Abhinavagupta remarks: śabdānām samatvāt samaḥ, cūrņa-padair a-samāsa-racanā yatra sātišayā na bhavati, ......... dīrgha-samāso' pyatyanta-samāsaś ca vişamatā, tad- viparyayeņa samatā, upakrānta-mārgāparityāga-rūpety uktam. 35 Abhinava explains: yasyārthasya abhiyuktaih pratibhānātiśaya- vadbhir viseşo'pürvaḥ svollikhita upapadyate sa samāhita-manaḥ- sampādya-viseşatvad artho visistah samadhih. In the second line of the text Abhinava reads parikīrtitah (and not parikīrtyate), and takes arthena as referring to the word samādhi: samādhi-śabdasya yo'rthah parihāra- lakşaņas tena parikīrtitaḥ paritaḥ samantād ākrāntyā uccāraņe sampan- nah. This explanation of sampanna is probably given to make the definition correspond to that of Vāmana's ārohāvaroha-kramaḥ samā-

14 SANSKRIT POETICS

v. mādhurya=sweetness, where a sentence heard or repeated many times does not tire or disgust36. vi. ojas=strength, which consists in the use of varied and dignified compounded words, having letters agreeable to one another37. vii. saukumārya=smoothness, where an agreeable sense is realised by means of agreeably employed words and well-connected euphonic conjunctions38. viii. artha-vyakti=explicitness, which describes the nature of things, as they appear in the world. by means of well known predicates39. dhiḥ (iii. 1. 13); for Abhinava goes on explaining ākrāntyoccārane ārohāvaroha-krama eva, the āroha and avaroha depending, as he discuss- es in detail, on uccārana. Hemacandra and Māņikyacandra explain Bharata's definition simply as arthasya guņāntara-samādhānāt samādhir iti bharata. 36 Abhinava reads śrutam and vākyam instead of krtam and kāvyam in the printed text (Kāvyamāla ed.); and this is supported by what Hemacandra and Mänikyacandra say with regard to this excellence of Bharata. The other eds. give the words correctly. 37 Abhinava reads bahubhih (instead of vividhaiḥ) and sānurāgaiķ (instead of sa tu svaraih), explaining the latter reading thus: yatra varņair varņāntaram apekșate tatra sānurāgatvam. Hemacandra, how- ever, attributes a different definition to Bharata, viz., avagītasya hīnasya vā śabārtha-sampadā yad udāttatvam nişiñcati kāvayas tad oja iti bharatah; and Manikyacandra says to the same effect (this follows the reading of Recension B): avagīto'pi hīno'pi syād udāttāvabhāsakaḥ/ yatra śabdārtha-sampattya tad ojah parikīrtitah/. 38 Abhinavagupta reads sukha-prayojyaiś chandobhiḥ (for sukha- prayojyair yac chabdaih) in the text. Hemacandra and Mānikyacandra: simply paraphrase: śukha-śabdārtham saukumāryam iti bharatah .- The "agreeable sense" (sukumārātha), which corresponds to the anisthu- rākşara-prāyatā of Dandin's Guna of this name, or to the ajarathatva or apārusya of Vāmana's twofold saukumārya, implies probably the avoidance of disagreeable or inauspicious (amangala) statements: Thus instead of mrtah, one should say kīrtiseşam gatah. It is on this ground that theorists after Mammata object that this is no Guna but a negation of the amangala dosa, which some writers include in the fault known as aślīlatva. 39 The text is obviously corrupt. Abhinava reads suprasiddhā-

FROM BEGINNINGS TO BHĀMAHA 15

ix. udāra=exaltedness. where there are superhuman sentiments, varied feelings, and the Erotic and the Marvellous moods40. . kanti=loveliness, which delights the ear and the mind, or which is realised by the meaning conveyed by graceful gestures (līlādi41). It will be noticed from this enumeration that in some cases it is difficult to see what Bharata means exactly by a particu- lar Guna, and that the classification is by no means exhaustive nor free from overlapping. On the other hand, some of the Guņas can be taken (as Abhinavagupta takes them) as appro- ximating roughly to the individual śabda-gunas and artha- gunas elaborated by Vamana and other later theorists. The development of the Guna-doctrine is intimately connected bhidhānā tu instead of suprasiddhā dhātunā ca, which phrase, how- ever, may mean use of well known verbs. Read also loka-dharma° instead of loka-karma° Hemacandra explains: yasminn anyatha- sthito'pi tathā-sthita evarthah pratibhāti so'rthavyaktih. He also points out that this Guna of Bharata corresponds to Vamana's artha-guņa arthavyakti (defined as vastu-svabhāva-sphutatvam iii. 2. 13), and would be equivalent to the poetic figure jāti or svabhā- vokti of Dandin and others. Cf. Mammata p. 583: abhidhāsya- māna-svabhāvoktyalaņkāreņa vastu-svabhāva-sphutatva-rūpārthavya- ktiḥ svıkrtā, but Viśvanātha would include it in prasāda-guņa. 40 The implication of adbhuta Rasa in this excellence and the characteristic that it deals with divya-bhāva indicate a certain utkarşavan dharmah, causing wonder, such as Dandin's udara would contain. See the illustration of this Guna given by Hemacandra p. 199. The inclusion of the śrngāra and adbhuta Rasas makes this Guņa of Bharata correspond partly to Vāmana's artha-guņa kānti. But Daņdin adds : ślāghyair viśesaņair yuktam udāram kaiścid işyate. There is no reference in Bharata to such 'praiseworthy epithets', although Hemacandra says: bahubhih sūksmaiś ca viseşaih sametam udaram iti bharatah. In the Agni-purana 346. 9, however, we. read : uttāna-padataudāryam yutam ślāghyair viśesanaih, which might be a direct echo of Dandin. 41 hīladi =līlādi-ceta. Abhinavagupta. This would be com- prehended by the dipta-rasatvam of Vāmana's artha-guņa kānti,- Kauțilya (ii. 28) mentions the following characteristics of the art of writing: artha-krama (arrangement of subject-matter), sambandha (relevancy), paripūrņatā (completeness), mādhurya (sweetness), audārya (dignity), and spaștatva (elearness).

16 SANSKRIT POETICS

with the central theory of the Riti-school and will be dealt with later ; but it may be pointed out here that although the definitions of the individual Gunas, given by Bharata, do not correspond exactly to those of later writers, there can be no doubt that here we have for the first time a definite statement, if not a proper theoretic treatment, of the doctrine. The disagreement between different theorists with regard to the definitions of individual Gunas is a well known fact in the history of Sanskrit Poetics, and one need not therefore be sur- prised that later authors give us definitions which do not agree with those of Bharata. To Bharata, again, the relation of the gunas to rīti, as elaborated by Vāmana, or to rasa, as first clearly enunciated by the Dhvanikāra and Anandavar- dhana, was probably unknown. So was also Vämana's distinc- tion between sabda-guna and artha-guna, although Bharata's gunas are mostly of the nature of artha-gunas, and some of them can be interpreted (as done by Abhinavagupta) as constituting śabda-gunas as well. But the number and nomenclature of the Gunas, as well as the substance of some of them, as outlined by Bharata, are conventionally adhered to by all later writers, excepting Bhämaha who, as we shall see, was a radical thinker in this respect, until we come to the Dhvanikāra and his followers who give a new interpretation to the Guna-doctrine. It is also important to note that Bharata takes the guņas, as well as the doşas and alamkāras, to be subservient to the purpose of awakening rasa, which is taken as the principal business of the drama. In this he anticipates and probably influences the view of the Dhvani- kāra and his school who, as we shall see, borrow Bharata's idea of Rasa from the case of the drama and apply it to that of poetry.

( 2 ) It has been noted above that Bharata makes all these elements, lakşaņa, guņa, doşa and alamkāra, subordinate to 'the principal purpose of awakening rasa in the drama.

FROM BEGINNINGS TO RHĀMAHA 17

'These elements constitute what he calls vācika abhinaya {defined in viii. 6,9), which is dealt with in chapters xiv-xx and which forms an important factor, the anubhāva (vii. 5), in calling forth the Rasa. Hence Bharata expressly considers (xvi. 104 f) the question of their employment in relation to Rasa. It is necessary, therefore, to consider here briefly Bharata's teachings regarding Rasa42, which is dealt with in the Nātya-śāstra, chapters vi and vii43. It must be observed at the outset that Rasa does not appear to be Bharata's principal theme, and that it is discussed only in connexion with his exposition of dramatic representation with which he is principally concerned. It is not surprising, therefore, that Rājaśekhara, probably following some current tradition, should regard Bharata as an authority on Rpaka (drama) rather than on Rasa, and mention one Nandikeśvara44 as the original exponent of the Rasa-doctrine which, if Rajaśekhara is right, Bharata must have borrowed and worked up into his dramaturgic system. That the Rasa- doctrine was older than Bharata is apparent from Bharata's own citation of several verses in the Arya and the Anustubh metres in support of or in supplement to his own statements ; and in one place, he appears to quote two Arya-verses from an unknown work on Rasa45. The idea of Rasa, apart from any theory thereon, was

42 The question has been dealt with briefly in S. K. De, Theory of Rasa in Sir Asutosh Mookerjee Silver Jubilee Commemoration Volume, Orientalia, vol. iii, 1922. reprinted in Some Problems of Skt. Poetics, Calcutta 1959, pp. 177-235. 43 In some other chapters Bharata deals with the cognate topics of the Nāyaka and Nāyika and their adjuncts and emotional states, which will be referred to in their proper place below (ch. viii). 44 See vol. 1 p. 1, 2, 19. 45 atrārye rasa-vicāra-mukhe, ed. Kāvyamālā p. 67. The line is wanting in Grosset's edition. Keśava Miśra, a comparatively recent writer of the 16th century. speaks of one bhagavan Sauddhodani who, according to him, was a sūtra-kūra on Rasa; but the opinions of this otherwise unknown writer (see vol. 1, p. 219), as recorded by Keśava Miśra, do not deviate materially from the conventional views

25

18 SANSKRIT POETICS

naturally not unknown to old writers; and Bharata's treatment would indicate that some system of Rasa, however un- developed, or even a Rasa school, particularly in connexion- with the drama, must have been in existence in his time. But the bearings of this doctrine on poetry were seldom discussed, and the importance of Rasa as one of the essential factors of poetry was indeed naïvely understood but was not theoreti- cally established.46 As Dramaturgy was in the beginning a separate study, from which Poetics itself probably took its cue, the Rasa-doctrine, which sprang up chiefly in connexion with this study, confined its activity in the first stage of its. development to the sphere of dramatic composition and exerted only a limited influence on poetic theories.47 The importance of this dramaturgic Rasa-system must have been somewhat overshadowed by the early dominance, in Poetics.

of those later writers of the new school who admit the essential- ity of Rasa. 46 We get the first definite exposition, as we shall see, of the idea of Rasa and its relation to poetry in the works of the Dhvani- kāra and Ānandavardhana : and its importance in poetry, as distin- guished from the drama, was probably understood from that time. Mägha in some verses (see vol. i pp. 61, fn 3) shows him- self conversant with some theory of Rasa; but it is to nātya- rasa (such as described by Bharata) rather than to kāvya-rasa that he appears to refer. It is not maintained that older Sanskrit Poetry was devoid of Rasa or that the earlier poets never possessed any idea of it; but it is suggested that the theory of Rasa was not critically set forth, nor its aesthetic importance in poetry properly understood until the Dhvanikara and Anandavardhana came into the field. The presence of Rasa is such a familiar fact in Sanskrit Poetry as well as in comparatively modern Sanskrit writers on Poetics that one is apt to lose sight of this fact of historical im- portance. 47 That the doctrine of Rasa was originally associated with dramaturgy and later on applied to Poetics is clear from the tradi- tion which survives even in very late writers, and makes them not only discuss the theory directly in connexion with the drama (e. g. Viśvanātha) but even borrow the illustrations mostly from dramatic poetry. Abhinava speaks of the Kāvya as loka-nātya- dharmi-sthānīya, and says : nātya eva rasah, kāvye ca nātyāya- māna eva rasah kāvyārthaḥ.

FROM BEGINNINGS TO BHĀMAHA 19

proper, of the Alamkāra and Rīti systems, whose traditions are carried on by the two earliest writers on Poetics, Bhāmaha and Dandin, both of whom allow a very subsidiary place to Rasa in their scheme ; but at the same time the comparative antiquity of such a Rasa-system in connexion with the drama, going back to a time even earlier than Bharata, cannot be doubted. The oldest known exponent of this system is Bharata, from whom spring all later systems and theories such as we know them, and whom even Anandavardhana himself (p. 181), in applying the Rasa-theory to Poetics, names as his original authority. It is necessary for this reason to take into account Bharata's doctrine of nātya-rusa as the original source of the doctrine of kāvya-rasa elaborated in later Poetics. But long before the new interpretation of the relation of Rasa to poetry, given by the Dhvanikāra and authoritatively established by Mammata, was dominant, Bharata's views on Rasa appear to have been discussed in some detail in drama- turgic systems with the result that divergent theories came to prevail under the names of Bhatta Lollata, Sankuka, Bhatta Näyaka and others, all of whom are supposed to have been commentators on Bharata's text48, or at least to have taken Bharata as their starting point. With Bharata, therefore, we arrive at a distinctly definite landmark in respect of the Rasa-doctrine. It is worth noticing, however, that although all theorists take Bharata as their starting point and build up their own theories round his authoritative, if somewhat meagre, text, Bharata himself, like all old masters, is tantalisingly simple in his statements ; for the subject does not appear to have yet been brought into the realm of scholastic speculation. Bharata's work is encyclopaedic in its scope, but its primary theme is the drama and its conception of poetry dramatic, a view which perhaps inspired Vamana's partiality towards

48 See vol. i pp. 31f.

20 SANSKRIT POETICS

dramatic composition already alluded to, and which is concisely put by Abhinavagupta by saying kāvyam tāvad daśa- rūpātmakam eva. In such a composition Rasa, according to Bharata, should be predominant, and there are numerous passages which clearly indicate that there can be no sense of poetry, in his view, without Rasa49. Although Bharata does not enter into technicalities, he seems to be of opinion that the vibhāvas and the anubhāvas, which later theory takes to be essential factors, call forth or evolve Rasa; but he is not clear as to what this process of evolution exactly is. He takes the bhava as the basis of Rasa and explains it generally as that which brings into existence the sense of poetry through the three kinds of representation, viz., through words, gestures, and internal feelings (vāg-anga-sattvopetān kāvyārthān bhāva- yantiti bhavah). This Bhava, which consists of an emotion- al state of the mind, reaches, when permanent and not transi- tory, the state of Rasa through the elements known as vibhāva and anubhāva. A vibhāva is explained thus: vibhāvo nāma vijñānārthaḥ, vibhāvyante'nena vāg-anga-sattvābhinayā ity ato vibhāvah. The term vibhāva, therefore, is used to connote knowledge or cognition, and is explained generally as denoting that which makes the three kinds of representation capable of being sensed. In the same way, the anubhāva is explained as that which follows upon and makes the three kinds of representation actually sensed. The third element of Rasa, the vyabhicari-bhava, consists of accessory emotional facts which help and strengthen it, and is etymologically explained as: vi abhi ity etāvupasargau, cara gatau dhātuh, vividham ābhimukhyena rasān carantīti vyabhicāriņah. As to what relation these elements bear to Rasa and how this state of relish is brought about, Bharata simply lays down in a cryp- tic formula: vibhāvānubhāva-vyabhicāri-samyogād rasa- nispattih, a formula which, in spite of his own explanation, is so ambiguous with respect to the exact significance of the

49 e. g. na hi rasād rte kaścid arthah pravartate, ed. Grosset p. 87=ed. Kāvyamālā p. 62. Cf. also vii. 7.

FROM BEGINNINGS TO BHĀMAHA 21

terms samyoga and nispatti that a great deal of controversy has centred round their interpretation, giving rise to a number of theories about Rasa. Bharata's own explanation, if it can be called an explanation, is that just as a beverage is accom- plished through various seasoned articles and herbs, so the permanent mood (the sthāyi-bhāva), reinforced (upagata) by various bhavas, attains the state of Rasa ; and it is so called because its essence consists in its taste or relish (āsvādyatvāt), this being the etymological meaning of the word rasa50. He also explains51 that the sthāyi-bhāva is the basis of Rasa be- cause it attains, as it were, mastery or sovereignty among forty-nine different bhavas mentioned by himself52, which naturally rest upon it as being presumably the principal theme or mood in the composition in question. Nothing definite can be concluded from all this except that, in Bharata's opinion, the sthāyi-bhava or the principal mood in a composition is the basis of Rasa, the essence of which consists in asväda or relish by the reader or spectator, while the vibhāva, anubhāva and the vyābhicāri-bhāva awaken this state of emotional realisation or 'relish' in the reader's mind. But this explanation by its very ambiguity or vague- ness taxed the ingenuity of theorists and commentators, its general trend anticipating theories like the utpatti-vāda of Lollața and the anumiti-vāda of Śankuka, and special terms in the passage in question like vyañjita and sāmānya-guņa- yoga suggesting specialised doctrines like the vyakti-vāda of Abhinavagupta and the bhukti-vāda of Bhatta Nāyaka. The original outlines of the theory, however, are accepted as fixed by Bharata. It is practically admitted on all hands, on semi-psychological considerations of poetry, that the Rasa 50 ed. Grosset p. 87=ed. Kāvyamālā p. 62. It should be noticed that all the terms which describe the essence of Rasa such as rasanā, carvaņā, or āsvāda, refer etymologically to the physical pleasure of taste; this point will be dealt with below. 51 ed. Grosset p. 102, ll. 7-19=ed. Kāvyamālā p. 70, 11. 13-22. 52 viz., 8 sthāyi-bhūvas, 8 sāttvika bhāvas and 33 vyabhicāri- bhāvas.

22 SANSKRIT POETICS

is a state of relish in the reader of the principal sentiment in the composition, a subjective condition of his mind which is brought about when the principal or permanent mood (sthāyi- bhava) is brought into a relishable condition through the three elements, the vibhāva, the anubhāva and the vyabhicāri-bhāva, exhibited in the drama. Of these elements, the first two are important, the vyabhicarin being only concomitant or acces- sory. Bharata's explanation of these terms is rendered with greater precision by his followers. By the sthāyi-bhāva in poetry and drama are meant certain more or less permanent mental states, such as Love, Grief, Anger or Fear. This permanent mood, constituting the principal theme of a com- position and running through all other moods like the thread of a garland, cannot be overcome by those akin to it or those opposed to it, but can only be reinforced. Those elements which respectively excite, follow and strengthen (if we may use these expressions) the sthāyi-bhāva are in poetry and drama known as vibhāva, anubhāva and vyabhicāri- (also call- ed samcāri-) bhāva53, corresponding in ordinary life (as opposed to the extraordinary world of poetry) to the mundane causes and effects (laukika kāraņa and kārya). Devoid of technicalities, a vibhāva may be taken as that which makes the permanent mood capable of being sensed, an anubhāva as that which makes it actually sensed, while a vyabhicāri-bhāva is that which acts as an auxiliary or gives a fresh impetus to it. In the case of Love as a permanent mood, the stock-examples given of a vibhāva are women and the seasons; of anubhāva, glance and embrace; of vyabhicārin, the transient subordinate feelings of joy or anxiety. Now Bharata says that the reader is enabled to realise or relish as Rasa the permanent mood of a composi-

53 Ballantyne renders these terms conveniently, if not adequately, as the Excitant, the Ensuant and the Accessory respectively, a nomenclature which is followed by Ganganatha Jha in his trans- lation of the Kūvya-prakāśa. Jacobi, however, uses (ZDMG. 1902, pp. 394 f) the terms Factor, Effect and Concurrent.

FROM BEGINNINGS TO BHĀMAHA 23

tion through a certain correlation of these elements with the permanent mood, the correlation taking place apparently for the purpose of manifesting it and bringing it to a relishable condition. The question, therefore, arises, to which Bharata himself gives no definite solution, viz. what relation these elements bear to Rasa, or in other words, how do they bring about this subjective condition of relish in the reader's mind, the solution depending, as we have noted above, upon the explanation of the two much-discussed terms samyoga (lit. correlation) and nispatti (lit, consummation) in the original sutra of Bharata. This is the central pivot round which all later theories move, and we shall take it up again in their connexion below (ch. iv). Bharata mentions eight different moods or Rasas in the drama, of which a detailed account is given in Nātya-śāstra ch. vi, which is the authoritative source drawn upon by all later writers, although they sometimes differ, as we shall see, in the enumeration of the orthodox number of eight. Properly speaking, the primary Rasas, according to Bharata, are only four in number, viz., śrngāra (the Erotic), raudra (the Furious), vīra (the Heroic) and bibhatsa (the Disgusting). The other four Rasas proceed from these, as follow: hasya (the Comic) from śrngāra, karuņa (the Pathetic) from raudra, adbhuta (the Marvellous). from vīra, and bhayānaka (the Terrible) from bībhatsa (xvi. 39-40). The eight sthāyi-bhavas or permanent moods, correspond- ing to the eight Rasas, are given categorically as (i) rati (Love) (ii) hāsa (Mirth) (iii) krodha (Anger) (iv) utsāha (Courage) (v) bhaya (Fear) (vi) jugupsā (Aversion) (vii) vismaya (Wonder) and (viii) śoka (Sorrow), forming the basis respectively of śrngāra, hāsya, raudra, vīra, bhayānaka, bībhatsa, adbhuta and karuņa. The vyabhicāri-bhāvas are mentioned as thirty-three in number and include the subordi- nate feelings of self-disparagement (nirveda), debility (glāni), apprehension (śankā), envy (asūyā), intoxication (mada), weariness (śrama), indolence (ālasya), depression (dainya),

24 SANSKRIT POETICS

reflection (cintã), distraction (moha), recollection (smrti), equanimity (dhrti), shame (vrīdā), unsteadiness (capalatā), joy (harsa), flurry (āvega), stupefaction (jadatā), arrogance (garva), despondency (vişāda), longing (autsukya), drowsiness (nidrā), dementedness (apasmāra), dreaming (supta), awakening (vibodha), impatience of opposition (amarsa), dissembling. (avahittha), sternness (ugratā), resolve (mati), sickness (vyādhi),. madness (unmāda), death (maraņa), alarm (trāsa) and doubt. (vitarka)54. The sāttvika bhāvas, which can be taken generally as involuntary evidences of internal feeling, are then specified as eight in number55, viz., stupor (stambha), perspiration.

54 The English equivalents follow generally Ballantyne's render- ings .- It must be borne in mind that the vyabhicāri-bhāvas are in- dependent Bhävas but occurring as accessory or concurrent to the principal mental state depicted, which is known as the sthāyi-bhāva. Sometimes it may happen that the vyabhicārin is principally mani- fested in a composition, and the sthayin is merely awakened; such cases later theorists would call bhavas (and not rasas) which are thus incomplete rasas. Attempts have been made to distinguish bet- ween Rasa and Bhäva, and this question will be discussed later on. It would appear from the enumeration of the vyabhicri-bhavas that the older theorists consider many conditions from the spiritual point of view, which we would regard from the standpoint of the body (e.g. vyūdhi or maraņa). See Jacobi in ZDMG lvi, 1902, p. 395 fn 2. 55 Regarding the sāttvika bhāva (which later theorists, e.g. Abhinavagupta, connect with the sattva guna of the Sāmkhya philo- sophers), Bharata says (ed. Grosset p. 129=ed. Kāvyamālā, p. 82): iha hi sattvam nama manah-prabhavam, tac ca samahita-manastvad utpadyate, manaḥ-samādhānāc ca sattva-nirvrtir iti; tasya yo'sau svabhāvo romāñcāsrādi-krtah sa na śakyate'nya-manasā kartum iti; loka-svabhāvānukaraņāc ca nātyasya sattvam īpsitam. Thus Bharata connotes by it certain tokens of mental feelings, delineated in the dramatic representation by an imitation of human nature through steady concentration of the mind. Bharata adds the illustration: iha hi nūtya-dharma-pravṛttāḥ sukha-duhkha-krto bhāvās tathā sattva- visuddāh kāryā yatha svarūpa bhavanti. For, how can sorrow, he says, which must be manifested by weeping, or joy which must be ex- pressed by laughter, be delineated except by these involuntary evidences ? This is apparently what Bharata means by sattvābhinaya or sāttvikābhinaya. In the Daśa-rūpaka, however, sattva is taken to mean 'a sympathetic heart' and sättvika is explained as sattvena'

FROM BEGINNINGS TO BHĀMAHA 25

(sveda), horripilation (romāñca), break of voice (svara- bhanga), trembling (vepathu), change of colour (vaivarnya), tears (aśru) and loss of consciousness (pralaya). This psycho-physical analysis, however formal it may appear to us, is taken up in detail in chapter vii, and each of these states is categorically defined and illustrated strictly from the standpoint of the drama; but in later literature they are established authoritatively for pcetry as well56.

( 3 ) This is a rough outline of the teachings, relevant to Poetics. proper, that we can gather from the somewhat meagre text of Bharata, and it may be taken as an outline of the discipline as it existed in the earliest known period of its history. With Bhāmaha, Dandin and the Dhvanikāra, on the other hand, begins the next period of its history, a comparatively brief but exceedingly important stage of extraordinary creative genius. Of the period anterior to Bharata our knowledge is extremely scanty ; and between Bharata and the definite formulation of poetic theories which begins with Bhamaha, lies, again, a long gap of which we do not possess much knowledge. It is clear, however, that certain poetical gunas, doşas, alamkāras and laksanas were known to Bharata and dealt with by him even as decorative devices of the drama- tic speech. It would not be wrong to presume from this fact that the study of Alamkāra-sästra, even if it was not yet fully

nirvrttah (Cf Sāhitya-darpaņa iii. 134). Bhānudatta, in his Rasa- taranginī gives a somewhat different interpretation, and takes sāttvika to mean physical gestures as evidences of natural feeling (sattvam jīva-śarīram, tasya dharmah sāttvikāh, ittham ca śārīra-bhāvāh stambhādayah sāttvikā bhāvā ity abhidhīyante). Later writers like the author of the Kāvyaprakūsa-pradipa bring in philosophical implications and interpret sāttvika as originating in the sattva-guna. Whatever difference there might be as to the meaning of the term sättvika itself, all the writers on this subject agree in applying the term to denote those gestures (enumerated above) which give an involuntary expression to internal feelings. 56 e. g., in Mammata.

26 SANSKRIT POETICS

developed and self-conscious, was probably older than Bharata himself. It follows from this conclusion that the tradition of opinion, which crystallises itself in the oldest available manuals of Bhamaha and Dandin, or in the memorial verses of the Dhvanikara, comes to us in a definite shape indeed at a date much posterior to Bharata, but it is probable that in substance, if not in actual formulation, it may have been much anterior to Bharata, who himself gives an indication of such teachings. Excepting what we get in Bharata, however, the history of this process of crystallisation .(for the different systems appear in a relatively developed form in Bhämaha and others) is not known to us ; but it must have covered a tentative stage whose productions, if they had been extant today, would have shown Bhamaha, Dandin and the Dhvanikara in the making, and would have partly filled up the presumably long gap between Bharata and these earliest formulators of Poetics. Even if they were the earliest formulators, neither Bhā- maha, nor Dandin, nor the Dhvanikāra, however, claims entire originality of having created the system which he individually represents. None of them can be taken as the absolute found- er of the particular doctrines of alamkāra, riti or dhvani with which they are respectively associated ; and with them we do not start at the absolute beginnings of the discipline. Indebtedness of these writers to their predecessors in the line, acknowledged by themselves, has already been noted in the first volume of this work (pp. 50, 67-68, 109) ; but apart from such explicit admissions, one can easily argue that certain fundamental concepts and formulas (such as vakrokti, rīti, guņa or alamkāra) appear in writers like Bhamaha without a preliminary explanation, as things traditionally handed down or already too well known to require any detailed dis- cussion. It is also unthinkable that these early writers could have, as they certainly do not claim to have, evolved by them- selves the relatively developed form and treatment of the main topics of Poetics in the absence of earlier tentative works.

FROM BEGINNINGS TO BHĀMAHA 27

It will be profitable, therefore, to pause and enquire if these works of Bhamaha and others give us any indication of the stage or stages through which the discipline might have passed in the interval between Bharata and themselves. For- tunately there are passages in these writers which would give us hints as to the existence of such intermediate stages. Jacobi has already shown57 that Bhamaha, in his treatment of the poetic figures, groups them in a curious but suggestive way which probably indicates the different stages in the growth and multiplication of such figures before his time. Unlike later authors who, adopting some definite principle of classi- fication, enumerate the poetic figures en masse58, Bhāmaha begins (ii. 4) by naming and defining first a group of five such figures only, and then goes on to enumerate other such limited groups of figures, taking 24 remaining figures in a final group. The first group of figures thus mentioned comprises anuprāsa, yamaka, rūpaka, dīpaka and upamā, recognised, as Bhāmaha says, by others (anyair udahrtah) and accepted by himself. These five correspond in reality to the four ancient poetic figures known to and defined by Bharata, viz. yamaka, rūpaka, dīpaka and upamā. The additional figure anuprāsa, mentioned by Bhämaha, can be taken as falling in the same class as yamaka59, the one being varņābhyāsa and the other padābhyā-

57 In Sb. der preuss. Akad. xxiv, 1922, pp. 220-222. 58 By the time of Dandin, for instance, a large number of poetic figures appears to have been recognised, and he does not find any necessity of 'reporting' them or mentioning them successively in groups as Bhamaha does; but he arranges them in his own way, taking the arthālamkāras first and the śabdālamkāras next, in two separate chapters. Udbhata, a follower of Bhamaha, deals with the first three groups of Bhämaha in the first three chapters of his own work (omitting, however, Bhämaha's phrases like anyair udāhrtāh, aparah, abhihitāh kvacit etc, with reference to these groups), the other three chapters taking up the remaining twenty-four figures of the last group. Although he follows generally the sequence as well as the definition of Bhämaha, he does not recite them in the manner of groups after Bhamaha 59 The distinction between yamaka and anuprāsa may be explained thus: in the anuprasa there is a repetition of one or more consonants,

28 SANSKRIT POETICS

sa, while both are what Bharata would call śabdābhyāsa. Abhinavagupta very significantly takes60 anuprāsa as implied in yamaka by Bharata; and the very fact that the anu- prāsa in Bhāmaha is thus clearly differentiated from yamaka may indicate further refinement in the analysis of these figures. and betoken a somewhat later stage. In course of time, six other figures appear to have been analysed and added, and Bhämaha mentions them next in a group in ii. 66. They are ākşepa, arthāntara-nyāsa, vyatireka, vibhāvanā, samāsokti and atiśayokti. Of these there is no trace in Bharata. This constitutes probably the second stage of development. in which can also be included a seventh figure värtta, which is referred to by Dandin in i. 85, but which is not accepted by Bhämaha as non-poetic utterance in which there is no Vakrokti (ii. 87)61. The third stage indicated by Bhämaha's treatment does not appear to have been very pro- ductive, for in it we have the addition of only two more figures yathāsamkhya and utprekșā (ii. 88), and possibly of a. third svabhavokti. In this connexion it is noteworthy that by Daņdin's time svabhāvokti (also called jāri, recognised by Bāņabhatta) is established as the primary or first figure (ādyā

sometimes but not necessarily along with the accompanying vowels; in yamaka, the consonants as well as the vowels are repeated strictly in the same order or sequence. In yamaka, the same group. of vocables is repeated but it need not have the same meaning, and may even be quite meaningless in itself; but in the repetition of the anuprusa one should consider the meaning. Nicitam kham upetya nīradaih priyahīna-hrdayavanīradaih would be an example of one variety of yamaka; while adri-droņī-kutīre kuhariņi hariņaratayo yāpayanti would be an example of anuprāsa. 60 tenānuprāsa-lāțīyāder anena ( =vamakena) evopasamgrahāt. 61 V. Raghavan (Some Concepts, p. 99f), however, argues that in Bhamaha Vārttā is no name of an Alamkāra .- Here Bhāmaha speaks of three other poetic figures hetu, sūksma and lesa, which he rejects as not involving that he calls vakrokti. These, however, appear not in direct connexion with this enumeration of poetic figures but in the immediate context of vakrokti alluded to in the previous verse. These figures, however, are illustrated (as interpreted by commenta- tors) by Bhatti, and were probably recognised before Bhāmaha's time,

FROM BEGINNINGS TO BHĀMAHA 29

alamkrti), while in Bhamaha it occupies a rather dubious position ; for the latter, in pursuance probably of his peculiar theory of Vakrokti, does not appear to favour this figure very much. With regard to utpreksa 62, which is indeed an im- portant addition to the Poetics of this period, Bhāmaha reports (ii. 88) that Medhavin called it samkhyana63. Is it possible that Bhämaha's predecessor Medhavin was the first to analyse and name this figure? Dandin, in a well known passage, considers in detail the question whether the word iva is indicative of utpreksā, a question which was apparently disputed by other rhetoricians between Bhamaha and Dandin; but in Vamana we find this is already an established fact. .From Bhamaha ii. 40, again, we learn that Medhāvin set up

62 The utpreksū and āksepa as poetic figures are expressly men- tioned by Subandhu, ed. Srirangam p. 146; while upamā and dīpaka are recognised by Bāņabhatta (introd. to Kādambarī). Both Subandhu and Bāna speak of composition enlivened by ślesa; but it is not perfectly clear whether they mean by it a prabandha-guna (as in Pharata) or the specific poetic figure of that name. Bhamaha not only speaks of ślesa but mentions three kinds or cases of its occurrence. Subandhu's boast of having used Slesa (as his commenta- tors interpret) in every word of his composition is not an idle one; and from his use of it, one can indeed incline to the view that Subandhu's Slesa is no other than our modern poetic figure of the same name, especially as Bharata's definition of Slesa as a Guņa is hardly applicable to Subandhu's case. 63 The text reads (Bhāmaha ii. 88): yathāsamkhyam athotpreksām alamkāra-dvam viduḥ/ saņkhyānam iti medhāvinotprekşābhihitā kvacitļ / It should be noted in this connexion that Dandin (ii. 273) gives samkhyāna (and krama) as alternative names, not of utprekșā, but of yathasamkhya, which seems to be more plausible. It is possible that the text is corrupt here. Kane HSP. pp. 61-62 suggests the emenda- tion : samkhyūnam iti medhuvī notprekşabhihita kvacit, which he translates as "Medhāvin (calls yathāsamkhya) by the name of sam- khyāna, and in some places (in works on alankāra) utpreksā has not been spoken of as an Alankara." But the difficulty in accepting this emendation lies in the fact that Dandin's elaborate treatment of utpreksa would indicate that this figure must have assumed enough importance, even in Bhamaha's time, to have been entirely overlooked in works on Alamkāra. Vāmana calls yathāsamkhya by the name of krama.

30 SANSKRIT POETICS

seven upama-dosas, and this statement appears to have been accepted by Nami-sadhu (on Rudrata xi. 24), who mentions and illustrates the seven Dosas by examples. Unfortunately we do not possess any other information about Medhavin (or Medhāvirudra)64; and to Nami-sādhu and Rājaśekhara, who cite him, he was possibly nothing more than a name. From Bhämaha's references, however, it is not unreasonable to presume that Medhavin was the first writer on Poetics. who, at an early period, gave an exposition of two very im- portant poetic figures like upamā and utprekşā.65 After dealing with these groups of figures, Bhāmaha takes up (iii. 1-4) in the next chapter the remaining poetic figures. recognised in his time, all in a body, without any further break. These constitute a long list as twenty-four figures. This may be taken as the fourth stage which brings us down to Bhamaha's own time, and which must have ended with the elaboration of a large number of figures, although the number is not as large as we find it in Dandin's time.66 This stage is also represented by a canto in the Bhatti-kāvya which illus- trates in all thirty-eight independent poetic figures, although.

64 See vol i p. 50. Nami-sādhu's quotation from Medhavin (on xi. 24) merely refers to Bhämaha ii. 40 which hardly adds anything to our knowledge. Rājaśekhara and Vallabhadeva (on Siśu xi. 6) cite Medhavin as a poet, the former coupling his name with that of Kumāradāsa, and the latter actually citing a verse from Medhavirudra (see ZDMG Ixxiii, 1919, p. 190 fn 1). 65 The problems regarding upumū-dosas and utpreksā appear to have engaged, to a great extent, the attention of early writers. See vol. i. p. 60, fn 1 on upamū-doșas. On utpreksā, see Daņdin ii. 226 -- 234, Vamana iv. 3. 9 vrtti, Udbhata ed. Telang, pp. 43-46. 66 The differentiation and multiplication of poetic figures with the growth of speculation is a familiar fact in Alamkāra literature; and it is not surprising that as the study advanced, the process of refining went on until a time came when the number became too unwieldy and the distinctions too fastidious ; and then they had to be systematised in the light of some central principle. Such attempts. were made from time to time, the earliest of which, as we shall see,. was perhaps that of Vāmana.

FROM BEGINNINGS TO BHĀMAHA 31

as we have already discussed 67) Bhatti appears to have made- u se of a text possibly unknown to Bhämaha, if not materially differing from the latter's sources. There can be no doubt,. however, that by this time the standpoint of the Alamkāra- śāstra was clearly defined and firmly established. With Bhämaha begins a new epoch in which the uncertain groupings of the earlier periods vanish with the setting up of a more or less authoritative standard,

67 See vol i pp. 52-56.

CHAPTER II

BHĀMAHA, UDBHAȚA AND RUDRAȚA (The Alamkāra System)

BHĀMAHA

( 1 ) In Bhāmaha's Kāvyālamkāra, the different topics of Poetics are formulated not incidentally, as in Bharata, but in such well-defined outline as would indicate that the Alamkāra- śāstra had already attained the rank of an independent disci- pline. We have seen that Bharata considered certain important elements of poetry as devices for embellishing dramatic speech and as subservient to the principal purpose of produc- ing the dramatic Rasa. In Bhamaha, on the other hand, the poetical embellishments form the principal object of study; and, while Dramaturgy and Rasa are entirely ignored, we find for the first time a definite scheme of Poetics more or less systematically elaborated and authoritatively established. Bharata's treatment would show that even before his time some of the older poetic figures, most of the Gunas and Dosas, had been recognised and clearly defined, even if no particular theory of Alamkara had been in existence. But Bhāmaha throws into prominence these poetic embellishments and the consideration of Gunas and Dosas in their connexion, in conformity perhaps to a tradition from which the whole discipline appears to have received the significant designation of Alamkāra-śāstra. We have attempted to explain in the preceding chapter that the comparative antiquity of this tradition or of this school of opinion cannot be denied ; and the presumption is not unlikely that a system of Alamkāra or an Alamkära school1 (if the term is allowed from the em-

1 The word 'School' is used here in a very general sense to indicate affiliation to a particular doctrine or system. Only Udbhata

BHĀMAHA, UDBHAȚA AND RUDRAȚA 33

phasis it puts on the consideration of alamkāra or the poetic figure as the principal element of poetry)2 existed side by side with the Rasa school or the dramaturgic Rasa-system, and influenced it, as it was to a limited extent influenced by it. But this doctrine or system of Poetics is represented to us by a comparatively late writer like Bhämaha, who was by no means its original founder. The general doctrine of this Alamkara-system is almost co-extensive with what appears to have been the original stand- point of the Alamkāra-śāstra itself as an objective, empirical, and more or less mechanical discipline ; for, despite the pre- vious or synchronous existence of a system which elaborated the idea of Rasa in the service of the drama, there is nothing to contradict the hypothesis, which we have already indicated and which is confirmed by the very early existence of the Alamkāra-system itself, that Sanskrit Poetics started apparent-

and Vamana, we are told, had their groups of followers respectively designated Audbhatas and Vamanīyas. But there is no evidence to show that the particular doctrines of Rasa, Alamkāra, Rīti and even Dhvani were worked out in proper schools, consciously or unconsciously founded by a great writer and supported by his followers. At the same time it can hardly be doubted that in spite of mutual and inevi- table contamination or appropriation, these doctrines or systems had a tradition and a history of their own which naturally differentiated them from one another; and each writer conformed, in his fundamental principles, to some theory which threw into prominence one or other of these doctrines. One could say, for instance, that Abhinava- gupta owes his allegiance principally to the Dhvani-system of Ananda- vardhana, although he acknowledges the importance of Rasa or brings in Vämana's ideas of Guna in his interpretation of those of Bharata. With these reservations, the word 'school' is meant here to denote the different systems which emphasise respectively the Rasa, Alamkāra, Rīti or Dhvani theories .- Samudrabandha (p. 4) speaks of five Pakșas or views about Kāvya, namely (i) of Udbhata (ii) of Vāmana (iii) of Vakroktijīvita-kāra (iv) of Bhatța Nāyaka and (v) of Anandavardhana. The Anumāna-paksa of Vyakti-viveka is said to be unworthy of serious consideration. 2 Referring to Bhamaha, Udbhata and other clder writers, Ruyyaka says: tad evem alamkāra eva kāvye pradhānam iti prā- cyānām matam (p. 7). 26

34 SANSKRIT POETICS

ly from some theory of embellishment (alamkāra) which took into consideration the whole domain of poetic figures and confined its energies to the elaboration of more or less mecha- nical formulas with reference to the technique of expression.3 Just as there may be a theory of painting consisting of a collection of information regarding the techniques of tempera, of oil-painting, of water-colour, of pastel, on the proportion of the human body and on the laws of perspective, the art of poetry was supposed to comprise a collection of precepts relating to the forms of expression, its structural beauty, its damaging faults and its rhetorical ornaments, without going further into the speculative aspects of the problems involved. The whole aesthetic judgment was directed to these means of externalisation, and aesthetic pleasure was regarded object- ively from the standpoint of extraneous facts which contribu- ted to it. It cannot indeed be dogmatically stated that the necessity and inevitability of postulating an ultimate principle did not trouble these older writers ; but the study must have begun with a method which resulted in the establishment of a series of more or less rigid definitions and categories elabo- rated to a degree of fineness. The question as to what consti- tutes poetry or poetic charm, the aesthetic fact, does not arise until Vämana and the Dhvanikara come into the field; for earlier authors like Bhāmaha (i. 23) and Dandin (i. 10) pro- pose to confine themselves chiefly to what they call the kāvya-śarīra3 or the 'body of poetry', as distinguished from its

3 Vol. i p. 7. Sanskrit Poetry, more than ever in this 'classi- cal' period of its history, appeared as the careful work of a trained and experienced specialist. The tradition of such poetry points naturally to the working of the rules and means of the art into a system. This, combined with a natural and characteristic love of adornment, which demanded an ornamental fitting out of thought and word, probably supplied the original motive-force which brought the study of Poetics into existence. The word "Alamkāra" (lit. embellishment), applied to the discipline itself, as well as to the poetic figure, which forms the main topic of discussion in the earliest extant works from Bhamaha to Rudrata, would indicate that Sanskrit Poetics had probably grown out of a theory and practice of

BHĀMAHA, UDBHAȚA AND RUDRAȚA 35

atman, its 'soul' or animating principle.4 The advantages of verbal arrangement with due regard to the expression of an agreeable sense and of clever clothing of the sense with

embellishment, which included and threw into prominence the whole domain of verbal and ideal figures of speech, those decorative devices by which poetic expression may be rendered attractive. It seems that originally Ars Poetica in India, as Jacobi suggests, did not go further than being a series of advice to the poet in his profession, and consisted more or less in formulating prescriptions for the practical working out of poetry. It pointed out the faults to be avoided and the excellences to be attained, and described the poetical embellishments which should enhance the beauty of expression, insomuch as the whole study came to receive the designation of Alamkara-sāstra, or the Science of Poetical Embellishment. This theme of the "education" of the poet became in later literature a separate study when the theoretical aspects of the problems involved began to occupy an important place in the discipline, and we find a group of writers devoting themselves entirely to the subject of kavi-siksa which, having the practical and somewhat mechanical training of the poet in his art for its object, really represents the original standpoint of Poetics. Sanskrit poetics was probably raised to the rank of an independent disci- pline almost about the time when Sanskrit 'Classical' poetry was already overstepping itself in its development. 4 The metaphorical expression kāvya-sarīra with its implied kāvyātman plays an important part in Sanskrit Poetics throughout its history. Its origin has been traced to the allegory of the Veda- purușa in Rgveda vi. 58. 3 (catvāri śrigā°), and Rājaśekhara indul- ges in an analogous poetic conception of the Kavya-purusa (Spirit of Poetry) and his bride Sāhitya-vidya. According to Rājasekhara's somewhat fanciful description, the body of this Kāvya-purușa is composed of word and sense (sabda and artha), the face made of Sanskrit, the arms of Prakrit, the hips and loins of Apabhramsa, the feet of Paisaca, and the breast of mixed languages. This is evidently from the linguistic point of view. It is further added that his speech is rich in different modes of poetic expression (ukticaņam) ; moods and sentiments (rasa) make up his soul; metres form the hair of his body ; his conversation consists of questions and answers and riddles; and he is adorned by poetic figures like alliteration and simile. The later writers attempt to arrive at greater precision, first indicated by Vāmana, who makes 'diction' (riti) to be the soul of poetry. That the external art of poetry can be systematised formed one of the fundamental postulates of Sanskrit Poetics; but at the same time the necessity of some deeper principle to explain the manifold character of its content

36 SANSKRIT POETICS

poetical or rhetorical ornaments absorb the attention of these writers; and whatever may be the theoretic basis of poetic charm, it is enough if it is realised by the objective beauty of ingenious expression. The two important factors, which go to make up the kāvya-śarīra, i.e. the 'body' or external framework of poetry,

could not be ignored. Hence the attempt to find the 'soul' or the animating principle in the 'body' or external framework of poetry. Bhämaha perhaps vaguely realised this when he proposed to take vakrokti as the underlying principle of artistic expression; but Dandin goes a step further and designates the Gunas as the 'life- breath' (prūņāh) of the diction (märga or riti), which he sets up as the most important part of poetry. Vamana is the first known writer io comprehend and state distinctly what this 'soul' is, and the Dhvanikära takes the last step in completing this figurative idea by defining systematically the mutual relation of the 'body' and the 'soul' of poetry. The Dhvanikāra implies in ii. 7 that vyangya artha is this atman, the gunas being compared to natural qualities like courage, and the alamkaras likened to external ornaments like bracelets which adorn the body. This view is apparently accepted by Mammata (viii. 1) and taken as authoritative by all subsequent writers, while Nami-sadhu (on Rudrata xii. 2) gives a similar but not accurate explanation of Rudrata's opinion on the subject. The final extension of this metaphorical conception is thus set forth by Viśvanātha : kāvyasya śabdārthau śarīram, rasādiś cātmā, guņāh saundar yādivat, dosāh kāņatvādivat, rītayo'vayava-samsthāna- viśeşavat, alamkāraḥ kataka-kundalādivat, thus comprehending all the elements of poetry, discussed by previous writers, into this ela- borate metaphor. Whatever may be the value of this metaphor as an index to the conception of poetry gradually evolved by Indian theorists, one point is clear, viz., that they all take, from Bhämaha to Jagannätha, the sabda and artha as constituting what they call the 'body' of poetry; and with this idea the theories start, ulti- mately ending in a search for its 'soul'. From another point of view, the sabda and artha form the central pivot round which all theories move (for they are all theories starting with expression) with particular reference to the question of the function par excellence operative in poetry. And as the study of Poetics itself, on the admission of some of its greatest exponents, drew its ori- ginal inspiration from grammatico-philosophical speculations on speech, it is not surprising that enormous emphasis should be put on these two elements.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 37

are supposed to be śabda (word) and artha (sense)5, and the alamkāras or poetic figures which adorn these are taken as forming the essential sign of a Kavya. In other words, poetry consists of a verbal composition in which a definite sense must prevail, and which must be made charming by means of certain turns of expression to which the name of poetic figure is given. This general standpoint is implied by Bbämaha at the very outset in connexion with his general discussion of the two kinds of figures based respectively on word and sense6: rūpakādim alamkāram bāhyam ācakșate pare/ supām tinām ca vyutpattim vācām vānchanty alamkrtim/ / tad etad āhuḥ sausabdyam nārtha-vyutpattir īdrśī/ śabdābhidheyālamkāra-bhedād istam dvayam tu naḥ//: This passage, quoted with approval in the Vakrokti- jīvita (on i. 8) and the Kāvya-prakāśa vi, is difficult to translate, but the meaning is clear and may be freely rendered

5 See above footnote 3. Sabārthau sahitau kāvyam, Bhāmaha i. 16, from which, as Kuntaka indicates, the name sūhitya was pro- bably given to poetry. The earliest use probably of this term sāhitya in Sanskrit Poetics occurs in Mukula (pp. 21 and 22) and in his pupil, Pratīhārendurāja, while Rājaśekhara expressly uses the term sāhitya-vidya. The orthodox etymology of the term, which derives it from the above definition of poetry, as the union of word and sense, is thus put by Rajaśekhara: śabdārthayor yathāvat saha- bhāvena vidyā sāhitya-vidyā, an interpretation with which Kuntaka agrees. This Sahitya or alliance of word and sense is admitted as a fundamental postulate from a very early time, and with proper modifications, by all schools and authors. Cf Dandin i. 10, Vāmana i. 1. 1 (vrtti), Rudrața ii. 1. Ānandavardhana admits as un- questioned: śabdārthau tūvat kāvyam, na vipratipattir iti darsayati. The view is alluded to by Māgha in ii. 86b, and apparently by Kālidāsa in the first verse of Raghu°. 6 This distinction between Alamkāras cf Sabda and of Artha began to be recognised, if not directly stated, from the time of Bhamaha. Although Dandin is not explicit, he has the same dis- tinctive view when he deals with Arthālamkāras in ch. ii and Sabdā- lamkāras in ch. iii. It is Bhoja who classifies Alamkāras into those of Sabda, of Artha or of both, defining and illustrating 24 of each in his Sarasvati-kaņthā°.

38 SANSKRIT POETICS

thus: "Others regard metaphor and the like to be external ornaments. They postulate that grammatical correctness adorns speech, and call it excellence of language, (implying that) there is no such corresponding correctness of sense. We, however, accept two kinds of ornaments, referring respectively to word and sense". Dandin, who does not strictly belong to this school but who substantially agrees on this point with Bhämaha, is more explicit in his statement, and lays down (i. 10) that the 'body' of poetry consists of a series of words regulated by an agreeable sense. Although Bhamaha is the oldest representative of this system whose work has survived, he was, as we have stated more than once, by no means its originator. The system, as we find it set forth in his work, is certainly not primitive, but indicates the clear existence of a developed teaching on the subject. Rājaśekhara, as we have noted (vol. i, p.1), gives us a long list of mythical names with which he identifies the origi- nal treatment of the various topics relating principally to the poetic figures. He assigns the elaboration of the poetic figures anuprāsa, yamaka and citra, śabda-śleşa, vāstava, upamā, atiśaya, artha-śleşa, and ubhayālamkāra respectively to Prace- tāyana, Citrāngada, Seșa, Pulastya, Aupakāyana, Pārāśara, Utathya and Kubera. While the antiquity of the distinction between śabdālamkāra and arthālamkāra generally (in spite of the above passage of Bhamaha's) and of śabda-ślesaland artha-śleșa in particular may be seriously doubted, some of these poetic figures may be allowed to have been very early recognised, as Bharata mentions some of them and Bhāmaha acknowledges nearly all (excepting citra which we find in Dandin and vastava which we find in Rudrata). Medhavin, cited by Bhamaha, probably belonged to this school, and his is the only authentic name of an early exponent of this System. With Bhamaha's work, however, we emerge from the region of conjecture and doubt, and arrive at the first classic statement of a definite doctrine of Poetics. We must not yet

BHĀMAHA, UDBHAȚA AND RUDRAȚA 39

look to his work, nor to that of his follower Udbhata, for a thoroughly critical system. We can deduce certain broad conceptions, but the practical object, underlying the specula- tion of this school in general, and its more or less norma tive character did not allow sufficient scope for purely theoretic treatment ; nor can we expect such treatment at this early stage. Bhämaha, therefore, nowhere attempts a formal definition of poetry, nor does he state clearly the theory of Vakrokti and Alamkāra which, as we shall see, was first systematically enunciated by his follower, the Vakroktijīvita- kāra. The first chapter of Bhamaha's work gives us some preliminary remarks about the general characteristics of poetry and its subdivisions, but a large part of it is taken up with the enumeration of the general faults which obstruct the proper expression of an idea. Bhämaha states at the outset the purposes of poetry (kāvya-prayojana) and the qualifications of a poet (kāvya- hetu), incidentally mentioning the 'sources' of poetry (kāvya- .yonayah). Regarding the first topic of the aim and purpose of poetry, it is not necessary to discuss the different views in detail, as they generally enumerate extraneous objects and throw little light on the general theory of poetry. In the older writers there is a more or less uncritical mention of fame (kīrti) for the poet and delight (priti)7 for the reader as the chief objects of poetry ; and herewith Bhāmaha (i. 2), Daņdin (i. 105), Vāmana (i. 1. 5), Rudrața (i. 21, 22) and Bhoja (i. 2), though belonging to different schools of opinion, seem to be content. But it became customary to add,8 from the poet's standpoint, 'wealth', 'social success' and 'escape from ills'. From the reader's point of view, poetry is said to bring "solace', 'instruction in knowledge' and 'proficiency in the arts and ways of the world'; and these are sometimes summarily comprehended by the term trivarga, viz., profit, pleasure and

7 Bharata had already laid down this pleasure-giving function of the dramatic art as krīdanaka (i. 11), vinodakāraņa (i. 86). 8 E. g. Mammata i. 2, Hemacandra p. 2 etc.

40 SANSKRIT POETICS

virtue, to which later on the caturvarga, anticipated by Bhāmaha (i. 2), adds mokşa or liberation of the soul. This is probably an attempt to bring poetry on a level with other arts or sciences which profess similar ends, and is in harmony with the deep-rooted idea of the functions of Sāstra. It is, however, pointed out by later theorists like Mammata and his followers,9 who in their turn develop Abhinavagupta's idea (°Locana p. 12), that the Kāvya, as distinguished from the scriptures and the sciences, is kāntā-sammita, i. e., like the teaching of a loving mistress, implying thereby that the pedagogic powers of poetry resolve themselves into a peculiar power of suggesting a condition of artistic enjoyment. The famous opening verse of the Kāvya-prakāśa makes this clear when it describes poetic speech as comprehending a creation ungoverned by nature's laws and consisting of pure joy. The caturvarga and the other material objects of poetry are mentioned in almost unbroken tradition; but with the elabora- tion of a full-fledged scheme of Poetics in connexion with the suggestion of Rasa, the purpose of poetry was brought into a level with the ultimate theory about its nature ; and poetry was supposed to create a peculiar mood of aesthetic pleasure, conveyed generally by the philosophic term ānanda10.

9 Abhinava uses the terms prabhu-sammita, jāyā-sammita and mitra-sammita, which are accepted by Mammata (ed. B.S.S, 1917, p. 9). Later writers distinguish (e. g. Ekāvalī pp. 13-15) between the Vedas which are prabhu-sammita, the Itihasa etc. which are mitra- sammita, and the Kāvya which is kāmā-sammita. 10 Abhinavagupta's comment on Bhämaha i. 2 on this point is in- teresting (°Locana p. 12, partially copied by Hemacandra in his commentary, p. 3): yathoktam-dharmārtha-kāma-mokșesu vaicakşaņ- yam kalāsu ca | karoti kīrtìm prītim ca sādhu-lāvya-nișevaņam / / iti, tathāpi prītir eva pradhānam. Anyathā prabhu-sammitebhyo vedā- dibhyo mitra-sammitebhyaś cetihāsādibhyo vyutpatti-hetuthyah, ko'sya kāvya-rūpasya vyutpatti-hetor jāyā-sammitatva-lakșano vișesa iti prā- dhānyenānanda evoktah. Caturvarga-vyutpatter api cānandah, pāryan- tikam mukhyam phalam .. The essence of Rasa, which came to be considered as the most important thing in poetry, is said to consist of this priti or ananda; naturally ānanda or priti became in later

BHĀMAHA, UDBHAȚA AND RUDRAȚA 41

Jagannātha completes the idea by defining it as a 'disinteres- ted' or 'dissociated' (alaukika) pleasure, which depends upon. a taste formed by repeated representation of beautiful objects, and which can be enjoyed by a man initiated into the poetic mysteries. With regard to the other two questions, viz. the equip- ment of a poet and the sources of poetry, Bhamaha appears. to be cognisant of their importance. His remarks on these points are, however, brief as compared with those of Vāmana, who deals with the subject elaborately for the first time. It may be pointed out in this connexion that Sanskrit Poetics, consistently with the original idea of its having been a more or less mechanical discipline, gives a long list of the essential qualifications which a poet should possess and lays down elaborate rules for his 'education'. With the advance of the theoretical aspects of the science, this theme was, no doubt, made the object of a separate study by a group of writers who make it their business to instruct the poet in his profession ; but all early writers on general Poetics, more or less, touch upon the point. We shall have occasion to deal with this school of kavi-śikşā; but it will be convenient to indicate here: briefly the earlier speculations on the subject. While not denying the supreme necessity of genius or poetic gift (sat- kavitva, Bhāmaha i. 4) which consists in pratibhā (poetic conception), all writers, early or late, agree in emphasising the necesssity of study and experience. Both Bhāmaha (i. 5) and Daņdin (i. 103-4) acknowledge pratibhā which is said to be natural (naisargikī) or inborn (sahajā) ; and Vāmana puts. it into a formula that in pratibha lies the seed of poetry, and defines it (i. 3. 16 Vrtti) as an antenatal capacity of the mind11 without which no poetry is possible, and if possible, it is only ridiculous, a dictum which is almost literally copied by

Poetics the chief object of poetry. Mammata calls it sakala-prayojana- maulībhūtam. 11 janmāntara-gata-samskāra-viseşah kašcit. Daņdin describes it as pūrvu-vāsanā-guņānubandhi.

42 SANSKRIT POETICS

Mammata who, however, uses the more general term śakti.12 Abhinavagupta (°Locana p. 29) defines it as intelligence (prajñā), capable of fresh invention (apūrva-vastu-nirmāņa- kşama), its distinguishing characteristic being the capacity of creating poetry possessed of passion, clarity and beauty (tasya viśeso rasāveśa-vaiśadya-saundarya-kāvya-nirmāņa- kşamatvam) ; and he quotes the authority of Bharata (vii, 2) who designates it as the 'internal disposition' of the poet (antargata bhava). This agrees with the definition of pratibha as prajñā nava-navollekha-śālinī, given in a verse cited anonymously by Hemacandra, but attributed by Ksemendra (Aucitya-vicāra° ad śl. 35) to Abhinava's guru, Bhatța Tauta ; and it is recognised as canonical by later writers, to whom Abhinava and Mammata were the final lawgivers, but who sometimes add that it is lokottara and capable of producing an indefinable charm variously termed vaicitrya, vicchitti. cārutva, saundarya, hrdyatva or ramaņīyatva. While these theorists believed in pratibha, they also be- lieved in "making a poet into a poet," and maintained the importance of what Dandin calls śruta and abhiyoga, but what later writers call vyutpatti (culture) and abhyāsa (prac- tice). Rudrata, therefore, thinks that pratibhā is not only sahajā or inborn, but also utpādyā or capable of attainment by vyutpatti or culture. The poet is thus required to be an expert in a long list of sciences or arts. The earliest of such lists is given by Bhämaha in i. 9, where mention is made of the following studies as 'sources' of poetry,13 viz., grammar, prosody, lexicography, stories based on ltihasa, ways of the world, logic and and the fine arts. This substantially agrees

12 Used by Rudrata (i. 14-15), who distinctly gives two alternative terms śakti and pratibhā and by Abhinavagupta (°Locana p. 137) who says: saktih pratibhānam, varņ anīya-vastu-visaya-nūtanollekha- śālitvam. 13 Read kāvya-yonayah (instead of kūvyayairvaśī in the printed text), as indicated by Vāmana i. 3. 1 (kāvyāngāni) and Rājaśekhara viii (kavya-yonayah). Cf. Jacobi in op. cit. p. 224.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 43

with Rudrata's list (i. 18), but Vämana deals with the topic in greater detail in i. 3. 21-22, and requires the poet to be conversant with grammar, lexicon, metrics, arts, morals, erotics, politics, and, above all, the ways and means of the world. It is also sometimes implied that the poet must have studied the theory of poetry and made himself proficient in poetical exercise (abhyāsa). He must be clever at weaving metaphors and other poetic figures, at the trick of producing a double meaning, at manipulating complicated schemes of alliteration and rhyming, at following up quick composi- tion, at making complete verses out of broken lines and sentences, and similar ingenious practices. When a new work is published, it is submitted to and approved by assemblies of experts, as we are told by Mankhaka, Rājaśekhara and others. It was obviously expected to answer all the demands of theory, although it was by no meaas an easy test ; for style, says an Indian stylist, is like a woman's virtue which cannot bear the least reproach. The public likewise possessed or were expected to possess a certain amount of theoretical knowledge ; for the rasika or sahrdaya, the man of taste, the true appreciators of poetry, must be, according to the con- ception of the Sanskrit theorists, not only well read and wise, and initiated into the intricacies of theoretic requirements, but also possessed of fine instincts of aesthetic enjoyment14. The poet naturally liked to produce an impression that he had observed all the rules, traditions and expectations of such an audience ; for the ultimate test of poetry is laid down as consisting in the appreciation of the sahrdaya. Thus, the poet is required to be true to his natural gifts and yet conform to the rigid demands of theory. The art of poetry in this way came to flourish in a learned atmosphere, and the theory of Poetics, as we shall see, naturally assumed a scholastic

14 On the subject of the "education" of the poet, see F. W. Thomas, The Making of the Sanskrit Poet in Bhandarkar Comme- moration Volume p. 375 f.

44 SANSKRIT POETICS

and dialectic character in common with the whole scientific literature of ancient India. It is true that a certain amount of inevitable difference is always to be found between theory and practice ; and, as on the one hand, we have gifted poets. aspiring to untrammelled utterance, so on the other, there is a tendency to degenerate towards a slavish adherence to rules, which naturally resulted in a strong overloading of a com- position by complicated or artificial expressions. With these general remarks we may now turn to a brief consideration of other topics in the work of Bhāmaha15. Bhā- maha rests content by taking the Kāvya to consist of sabda and artha (śabdārthau sahitau kāvyam), giving equal promi- nence to word and sense in poetry. But he implies by his treatment that the Kāvya should also be faultless (nirdosa) and embellished by poetic figures (sālamkāra). Then follows the classification of poetry (i) according to form, into verse and prose, (ii) according to the language employed, into Sans- krit, Prakrit and Apabhramsa, (iii) according to the subject- matter, into fourfold division, so far as it deals respectively with incidents human or divine, incidents invented by the poet's imagination, or incidents based on the several arts or sciences, (iv) according to the conventional way of grouping compositions into fivefold recognised division, viz. sarga- bandha (mahākāvya), abhineyārtha (drama), ākhyāyikā, kathā and anibaddha- kāvya (i. e. detached poems like gāthās or individual ślokas). Bhāmaha's definition of mahākāvya is. more or less conventional, and practically agrees with that to be found in Dandin (i. 14f) and in the Agni-purāna (336. 24-32). With regard to abhineyārtha, he omits its treatment because others have treated of it already. Bhāmaha lays down a somewhat hard-and-fast line of demarcation between the kathā and the ākhyāyika, a view which is not accepted

15 Bhämaha's work consists of six chapters: ch. i on poetry generally (60 verses), ch ii and iii on Alamkāra (160 verses), ch. iv on Doșa (50 verses), ch. v on logical correctness or Nyāya (70 verses) and ch. vi on grammatical correctness (60 verses).

BHĀMAHA, UDBHAȚA AND RUDRAȚA 45

by Dandin and which Vamana does not enter into as being too fine16. With regard to the formal classification of verse and prose, it may be remarked that poetry, according to the view of Sanskrit theorists, has a twofold aspect according as it consists of verse (padya) or prose (gadya), although some authors speak of a mixed kind (miśra), e. g., the drama in which both prose and verse occur. The doctrine that prose is the opposite, not of poetry but of verse, which began to be realised rather late in European critical theories, was very early admitted without question by Sanskrit authors with whom metre does not play the same part as it does in Euro- pean poetry ; for in India from the earliest time, it was usual to put down even the driest teachings in a metrical form17. Bhämaha, however, appears (i. 31-35) to be indifferent

16 Vol. i. p. 65. We have tried to show (The Kathā and the Akhyūyika in Classical Sanskrit in BSOS, vol. elsewhere

iii (reprinted in Some Problems, pp. 65-79) that two or three well-defined stages are distinguishable in the development of these two species of the Kävya in Classical Sanskrit, the earliest being represented by the characteristics given by Bhamaha and the latest by those given by Rudrata; and that while Bhämaha cannot be taken as having accepted Bana's two masterpieces as his prototype, Rudrata has only generalised their important features into universal definitions of the kathā and the ākhyāyikā. Lacóte in his Essai sur Guņādhya et la Brhatkatha suggests (p. 282) that Dandin must have found that Guņādhya did not observe the traditional distinction between the kathā and the ākhyāyika (e. g. in the original Brhatkatha, on Lacôte's showing p. 220, there was a narration by Naravahanadatta of his own victories, which is contrary to the rule laid down by Bhämaha). He was, therefore, led to reject it altogether. Bhamaha, no doubt, refers to a kathā in Apabhramsa in i. 28 but it is not known whether he was aware of the existence of the Brhatkatha. 17 One need not emphasise the point that Sanskrit theorists define poetry so as to include any literary work of the imagination in its scope, and absolutely refuse to make of rhyming or versing an essential. This tradition is so well established that the question is nowhere discussed and never doubted. Thus, the theorists include under the head of poetry romances like Kadambari or Harsa-carita which are written for the most part in prose. Vamana even quotes a dictum which says that prose is the touchstone of the poets (gadyam kavīnām nikașam vadanti, cited in Vrtti on 1. 3. 21).

46 SANSKRIT POETICS

to the literary value of riti (roughly, 'diction'), to which Dandin and his followers of the Riti school attach so much importance. He thinks that the distinction made by the Rīti- theorists between vaidarbha and gauda is meaningless ; and though he does not use the terms mārga (Dandin) or rīti (Vämana), his statement would imply that he is speaking of vaidarbha- and gauda-kāvyas respectively, in which some had apparently seen differences of manner and treatment.18 As a necessary corollary apparently to this view, he does not think it worth while to devote much attention to the gunas, which the Riti-theorists take as forming the constituent excellences of riti, and summarily mentions in another context (ii. 1-3) only three Guņas, viz., mādhurya, ojas and prasāda, apparently rejecting Bharata's ten orthodox excellences. He does not, however, connect them directly with the Rīti, and thinks that they are distinguished according to the presence or absence, in varying degrees, of compound words, ojas employing long compounds, and mādhurya and prasāda not doing it. These Gunas, in his opinion, are not qualities of any particular diction, but of the good Kāvya generally. It is noteworthy that this brief description of the Gunas precedes. in context the treatment of Alamkāras, implying probably that they are analogous to each other. It is also noteworthy that Bhamaha does not employ the term Guna at all, except in another context in connexion with the Bhāvika Alamkāra which he, like Dandin, designates as a prabandha-guna. Bhämaha then proceeds to define and discuss, with illustrations, the poetic figures or alamkāras, to which he devotes two long chapters (ii. 4-95, iii. 1-56), consisting of nearly one hundred and fifty verses. Then come (ch. iv) the dosas or demerits of composition (some of which are already dealt with in i. 37-56), and the whole subject is wound up

18 As each of these types have certain distinguishing features. Bhämaha is of opinion that one need not condemn Gauda nor praise the. Vaidarbha; but he himself does not deal with these types which must. have been well known in his time.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 47

with two chapters (v-vi) on the logical19 and grammatical20 correctness of poetry respectively. Now leaving aside these two requirements and the Dosas, all of which are in a sense negative requirements, the only thing of the highest impor- tance in poetry, in Bhamaha's view, is apparently the Alamkāra or the poetic figure, which takes up the bulk of his treatment.21 Bhamaha attempts to classify poetic expression into fixed rhetorical categories; and from this point of view his work possesses the general appearance of a technical manual, comprising a collection of definitions with illustrations and empirical canons for the benefit of the artist desirous of externalising his ideas. But in the course of his enquiry, it probably struck him that a philosopbical or scientific classification of expressions is not possible ; for, although every single expressive fact may be grouped together generically, the continuous variation of the individual content results in an irreducible variety of expressive facts. He attempts, therefore, to arrive at a synthesis by holding that there may be modes or grades of expression, of which the best mode is that which involves vakrokti22, by which a certain peculiarity or charm of expression is posited as the essential principle of all Alamkāras.

19 In which are discussed such Nyāya-vaisesika topics as the pramāna, prajñā, hetu, drstānta etc. 20 Giving practical hints for attaining grammatical correctness (sauśabdya), and corresponding to the last adhikarana of Vämana's. work. 21 Excluding subvarieties the Alamkāras mentioned or defined by Bhamaha are 39 (+4) in number, viz. in this order: anuprasa (two varieties), yamaka (five kinds), rūpaka (two varieties), dīpaka, upamā, prativastūpamā (as a variety of upamā), āksepa (two kinds), arthāntara-nyāsa, vyatireka, vibhāvanā, samāsokti, atišayokti, yathā- saņkhya, utpreksā, svabhāvokti, preyas, rasavat, ūrjasvi, paryāyokta, samūhita, udātta (2 kinds), ślista, apahnuti, viseșokti, virodha, tulya- yogitā, aprastuta-prašaņısā, vyāja-stuti, nidaršanā, upamā-rūpaka, upa- meyopamū, sahokti, parivṛtti, sasamdeha, ananvaya, utpreksāvayava, samsrsti, bhūvika, asīh (according to some), as well as hetu, sūkșma lesa and vārttā (to which the status of Alamkāra is denied). 22 It is true that in one place Bhämaha speaks of the figure

48 SANSKRIT POETICS

The etymological meaning of the term vakrokti is "crooked speech"; and this meaning appears in the verbal poetic figure defined by Rudrata (ii. 13-17) and, after him, by all later theorists, who connote by this figure a kind of pretended speech based on paronomasia (ślesa) or peculiarities of intona- tion (kāku). In Vāmana, on the other hand, vakrokti appears not as verbal figure (śabdālamkāra) but as a figure based on the sense (arthālamkāra) ; and it is defined as a metaphorical mode of speech based on "transference of sense" (lakşanā). Bhämaha, while admitting it, apparently in common with Dandin (ii. 363), as a collective designation of all Alamkāras23, uses the term to imply a selection of words and turning of ideas peculiar to poetry and abhorrent of matter-of-fact speech. Kuntaka, who develops this idea and builds a unique theory of alamkāra on its basis, makes this meaning clear when he indicates by such vakrata the peculiar charm (vicchitti) or strikingness (vaicitrya) which can be imparted to ordinary expression by the conception of the poet (kavi- pratibha). When words are used in the ordinary manner of common parlance, as people without a poetic turn of mind use them, there is no special charm, no strikingness; and

bhavika as the characteristic excellence of a composition as a whole, a view which coincides with that of Dandin, as well as of Bhatti who (according to commentators) illustrates this figure in one whole canto (canto xii). It is defined as the representation of objects, whether past or future, as if they were present, the condition of the representation being that the story or theme must have picturesque, strange and exalted meaning (citrodattadbhutārthatvam kathayah) and must be capable of being enacted well (svabhinītata), and that the words used must be agreeable (sabdānukulata). Bhamaha, however, does not appear to lay any special emphasis on bhavikatva in Poetry, but deals with it as he does with any other poetic figure. No doubt, he speaks of bhavika as a prabandha-guna, but Bhamaha does not seem to have made any theoretical distinction between Guņa and Alaņkāra as such, and the word guna here should not be taken in any technical sense. 23 On this subject, see S. K. De, introd. to Kuntaka's Vakrokti- Jivita, 2nd Ed. pp. xiv- XXV.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 49

consequently it is not poetical in the sense in which Bhämaha and his follower understand it. Such svabhāvokti or 'natural' mode of speech, which Dandin calls ādyā alamkrti and distinguishes from vakrokti, is not acceptable to Bhāmaha24 and to Kuntaka, who refuse to acknowledge svabhāvokti as a poetic figure at all ; for these theorists apparently imply a distinction between the "artistic", or ornamental or extra- ordinary, on the one hand, and the "naturalistic", or unadorned or ordinary expression, on the other25.

24 See Jacobi in ZDMG lxiv 130 f, and in Sb. der preuss. Akad. xxiv, 1922, pp. 224f. Kuntaka also uses the term vakrokti as almost co-extensive with the term alamkara (p. xxx), and regards the so- called roetic figures as aspects of vakrokti .- Bhāmaha does not, as Kuntaka does, elaborately argue against Svabhāvokti; he acknowledges it, but the terms by which he refers to it (ii. 93-94) would make one think that it is not as acceptable to him as it is to Dandin. Dandin would divide Vanmaya into Svabhāvokti and Vakrokti, the latter including in its scope all the poetic figures. Bhoja (Sarasvatī-k.) would divide Vanmaya into Svabhāvokti, Vakrokti, and Rasokti. 25 To Bhämaha and Kuntaka, svabhāvokti, which consists in a description of the natural disposition (svabhāva) of an object, is obviously wanting in the requisite strikingness to be poetical; for they take it to be merely plain or unadorned description and imply that a poet should express things or ideas differently from the bana- lity or prosiness of the Sastras or of common life. But Dandin and later theorists, on the contrary, reckon jāti or svabhavokti among the poetic figures. On this point we quote what we said elsewhere (Introd. to Vakrokti-jīvita, p. xix, fn 19). "Though formally the ex- pression of the svabhāvokti may not differ from a statement or description in common life there is still a substantial difference. For the poet sees or conceives the very same thing noi in the same way as common people. In the case of the latter, all things stand in some relation to his personal interests, which should be understood to connote also scientific interest in them as objects of knowledge. But for the poet the object has no connexion with his or anybody's interests, not even as an object of knowledge; he has a vision of the thing in itself in its true nature. This is what is partially under- stood by lokātikrānta-gocaratā, and Jagannātha makes it clear (ed. Bombay, 1915, p. 4) while explaining the term lokottaratva as an ele- ment of poetic charm. Literally lokottara means supermundane, but in the sense indicated above it may be translated roughly as 'dis- interested' or 'dissociated'. Now, Dandin, adopting the traditional term alamkāra and applying it to the svabhāvokti, could not very

27

SANSKRIT PUETICS

Bhamaha, therefore, lays down, in his classification of the different kinds of Kävya, that the subdivisions of poetry mentioned by him are admissible to that designation in so far as they possess vakrokti (i.30); and this is made more explicit by declaring later on that whatever value might be attached to the function of Rīti in poetry, the vakrokti is desirable as an embellishment of poetic speech (i. 36), which he charac- terises elsewhere as vakrā (vi. 23). Therefore he calls upon the poet to be diligent in accomplishing this, as the vakrokti manifests the sense of poetry and as no embellishment of poetry is possible without vakrokti (ii. 85). It is not surpris- ing, therefore, that he rejects figures like hetu, sūkşma and. lesa on the ground that they do not involve vakrokti. It is curious, however, that Bhamaha nowhere explicitly defines or explains the word vakrokti. Perhaps here we have the work of early theorisers who have not yet learnt to. theorise systematically, but who are carried away more or less by their practical object of establishing definite norms. and prescribing general formulas as a means of attaining literary expression. Or, perhaps the idea of vakrokti was traditional or already too well known in his time to require- detailed explanation. At any rate, after enumerating and defin- ing the poetic figures up to and including atiśayokti, he, says generally saişā sarvaiva vakroktiḥ (ii. 85), with a hint (as it appears from the context) of identifying the vakrokti in sub- stance with the idea involved in atiśayokti. Kuntaka appears.

well accept Bhämaha's statement that vakrata is the characteristic of all poetic figures, because vakrokti excludes the svabhāvokti (Daņdin ii. 362); but he tries to reconcile his own view with Bhamaha's opinion by extending the latter's remarks regarding the atisayokti (ii. 81) to all poetic figures, thus including the svabhavokti. The Indian' theorists have almost neglected an important part of their task, viz.,. to find a definition of the nature of the subject of a poem as the product of the peet's mind ; this problem is the main issue of Western Aesthetics. Only svabhāvokti and bhāvika can be adduced as a proof that the- Indian theorists were conscious of the problem, but did not attack it in its entirety, treating it only in some of its aspects". See also,. on this point, Jacobi in Sb. der preuss. Akad. cited above, pp. 224 f.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 51

to agree with Bhämaha that some kind of atisaya is involved in vakrokti, and thinks that the atisaya is a necessary element in what he calls vicitra-mārga, where vakrokti-vaicitrya pre- vails (i. 27). Dandin probably arrives at the same conclusion in a different way when he speaks of all poetic figures as depending on atiśayokti, a view which is explained thus by one of his commentators: alamkārāntarāņām api eșa (=ati- śayoktyalamkāraḥ) upakārī bhavati, atiśaya-jananatvam vinā bhūşaņatayā na syād ity abhiprāyah. Ānandavardhana's re- marks in this connexion are illuminating. He says that it is possible to include atisaya in all poetic figures, as it has been successfully done by great poets, for the purpose of increasing the beauty of poetic composition ; and citing Bhamaha's idea of atiśayokti and vakrokti he remarks (pp. 208)26: "There is an excellence of charm in that poetic figure in which the atiśayokti is established by the imagination of the poet ; other figures are merely so called. Since it is able to enter into the body or composition of all poetic figures, it is, by assum- ing it to be identical with them, called their essence". The atiśayoktì, therefore, is taken, in the words of Abhinavagupta's explanation, as the common token or generic property of all poetic figures (sarvālamkāra-sāmān) a-rūpam), or as Mammța puts it, as their life-breath or essence (prānatvenāvatisthate, p. 743). One can realise from this the close connexion bet- ween this important figure and Bhamaha's notion of vakrokti. Bhāmaha defines atiśayokti as nimittato vaco yat tu lokātikrānta-gocaram (ii. 81), which Dandin paraphrases as vivakşā yā viśeşasya loka-sīmātivartinī (ii. 214). It would seem, therefore, that the atisaya in the vakrata of poetic figures consists essentially in this lokātikrānta-gocaratā, and Abhi- nava makes this clear when he explains in this connexion (°Locana p. 208): śabdasya hi vakratā abhidheyasya ca vakra-

26 tatrātisayoktir yam alamkūram adhitisthati kavi-pratibhā-vaśāt tasya cārutvātisaya-yogah. anyasya tvalamkāra-mātrataiveti. Sarvālam- kāra-sarīra-svīkaraņa-yogyotvenābhedopacārāt saiva sarvālaņ kāra- rūpety ayam evārtho'vagantavyah.

52 SANSKRIT POETICS

tā lokottīrņena rūpeņāvasthānam. From this it is reasonable to conclude that by vakrata Bhäma ha implies a kind of height- ened or extraordinary turn given to expression (what Kuntaka would call bhangi or vicchitti), which constitutes the charm or strikingness of poetic expression, as distinguished from common speech where facts are simply stated. We shall see that Kuntaka elaborates this idea by the peculiar theory of vaicitrya or vicchitti (which is taken as almost equivalent to the term vakrat) of word and sense as forming the basis of all poetic decoration (the so-called poetic figures being mere aspects of it), whereby the poet lifts ordinary speech to the level of extraordinary poetic utterance. As a necessary corollary from the prominence given to vakrokti or alamkāra in poetry by this system, it follows that ideas of Rasa should be included in the scope of particular poetic figures. We shall see that Bhamaha actually assigns this function to the particular figure rasavat, and if we are to accept Udbhata's position as indicative of that of Bhāmaha, also to the figures preyas and urjasvin. By putting a technical interpretation on the word vibhāvyate in Bhāmaha ii. 85, Abhinavagupta attempts to make out that Rasa as well as Alaņkāra originates in vakrokti; but this is probably an instance of the not-unusual but rather far-fetched ingenuity of the commentator. Regarding vyangyartha or dhvani, the "suggested sense", which plays such an important part in later theories, Bhamaha nowhere expressly alludes to this idea ; but we can never dogmatically affirm that some kind of suggested sense was not known to him. He defines figures like paryāyokta, vyāja-stuti, aprastuta-praśamsā and samāsokti, in all of which there is an indication of an implied sense. The paryāyokta, for instance, is defined as paryāyoktam yad anyena prakāreņābhidhīyate (iii, 8), and Udbhața expands this with vācya-vācaka-vrttibhyām śūnyenāvagamātmanā, in which there is a clear indication of an avagamyamāna artha27. This

27 Cf. Ruyyaka's remarks on this figure. See also Bhāmaha's definition of samāsokti, ii, 79.

BHĀMAHA. UDBHAȚA AND RUDRAȚA 53

is also apparent from the criticism of Anandavardhana, who does not agree, however, that in Bhāmaha's paryāyokta there is a predominance of the suggested sense, inasmuch as the expressed sense is not intended there to be merely subservient (pp. 39-40). In another place (p. 108), Ānandavardhana further remarks that Udbhata has shown in detail that expressed poetic figures like rūpaka can sometimes be a suggested element, a case of what is explained by Dhvani- theorists as alamkara dhvani. Thus, in the opinion of the great exponent of the dhvani-theory himself, Bhamaha as well as Udbhata (cf °Locana p. 10) is not an abhāva-vādin or one who denies the existence of dhvani (as Mallinātha, p. 24, wrongly considers him to be), but an antarbhāva-vādin who includes the idea of dhvani in other elements of poetry. Discussing this point, Pratīhārendurāja appears to agree with Änandavardhana ; for, in his opinion, the dhvani, which is considered by some school to be the 'soul' of poetry, is not separately dealt with by these early writers because they include it in poetic figures (p. 79). In the same way, Jagannatha (pp. 414-15) remarks that although Udbhata and others, who were earlier than the author of the Dhvani- system, never use the term dhvani, it is yet unreasonable to hold on that ground that the concept of dhvani was unknown to them, because they indicate some of its aspects in their definitions of figures like paryāyokta, samāsokti, vyāja-stuti and aprastuta-praśamsā. To the same effect is the general statement of Ruyyaka (p. 3), who says that Bhämaha, Udbhata and other ancient writers would comprehend the suggested sense in the Alamkāra as an adornment of the expressed sense ; in other words, they do not take it indepen- dently but as an accessory to the expressed sense, in the same way as they take Rasa as an accessory element. Following perhaps the tradition of Bhämaha's paryāyokta, the younger Vāgbhața defines (pp. 36-37) the figure as dhvanitābhidhānam, and refers the curious reader to the treatise of Anandavar- dhana for a detailed treatment of dhvanitokti ; while Hema-

54 SANSKRIT POETICS

candra defines (p. 263) it more briefly as vyangyasyoktiḥ. In all this, one can perceive an attempt to read the idea of dhvani into older authors like Bhāmaha and Udbhata and thus to find an orthodox authority for it from an early time ; but it is not unlikely that the general notion of a suggested sense, like the general notion of Rasa, was not unknown to these ancient authors, although it was only naively understood and never independently treated, being uncritically included as an element of some poetic figures.

( 2 ) UDBHAȚA

The only writer of later times who develops Bhamaha's notion of vakrokti is Kuntaka, the author of the Vakrokti- jīvita ; but for this exposition, it disappears from the writings of this school. Udbhata, one of the earliest avowed followers of Bhämaha, nowhere mentions it, although it is quite possible that we would have got a much more comprehensive idea of Udbhata's standpoint from his lost Bhāmaha-vivarana or Kāvyālamkāra-vivrti28 than from his existing brief com- pendium of poetic figures. His Alamkāra-samgraha, as its name implies, consists merely of a collection of verses defining forty-one poetic figures (including three varieties of anuprāsa), and we are left absolutely in the dark regarding his views on general problems. In his treatment of these poetic figures, Udbhata follows

28 See vol. i. p. 46. The six chapters of Udbhata's existing work deal exclusively with the poetic figures in the following order and divisions: I. punaruktavad-ābhāsa, chekānuprāsa, vrttynuprāsa (with three vrttis), latānuprāsa, rūpaka, dīpaka (3 kinds), upamā, prativastūpamā. II. ākşepa, arthāntara-nyāsa, vyatireka, vibhāvanã, samāsokti, atiśayokti (4 kinds). III. yathāsamkhya, utpreksā, svabhā- vokti. IV. preyasvat, rasavat, ūrjasvin, paryāyokta, samāhita, udātta (2 kinds), śleşa. V. apahnuti, viśeşokti, virodha, tulyayogitā, aprastuta- praśamsā, vyāja-stuti, vidarśanā, samkara (4 kinds), upameyopamā, sahokti, parivrtti. VI. sasamdeha, ananvaya, samsrsti, bhāvika, kāvyalinga (hetu) and kāvya-drstānta (drstānta).

BHĀMAHA, UDBHAȚA AND RUDRAȚA 55

Bhämaha very closely, enumerating the figures in the same order and even borrowing literally the definitions of a large number of them. In the case of a few particular poetic figures, however, Udbhata enters into distinctions which were probably unknown to Bhämaha. Thus, he speaks of four forms of the atisayokti, which Bhamaha does not mention, but which agree substantially with the four out of the five varieties of that figure recognised by later writers. Bhāmaha spoke (ii. 6-8) of only two kinds of anuprāsa, viz., grāmyānu- prāsa and lāțīyānuprāsa, which classification, Pratīhārendurāja thinks, is based on a tacit admission of the two Vrttis, viz., grāmyā and upanāgarikā. Udbhața, on the other hand, dis- tinguishes three varieties of anuprāsa, called chekānuprāsa, lātānuprasa and vrttyanuprāsa, the last of which appears to be classified again on the basis of the three Vrttis, viz., grāmyā (or komalā), paruşā and upanagarikā. These Vrttis, which consist primarily of suitable sound-adjustment with a view to alliteration, appear to have been first recognised, as Abhi- navagupta points out, by Udbhata, and from him known to Änandavardhana (pp. 5-6). We shall see presently that Ru- drața mentions five Vrttis (ii. 19 f) ; but we find Udbhata's views accepted by later theorists like Mammata and Ruyyaka who, however, consider the whole question from the point of view of Rasa29. Again, the grammatical basis of the divisions of upama (of which there is only a hint in Bhamaha ii. 31-33) first appears in Udbhata (i. 35-40) in a form which establishes itself in later theory. It is true that Udbhata does not in the present treatise devote, as Bhāmaha does, a special chapter to the question of grammatical correctness, nor does he allude to the theories regarding functions of words already hinted at by Bhämaha (vi. 6f), yet in deference to the grammatical analysis of speech, he discusses at some length the various

29 These Vrttis refer primarily to anuprāsa (alliteration, or sound- arrangement of letters), and has nothing to do with the four dramatic Vrttis mentioned by Bharata (vi. 25, xx. 24f).

56 SANSKRIT POETICS

subdivisions of upamā, due to suffixes like vat, kyac, kyan, kvip, kalpap and the like, indicative of resemblance ; and this analysis became almost standardised in later literature. Regarding definitions of individual figures, minor differ- ences, as well as further elaboration, are noticeable. Thus, Udbhața's tulyayogitā corresponds to that of Mammata, but Bhämaha's figure of the same name is perhaps equivalent to Mammata's dīpaka. The figures drstānta and kāvya-linga (also called kāvya-drstānta and kāvya-hetu respectively) are omitted by Bhämaha, but defined and illustrated by Udbhata for the first time. But Udbhata is the only older writer who entirely omits the treatment of yamaka. Again, Bhāmaha recognises śleşa involved in sahokti, upamā and hetu, and Dandin speaks of ślesa as coming in and increasing the charm of all figures. But the well-known controversy regarding the division of ślesa into śabda-śleșa and artha- ślesa, together with the question of its relations to other poetic figures in which it may appear, seems to have started, as Ruyyaka notes, from Udbhata's time ; and Udbhata declares that in cases of combination, the ślesa is stronger than the other figures to the extent even of dispelling their apprehen- sion. We shall also see that Udbhata is certainly more advanced in recognising Rasa and defining its place in the poetic figures, if not in poetry as a whole ; and he even goes so far as using the technical terms bhāva and anubhāva, which cannot be traced in Bhamaha. The samsrsti of two or more independent poetic figures is found indeed in Bhā- maha and Bhatti (as also in Dandin and Vāmana), but Ud- bhata does not refer to the two cases of such samsrsti mention- ed by Dandin (ii. 360) and distinguishes it definitely from samkara (pp. 63 and 72), of which he mentions four cases30. 30 Vämana gives the samsrsti a limited scope, recognising only two varieties, upamā-rūpaka and utprekșāvayava, in opposition to Daņdin ii. 258-60. Dandin does not mention samkara. Possibly Daņdin's angāngibhūva-samsthāna variety of samsrsti comes, as Pratīhārendu indicates, under Udbhata's anugrāhyānugrāhaka variety of samkara.

BHĀMAHA. UDBHAȚA AND RUDRAȚA 57

All this, however, means an advance, and not a devia- tion; it indicates an aspect of the growth of scholastic activity, which delighted in indulging in fine distinctions and minute classifications, and not a departure from the original standpoint. To later writers, however, it is not Bhämaha but Udbhata who is the authoritative exponent of this system and whose views are entitled to great respect from all schools of opinion. Bhamaha indeed commands veneration due to his antiquity ; but he was, in course of time, eclipsed by his commentator, and later theorists turn to Udbhata's work as embodying the standard opinions on the subject. We have the testimony of Anandavardhana, Abhinavagupta and Ruyyaka that some of the later specula- tions and controversies (e. g. those regarding upama-divisions. or śleșa) started from Udbhața's time; and Udbhața (as also Dandin and Rudrata) probably showed the way to minute analysis and differentiation of poetic figures, which play such an important part in later theories. We can understand what influence Udbhata's teachings exerted in this respect when we bear in mind that they guided very considerably the enquiries of two important later lawgivers in Poetics, Mammata and Ruyyaka, who fixed for the last time the definitions of most figures, analysed and arranged them on some general principle, and systematised their underlying doctrine. Al- · though Kuntaka elaborated one part of Bhämaha's teaching which he took as the basis of his own peculiar system of vakrokti, it was Udbhata who properly carried on Bhämaha's. tradition and gave a systematic exposition of his work. Along with his contemporary Vāmana, Udbhața may be taken without exaggeration to have been the founder of the Kashmirian school of Poetics which produced its finest fruit in Anandavardhana ; for Udbhata in Kashmir establish- ed the alamkāra-doctrine in Poetics at a time when Vāmana was skilfully constructing a theory of riti on the basis appa- rently of Dandin's teachings, and both of them prepared the way for Anandavardhana.

58 SANSKRIT POETICS

Pratīhārendurāja's interpretation of Udbhata is not always reliable as an indication of Udbhata's standpoint, for the commentator flourished a little over a century later than the text-writer and frequently reads his own notions into the text. For instance, Udbhata can be taken, as we have seen, to have been cognisant of a suggested sense, though he never speaks of dhvani or deals with it directly; but Pratībārendurāja refers to it in clearest terms and attempts by forced interpreta- tion to make out that Udbhata deliberately included it in the treatment of poetic figures. There is no doubt, again, that Pratīhārendurāja was a great deal influenced by the views of the Riti school of Vamana. Discussing the mutual relation of guņa and alamkāra (pp. 75 f), Pratīhārendurāja not only cites Vamana but closely follows his exposition. To Bhamaha, the distinction between guņa and alamkāra was hardly of any theoretic importance, and Udbhata appears to have been of the same opinion ; for Ruyyaka distinctly states udbhatādibhis tu guņālamkārāņām prāyaśah sāmyam eva sūcitam (p. 7), and Hemacandra adds in the same way: tasmad gadarikā-pravāheņa guņālamkāra-bheda iti bhāmaha-vivaraņe ... bhattodbhato'bhyadhāt (p. 17). Vāmana, on the other hand, putting greater stress on rīti elaborately distinguishes between the gunas and the alamkāras ; and Pratīhārendurāja apparently reads Vamana's views into Udbhata. Udbhața omits all mention of riti which Bhamaha had only referred to in passing; but Udbhata speaks of three vrttis, which are connected in particular with the figure anuprasa but which correspond roughly to the three ritis of Vamana, and like the latter, again, to the three gunas recognised by Anandavardhana and his followers.31 But even then it cannot be said that Udbhata's vrttis cover the same ground or possess the same functional value as the three ritis of Vamana or the three gunas of Anandavardhana. Udbhata, according to Abhinava- gupta (p. 134), regards the gunas, again, as the properties of

31 rīter hi guņesveva paryavasāyitā, ° Locana p. 231.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 59

samghatanā, but this samghatana cannot be taken as equivalent to Vamana's riti32. In the same way, Pratīhāren- durãja speaks of rasa as the 'soul' of poetry (p. 77), although there is nothing to warrant the supposition that Udbhata, fully aware as he appears to be of the importance of this element, would regard it as anything but a subservient factor in some special poetic figures.33

( 3 )

RUDRAŢA Although influenced considerably by the Rasa-doctrine, Rudrata belongs properly to the Alamkāra school. He recognises the Rasas and devotes two fairly long chapters to it; but, as we shall see later on, the function he assigns to Rasa is more or less extraneous. On the other hand, what

32 Anandavardhana speaks of samghaiana as threefold, viz., a-samāsā, dīrgha-samāsā and madnyama-samāsā, according as there is the presence or absence, in varying degrees, of compound words. Each of these is suited, though not invariably, to a particular Rasa. But he thinks that the Gunas are not of the nature of samghatanā, nor are they dependent upon samghatana, but that the appropriateness of the samghatan is determined by the Rasa and by the speaker and the subject (pp. 133-5). See Jacobi in ZDMG, lvi, 1902. p. 779, fn 6, and S. K. De, Anandavardhana on samghatanā in Some Problems pp. 91-94. 33 This point will be discussed in ch. iv below. Rājasekhara attributes some other doctrines to Udbhata and his school (audbhatah), which cannot be traced in Udbhata's existing work: (1) that a sentence has a threefold denotation (vākyasya tridhābhidhā-vyāpāra iti audbhatāh), (2) that artha is of two kinds, viz., vicarita-sustha and avicārita-ramanīya, the first found in the Sastras and the second in Kāvyas. The Vyaktiviveka-vyākhyāna attributes a similarly untraceable Siddhänta at p. 4. Such citations or association of earlier authoritative names with a particular view may be merely pūjartha, which is not an unusual procedure with later commentators, as Sukthankar in ZDMG Ixvi, 1912 discusses. Pratīhārendurja, for instance, attributes a strange opinion to Bharata that grammatical works and the like do not deserve the name of poetry because thev are not acceptable as such in the absence of the necessary Gunas.

6G SANSKRIT POETICS

he appears to consider as important in poetry is the alamkāra or poetic figure, to which he devotes ten chapters which form indeed the bulk of his work. His work itself is named Kāvyālamkāra, apparently after the works of Bhāmaha and Udbhata, and is so designated, as his commentator Nami- sādhu admits (on i. 2), from the undoubted emphasis laid on kāvyālamkāras or poetic figures as elements of poetry. Rudrata, like other writers of this school, does not seem to attach much importance to the Rīti or its constituent Gunas. He speaks, no doubt, of four Rītis (and not two, after Dandin, or three, after Vāmana) viz., pāñcālī, lāțīyā, gaudīyā and vaidarbhi ; but in his exposition he is not influenced by the views of the Riti school. The classification of 'diction', he thinks, depends on the presence in varying degrees of short (laghu), middling (madhya) and long (āyata) compound words, or on their entire absence as in the case of the Vaidarbhī which is, apparently for this reason, considered to be the best type. Bhāmaha, we have seen, adopts a similar principle of classification, not with regard to the Rīti but to the three Gunas admitted by himself. The notion of Rīti, therefore, belongs, in Rudrata's opinion, altogether to the province of śabda, governed by fixed rules of verbal arrangement, or rather, of using compounds, and is therefore called the samāsavatī vrtti of śabda. Rudrata does not speak of dhvani, nor does he appear to have been cognisant of its function ; but he implies a suggested sense (as also Bhāmaha and Udbhata do) ancillary to the expressed sense in a limited. number of poetic figures, e. g., in figures like paryāya or paryāyokta and in the figure bhāva vii. 38-41.34

34 The two illustrations that Rudrata gives under the figure are quoted in the Kūvya-prakūsa, and in the Locana p. 45. Abhinava distinctly refers to Rudrata's bhāvālamkāra as a case in which the vyangya sense is subordinate. Abhinava thinks that Udbhata would take bhāvālamkāra as preyas (pp. 71-72). It is remarkable that Ruyyaka, in his review of Rudrata's opinion on this point, states that Rudrata admits the three kinds of suggestion mentioned by the Dhvani-theorists. He says that Rudrata implies vastu-dhvani in the figure bhāva;

BHĀMAHA, UDBHAȚA AND RUDRAȚA 61

Rudrața's detailed treatment of the poetic figures or alamkaras, however, is the distinguishing feature of his work and indeed justifies its title. It is not only elaborate and exhaustive, but also presents considerable difference of method and treatment, which distinguishes him from other earlier writers of this school, and which may lend plausibility to the supposition that he is not only later in time but is also probably following a tradition other than that of Bhāmaha and his followers. To Udbhata's limited number of poetic figures Rudrata adds nearly thirty more independent figures35, besides enumerating several subvarieties of most of the im- portant ones, and devoting an entire chapter to citra (already discussed by Dandin). Udbhata (not to speak of Bhāmaha and Bharata) nowhere treats clearly of the distinction between ideal and verbal figures (i. e. figures relating to word and sense respectively), although such a distinction is implied by

alamkūra-dhvani in rūpaka etc; rasa-dhvani in rasavat and preyas. But it may be pointed out that excepting what is stated with regard to vastu-dhvani being traceable in Rudrata's bhāva, the remark does not apply. Rudrata does not mention, define or otherwise deal with the figures rasavat, preyas etc, nor does he speak of pratiyamānā utpreksū referred to by Ruyyaka in this connexion (although he gives an example of implied utpreksā in ix. 13). See on this point Jacobi in ZDMG 1xii, 1908, p. 295 fn 5. 35 The number in Udbhata is 41, in Rudrata 68 (excluding sub- varieties). The sixteen chapters in Rudrata deal with the following subjects: (I) the purpose and object of poetry, the qualifications of a poet etc. (II) the four ritis (pūñcūlī, lāțīyā, gaudīyā and vaidarbhī), the six bhāsas (Prakrit, Sanskrit, Māgadha, Piśāca, Saurasenī Apa- bhramsa), and five alamkāras of śabda, of which vakrokti and anu- prasa are here treated, along with five Vrttis of anuprāsa. (III) yamaka. (IV) sleșa and its eight varieties, (V) citra. (VI) śabda-dosas, including doşas of pada and vākya. (VII) four bases of arthūlamkāra (vāstava, aupamya, atiaya and śleșa), and 23 figures based on vāstava. (VIII) 21 figures based on aupamya. (IX) 12 figures based on atiśava. (X) 12 figures based on ślesa, śuddha and samkīrņa. (IX) nine doșas of artha, and four upama-doșas. (XII) ten rasas, and treatmuent of śrngāra. (XIII) sambhoga-śrngāra etc. (XIV) vipralambha-srigāra, and the upūyas. (XV) characteristics oi other rasas. (XVI) kinds of poetic composition, such as kathā and ākhyāyikā and their characteristics.

62 SANSKRIT POETICS

his treatment of four śabdālamkāras first, followed by an ex- position of the arthālamkāras. Daņdin also implies a similar distinction by a similar separate treatment without expressly stating it. Rudrata, on the other hand, classifies the figures,. like Vamana, clearly into two groups according as relative prominence is given to śabda and artha. He also gives us for the first time a basis or principle of arranging the individual figures in groups in respect of their general nature or charac- teristic. The śabdālamkāras are arranged under five broad heads, viz., vakrokti (equivocation), śleșa (paronomasia), citra (tricks of pictorial effects, like conundrum etc.), anuprāsa (alliteration) and yamaka (repetition of sounds or rhyming); while the arthālamkāras are classified on a principle of his own, viz., under vāstava (reality), aupamya . (comparison), atiśaya (elevatedness) and śleșa (coalescence)36. The figures. mentioned under arthālamkāras are: (1) VĀSTAVA. Sahokti, samuccaya, jāti, yathāsamkhyā, bhāva, paryāya, vişama, anu- māna, dīpaka, parikara, parivrtti, parisamkhyā, hetu, kāraņa- mālā, vyatireka, anyonya, uttara, sāra, sūkma, leśa, avasara. milita and ekāvalī (23 figures), (2) AUPAMYA. Upamā, ut- prekşā, rūpaka, apahnuti, samśaya, samāsokti mata, uttara, anyokti, pratīpa, arthāntara-nyāsa, ubhaya-nyāsa, bhrāntimat, ākşepa, pratyanīka, drstānta, pūrva, sahokti, samuccaya, sāmya and smaraņa (21 figures). (3) ATIŚAYA. Pūrva, viśeșa, ut- prekşā, vibhāvanā, tadguņa, adhika, virodha, vişama, asamgati, pihita, vyāghāta and hetu (12 figures). (4) ŚLEȘA. Two kinds-śuddha and samkīrna, the former subdivided into aviśeşa, virodha, adhika, vakra, vyāja, ukti, asambhava, avayava, tattva and virodhābhāsa, and the latter comprising. two varieties (10+2=12 figures).

36 Vämana had already taken aupamya as the basis of his classi- fication, for he would regard all figures as upamā-prapañca, implying. that all figurative expression forms nothing more than aspects of metaphorical expression. Rudrata. however, thinks that all figures do not imply comparison ; and in this he is in agreement with all writers excepting Vämana who is unique in his extreme view.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 63

Dandin, who himself enumerates a very large number of poetic figures37, very sagely remarks that if for some slight difference, a different figure is to be defined, there would be hardly any end to their infinite multiplication. This remark partly applies to Rudrata whose general scheme, as well as particular definitions, is open to such an obvious objection. One of the curious results of rigorously following this classi- fication is that the same figure reappears as an alamkāra under different groups. Thus sahokti and samuccaya have two aspects, based respectively on vāstava and aupamya, while the figure utpreksa appears similarly under aupamya and atiśaya respectively. Some of Rudrața's figures have been abandoned by later writers, while some have changed their names or have been modified38, later speculation inclining rather towards the more orthodox expositions of Udbhata or even of Dandin ; yet the general merit of Rudrata's analysis and definitions, testified to by the more or less implicit acceptance by later theorists like Mammata39, cannot be denied. They indicate not only a considerable advance in scholastic activity on Bhamaha and Udbhata, but also remarkable independence, and consequent divergence of treatment in several notable cases. Taking the verbal figures, Rudrata's vakrokti, based on paronomasia (śleșa) and intonation (kaku), has nothing in common with that of his predecessors. The intonational vakrokti is indeed not accepted by some of his successors

37 Viz. 35 in ch. ii. and yamaka, citra and prahelikā in ch. iii. 38 E.g., Rudrata's bhāva, mata. sāmya and pihita are not defined by later writers (excepting Vāgbhata in his Kāvyānuśāsana), while his hetu is not admitted by Mammata. Rudrata's avasara and pūrva (mentioned by the younger Vägbhata) appear to be the same as the second variety of Mammata's (and Udbhata's) udutta and the fourth variety of Mammata's atišayokti respectively. 39 Mammata's indebtedness to Rudrata is discussed by Sukthankar in ZDMG, lxvi, 1912, p. 478, as well as in many places in Nobel's Beiträge already cited. Ruyyaka, on the other hand, while drawing. largely on Mammata himself, is more indebted to Udbhata.

64 SANSKRIT POETICS

(e g. Rājaśekhara p. 31 and Hemacandra p. 234), inasmuch as it is supposed to depend on mere peculiarities of reading (pāțha-dharmatvāt): but on the whole, Rudrata's definition of the figure replaced that of Vāmana's metaphorical vakrokti, survived Kuntaka's broader interpretation of vakrokti, and established itself as the only recognised figure of that name in later literature from Mammata onwards. Again, Rudrata's classification of anuprasa is somewhat different from that of Udbhata, the former basing it on the five vrttis of letters (varņa), viz., madhurā, paruşā, praudhā, lalitā and bhadrā, and the latter admitting only three vrttis (parusā, upanāgarikā and grāmya or komala) only in connexion with one of his three kinds of anuprāsa, viz. vrttyanuprāsa. The later writers follow Udbhata, on this point. Udbhata, again, omits the treatment of yamaka, in spite of the examples of Bharata and Bhämaha before him, and in spite of the fact that Dandin had already given one of the fullest treatments of that figure in the whole realm of Alamkāra literature. Rudrata perhaps ranks next to Dandin in the fulness of his treatment, though there is considerable divergence in the details of classification of these two writers. In the same way, there is no reference to citra in Bharata, Bhāmaha or Udbhata, although Mägha says (xix. 41) that it was in his time a figure indispensable in a Mahākāvya. Daņdin dilates upon some of its varieties, but Rudrata gives a much fuller exposition ; and it is noteworthy that although Mammata does not attach much value to such verbal ingenuity, yet in his discussion of this figure he quotes almost all the illustrations from Rudrata. In connexion with the faults concerning verbal figures, Rudrața points out several cases (vi. 29-33) where punarukta or tautology is not a fault; Udbhata, as Nami-sādhu also notes, includes all these cases in the figure punaruktavad-ābhāsa or 'semblance of tautology'. Udbhata, again, speaks of śleşa apparently as an arthālamkāra, and divides it into śabda śleșa and artha-ślesa, which correspond respectively to abhanga- and sabhanga-śleşa of later writers. Rudrata, on the other

BHĀMAHA, UDBHAȚA AND RUDRAȚA 65

hand, speaks of śleşa as a śabdālamkāra which he carefully distinguishes (ii. 13) from the arthālamkāra of the same name, which he deals with separately in ch. x and which forms the basis of twelve independent figures. The verbal figure ślesa, on the contrary, is elaborately classified according as it relates to varņa, linga, prakrti, pratyaya, vibhakti and vacana (iv. 12), Rudrata thus avoiding the controversy carried on by later theorists as to whether the ślesa is a figure of śabda or of artha. Regarding arthālamkāras, Rudrața mentions only four upamā-doșas (xi. 24), in contradistinction to seven of Bhāma ha and Medhāvin and six of Vāmana40, viz., vaisamya, asambhava, aprasiddhi and sāmānya-śabda-bheda, the last defect including all cases of change of a word signifying common property (as construed with the upameya and the upamāna), due to the difference of linga, vacana, kāla, kāraka and vibhakti. We have already noted that Udbhata enters rather minutely into the grammatical subdivisions of upamā, but this finds no place in Rudrata's treatment, which includes them in a lump in samāsopamā and pratyayopamā. Again, Bhāmaha positively rejects hetu as a poetic figure, although Daņdin speaks of it as vācām uttama-bhūşaņam, including it under kāvya-linga (kāraka-hetu) and anumāna (jñāpaka-hetu). Udbhața recognises only kāvya-linga, calling it also kāvya- hetu and distinguishing it from drstanta which he calls kāvya- drstānta. It is Rudrața (vii. 82) who first defines and fixes its characteristics as finally accepted in Poetics. It is needless to cite any more instance ; but what is said above will be enough to indicate, in the first place, that there is a considera- ble divergence of view between Rudrata and his predecessors with regard to the nature and scope of individual figures and their classification ; and what is more important to note, it is easy to demonstrate that most of these differences are funda- mental. We can reasonably assume, therefore, that Rudrata, possessed as he is of great inventive power, either follows a 40 See vol. i. p. 60, fn 1. 28

66 SANSKRIT POETICS

system of classification and definition peculiar to himself, or follows a tradition of opinion different in some respects from that of Bhamaha and his followers, although in general theory he belongs to a common school. Although Rudrata's work is remarkable indeed for its careful analysis, systematic classification and apposite illustra- tion of a large number of poetic figures, some of which have become more or less standardised, his direct contribution to the theory of Poetics cannot be valued too highly. Indeed, the practical nature and scope of his work, like that of Udbhata's, leave hardly any room for discussion of general principles or of speculative aspects of the questions involved. Rhetoric rather than Poetics appears to be his principal theme, as it is of most writers of this system who concern themselves. entirely with the elaboration of rhetorical categories in which they suppose the whole charm of poetry lies. Partly perhaps. to his novelty of treatment and partly to his omission of dis- cussion of ultimate principles is due the fact that Rudrata has hardly any direct follower in later literature, unless one cites Rudrabhatta who, however, utilises only his Rasa- chapters. Rudrata's name is not associated with the establish- ment of any particular system, although he shows great fertility and acuteness in his treatment of individual figures, which, in some cases, have been implicitly accepted by later writers on the subject. Rudrata is the last great exponent of the Alamkāra school, strictly so called ; for after him the school began to decline and merge ultimately, like the two other sister scho ols relating to Rasa and Rīti, into the finally. dominant Dhvani school.

( 4 ) The decline of the Alamkara-system was probably synchronous with and perhaps hastened by the rise of the rival Rīti-doctrine. The first step towards this is indicated by the general trend of Dandin's work. Dandin who stands, as it were, midway between the Alamkāra and the Rīti

BHĀMAHA, UDBHAȚA AND RUDRAȚA 67

schools, admits, no doubt, the great importance of poetic figures (alamkāras) in a scheme of Poetics (ii. 1); but he takes them, along with the gunas, as constituting the essence of what he calls the märga (or riti). As the characteristic of 'embellishing' poetic speech is possessed by both Alam- kāras and Guņas, both are, in his opinion, Alamkāras in a wider sense, the Gunas being special embellishments of the Vaidarbha Märga, while the Alamkāras are common to both the Vaidarbha and the Gauda Mārgas. Vāmana, who system- atically established the Riti-theory, goes further and lays down that the Gunas are essential (nitya) excellences, while the Alamkāras are not essential (anitya) but serve only to increase the beauty of a poem already brought out by the former. With the advent of the Dhvani-theory, there was an ela- boration of the concept of Rasa as the principal suggested element not only in the drama but also in poetry ; and both the Guna and the Alamkära naturally came to be subservient to it. But the Gunas were supposed to reside in intimate relation to the Rasa, without which they could not exist and existing with which they only served to heighten its beauty. The Alamkäras, on the other hand, were supposed to be extra- neous and artificial sources of beauty, just as ornaments are to the body. We shall have occasion to deal with these questions in greater detail in connexion with the views of the Rīti- and the Dhvani-theorists ; it would be enough to indicate here that the later theorists, in their search for a fundamental principle, could no longer regard the discipline as co-extensive with an external theory of embellishment; and necessarily the Alamkāra, as well as the Guna, which appertain more or less to the objective beauty of representation, came to occupy a. subordinate position as an element of poetry. It was held that the term alamkara 'embellishment' should explain the question as to what is to be 'embelhshed' alamkārya ; and as such it must confine itself, as the early formulators of the Alamkāra- system modestly yet wisely held, to the 'body' or framework of poetry ; it must not attempt to explain its 'soul' or essence.

68 SANSKRIT POETICS

The Alamkara-system, however, left its undoubted impress on later theories. The Rīti-systems of Dandin and Vāmana amply recognise its influence by devoting considerable atten- tion to the detailed discussion of various poetic figures ; and although no writer after Anandavardhana seriously contends that the poetic figure is the only element worth considering in poetry, yet all of them acknowledge its importance and as- sign to it a place in their system. In spite of the emphasis which they put on Dhvani and Rasa, the new school, begin- ning from Mammata, devote a large section of its work to the elaboration of various poetic figures, and the Alamkāra- chapter may justly claim to have been a thoroughly worked out theme. Here was given to Indian scholars rich material for subtle distinction and endless classification ; and with a hair- splitting care, befitting scholastic minds, all kinds of meta- phors, similes, alliterations and other figures were minutely analysed and defined. Indeed, the multiplication of limitless varieties of poetic figures41, based on minute differences, as well as the making of a large number of subvarieties of each figure, went on through the whole course of the history of the discipline ; and down to the latest times, we find traces of new and ever new poetic figures. The extent to which this specialisation is carried will be understood by taking a typical example. The beauty of a lady's face is described ; this can be done in several ways, resulting in a number of poetic figures, by taking the familiar comparison of the Sanskrit poet as the starting point. "Your face is like the moon"-upama ; "the moon is like your face" -pratipa ; "your moon-face"-rūpaka ; "is this your face, or is this the moon"-sasamdeha; "this is the moon, and not your

41 Excluding subvarieties, Bharata mentions only 4 Alamkāras, but in Vişnu-dharmottara we find 18, Bhāmaha 39 (+4), Dandin 38, Udbhata 41, Rudrata 68, Vāmana 31, Mammața 61 (+6), Ruyyaka 75 (+4), Vāgbhața II 63 (+6), Viśvanātha 77 (+7), Jayadeva (Candrāloka) 100, Kuvalayānanda 115. Hence Änandavardhana says (p. 8): sahasraśo hi mahātmabhir anyair alamkāra-prakārāh prakāsitāh prakāsyante ca !

BHĀMAHA, UDBHAȚA AND RUDRAȚA 69

face"-apahnuti ; "the moon is like your face, and your face is like the moon"-upameyopama; "your face is only like your face"-ananvaya ; "having seen the moon I re- membered your face"-smarana; "thinking it to be the moon, the cakora (a bird which is said to feed on moon- beams) flies towards your face"-bhrantimat; "this is the moon, this the lotus, thus the cakora and the bee fly towards your face"-ullekha; "this is verily the moon"-utpreksā; "this is a second moon"-atiśayokti; "the moon and the lotus are vanquished by your face"- tulya-yogita ; "your face and the moon rejoice in the night"- dīpaka; "the moon shines in night, but your face always shines"-vyatireka; "in the heavens the moon, on earth your face"-drstanta ; "the moon reigns in heaven, your face reigns on earth"-prativastūpama; "your face bears the beauty of the moon"-nidarsana; "the moon is pale before your face" -aprastuta-praśamsā ; "by your moon-face the warmth of passion is cooled"-pariņāma ; "your face beautifully spotted with black eyes and adorned with the light of smile"- samäsokti. Some of these turns of expression lose their force or point in the translation, but this will roughly indicate the varieties of figures arrived at by nice distinctions, although these constitute only a few, of which comparison forms the basis. They are sharply separated from one another ; and although some of the distinctions may appear to us to be somewhat trivial or formal, we cannot refuse to recognise the amount of ingenuity shown in the matter. Even taking only one figure upama, they subdivide it into a number of inferior varieties, most of which, however, are based on peculiarities of grammatical construction, but which Appayya Dīkșita, one of the latest writers on the subject, refuses to admit on this very ground in his treatise on poetic figures42.

42 evam ayam pūrņa-lupta-vibhāgo vākya-samāsa-pratyaya-viśeşa- gocaratayā śabda-śāstra-vyutpatti-kauśala-pradarsana-mātra-prayojano nātīvālamkāra-śāstre vyutpādyatām arhati, Citra-mīmāmsā p. 27.

70 SANSKRIT POETICS

But the different theorists are not agreed in their exposition of the exact nature and scope of individual poetic figures. This difference is partly due to the inevitable change of view- points and gradual growth of ideas consequent upon the progress of the study itself, and partly to the favourite refining process which loved to indulge itself in niceties of distinction. The development of the conceptions of the different poetic figures in the writers of different schools affords an interesting field of study in itself, and cannot be comprehended in our limited scope43; but one or two instances will make the process clear. The figure aksepa, which (generally speaking) consists of an apparent denial of something which is intended to be said for the purpose of conveying a special meaning, is variously analysed by different writers. Vāmana defines it as the repudiation of the standard of comparison, upamānāksepaś cākşepaḥ (iv. 3. 27). One interpretation of this, as given in Vāmana's own Vrtti, is upamānasyākşepaḥ pratişedha upamānākşepah, tulya-kāryārthasya nairarthakya- vivakşāyām; that is to say, the standard of comparison is rejected for the purpose of indicating that it is useless in the presence of the object described. This would be equivalent to the figure prafipa of later writers. But Vāmana adds another explanation which indicates that the figure can also occur when the standard of comparison is only hinted at (upamānasyākşepatah pratipattir ity api sūtrārthaḥ). This

43 No complete attempt has yet been made to study the development of the different conceptions of individual poetic figures from the earliest time to that of Jagannatha. Much material, however, will be found in Trivedi's and Kane's notes to their learned editions of Ekāvall and Sāhitya-darpana respectively. J. Nobel has published a series of articles on some of the Alamkaras studied in their development. His Beiträge zur älteren Geschichte des Alamkāra-śāstra (Diss. Berlin 1911) deals with the figures dīpaka, tulya-yogitā, vibhāvanā, viśeşokti, aprastuta- prasamsā, samāsokti, nidarśana, and arthāntara-nyāsa ; while his articles in ZDMG lxvi, 1912, pp. 283-93 and lxvii, 1913, pp. 1-36 deal with vyāja- stuti, and sahokti and vinokti respectively, and in lxxiii, 1919, pp. 189f with prativastūpamā and drstānta.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 71

would be equivalent to the samāsokti of some writers. Dandin's definition of āksepa, on the other hand, is very wide; for, according to him, the denial (pratisedha) need not be of what has been said (ukta) or of what is about to be said (vakşyamāna), but it may be of anything whatever. Bhāmaha, Udbhata and Mammata limit the denial in so far as it concerns the ukta or the vaksyamana. They are followed by Ruyyaka, Vidyādhara and Viśvanātha, but a second kind is added, viz., the apparent permission of what is not wished for. Jagannātha refers (p. 421f) to both the views of Vāmana and Udbhata, but adds that, according to a third view, which he takes to be the view of the Dhvanikara, all suggestive negation or denial is the province of aksepa. This is supported by the fact that the visesa or special meaning to be conveyed by the apparent denial is never expressed but always left to be understood. It would, therefore, be classed by the Dhvanikara under "poetry of subordinate suggestion" (gunībhūta-vyangya), for the expressed sense itself is charming here and the suggested sense is subordinated to it. It is probably in reference to such views that the Agni-purāna łays down: sa ākşepo dhvaniḥ syāc ca dhvaninā vyajyate yatah. The case of āksepa will exemplify, to some extent, the way in which each poetic figure is not only minutely analysed, but elaborately classified into subvarieties by taking into account the different cases of its occurrence. Thus, upama (simile) is classified into six complete and twenty-seven incomplete forms ; the figure utpreksa into thirty-two varieties, the vyatireka into forty-eight, the virodha into ten. The number of self-standing figures, together with their innumera- ble adjunct of subvarieties, goes on increasing as the study progresses, until it reaches to a number exceeding one hundred ; and it is not surprising that in the later stages of its history, whole volumes like Ruyyaka's Alamkāra-sarvasva, Jayadeva's Candrāloka or Appayya's Kuvalayānanda are dedicated exclusively to the special purpose of analysing, defining and illustrating the various poetic figures.

72 SANSKRIT POETICS

The simple basis of classfying the poetic figures, according as they appertain to the word or the idea, into sabdālamkāra and arthālamkāra (verbal and ideal figure) obtained through- out from Rudrata's time44, but some writers add figures which are both of the word and the idea (sabdārthālamkāra). The Agni-purana appears to be one of the earliest known works to. mention this third division, and the position is taken up by Bhoja in his Sarasvatī-kaņthābharaņa and Śrngāra-prakāśa. A long controversy, however, has centred round the propriety of such a classification ; and it has been held that although,. generally speaking, all figures are both of sabda and artha, the raison d'être of such divisions is the relative prominence given respectively to śabda, artha or both, on the dictum yo'lamkāro yadāśritah sa tad-alamkārah. But this relation of āśraya and aśrayin (i. e. interdependence) is not accepted by all, and Mammața maintains that anvaya (connexion) and vyatireka (disconnexion or contrast) must form the test, which consists in considering whether the particular figure does or does not bear a change of synonymous words (parivrtti-sahatva). If the figure disappears with the change of the word by its synonym, it is a verbal figure or śabdālamkāra; if not, it is. an ideal figure or arthālamkāra. The number of independent Sabdālamkāras has never been large, the largest being probab- ly the twenty-four mentioned by Bhoja. The older writers. pay, as a rule, greater attention to this class, which apparently afforded ample scope to decadent classical poets for mere verbal jugglery ; but more recent writers like Anandavardhana or Mammata do not deal with it with so much care, on the- ground that though such word-play brings about variety or vai- citrya of expression to some extent, it does not help but only retards the comprehension of Rasa in a composition by entirely engrossing the reader's mind. The Arthālamkāras, on the other hand, have always engaged more care and attention, and the favourite refining process has been systematically and

44 See above p. 62.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 73

untiringly pursued in this sphere, Their number, however, has always been subject to fluctuation; but it can be gererally stated that while in the older writers the number is limited, in comparatively recent authors the multiplication is more marked. On the one hand, Bharata speaks only of four Arthālaņkāras ; Bhāmaha, Udbhata and Vāmana deal with about thirty to forty different figures; the number reaching its. maximum perhaps in Rudrata's sixty-eight. In Bhoja, Mam- mata and Ruyyaka there is a reaction towards restriction and decrease ; but in the latest stage of our history, the Candrā- loka gives about one hundred, while the Kuvalayānanda adds. a score more. Dandin, himself a great sinner in this respect, very early protested against such endless differentiation, and Anandavardhana agrees with him; but it appears to have afforded endless scope to the scholastic ingenuity of later theorists who, after the time of Anandavardhana, finding hardly anything to systematise in respect of the essentials of theory, occupied themselves in elaborating the details. As to the classification of the ideal figures (arthālamkāra) according to their essential characteristic, Dandin divides all figures into svabhāvokti and vakrokti ; Vāmana makes an early attempt to take aupamya as the central principle; while Rudrata groups them systematically under vāstava, aupamya, atiśaya and śleșa45. Mammata has no definite principle of classifica- tion. Ruyyaka suggests one based on (i) aupamya (comparis- on) (ii) virodha (incongruity) (iii) śrnkhalā (linked succession) (iv) nyāya (logical reason) (v) gūdhārtha-pratīti (understanding of a concealed sense) (vi) combination of figures (samsrsti or samkara). Vidyādhara and Viśvanātha substantially follow this classification, only splitting up nyāya into tarka-nyāya, vākya- nyāya and loka nyāya; but Vidyānātha substitutes the word sādharmya for aupamya (or sādrśya) and speaks of adhya- vasāya (complete identification) and viśeşaņa-vaicitrya (strik- ingness of adjectives or attributes) as two other bases of

45 iii. 9. See above p. 62.

74 SANSKRIT POETICS

elassification. Perhaps none of these classifications would be regarded as strictly scientific, for they mix broad heads indicating psychological factors (like similarity, contrast or contiguity) with mere formal bases of classification as gūdhār- tha-prafiti or apahnava. A development is also noticeable in the general conception of a kāvyālamkāra or poetic figure. It is true that a "poetic figure" corresponds to a certain extent to a "speech-figure" or to what is known as a figure of speech in a formal scheme ; but later theorists explain that something more belongs to a poetic figure as such. The special charm, known as vaicitrya or vicchitti, peculiar to each composition, which rests ultimately on the conception or skill of the poet (kavi-pratibhā or kavi- kauśala) makes up the kāvyālamkāra as such, and gives it its distinguishing characteristic. This view would be entirely omitted in a treatise on rhetoric merely ; and with this point of view it is misleading to describe the theory of Alamkāra as a theory of rhetorical categories only. Originally it might have been, more or less, a theory of externals, but the prob- lem was complicated by the appearance of this new factor of thought, first introduced by Kuntaka and then elaborated in the sphere of individual figures by Ruyyaka, Jayaratha and others, the development of which will be traced hereafter in its proper place.

CHAPTER III

DAŅŅIN AND VĀMANA

(The Riti System)

( 1 )

DAŅŅIN Dandin comes chronologically after Bhāmaha; and Vā- mana, who comes after Dandin, was contemporaneous with Bhämaha's commentator Udbhata ; but the Rīti-system, which Daņdin and Vāmana represent in Poetics, was probably older in tradition than Bhämaha himself, who would not seriously concern himself with the distinction between vaidarbha and gauda Kāvya. It can be traced back to the time of Bāņabhatta {first-half of the 7th century) who tells us that the Gaudas were already notorious for aksara-dambara, to which fact Dandin also alludes in his depreciation of the gauda mārga. It will be seen that although the terms märga or riti in the technical sense might not have been very ancient, both Dandin and Vämana themselves indicate that some such system as they advocate was traditionally existent ; and they appear to refer to and sometimes actually quote from unknown exposi- tors of the past1. It is probable that the Riti school, if we use this term to separate those writers who put an emphasis on riti as the most important element of poetry, had an independent origin and history, and existed for a long time side by side with the sister schools, which threw into promi- nence the elements of rasa, alamkāra or dhvani respectively. We have already stated that Dandin is influenced, to some extent, by the teachings of the Alamkāra school, and as such

1 e. g. under Vāmana i. 2. 11, 12-13; 3. 15, 21; iii. 1. 9, 25; iv. 1. 7 etc.

76 SANSKRIT POETICS

stands midway in his view between the Alamkāra-system of Bhamaha and the Riti-system of Vamana. At the same time, there can be no doubt that in theory he allies himself distinct- ly with the views of Vamana. In Vämana, however, we find the system in its completely self-conscious form ; and here we have nothing of that vagueness or indefiniteness which charac- terises the rival Alamkāra-system of Udbhata. With a clear- cut scheme and a definite central principle, Vämana proceeds. to set forth his system in the brief but concise sūtra-form ; and whatever may be the value of his speculations, there can be no doubt that Vämana was the first writer to enunciate a definite theory which, before the Dhvanikāra, must have had great influence on the study of Poetics. With regard to the commonplaces of poetic speculation,2 Dandin's standpoint does not differ much from that of Bhämaha; and both start with the same notion of embellish- ing sound and sense, which, in their opinion, should con- stitute the 'body' of poetry. The classification of poetry into species of composition like the sarga-bandha (mahākāvya) etc., is almost identical,3 the only remarkable divergence oc- curring, as already noted, in the case of kathā and ākhyāyikā, which rigid distinction is not admitted by Dandin.4 Poetry

2 Daņdin's Kāvyādarśa consists of three chapters devoted res- pectively to (i) Mārga-vibhāga, (ii) Arthālaņkāra and (iii) Śabdālaņkāra and Dosa (the last topic is separately given as ch. iv in the Madras ed.). 3 Under verse, Daņdin mentions sarga-bandha (=mahākāvya), muktaka (single verse), kulaka (five verses), kośa (unconnected verses of different authors), and samghata (unconnected verses of the same author) ; under prose, he speaks kathā, ākhyāyikā and campū. Vāmana gives two divisions of both prose and verse, viz., nibaddha and anibaddha (connected and unconnected). He thinks that verse is of various kinds, but divides prose into three varieties, vrtta-gandhi (savouring of verse), cūrna (having no long compounds and possessing sweet vocables), and utkalikā-prāya which is the opposite of cūrna (i. 3. 21-26). 4 The varieties of kathā mentioned by Hemacandra are ākhyāna, nidarsana, pravahlikā, matallikā, maņikulyā, parikathā, brhatkathā,

DAŅDIN AND VĀMANA 77

is classified, according to its form, into prose, verse and mixed (miśra), while language furnishes another fundamentum divisionis yielding four distinct groups, viz., samskrta, prākrta, apabhramśa5 and miśra,6 the last kind in both these divisions not being mentioned by Bhamaha. The effect of combining these two principles of classification gives us four species of composition, viz. sarga-bandha (mahākāvya) in Sanskrit, skandhaka in Prakrit, osara in Apabhramśa, and nāțaka in mixed languages, although it is not thought necessary to deal with them all in detail. The old division into śravya and preksya Kāvya,7 according as a composition appeals to the eye or the ear, is also referred to in i. 39; but regarding preșya kāvya, by which dramatic composition is generally meant, Dandin summarily refers to specialised treatises on the subject. These speculations, of course, constitute the common stock-in-trade of Poetics, and find themselves repeated in a more or less similar form in most writers irrespective of the school or tradition to which they belong. Thus, Vamana also gives us preliminary chapters on the divisions of poetry

khaņda-kathā, sakalakathā and upakathā (pp. 338 f). The last three are also admitted by Anandavardhana (p. 141) and defined by Abhinava. The Agni-purāņa 337. 20 defines kathānikā. 5 It is not known what Bhamaha signifies by the term apabhramsa, but Dandin gives to it a definite connotation as the language of the Abhiras and others in the Kāvya, as distinguished from the Sastra where it is the name applied to all languages other than Sanskrit. Nami- sādhu sententiously says: prākrtam eva apabhramsah. Hemacandra adds to Apabhramśa another kind called grāmyāpabhramśa. Cf. Bharata xvii. 49, which makes it clear that the Apabhramśa was a jāti-bhāşā and not a desa-bhāşā. The Abhīras were ancient settlers in the land and are mentioned by Patañjali (i. 252). See IA, 1918, p. 26. 6 The mixed variety of Kāvya is cryptically summed as nāțakādi tu miśrakam. But if mixed language is meant, it probably refers to what is now called Hybrid Sanskrit or mixed Sanskrit. 7 This division occurs again in Hemacandra, who divides preksya Kavya into pathya and geya.

78 SANSKRIT POETICS

(kāvya-viśeșa), on che auxiliary aids or sources of poetry (kāvyānga), on the persons entitled to study the science (adhi- kāri-nirūpaņa). Although belonging to a different school, Rudrata in the same way devotes two chapters (i and xvi) to these general topics of Poetics. But in respect of the attention which Dandin pays to the elaboration of poetic figures, his sympathy obviously allies. him with the standpoint of the Alamkara school. He shares the views of this school in his general opinion that a good Kāvya should be embellished by those decorative devices which go by the name of alamkāra. At the same time, it is important to note that while Dandin believes, with all early writers, in the theory of embellishment, he differs in his view as to the means by which this embellishment should be realis- ed; for he apparently holds that it is not the poetic figures only but the several literary excellences, the gunas (which are also designated alamkāras by him), that constitute the essence of the poetic manners (mārgas) or poetic diction, in the realisa- tion of which alone the essence of poetry lies. Indeed, the marked emphasis laid on the Märga, which is almost equiva- Ient to Vamana's Rīti,8 and on its constituent excellences, known as Gunas, to which the Alamkāra school is apparently indifferent, is a distinct feature of Dandin's work, and places Dandin in his fundamental theoretic attitude in the Riti school. Although he does not go so far as Vämana in setting up the Rīti as the essence of poetry, there can be no doubt that he attaches special importance to its literary value. It is true that Dandin never uses the term Riti throughout his work, but his employment of the term Mārga (i. 9, 40, 42, 67, 75, 101) or Vartman (i. 42, 92), implying 'mode,' 'manner', or style in the objective sense, may be taken as almost synonymous. His general definition of poetry, or rather its 'body' or frame- work, as iştārtha-vyavacchinnā padāvalī (i. 10)-a series of words characterised by an agreeable sense or idea-naturally

8 Vāmana also uses the term mārga in iii. 1. 12.

DAŅDIN AND VĀMANA 79

leads him to consider, first of all, the question of appropriate expression of appropriate ideas, or in other words, to discuss. the suitable arrangement of sound and sense for the purpose of producing poetic effect, which is technically denoted by the term märga or riti. Speech, he says, is diversified in its mode of expression (vicitra-mārga, i. 9 ; also i. 40), and he is. aware of the fine distinctions which mark off one mode from another (i. 40) and result in a multifarious variety of modes. All these he broadly divides into two clearly distinguishable types, called the vaidarbha and the gauda, to the critical study of which his whole work is avowedly devoted. This classi- fication is probably not Dandin's own but derived from some recognised tradition (i. 40), although Bharata, who mentions the kāvya-guņas, does not refer to mārga or rīti, and Bhāma- ha's analysis of the two types9 (as well as of the gunas) is. somewhat different. Of the two types, Dandin gives pre- .ference to the vaidarbha mārga, which, in his opinion, results. from a harmonious unification of the ten gunas or excellences of composition, the gauda being the exactly opposite type. The ten Gunas, which are spoken of as the prānāh or life- breath of the Vaidarbha Märga and which are said to be generally wanting in the Gauda,10 are therefore essential in a good composition. They are thus enumerated by Dandix:

9 Bhāmaha does not use the terms mārga or rīti but distinguishes- between the vaidarbha and the gauda Kāvya as two types of poetry obtaining in two different places. Bana has already told us that people of different places liked different poetic devices (Harsa-carita i, śl. 7),. with which statement Dandin's remarks regarding the partiality of the Gaudas to certain tricks and excellences agree. As Vāmana distinctly says, the Ritis, which took their names from localities, were probably analysed empirically from the styles which prevailed in these localities. With regard to Gunas, Bhämaha mentions them independently of the Riti. They are madhurya and prasada marked by the absence of compounds, and ojas by their presence. They would thus correspond to the Ritis of Rudrata. Bharata mentions ten Gunas only as essential. to a good Kāvya, and not in relation to Rīti. 10 Dandin says: esām (i.e. of the ten Guņas) viparyayah prāyo.

80 SANSKRIT POETICS

(i) ślesa, the quality of being well knit, the opposite being śithila or looseness. (ii) prasāda, or lucidity, the opposite being vyutpanna or far-fetchedness. (iii) samatā, or evenness (in the grouping of word-sounds), the opposite being vaisamya or unevenness. (iv) mādhurya, or elegance, consisting of alliteration of similar sound (śrutyanuprāsa11) and absence of vulgarity (agrāmyatva)12, respectively termed vāg-rasa and vastu-rasa. The name of the opposite of mädhurya is not given, but the opposite of śrutyanuprāsa is ulbana-varņāvrtti (i. 35)13 and that of agrāmyatva is grāmyatva, the latter rejected in both the Mārgas. (v) sukumāratā, or absence of harshness due to the use of soft vocobles, the opposite being termed nisthura or dipta.14 (vi) artha-vyakti, or explicitness of sense (common to both the Märgas, the opposite neyatva or neyarthatva being rejected in both).

drsyate gauda-vartmani. It implies that in his opinion the Viparyayas are generally found in the Gauda Marga and not the ten Gunas. The word präyas is significant, and the older commentators draw attention to it. Thus, Taruņavācaspati: prayah-śabdaḥ arthavyakty-audārya- samādhyādayo guņā ubhaya-sādhāraņā iti darśayati; Hṛdayangama: prāyo-grahaņam sākalya-nivrtyartham, tena arthavyakty-audārya- samādhi-guņā ubhaya-mārga-tulyā iti gamyate. Some of the Guņas, therefore, are common to both the Margas. 11 The śrutyanuprāsa is the name given to the grouping of similar sounds which exist between letters belonging to the same sthana, such as kantha, tālu, mūrdhan, danta etc. It is thus distinguished from varņānuprāsa. The subdivisions of anuprāsa are infinite ; but Bhoja (ch. ii. 71f) gives an elaborate scheme of classification containing six main varieties, viz. śruti, vṛtti, varņa, pada, nāma-dvirukti and lāta. 12 For the meaning of the terms gramyatva and rasa occurring in this definition, see below ch. iv. 13 This implies "an excess of a wild variety of Anuprāsa" as V. Raghavan interprets it (Šrngāra-pr. pt. ii, p. 283). For an historical analysis of the Gunas of Dandin and Vämana, see this work of V. Raghavan (pp. 282-299). 14 Dīpta is explained by Dandin as Krcchrodya=hard to pronounce.

DAŅDIN AND VĀMANA 81

(vii) udāratva, or elevation consisting of the expression of some high merit (common to both the Mārgas). (viii) ojas or force due to the presence of compounds (common to both the Märgas, but the Vaidarbha attempts a simpler kind of prose, while the Gauda attempts a hightened style both in prose and in verse, long compounds prevailing in the latter case). (ix) kānti, or agreeableness due to conformity to general usage ; in other words, absence of the unnatural, the exaggera- ted or the grotesque, the opposite being atyukti. (x) samadhi, or transference of the qualities or actions of one thing to another, i. e. metaphorical expression generally (common to both the Margas) It will be noticed that the above enumeration of the liter- ary excellences of diction, which differs greatly from that of Bharata15, is neither exhaustive nor strictly logical. The artha- vyakti, for instance, may well be included in the prasāda. The definition of udāratva is rather vague, so also is that of kānti, in both of which Dandin apparently admits subjective valuations not clearly indicated. Again, the madhurya, though defined primarily as a particular mode of word-arrangement, is regarded more or less as a subtle excellence which defies analysis16. A similar indefinable psychological factor is apparently admitted in the samadhi, the definition of which makes it difficult to distinguish from it poetic figures like rūpaka or metaphor, where there is also poetic superimposi- tion of an object or its qualities on another. It is quite possi- ble that from Dandin's point of view, the difference between the samādhi-guna and the rūpaka-alamkāra may consist in the fact in the Guna there is a transference only of the qualities or actions of one thing to another, while in the Alamkāra

15 See above pp. 12-15. 16 Dandin's treatment implies that he considers it both as a sabda- guņa (illustrated in i. 53) and as an artha-guna (illustrated in i. 64), although this distinction is not mentioned by him (but cf. the word vibhakta in i. 68). 29

.82 SANSKRIT POETICS

either one dharmin itself is substituted for another, or the new dharma entirely supplants the existing dharma. But this process of poetic transference is essentially a mode of figurative expression resting finally on laksaņā, and Vāmana would regard Dandin's definition of the samādhi-guna as constituting the figure vakrokti which, in his opinion, consists in a similar transference based on resemblance. It must also be pointed out that some of Dandin's Gunas refer to śabda (word), some to artha (sense), while others to both these elements. The mādhurya and sukumāratā are primarily śabda-gunas, depending on the use of specific voca- bles, while between themselves the mādhurya requires sound- alliteration, and sukumārata the prevailing use of tender vocables. But the use of words or syllables suggesting a vulgar sense must be avoided in mādhurya, a fact which precludes us from taking it strictly as a sabda-guna. Dandin does not make a hard-and-fast distinction between śabda- gunas and artha-gunas, as his successor Vāmana does; but taking his definitions as they stand, it would appear that he probably regards some of them (e. g. ślesa, samatā, sukumā- rata or ojas, all of which refer to sound-effects) as constituting what later theorists would call śabda-gunas, and others (e. g., prasāda, artha vyakti, udāratva, kānti or samādhi) as con- stituting artha-gunas; while Gunas like madhurya he would in the same way classify as belonging both to śabda and artha. Dandin himself was perhaps conscious of the defective nature of his classification, and consequently added (i. 101- 102) that in the midst of a general agreement regarding the usage of Gunas, there always exist differences between writers and writers as regards the emotional value of their composi- tion or their artistic presentation. After dealing with the Gunas in relation to the two oppo- site types of Märga, Dandin begins (ch. ii) the treatment of those poetic embellishments which are specifically called Alamkāras or poetic figures. It must be distinctly understood that the word alamkara is used by Dandin in the general sense

DANDIN AND VĀMANA 83

of that which causes beauty in poetry, kāvya-śobhākarān dharmān alamkārān pracakșate, ii. 1. It appears to include in its wide scope both Guņas and Alamkāras properly so called. Referring to his own discussion of the Gunas in the previous chapter, in relation to the Vaidarbha Märga of which they constitute the essence, Dandin speaks of them in ii. 3 as alamkāras, and goes on to mention the figures as sādhāranam alamkāra-jātam. In other words, poetic figures are Alamkāras common to both the Mārgas (sādhārana), while Guņas are Alamkaras belonging exclusively to the Vaidarbha. He says, therefore, at the outset of his treatment of the poetic figures (ii. 3) kāścin mārga-vibhagārtham uktāh prag apy alamkriyāḥ/ sādhāraņam alamkāra-jātam adya pradarśyate | /. "For the purpose of classifying the mārgas, some alamkāras have been already spoken of (by me in the previous chapter) ; now are shown those alamkāras which are common (to both the mārgas)". Taruņavācaspati rightly comments, on this verse17: "The ślesa and the like are already spoken of as the ten gunas. If it is objected, therefore, that they cannot also be called alamkāras, the reply is that the characteristic of an alamkāra consists in its capacity of embellishing, and that on account of this characteristic they (gunas) are also alamkaras. The learned teachers have already said that the gunas are indeed alamkāras. Hence alamkāras, like the śleșa, which are essentially gunas, were mentioned before to indicate the difference between the (two) märgas ; but now are enumerated those alamkāras which are common to the two margas." Dandin, it may be pointed out, uses the word alam-

17 pūrvam śleşādayo daśa guņā ity uktam. Katham te'lamkāra ucyante iti cet, sobhākaratvam hi alamkāra-lakşaņam, tallakşaņa-yogāt te' pyalamkārāḥ ......... guņā alamkāra eva ity ācāryāḥ ......... tataḥ śleşā- dayo guņātmakālamkārāh pūrvam mārga-prabheda-pradarsanāya uktāh, idānīņ tu mārga-dvaya-sādhāraņā alamkārā ucyante.

84 SANSKRIT POETICS

kriyā in same general sense in iii. 137 (or iv. 14 in Madras ed.), From what is said above it follows that Dandin does not make a fundamental theoretic distinction between the guna and the alamkāra as such (as later writers from Vāmana onwards do), but apparently regards them both essentially as alamkāra, taking the word in its wider sense of that which embellishes ; the gunas being of primary importance as essen- tials of a good diction, and the alamkāras (i. e. poetic figures) of subsidiary value as constituents of diction, both good and bad. It is noteworthy that Dandin never makes a confusion in the use of the two terms, but invariably applies the former term to denote (except in one case in ii. 364) the literary ex- cellences of diction (i. 42, 76, 81, 100), and the latter to desig- nate the poetic figures to which the name is traditionally restricted (ii. 7, 116, 214, 220, 268, 300, 340, 359 ; iii. 141, alamkāratā in ii. 237, 287, 367). In this way he practically foreshadows, if he does not theoretically develop, the rigid differentiation of the Guna and the Alamkāra of the Rīti school. The subsidiary alamkāras, consisting of poetic figures, are dealt with by Dandin in two chapters (ii and iii), devoted respectively to the treatment of the verbal (śabda-) and ideal figures (artha-alamkāras). He does not expressly state this distinction, which is implied in his treatment, but he gives his general opinion that verbal tricks like those of yamaka are not especially attractive (naikānta-madhuram). He deals with them, however, in greater detail than his predecessor Bhämaha. The prahelika (conundrum),18 for instance, which is merely alluded to in one verse by Bhämaha, is elaborately discussed and illustrated by Dandin (iii. 96-124) who mentions sixteen different kinds of this figure. He also gives a detailed treatment of yamaka, and defines with illustrations such

18 Some prahelikās are already mentioned by Bāņa. Bhāmaha in a somewhat obscure verse states that a prahelikā is a serious compo- sition possessing varied constituent meanings as well as the tricks of yamaka, and is so called in Rāmasarman's Acyutottara. Dandin men-

DAŅDIN AND VĀMANA 85

difficult tricks as gomutrikā, ardha-bhrama and sarvato- bhadra. To the Arthālamkāras, however, Daņdin naturally pays greater attention. He mentions by name (ii. 4-7) only thirty- five poetic figures19 ; but the special feature of his treatment, as contrasted with those of Bhamaha and Vāmana, consists in his attempt to make a large number of subordinate varieties, the most remarkable instance being that of upama, which has thirty-two subdivisions, of which eight at least have the value of independent figures to later writers. Dandin also anticipates Udbhata in pointing out the importance of ślesa as the cause of special charm in other figures (ii. 362), and agrees with Bhämaha in holding that the atiśayokti is essential in all poetic figures (ii. 220). The term vakrokti is used only once and is reserved by him as a collective name of all poetic figures barring the svabhāvokti, According to Daņdin's scheme (as indicated in ii. 362), the whole realm of poetic figures can be divided into two distinct groups, consisting of svabhavokti, on the one hand, and vakrokti, on the other. By the former, which he characterises as the first or primary figure (ādyā alamkrtiḥ), he implies a plain and direct descrip- tion of things belonging to a genus (jati), or of an action

tions 16 kinds of prahelika. Rudrata, like Dandin, deals with the prahelikā as well as the citra-bandhas in some detail. But Vāmana excludes them. Some of the citra-bandhas are mentioned by Bana and Mägha. Under the general designation of citra-kāvya, they are dis- credited by Anandavardhana, and their importance diminished in later Poetics. They became the subject of specialised treatises like the Vidagdha-mukha-mandana of Dharmadāsa Sūri (vol. i, pp. 283-84). 19 The figures dealt with in their order of treatment are: svabhā- vokti, upamā, rūpaka, dīpaka, āvṛtti, ākşepa, arthāntara-nyāsa, vyatireka, vibhāvanā, samāsokti, atiśayokti, utprekşā, hetu, sūkşma, leśa (or lava), yathā-samkhya (or samkhyāna or krama), preyas, rasavat, ūrjasvin, paryāyokta, samāhita, udātta, apahnuti, śleşa, viśeşokti, tulya-yogitā, virodha, aprastuta-praśamsā, vyāja-stuti, nidarśanā, sahokti, parivṛtti, āsīs, samkīrņa and bhāvika. The verses ii. 4-7, which give a prefatory list of figures, are suspected to be an interpolation, but the list sub- stantially agrees with the poetic figures dealt with in ch. ii.

86 SANSKRIT POETICS

(kriyā), of a quality (guna), or of an individual (dravya)20. In this so-called natural description, there is apparently no scope for any artificial or ingenious mode of expression, and it should, therefore, be distinguished from all other poetic devices, figurative or otherwise, collectively designated as the vakro- kti21. Among other figures, defined by Dandin for the first time, may be mentioned the āvrtti, leśa (=vyājokti or vyāja- stuti), sūksma and hetu (the last included by Udbhata in his kāvya-linga). He does not define ananvaya and sasamdeha, calling them asādhāraņopamā and samśayopamā respectively, and includes upamā-rūpaka and utprekşāvayava under rūpaka and utpreksā respectively. With Bhamaha, he alludes to vārttā (i. 85), which is apparently illustrated by Bhatti, but which disappears from later Poetics, being included perhaps in the scope of svabhāvokti. The prativastūpamā is not an independent figure in Dandin but a variety of upamā, while the samahita of Dandin is different from the same figure of Udbhața and Vamana. These few instances would indicate that, compared with the work of his predecessor, Dandin's work attempts to present many new ideas. Possessing great inventive powers and gift of lucid exposition, as well as a notable degree of scholastic acumen, he endeavoured not only to refute and correct in many places the earlier views, but sometimes gave a new shape to them. It will be convenient to examine here briefly the doctrine of Dosa, which forms a counterpart of the doctrine of Guna enunciated by the Riti school. Dandin mentions, after Bha- rata, ten flaws or Dosas of literary composition (ch. iii. 125f, or ch. iv Madras ed.), but he defines them differently in most cases. They are in name and substance identical with Bhā-

20 These terms are to be taken in the sense they have in grammar (and not as they have in philosophy). 21 This interpretation is supported by Dandin's commentators (see Madras ed. pp. 201-2). We have already dealt with the question in our Introd. to Vakrokti-jivita, 2nd Ed. pp. xiv f. See above pp. 48f. 22 See above pp. 9-11. Also Jacobi in Sb. der preuss. Akad. xxiv,

DAŅDIN AND VĀMANA 87

maha's first list of Dosas noted above22, with the only excep- tion of the eleventh fault of defective logic, which is recognised by Bhamaha but vigorously rejected by Dandin as a fault difficult to judge and unprofitable to discuss. But even with reference to this fault, Dandin agrees with Bhämaha in the enumeration of its six subdivisions. With regard to Bhamaha's second list of faults, which concern the inner essence of poetry they would correspond in general to the Dosa (or rather the opposites of Guna) which Dandin mentions as being absent in the Vaidarbha Marga and as generally characterising the opposite Gauda Marga. We have noted that some of these Guņa-viparyayas are expressly named by Dandin. They are (i) the opposite of ślesa, called śithila (ii) the oppisite of prasāda, called vyutpanna (iii) the opposite of samatā, called vaişamya (iv) the opposite of sukumāratā, called dīpta (v) the opposite of kānti, called atyukti (vi) the opposite of artha-vyakti, called neyatva and (vii) the opposite of mādhurya (unnamed). These form seven faults as against ten of Bhamaha, but Dandin speaks of the excellences udāratva, samādhi (and probably ojas) as having no opposites (or corresponding faults), inas- much as they are common to both the Mārgas. Dandin does not regularly deal with Upamā-doșas. But Dandin does not enter into the question first raised by Bharata as to whether the Doșas in Poetics are positive entities or mere negations of Gunas. Bharata holds that Gunas signify nothing more than the negative condition of doşābhāva, so that Dosas are, in his opinion, positive entities, from which the Gunas are known by implication. It is clear from Dandin's treatment, however, that he mentions in ch. iv the external faults apparently as positive entities, after the manner of Bhämaha ; while the essential faults are taken as negations of some of the Gunas of the Vaidarbha Marga and consequently as positive characteristics of the Gauda Mārga. He attempts to avoid the controversy by making use of the distinction of 1922, pp. 222-3. On Dandin's idea of Dosa see V. Raghavan, Śrńgāra- prakāśa, pt. ii, p. 234f.

88 SANSKRIT POETICS

the two opposite types of diction, making the so-called Guņas the characteristics of the Vaidarbha type and some of the so- called Dosas the characteristics of the Gauda type. Vāmana, on the other hand, in conformity to his clear-cut theory of Rīti, goes directly against the opinion of Bharata and expressly makes the Gunas positive entities, defining the Dosas as op- posites of Gunas and as known from the latter by implication (guņa-viparyātmano dosāh, arthatas tad-avagamaḥ). But he adds that the Dosas should be dealt with separately for the sake of clear understanding. He, therefore, divides the Dosas into four classes (i) defects of words (pada dosas) (ii) defects of the meaning of the words (padārtha-dosas) (iii) defects of sentences (vākya-dosas) and (iv) defects of the meaning of sentences (vākyārtha-doșas)23.

23 Rudrata, apparently accepting both Guņas and Dosas as indepen- dent entities, enumerates and classifies Dosas on a different principle. Taking sabda and artha as the two elements of poetry, he mentions in two series (1) sabda-dosas or defects of words and (2) artha-dosas or defect of sense. The first series includes eleven faults, viz. (i) pada-doșas like asamartha, aprafīta, visamdhi, viparīta-kalpanā, grāmya, avyutpanna and desya (7 kinds) (ii) vākya-doşas, such as samkīrņa, garbhita, gatārtha and analamkāra (4 kinds). The second series comprehends. (besides four upamā-doșas) nine faults, viz., apahetu, apratīta, nirā- gama, bādhayat, asambaddha, grāmya, virasa, tadvat and atimātra. Rudrața recognises (xi. 24) only four Upamā-doșas, viz., sāmanya- sabda-bheda, vaisamya, asambhava and aprasiddhi, but Bhāmaha. mentions seven. (On Rudrata's general idea of Guņa and Doșa respec- tively see V. Raghavan, Šrngāra-prakāśa, pt. ii, pp. 302 f and 239 f). Like Bhāmaha and Dandin, Rudrata believes that with change of conditions Dosas become Gunas. After the advent of the Dhvani- theorists, the Dosa (like the Guna) came to be related to the Rasa, the poetic mood in a composition, and began to be defined as that which depreciates or hinders the awakening of Rasa. The doctrine of Dosa was taken along with the doctrine of Guna, of which it formed the counterpart, and was considered from the staudpoint of Rasa alone. They were no longer absolute entities, but attributes or absence of attributes relative to the development of Rasa, and must therefore be governed by the theory of aucitya or propriety which these theorists put forward in their treatment of Rasa. The Dosas

DAŅDIN AND VĀMANA 89

( 2 )

VĀMANA

Vāmana's work24, in comparison with Dandin's, shows: further progress and elaboration of the ideas discussed above. Indeed, what is vague and unsystematic in Dandin appears fully developed and carefully set forth in Vamana, who may thus be fittingly regarded as the best representative of the Riti-system. To Vamana belongs the credit of being the first writer on Poetics who, before the Dhvanikāra and Änandavardhana, gave us a well thought-out and carefully outlined scheme of Poetics, no longer naïve or tentative, which

came to be defined generally as rasāpakarşaka (Viśvanātha) ; but specific rasa-dosas also came to be defined and discussed. The question whether the Dosa is nitya or anitya (already raised by Bhāmaha and Rudrata) is solved by supposing that a Dosa may sometimes become a Guņa if it helps (and not hinders) the development of the Rasa. Mammata and most later theorists accept the distinction of Dosas relating to pada, vākya and artha ; but they speak also of rasa-dosas and alamkāra- doşas. The later opinion regarding the respective character of Guņa and Doșa appears to be that each of them conveys a positive meaning, in spite of the fact that some Dosas approach the condition of gunā- bhava and some Gunas approach the condition of dosabhava. Even Dandin in ch. iv separately mentions ten Dosas which he does not regard as constituting the opposite of any of his Gunas. 24 Instead of having adhyāyas first and adhikaranas thereunder, Vämana reverses the order of older sūtra-writers and divides his work into five adhikaranas, each of which consists of two adhyāyas (except- ing the first and the fourth which contain three each), the whole work thus having five adhikaranas and twelve adhyayas. The arrangement of these adhikaranas is as follows. I. sarira: dealing with the object of poetry, persons qualified to receive instructions in the subject, the Rīti and its subdivisions, the subsidiary aids and divisions of poetry. II. doșa-darśana: treating of the defects or flaws of composition. III. guna-vivecana : considering the Gunas or excellences of composition. IV. ālamkārika: devoted to the definition and illustration of poetic figures. V. prāyogika: setting forth the poetic conventions and pro- priety of poetic usages, and sabda-uddhi (corresponding to the last chapter of Bhamaha's work), explaining grammatical solecism.

90 SANSKRIT POETICS

in spite of its theoretic defects, is in some respects unique and valuable. The enquiry as to what is the 'soul' or essence of poetry is for the first time definitely posed and systematically worked out by Vamana; his predecessors, to whom the 'body' of poetry was more important, never having troubled themselves with this question. Vamana lays down in clear terms: ritir ātmā kāvyasya 'the Rīti is the soul of poetry' (i. 2. 6) ; and working out this figurative description he points out (on i. 1.1) that the word (sabda) and its sense (artha) constitute the *body', of which the soul is the Rīti25. He defines the Rīti as viśistapada-racanā or particular arrangement of words. This particularity (vaśistya) of arrangement, again, rests upon certain definite combination of the different Gunas or fixed excellences of composition. For instance, of the three kinds of Rīti proposed by Vamana, the Vaidarbha unites all the ten Guņas, the Gaudī abounds in ojas and kānti, the Pāñcālī is endowed with mādhurya and saukumārya. This is how Vamana would distinguish the different Rītis from one another. On these three Rītis poetry takes its stand, just as painting has its substratum in the lines drawn on the canvas (on 1. 2. 13). The Vaidarbhi is of course recommended, for it contains all the excellences; and as the genius of each diction is peculiar to itself, Vämana rejects the view that the other two inferior dictions ought to be practised as steps leading up to the Vaidarbhi. He argues that the proper

25 Vāmana says (i. 1. 1) that the word kāvya applies in strictņess to word and sense embellished with Guņa and Alamkāra, but it is employed in the secondary sense (bhaktya) to mere word and sense. By Kāvya-śarīra he means elsewhere (i. 3. 10) itivrtta or the content or subject-matter of poetry. But the first chapter of his work is designated Sārīra; and on i. 2. 6 he says that the word sarira must be understood after the word kavya in the Sūtra, meaning thereby that Kāvya con- sisting of word and sense is the Sarira, of which the Atman is Riti. For the history of the concept of Rīti see V. Raghavan, Some Concepts, pp. 131-172; Prakash C. Lahiri, Concepts of Rīti and Guna, Dacca University 1937.

DAŅDIN AND VĀMANA 91

diction cannot be attained by one who begins with the im- proper. If the weaver practises weaving with jute, he does not attain proficiency in the weaving of silk. It will be seen from this analysis of the three kinds of diction that the Vai- darbhī is the complete or ideal one which unifies all the poetic excellences, whereas the other two encourage extremes. The one lays stress on the grand, the glorious or the imposing, the other on softness and sweetness, whereby the former loses itself often in bombast, the latter in prolixity. It will be noticed also that the names of the different Rītis are derived from those of particular countries, and Vāmana expressly says in this connexion (i. 2. 10) that the names are due to the fact of particular excellence of diction being prevalent in the writings of particular countries26. This makes it probable that the theory of diction, peculiar to this school, originally arose from the empirical analysis of the prevailing peculiar- ities of poetic expression in different places, and furnishes another proof of the general a posteriori character of the discipline itself.27

26 vidarbha gauda-pāñcāleşu tatratyaih kavibhīr yathā-svarūpam upalabdhatvād tat-samākhyā, na punar deśaih kimcid upakriyate kāvyānām. 27 In the absence of proper data, it is not possible to determine when the distinction between the Eastern and Southern styles-Gauda and Vaidarbha-was first recognised. We have already noted that Bāna speaks of people of different localities affecting different tricks of style, some putting stress on sound, some on sense, while others indulging in a play of fancy. In Dandin's time the distinction must have been fully established. Jacobi (Māhārāstrī, pp. xvi) suggests that the simpler Vaidarbha style was a reaction against the older and more ornate Gauda- style (which Dandin disfavours), and that it came into existence pro- bably in the 3rd century A. D., being known to us from Hāla's Saptaśatī (5th century A.D.). It is possible to argue, on the contrary, that the Gauda-style itself is a sign of further development or decadence, exhibiting a tendency to a more elaborate style (as opposed to an earlier and simpler Vaidarbha-style) which we find, as a matter of practice, asserting itself more and more in later decadent Sanskrit Kāvya. Cf Keith, Classical Sanskrit Literature p. 50.

92 SANSKRIT POETICS

It should be observed that the term Rīti is hardly equiva- lent to the English word 'style', by which it is often rendered but in which there is always a distinct subjective valuation. Although artha (i. e. sense or idea) is admitted as an element by Sanskrit writers, the Riti consists essentially of the objec- tive beauty of representation (of the intended idea), arising from a proper unification of certain clearly defined excellen- ces, or from an adjustment of sound and sense. It is, no doubt, recognised that appropriate ideas should find appro- priate expression ; or in other words, the outward expression should be suitable to the inward sense. Bharata goes further and formulates that in the drama the expression should also be in keeping with the temperament and character of the speaker to whom it is attributed. But at the same time, the Rīti is not, like the style, the expression of poetic individua- lity, as it is generally understood by Western Criticism, but it is merely the outward presentation of its beauty called forth by a harmonious combination of more or less fixed literary 'excellences.' Of course, the excellences are supposed to be discernible in the sense or import. as much as in the verbal arrangement, but this subjective content is not equivalent to the indefinable element of individuality which constitutes the charm of a good style. If we accept the nomenclature of a modern analyst of style28, we may say that the Sanskrit authors admit what he would call the 'mind'-in-style, as a subject of technical formulation, but not the 'soul'-in-style, which is elusive and which they leave to individual writers to evolve in their own way.

28 Walter Pater's essay on Style in his Appreciations. V. Raghavan (Some Concepts, p. 140) demurs to our views; but it does not help to refer to Demetrius or Aristotle; for we are not concerned here with European classical literature but with modern literature, as viewed by Walter Pater or Benedotte Croce. It is not until we come to Kuntaka that poetry is brought in strict relation to poetic individuality (kavi- svabhāva) or poetic conception (kavi-pratibha). This is admitted by Raghavan himself.

DAŅDIN AND VĀMANA 93

Vämana, therefore, teaches that the Gunas are essential in poetry, as they go to make up the Rīti, which is the 'soul' of poetry. The objection of the author of the Ekāvalī (p. 51) is that to conceive the Gunas, on the one hand, as the principal element and therefore as something fit to be adorned (upas- kārya) and to call them, on the other hand, properties that adorn poetry (upaskāraka,) involves contradiction in terms. This raises only a scholastic quibble which does not bear serious examination. The Gunas are, no doubt, spoken of as śabdārthayor dharmāh, but this, as the commentator points out, is a loose or popular use of the term ; for, strictly speak- ing, they are concerned directly with Rīti (gunā vastuto rīti- nişthā api, upacārāc chabda-dharmā ity uktam p. 69, ed. Benares), which is described as gunātma. To the objection that these entities have no absolute existence, Vāmana replies that their existence is vouched by their cognition as such by men of taste (samvedyatvāt iii. 1. 26, on which comm. : sahrdaya- samvedanasya vişayatvāt), and that these excellences are not found in all cases of recitation but depend upon the presence or absence of certain well-defined characteristics (iii. 2. 28 and Vrtti). Like Dandin, Vāmana enumerates the Gunas as ten, which appears to have been the standard number from Bharata's time, but he really doubles the number by clearly differentiat- ing between the sabda-gunas and the artha-gunas, and regard- ing each Guna as belonging respectively to śabda and artha. In other words, each Guna is looked at from two different points of view, and the distinction thus proposed between verbal and ideal excellences comes in, as technically put, according as the word or the idea is the denoter (vācaka) or the denoted (vacya). We find in Vämana, for the first time, the definite classification of Gunas of Sabda and Artha respectively. This sharp distinction, no doubt, clears away some of the vagueness surrounding definitions of Bharata's and Dandin's individual Gunas ; and Vāmana, though widely differing from his predecessors in the peculiar connotation

94 SANSKRIT POETICS.

he attaches to some of them, is careful in distinguishing the allied Gunas from one another. In most essentials Vāmana undoubtedly continues and expands Dandin's somewhat unsystematic scheme, but his definitions bear in some cases an altogether different complexion, and justify us in presum- ing that Vamana develops his ideas from elsewhere, as he himself often supports his analysis by verses quoted from unknown sources (e. g. under iii. 1. 9, 25 ; 2. 15 etc.). His scheme of the Gunas may be tabulated thus: Śabda-guņa Artha-guņa i. ojas, or compactness i. ojas, or maturity of of word-structure (gādha- conception (arthasya prau- bandhatva, where bandha= ḍhiḥ) pada-racanā iii. 1. 4) ii. prasāda, or laxity of ii. prasāda, clearness of structure (śaithilya) meaning (artha-vaimalya) by avoidance of superfluity (anu- payogi-parivarjanāt, as Abhi- navagupta explains) iii. śleşa, or coalescence iii. śleşa, or coalescence: of words resulting in smooth- or commingling of many ness (masrnatvam, yasmin ideas (ghatanā) sati bahūny api padāny ekavad bhāsante) iv. samata, or homo- iv. samatā, or non- geneity of manner, i, e., of relinquishment of proper construction (mārgabhedah, sequence of ideas (prakra- yena mārgeņopakramas tasyā- mābheda) tyāgaḥ) v. samādhi, or symmetry due to orderly ascent and v. samādhi, or grasping: of descent, i.e. when the heigh- the original meaning

tening effect is toned down arising from concentration

by softening effect, and vice of the mind (artha-drstih samādhi-kāraņatvāt) versa (ārohāvaroha-krama)

DAŅŅIN AND VĀMANA 95

Śabda-guņa Artha-guņa vi. mādhurya, or distinct- vi. mādhurya, or striking- ness of words (prthak-padatva) ness of utterance (ukti-vaici- due to absence of long trya), i. e, in an impressive compounds (samāsa-dairghya- periphrastic manner for nivrtti) special charm vii. saukumārya, or free- vii. saukumārya, or free- dom from harshness (ajara- dom from disagreeable or thatva) inauspicious ideas (apārusya) viii. udāratā, or liveliness (viii. udāratā, or delicacy in which the words seem as i. e. absence of vulgarity if they are dancing (yasmin (agrāmyatva) sati nṛtyantīva padāni) i. e. pada-vicchedāt? ix. artha-vyakti, or ex- ix. artha-vyakti, or expli- plicitness of words whereby citness of ideas which makes the meaning is easily oppre- the nature of things clear hended (jhatitya-artha-prati- (vastu-svabhāva-sphuțatva) patti hetutva) x. kānti, or brilliance, x. kānti, or prominence i.e. richness of words of the rasas (dipta-rasatva) (aujjvalya) It will be seen from this brief enumeration that Vama na's Guņas differ considerably from those of Bharata or Dandin. Vamana's ojas, for instance, would correspond to Dandin's śleşa, while Daņdin's mādhurya is split up into prthak-padatva and agrāmyatva. Vamana brings in the idea of Rasa in the artha-guna kānti (cf. Bharata's kānti), thereby admitting it in one of the essentials of poetry, while Dandin acknowledges it in some of the non-essential poetic figures. Vamana's artha-guna artha-vyakti would be regarded by Dandin as an instance of svabhāvokti alamkāra. But it must be observed that Vamana's scheme, while being more systematic and thus marking an advance on the speculation of Dandin's, is hardly satisfactory as a whole. The somewhat pedantic classification of Gunas, into external and internal, verbal and

96 SANSKRIT POETICS

ideal, is in itself open to objections and has been controverted by later theorists29. The distinctions are sometimes uncon- vincing, and it is natural to suspect that they are made for the sake of symmetry of having two sets each of ten excell- ences. Taking the individual Gunas, it appears that Vamana himself was perhaps conscious of the defective nature of some of his definitions, although he must have elaborated the ten Gunas according to some current convention. For instance, his śabda-guna prasāda is merely a negation of his ojas ; Vāmana himself admits this and adds the qualification that the prasada is an excellence when appearing along with ojas and not by itself, for when it appears by itself it is clearly a defect. If it is objected that there cannot be a combination of two such contradictory Gunas, Vāmana replies that such a com- bination is a fact of common experience. Vamana's śleşa is rejected by Mammata as an independent excellence, inas- much as it is only a particular form of ojas. In the same way, the saukumarya is not admitted by later theorists on the ground that it is merely the negation of the defect of harsh- ness, just as Vāmana's udāratā is a negation of vulgarity (grāmyatva). His samādhi as an artha-guņa is different from Dandin's samādhi, but it is hardly an 'excellence.' It is supposed to consist in comprehending the original meaning, but this happens in the case of all compositions; for there can hardly be any poetry, worth the name, of which the meaning is not comprehended. These and other objections are brought forward by Mammata and his followers, who protest against this needless multiplication and differentiation of the Gunas, and who sum up the literary 'qualities,' which they admit as an embellishment of the principal sentiment (Rasa) of the composition, under three broad categories, differentiated on entirely psychological grounds, viz., ojas (energy), prasāda (lucidity) and mādhurya (sweetness).

29 e. g. Mammața viii, Hemacandra pp. 195-200, Māņikya-candra pp. 191f, Jagannatha 62f etc.

DANDIN AND VĀMANA 97

It is also noteworthy that the Rīti-system assigns to some of the Gunas functions which other systems assign to Alamkāras or poetic figures. Vāmana's artha-guņa kānti corresponds to figures like rasavat of the Alamkära-system,30 while Dandin's samadhi in some cases will be equivalent to the rupaka and analogous metaphorical figures. Vāmana's artha-guna artha-vyakti is nothing more than the svabhavokti of Dandin, and Dandin's kanti only defines the limit, as Hemacandra puts it, to the figure atiśayokti (seyam atiśa- yokter yantraņā, na punar gunāntaram). On the basis of Bhāmaha's vakrokti, the later writers of the Alamkāra school, we shall see, postulate ukti-vaicitrya as the fundamen- tal principle of all figurative expression, but this is compre- hended merely as one of the excellences by Vāmana's defini- tion of the artha-guna mādhurya. It will be clear from the above, as well as from the elaborate review of the Guna-doctrine by Hemacandra (pp. 195-200) and Mānikyacandra (pp. 191f) that the writers of the Rīti school, especially the followers of Dandin and Vāmana, differ widely in their attempts at defining and classifying the Gunas, and that such attempts are open to criticism. Some of Bharata's definitions, on the other hand, do not always agree with those either of Dandin or Vamana. For instance, Bharata's ojas, which consists in the use of high-sounding compound words of varied strikingness, may correspond to Dandin's ojas, but Hemacandra states that the essence of this Guna consists in imparting loftiness to an object which is low or treated with contempt. Again, Bharata's prasāda, corresponding to Daņdin's samādhi, is a metaphori- cal mode of expression, which Vamana would include in his peculiar definition of the vakrokti-figure, and which comes generally under laksanā or upacāra of later theorists. Bhara- ta's udārà differs considerably from Vāmana's udāratā which does not comprehend, as Bharata's udāra does, the Rasas and Bhavas in its scope ; but it corresponds partially to Vāmana's

30 For this reason Vamana does not define the rasavat as a figure. 30

98 SANSKRIT POETICS

artha-guna kanti. At the same time, it would seem that, in spite of such minor discrepancies, Bharata's scheme of the Gunas as a whole is developed to its furthest possibili- ties by Vamana. That there are inevitable differences in the definition of particular Gunas in the earlier writers on the subject, and that the whole doctrine, despite the care of its exponents, is still unsatisfactory, only indicate the fruitless- ness of the efforts of early theorists in comprehending all the literary excellences of a composition within the hard-and- fast limits of a few categories, on the interpretation of which they spend so much ingenuity but on which they cannot in the nature of things arrive at any absolute agreement. Hemacandra and Mānikyacandra in their review of this doctrine cite the opinion of an authority31, called Mangala who, we are told, agrees with Bharata in his defini- tion of ojas and maintains with Vamana that Dandin is not right in emphasising it in the Gaudi Riti inasmuch as it is common to all Ritis. The only other reference to this writer is made by Rājaśekhara, who cites some opinions of this writer agreeing substantially with those of Vamana, and quotes at p. 14 a dictum from Mangala which occurs in a slightly different form in Vāmana i. 2. 1. From these citations by Rājasekhara it appears that Mangala, if he is not earlier in date than Vamana, belongs most probably to the same school of opinion. There can hardly be any doubt that the system existed even before Vämana gave a definite form to it; and Vamana's systematic formulation certainly obtained for it a large number of adherents and followers, so that important later authors like Rājaśekhara, Hemacandra and Jayaratha cite with respect the opinions of the Vamaniyas, just in the same way as they cite the Audbhatas, the followers of Vamana's contemporary and rival Udbhata.

31 Vāmana himself quotes some verses giving us ancient · definitions of the various śabda-guņas (under iii. 1. 25).

DAŅDIN AND VĀMANA 99

After the Gunas, Vämana deals with the poetic figures or Alamkāras as elements of subsidiary importance. This definite differentiation of Guna from Alamkāra we meet for the first time in Vamana; for Bhamaha was indifferent to it, Dandin does not accept it, and Udbhata appears to have denied any difference32. At the outset Vamana states, no doubt, that poetry is acceptable from embellish- ment (alamkāra); but he is careful to explain embellish- ment, not in the narrow sense of poetic figure, but in the broad and primary sense of beauty or charm (kāvyam grāhyam alamkārāt, saundaryam alamkārah). He also points out that it is only in the secondary instrumental sense that the term alamkāra or embellisment is applied to simile and other poetic figures (alamkrtir alamkāraḥ, karaņa-vyutpattyā punar alam- kāra-śabdo'yam upamādișu vartate). In this view, Vāmana apparently develops logically Dandin's teaching ; but Vāmana does not make the presence of poetic figures a necessary condition. What makes poetry acceptable, in his opinion, is the presence of charm or beauty (i. e. Alamkära in its broad sense of Saundarya) which he does not define and which is in some respects undefinable33. The Rīti and its constituent Guņas come in as a sine qua non in the production of this beauty, but the poetic figures only contribute to its heighten- ing. This distinction between the Guna aud the Alamkāra as to their respective position in a formal scheme of Poetics, which is vaguely hinted at by Dandin, is fully developed for the first time by Vamana (iii. 1. 1-3). The Gunas, being essential to the Rīti, are defined as those characteristics which create the charm of poetry (kāvya-śobhāyāḥ kartāro dharmāḥ) -a function which is assigned to both Guņas and Alamkāras by Dandin-but Alamkāras are such ornaments as serve to enhance the charm already so produced (tad-atiśaya-hetavah).

32 The distinction was established by the Dhvanyāloka. 33 Vāmana simply says that this beauty is to be attained by avoiding Doşas and employing Guņas and Alamkāras (sa doşa-guņā- laņkāra-hānādānābhyām).

100 SANSKRIT POETICS

The Gunas are said to be nitya (permanent), implying that the Alamkāras are anitya (punar alamkārā anityā iti gamyate eva, Kamadhenu com. p. 71), for there can be charm of poetry without the Alamkaras but no charm without the Gunas (tair vinā kāvya-śobhanupapatteh). In other words, the Guņa stands to poetry in the samavāya-relation (see Kāmadhenu on iii. 1. 4), while the Alamkāra in samyoga-relation, samyoga being explained as mere conjunction and samavāya implying inseparable connexion or inherence (nitya-sambandha)34. To put it in the usual figurative language, the Guna is related to the 'soul' of poetry (viz. Rīti), while the Alamkāra rests merely on the 'body' (viz. sabda and artha). The Alamkāra, without the Guna, cannot of itself produce the beauty of a poem, but the latter can do so without the former. But Vāmana justifies at the same time the existence as such of the Alamkara as an element of poetry, and supports a phase of poetry, which is indeed admitted by Anandavardhana but not properly dealt with by him, and which is elaborated only by his follower Ruyyaka who, however, takes his inspiration on this point from the Vakroktijīvita-kāra. Vāmana's treatment of the poetic figures is in some respects peculiar to himself in its general outline, as well as in the specific definition of individual figures. Vāmana is the only old writer who deals with the smallest number of poetic figures35. He recognises only two kinds of śabdāla- mkāra, viz., yamaka (rhyming) and anuprāsa (alliteration). He rigidly excludes prahelikas and bandhas. With regard to

34 This is criticised by Mammata viii, p. 470 ; for he maintains that Guņas like ojas and Alamkāras like anuprāsa and upamā reside in the relation of inherence (samavāya-sthiti). 35 They are in their order of the treatment: yamaka, anuprāsa, upamā, prativastūpamā, samāsokti, aprastuta-praśamsā, apahnuti, rūpaka, śleşa, vakrokti, utprekşā, atišayokti, samdeha, virodha, vibhā- vanā, ananvaya, upameyopamā, parivrtti, krama, dīpaka, nidaršana, arthāntara-nyāsa, vyatireka, viśeşokti, vyāja-stuti, vyājokti, tulya-yogitā, ākşepa, sahokti, samāhita, and samsrsti (including upamā-rūpaka and utprekşāvayava) :- 30 figures excluding Samsrsți.

DAŅDIN AND VĀMANA 101

the arthalamkaras, he lays down in general terms that the upama or com parison lies at the root of all poetic figures, which are defined in relation to it and to which is given the collective name of upama-prapañca36. The importance of upama, involved in other figures, is recognised from Bhämaha's time ; and consequently this figure, which is the source of all the figures grouped together by later writers as sādrśya-mūla or aupamya-garbha Alamkāras, is always given a place of honour at the beginning of most treatises on Sans- krit Poetics37; but Vāmana goes to the extreme of defining all figures with reference to the idea of comparison, or in terms of the relation of the upamana and the upameya. On account of this fundamental postulate, his definitions of some of the figures differ widely from those given by other writers ; and he has also to exclude such figures as paryāyokta, preyas, rasavat, ūrjasvin, udātta, bhāvika and sūkșma which he does not define. He gives a peculiar definition of the figure vakrokti as a mode of metaphorical expression. His viśeşokti would correspond to the rūpaka of Jagannātha, and his ākșepa to the pratīpa or samāsokti of some later writers. Vāmana defines apahnuti as the concealment of one thing by a similar thing with a view to impose the character of the latter on the former ; in other words, the upameya is denied its nature and the upamana is established in its place. Dandin states that denial of something and the representing of some other thing in its place constitute the figure which need not be based on aupamya ; and following him, some later writers (e. g. Viśvanātha) speak of a second variety of

36 The commentator explains: prativastu-pramukhānām alamkārā- ņam upamā-garbhatvād upamā-prapañca iti vyapadesah krtah (on iv. 3. 1). 37 upamaivāneka-prakāra-vaicitryeņālamkāra-bījabhūteti pratham- am nirdistā, Ruyyaka p. 26; sādrśya-vicchitti-viseşaī rūpaka-dīpakā- dyanekālamkāra-bījatayopamāyāh prathamam nirūpaņam, Mallinātha p. 195. Although Upama need not be involved in all poetic figures, Vāmana appears to accept only those figures which are based on poetic comparison, as it imparts a special charm to them.

102 SANSKRIT POETICS

apahnuti in which there is no gamyamana aupamya or implied comparison38.

( 3 ) This brief summary of the views of the leading authors of this school will shew that the Riti-teachings mark a great advance on the Alamkara-doctrine in many respects. There are many points which are common to both these systems; but, by clearly defining and working out the doctrine of Rīti as that distinct characteristic of poetry which sharply separates it from dry philosophical or technical writings, the Rīti school seems to have first suggested and started the enquiry (only hinted at by Bhämaha's theory of vakrokti) as to what constitues the essential charm of poetry, and anti- cipated the theory of vicchitti (or ukti-vaicitrya) elaborated later by Kuntaka and other adherents of the Alamkāra- doctrine. The Dhvanikāra pays an indirect compliment (iii. 52) to the Rīti school for having first perceived, however dimly, the true nature of poetry, although he does not agree with its peculiar theory of Rīti. The Rīti school also goes a step further than the Alamkāra school in including Rasa among the necessary characteristics (in Känti as an Artha-gua). It is possible that Vämana's partiality for the drama, which he considers to be the best form of composition and from which he supposes other forms of poetry to proceed (i. 3. 30-32), led him to realise the importance of Rasa, already worked out as fundamental in the drama by the dramaturgic Rasa school, and to incorporate it in one of the essential properties of poetry (iii. 2. 15 and Vrtti). But, at the same time, it was perhaps his idea to make bis definition of poetry comprehensive enough to cover a larger field and include those instances, e. g., which develop no Rasa. The kāvya-sobha, a term which he probably borrows from Dandin (ii, 1), or saundarya which conveys the same general idea as

38 Cf Uddyota, ed. Candorkar, p. 39.

DAŅDIN AND VĀMANA 103

the word 'beauty', is regarded as the ultimate test of all poetry ; and this beauty, in his opinion, agreeing with the common-sense view of the matter, is realised by carefully worked-out diction, which avoids the damaging flaws by adopting primarily the so-called literary excellences, as well as the poetic figures for the secondary purpose of heightening the effect thus produced. But the Rīti-system, in spite of Vamana's well- reasoned formulation and the advocacy of his followers, never appears to have wielded very great influence, and its existence was comparatively short-lived39. There is no doubt that like the Rasa and the Alamkāra- systems, it left its impress on later theories, but it never found a serious champion after Vamana among latter-day writers, and its theories never found unqualified acceptance. Its general doctrine of Rīti began to be discredited and severely criticised from Anandavardhana's time as too crude an ex- planation of the nature of poetry; and Mammata, the fore- most authority of the latest school, ingeniously combats and sets aside the leading views of Vāmana. It may, however, be noted that some of the broad princi- ples enunciated by the Rīti school have been tacitly recognis- ed by later theorists. The importance of Rti or diction, as such, became established as a stock-idea in Poetics: but it was accepted with grave modifications. It was accepted by the Dhvani school in so far as it contributed to the develop- ment of the rasa-dhvani, and its chief characteristic was supposed to consist in an arrangement or disposition of words or letters for that purpose40. This modification natur- ally diminished the value of all discussion and elaborate

39 A commentator on Vamana, named Sahadeva, tells us that Vämana's work went out of vogue, and its tradition was restored by Bhatta Mukula (!) who obtained a copy of the work. See notes to the Kāvya-mīmāņsā, ed. Gaekwad Oriental Series, p. 5. 40 varņa-samghatanā-dharmatva, Ānandavardhana p. 5; see also ii. 8-11.

104 SANSKRIT POETICS

classification of the Rītis into different types, and the function of the three Rītis of Vamana was made practically equiva- lent to the three Gunas admitted by the authors of the Dhvanyāloka ; but they do not yet appear to have lost all interest with later theorists. Even those writers, who do not subscribe to the doctrine either of the Riti school or of the Dhvani school, pay considerable attention to this question. Thus, Rudrata adds lāti to the enumeration of the three Rītis of Vamana, although by Rīti Rudrata means a definite usage of compound words. The Agni-purana accepts this fourfold classification, but the distinction is supposed to lie not only in the length or shortness of the sentences but also in the qualities of 'softness' or 'smoothness', as well as in the pro- minence of metaphorical expression (upacara). Bhoja, who carries the elaboration still further, adds two more types of Rīti to the Agni-purāņa's four, viz. māgadhī and āvantikā, the former being an intermediate diction between Vaidarbhi and Pāñcālī, and the latter forming only a khanda-rīti, i. e. a de- fective or incomplete type. Rājasekhara in his Kāvya-mī- māmsā gives us the same three Rītis as Vamana does, but in his Karpūra-mañjarī he appears to speak of three Rītis, respectively named vacchomī (from Vatsagulma41), māahī (māgadhī) and pamcāliā (pāñcālī). The older Vāgbhața men- tions only two Rītis, viz., pāñcālī and lāțīyā, the one having some compound words and the other having none ; but the younger Vagbhata accepts the three Rītis of Vamana and cl assifies them on the basis of the three excellences, mādhurya, ojas and prasada, which were the only three Gunas recognised since Mammata's time. The Dhvanikāra does not discuss. this question, but Anandavardhana appears to assign equal functional value to the three vrttis of Udbhata and the three rītis of Vamana, a view which Mammata enunciates thus: etās tisro vṛttayo vāmanādīnām mate vaidarbhī-gaudī-pañ- cālyākhyā rītayo matāḥ.42 41 In Vidarbha; so this is really Vaidarbhī. 42 But a distinction has always been theoretically maintained

DAŅDIN AND VĀMANA 105

Abhinavagupta, however, thinks (p. 6) that the three Rītis of Vāmana, which, in his opinion, characterise an elevated, soft and middling theme respectively through a peculiar combination of the Gunas, have, together with Udbhața's Vrttis, no separate existence from Guņas and Alamkāras. Attention, therefore, was naturally directed to the Gunas and Alamkāras more than to the Rīti itself, of which they formed, in Vamana's opinion, the constituent elements ; and the Rīti, if recognised at all, was recognised as consisting in a particular disposition of words, letters or syllables which favours the development of Rasa, and stands in the same relation to it as (in the usual figurative conceit) the conformation of the 'body' to the 'soul' (pada-samghatana

between the vrtti and the riti. The Vrttis, originally styles of dramatic composition (Bharata iii. 25), have been included by Udbhata (i. 4 f) under anuprāsa or alliteration, as they are formed by a special arrange- ment of letters for conveying different ideas, suitable (Abhinavagupta adds, °Locana pp. 5-6) to different Rasas. Ruyyaka, therefore, says: vṛttis tu rasa-vişayo vyāpārah, tadvatī punar varņa-racaneha vrttih, pp. 20-21. The Rīti, on the other hand, is mostly a matter of objective adjustment of the different 'excellences' of a composition, although admitting artha as an element of consideration ; while the Vrtti con- cerns itself with the psychological effect produced by the arrangement, as well as by the sense of which that arrangement is a vehicle. One and the same Riti may conceivably produce different Vrttis, and the same Vrtti may be produced in different Rītis, although a fondness. for symmetry led the theorists to assign a distinct Vrtti to each Rīti. Änandavardhana expressly distinguishes between the dramatic Vrtti and the poetical Vrtti by saying that the former is dependent upon the sense, the latter upon the expression : vācyāśrayo yo vyavahāras tā etāḥ kauśikādyā vrttayaḥ, vācakāśrayaś copanāgarikādyāḥ. They enhance the beauty of drama and poetry respectively by being used according to the drift of the Rasa in the composition: vrttayo hi rasādi-tātparyeņa samnivistah kām api nātyasya kāvyasya ca chāyām āvahanti p. 182. Rājaśekhara (Kav. mīm. p. 9) sums up the differences wittily and concisely, if not accurately: tatra veśa-vinyāsa-kramah pravrttih, vilāsa-vinyāsa-kramo vrttih, vacana-vinyāsa-kramo rītiḥ. On the concept of Vrtti generally see V. Raghavan, Some Concepts p. 182-93 ; Śrngāra-pr. pp. 196-215.

106 SANSKRIT POETICS

rītir anga-samsthā-višeşavat/ upakartrī rasādīnām, Viśvanātha ix. 1). It follows from this that the respective functions assigned to Guņa and Alamkāra are not in relation to the Rīti but to Rasa, which is one of the fundamental elements of poetry with the later schools. We have already noted that, after Anandavardhana, the Gunas are taken as inseparable attributes and causes of excellence (angino rasasya utkarsa- hetavah acala-sthitayo guņāh, Mammața viii. 1) of the Rasa or the principal poetic mood in the composition (and not, as Vāmana thinks, of the Rīti). The poetic figures, on the other hand, are only attributes of sabda and artha43, which constitute the 'body' of poetry, and therefore heighten the poetic mood or Rasa in an indirect way (ibid, viii. 2). Vāma- na's ideas about 'poetic charm' are also taken as axiomatic, but they appear in later theories in a somewhat different form as the vicchitti, or vaicitrya, or kavi-praudhokti underlying all figurative expression. In spite of these and other important contributions to the general theory of Poetics, it is obvious that the fundamental doctrine of the Riti school could not have been accepted in its entirety. Nor could it have competed against that of the Dhvanikāra, because Vāmana comprehended poetry only from the formal point of view, whereas the former showed a deeper insight into its inner nature. The more or less objective definition of the Rīti, given by this school, was hardly enough to satisfy the search for ultimate principles. Viśvanatha, following the Dhvanikara and Anandavardhana in this respect, states this objection when he says that the Rīti is a particular kind of formal arrangement, a peculiar disposition or posture of parts; what is called the 'soul' or

43 Mammata, we shall see, takes the Guna as directly related to the Rasa as the angin, and if we sometimes speak of them as belonging to śabda and artha, we do so only figuratively (upacāreņa) ; but Jagan- nātha (pp. 33-35), going back to Vamana's old position, combats this view and thinks that this usage is not figurative.

DAŅDIN AND VĀMANA 107

essence of poetry is something quite different44. Again, the analysis of the several types of diction shows considerable ingenuity indeed, but it was found almost impossible, as Dandin himself admitted very early (i. 101-2), to label and classify all the modes of poetic expression with definite and unalterable characteristics. As the Rīti school, therefore, tended to make invidious and essentially unprofitable (except as empirical facts) distinctions between the Vaidarbhī, Gaudī and other kinds of diction with regard to whose exact significance there was bound to be inevitable difference of opinion, it naturally provoked criticism and opposition. In the same way, the endeavour to exhaust and classify all the literary excellences and flaws within clear-cut bounds on the basis of more or less formal analysis, was sure to prove unconvincing ; and a protest against minute differentia- tion or endless multiplication of the Gunas was rightly and definitely propounded by Mammata who (following Ānandavardhana, pp. 79 f) reduced the number to three only, viz. mādhurya, ojas and prasāda, in relation to the ultimate factor of the poetic sentiment. The attempt, therefore, to stereotype the entire poetical output into so many ready-made dictions and fixed excellences, was bound ultimately to be discarded in favour of other and more penetrating principles. We shall see in the next few chapters that such a principle in poetry came to be recognised in the suggested element of Rasa, which was being already elaborated in the service of the drama and which was utilised by the Dhvani-theorists as one of the most important aesthetic foundations of poetry.

44 yat tu vāmanenoktam-rītir ātmā kāvyasya iti, tan na, rīteḥ samghatanā-viśeşatvāt, samghațanāyāś cāvayava-samsthāna-rūpatvāt, ātmanaś ca tad-bhinnatvāt, p. 18 ed. Durgaprasada, 1915. (It will be noticed that Viśvanātha here speaks of Rīti as a kind of samghatanā).

CHAPTER IV

LOLLAȚA AND OTHERS

(The Rasa System)

( 1 ) While the orthodox schools of Poetics were elaborating systems of Alamkāra and Rīti, there flourished several writers who discussed the question of the dramatic Rasa after Bharata, and formulated explanations of the latter's much discussed sutra on the subject1. Their exposition, however, concerned the dramatic art, and their theories did not as yet come properly within the sphere of Poetics, which was. entirely dominated by the Alamkāra- and Rīti-systems. The aesthetic importance of the Rasa, therefore, was never realised, as we shall see, until it was taken up and worked into Poetics by the Dhvanikara and his followers. The dramaturgic Rasa school, however, had in tbe meantime elaborated several theories of Rasa and brought it into prominence as an element of the drama; and in this way it succeeded, to a certain extent, in reacting upon and influencing the orthodox theories of Poetics, which appear to have possessed, even at an early period a limited acquaintance with Rasa, and which actually accord it a place, however small, in their general systems of Alamkāra or Rīti. This will be obvious from a reference to the views of Bhämaha and Dandin on this subject. To Bhämaha, the most important element in poetry is alamkāra or vakrokti. He does not seem to possess any clear notion of the function of Rasa in poetry, the only direct reference to it occurring in the definition of the figure rasavat which, in his opinion, must manifest the Rasas clearly (rasavad darśita-spasta-śrngārādi- 1 See above p. 20. The Sūtra runs thus: vibhāvānubhāva-vya- vhicāri-samyogād rasa-nişpattiḥ.

BHĀMAHA, UDBHAȚA AND RUDRAȚA 109

rasam, iii. 6). The Rasa is thus included in the scope of a particular figure only and given a very subordinate place in his system2. Bhämaha seems to have been aware of the existence of the srngara and other dramatic Rasas; but the speculations regarding the origin and function of Rasa do not appear to have started in his time; and, in common with Dandin, he never thought it necessary to use the technical terms vibhāva, anubhāva etc., so familiar to later writers on this subject. In Bhämaha's opinion, the Rasa need not be invariably present in poetry, but it may sometimes be delinea- ted in some poetic figure. In i. 21, no doubt, Bhämaha lays down that a Mahākāvya must separately depict all the Rasas ; and in v. 3 he speaks of the kāvya-rasa as mitigating the rigour of the Sästras, a sentiment which is endorsed by Ru- drața (xii. 1-2) and which probably inspired the dictum of Abhinavagupta that the Sāstra is prabhu-sammita, while the Kāvya is jāyā-sammita. The term kāvya-rasa is used here probably to indicate 'the flavour of poetry' in an untechnical sense ; but even if we read, with Abhinavagupta (°Locana p. 182), a technical meaning into it, it only shews that the ear- lier authors were content with assigning a pleasing but ex- traneous function to Rasa in poetry,3 although all the Rasas may be developed in the drama (i. 21).

2 Commenting of Bhämaha's central verse on vakrokti (ii. 85), saişā sarvaiva vakroktir anayārtho vibhāvyate, Abhinavagupta attempts to read into it his own idea of the importance of Rasa and interprets vibhāvyate technically as pramadodyānair vibhāvatām nīyate, viśeşeņa ca bhāvyate, rasamayīkriyata iti. He apparently makes Bhāmaha mean that by vakrokti, the sense of poetry is rendered into a suitable factor of Rasa, so that by using the word vibhavyate with the technical mean- ing given to it Bhämaha would imply that the Rasa as well as Alamkāra originates in vakrokti. 3 Such a distinction appears to be implied also in Dandin ii. 292; and one can differentiate two meanings of rasa (see below): (1) kāvya- rasa, the flavour of poetry or the aesthetic delight produced by it, and (2) rasa in the technical sense of natya-rasa. It is worthy of note that in the latter sense it is very rarely used in earlier classical poetry (except

110 SANSKRIT POETICS

The same remarks with regard to the recognition of Rasa apply more or less to Dandin; but Dandin seems to have been more alive to its importance than Bhamaha. Like Bhāmaha, Dandin allows the Rasas to be included in figures like rasavat, which appear to have been the only means by which they could permit the Rasa to play any rôle in their systems. It may be contended4 that Dandin gives prominence to Rasa by including it in one of the essential excellences (Gunas) of the diction (Rīti), viz., in mādhurya, which is defined as the establishment of Rasa in the word and in the object (vāci vastuny api rasa-sthitiḥ, i. 51) ; but from ii. 292 it appears that Dandin means by the term Rasa in the mādhurya- guņa to connote absence of vulgarity (agrāmyatva)5, and does not contemplate the inclusion of Rasa in the technical sense6. The mädhurya-guna, according to Dandin (i.51-7), may appear in two different aspects, in so far as it creates vāg-rasa and vastu-rasa, the former consisting of repetition of similar sounds (śrutyanuprasa)7 and the latter connoting absence of

perhaps in Māgha). This kāvya-rasa in its essence is not very different from Bhämaha's vakrokti ; for it is possibly a kind of heightened expres- sion inconsistent with commonplace utterance. The word rasa must be taken in this untechnical sense when one meets with it in early. classics and in such expressions as rasavat or rasāvaha in Dandin. 4 Jacobi in ZDMG lvi, 1902, p. 401 fn. 5 Strictly speaking, gramya is not 'vulgar' in its usual restricted sense, although Dandin seems to bring under its connotation the aślila or the indecent. It means 'the low', 'the rustic', or 'the common', as opposed to the noble or the refined, and probably corresponds to the French word vulgaire. 6 This is made clear by the Hrdayangama commentary on this point: mādhurya-guņe pradarsitah sabdārthayor a-grāmyatayā jato raso vākyasya bhavati, alamkāratayā nirdistam rasavattvam asta- rasayattam (p. 167), the last part of the passage calling attention to the fact that the only cases where the eight (dramatic) rasas are admitted by Dandin occur in connexion with his inclusion of the rasas in figures like rasavat. In i. 64 Daņdin speaks of agrāmya artha as rasāvaha. 7 See above p. 80, fn 11. Dandin speaks of anuprāsa as rasāvaha in this sense in the same context (i. 52).

LOLLAȚA AND OTHERS 111

vulgarity (agrāmyatva). Thus, Hemacandra explains (p. 198) the Rasa in Dandin's madhurya, according as it resides in the vāk or vastu in this way: śruti-varņānuprāsābhyām vāg- rasaḥ ........ agrāmyābhidheyatayā tu vastu-rasaḥ8. Thus, the Rasa in Dandin's madhurya has a distinct connotation which separates it from the technical dramatic Rasa of the Rasa school9. At the same time it cannot be affirmed that Dandin was entirely ignorant of the concept of Rasa as elaborated by Bharata and his followers. He declares (i. 18) that a Maha- kāvya should invariably depict rasa and bhāva; so does Bhāmaha (i. 21). His treatment of the figures rasavat, preyas and ūrjasvin (ii. 280-87) gives a much clearer indication of his undoubted acquaintance with the eight recognised Rasas, all of which he enumerates by their respective names, and four of which (viz. śrngāra, raudra, vīra and karuņa) he illustrates as elements of the poetic figures under dis- cussion. If we are to accept Abhinavagupt's statement10,

8 With this explanation Māņikyacandra agrees: śruti-varnānuprā- sābhyām vāg-rasah, agrāmyatayā tu vastu-rasah, ittham raso dvedhā (p. 189, ed. Ānandāśrama). 9 A similar untechnical use of the term rasa (which, however, Dandin does not explain, as he does in this case) is to be found in iii. 149 (or iv. 26, Madras ed.) where the phrase giram rasah (=vāgrasah) is interpreted by Tarunavācaspati merely as sādhutvam. The modern commentators, misled, no doubt, by their own idea of the importance of Rasa, read into Dandin their own ideas on the subject. Dandin no- where speaks of the suggestion of Rasa (rasa-dhvani) as the 'soul' of poetry, but Premacandra, commenting on i. 10, reads this into Dandin. The artha-rasa in i. 62 seems also to have a distinct reference to agrāmyatā, 10 On Bharata vi, partially reproduced by Hemacandra p. 57f. The text of Abhinava's valuable commentary on Bharata's rasa-sūtra, so far as it is relevant to the theories on Rasa discussed below, has been published as an appendix to S. K. De, Theory of Rasa in Sir Asutosh Mookerjee Silver Jubilee Commemoration Volume (Orientalia, vol. iii 1922). It is cited below as "Abh. on Bh.," with references to the pages of the article, which has been largely utilised in this chapter. It is reprinted in S. K. De, Some Problems, pp. 219-35.

112 SANSKRIT POETICS

Dandin's conception of Rasa is similar to that of Bhatta Lollata (which we shall deal with presently), and he believes therefore in the development of Rasa as an effect from the vibhāvas and anubhāvas. Dandin's some- what meagre treatment hardly justifies us in making a definite assumption on this point, but we may assume that Daņdin apparently speaks of Rasas like śrngāra or raudra being developed as effects from such permanent moods as rati or krodha, For, speaking of the figure rasavat which should possess the characteristic of manifesting the Rasas, he gives an example of the manifestation of śrngāra in such a figure with the remark: ratih śrngāratām gatā / rūpa- bāhulya-yogena (ii. 281). Similarly, with reference to the development of raudra from krodha, he says: ity āruhya parām koțim krodho raudrātmatām gataḥ (ii. 283). But the Rasa in these figures is subordinate to the expressed figure itself of which it serves as a means of embellishment (alamkāratayā smrtam) ; in other words, the Rasa is deve- loped not for its own sake but as increasing the beauty of expression. It would seem, therefore, that Dandin was, to some extent, cognisant of rasa and bhäva, but be could not give it a place in his system except as an embellishment of the language or of the sense; and this objective view of the function of Rasa, if we may presume it in these early authors, was apparently responsible for this subordi- nate position given to it by the Alamkāra and Rīti-systems. Although Vamana improves upon Dandin's system in other respects, he does not seem to have gone further in the treatment of Rasa. His idea of making all poetic figures an aspect of metaphorical expression precludes him from defining the figure rasavat, but he attempts to include Rasa in one of the essential (nitya) character- istics of poetry, viz., in the artha-guna kānti, which he defines as an 'excellence of sense' in which the Rasas should be conspicuously present (dipta-rasatvam kāntih, iii, 2. 15). The suggestion for this inclusion of the Rasas

LOLLAȚA AND OTHERS 113

in one of the excellences was probably found in Bharata's definition of the kanti-guna, or more directly in Bharata's peculiar definition of the udara-guna; but it is clear that this certainly marks an advance upon the treatment of Dandin and Bhämaha, who include the Rasa in some of the non-essential figures. Udbhata adheres in the main to the views of Bhamaha and deals with Rasa as an element of the some of the figures like rasavat. One verse, however, which occurs in the text of Udbhata published by Jacob in JRAS, 1897, p. 847: rasādyadhisthitam kāvyam jīvad-rūpatayā yatah/ kathyate tad rasādīnām kāvyātmatvam vyavasthitam/ /. apparently designates Rasa as the essence or 'soul' of poetry, without, however, setting up an aesthetic system on its basis. But the verse is quite out of place in the context in which it occurs11, and in the text published by the Nirnay Sagar Press, it is wanting, although given as a quotation (with a tad ähuh) in the accompanying commentary of Pratīhārendurāja (p. 77). Misled by Jacob's text, Jacobi supposes12 that Udbhata was the first writer to consider the question as to what constitutes the 'soul' of poetry and to regard Rasa as such. It is clear, however, that the verse in question is not Udbhata's, and cannot be reconciled either with its immediate context or with Udbhata's general standpoint, as well as with his definition of rasavat. It cannot be denied at the same time that Udbhata betrays an acquaintance with some theory of

11 This verse (vi. 17) occurs after the figure kāvya-linga (vi. 16), after the definition of which one should expect its illustration, which is given in the verse vi. 18 next after the verse in question. If Jacob's text is accepted, then the verse rasādyadhisthitam (vi. 17) would be abruptly thrust in between the definition of kāvya-linga and its illustration. As a matter of fact, it occurs in Pratīhārendurāja's com- mentary on kāvya-linga and is erroneously incorporated in Udbhata's text by the editor. 12 ZDMG, 1902, p. 396. 31

114 SANSKRIT POETICS

Rasa and its technicalities using, as he does, terms like vibhāva. sthāyin, sañcārin (iv. 4) and anubhāva (iv. 2), and enumerat- ing, after Bharata, the eight orthodox nātya-rasas with the addition of a ninth Rasa (viz. santa) in the cate- gory. But Udbhata takes all this into account as an embellishment of an expressed figure like rasavat13 ; the Rasa is not considered on its own account, but because it helps to emphasise or constitute the charm of a particular figure. Hence Pratīhārenduraja remarks that the question as to the nature of Rasa and Bhäva, and as to how far they may stand as the very 'soul' of poetry is not discussed by Udbhata at all, partly for fear of prolixity and partly because it is irrelevant14. Rudrata, on the other hand, seems to be the earliest writer who explicitly includes Rasa in his treatment of Poetics, and devotes four chapters to its discussion. At the beginning of his work, he praises the poets who have won eternal fame by composing Kävyas enlivened by Rasa. In ch. xii he speaks of ten Rasas (adding preyas and śānta to the orthodox eight of Bharata)15, describing śrngāra (love) and the charac-

13 Udbhata's definitions of rasavat etc. differ from those of Bhamaha and Dandin; and he admits a new figure, namely, samāhita. In his view, preyas consists of development of emotions like love to the extent of a Bhava and not Rasa; rasavat occurs when Rasas are fully developed through sva-sabda, sthāyin, sañcārin, vibhāva and. anubhāva; ūrjasvi is development of improper or incongruous (an- aucitya-pravrtta) Rasa or Bhava, which would correspond to Rasa- bhasa of later writers; samahita=calming down of Rasa, Bhava or their Abhasa. 14 Udbhata might have been one of the commentators on Bharata, and was probably conversant with Bharata's text, as his citation of a half-line (iv. 5, though it is doubtful if this is at all a kārikā-verse of Udbhata's) from Bharata vi. 15 and use of technical terms like vibhāva etc. would indicate. But it does not prove that Udbhata belonged to the school of Bharata. On the other hand, Udbhata in his theoretical standpoint was undoubtedly a follower of Bhamaha. 15 If Udbhata iv. 5 is a kārikā-verse (and not wrongly incorporated into the text from Pratīharenduraja's commentary), then Udbhata was

LOLLAȚA AND OTHERS 115

teristics of the hero and heroine in that connexion. The next two chapters take up the two kinds of śrngāra (sambhoga and vipralambha-love in union and in separation) and questions cognate to each. This treatment is rounded off by a short chapter describing the nature and character of the diction (riti) suitable to each16. It is not clear, however, as to what significance Rudrata attaches to Rasa as an element of poetry, for he is entirely silent with regard to the theoreti- cal aspect of the question. Out of the sixteen chapters into which his work is divided, only four chapters deal with Rasa not theoretically but descriptively, while the rest of his work is taken up with the details of the poetic figures on which obviously he puts greater emphasis. Speaking of the necessity of making a poem sarasa from the standpoint of the reader, he says (xii. 1) that to those, who enjoy Rasa but fight shy of Sastra, instruction in the caturvarga is easier to impart through the medium of delectable writing; and this is the chief motive, in his opinion, for inspiring the sense of poetry with Rasa. Rudrata starts with sabda and artha as the two constituents of poetry, and elaborates his views about poetic figures as embellishment of these elements; but he does not discuss how the Rasa comes into his system, a fact which may lead one to suspect that these chapters on Rasa were probably later engraftment extraneous

the first writer to admit santa into Bharata's category of eight Rasas. The preyas Rasa of Rudrata is probably suggested by the poetic figure preyas admitted by Bhāmaha, Dandin and Udbhata. Rudrata appears to allude (as Nami-sadhu rightly comments) to Bharata by the term ācārya in xii. 4. 16 Rudrața defines Rīti, as we have already seen, with reference to the employment of compound words. He recommends the Vaidarbhi and Pañcalī Rītis in the cases of the four Rasas, preyas, karuna, bhayānaka and adbhuta, and the Lāțīyā and Gaudīyā in the case of raudra, there being apparently no fixed rule with regard to the remain- ing Rasas. He uses the term aucitya in this context, which anticipates the theory of Aucitya first elaborated by Anandavardhana in connexion. with the delineation of Rasa.

116 SANSKRIT POETICS

to, if not inconsistent with, his general standpoint. Rudrata, in his theoretical tendencies, has no affinity with the Rasa school, but belongs to the Alamkāra school, a fact which would distinguish him from Rudrabhatta, the keynote of whose system is to be found in the idea of Rasa. We have to this effect the testimony of Ruyyaka and Jayaratha. Ruyyaka says17 that Rudrata laid special stress on alamkāra, in which were comprised the three kinds of suggestion (dhvani), in- cluding the suggestion of Rasa, and that in figures like rasavat, the rasa and bhava implied are taken as elements which heighten the charm of the expressed idea18 The older writers on Poetics, therefore, before the advent of the Dhvani-theorists, content themselves with the working out of the outward form of expression, the 'body' of poetry, and hardly trouble themselves with the question of an ulterior aesthetic principle, the 'soul' of poetry ; nor do they identify, as some later writers do, this 'soul' with the psychological factor known as Rasa. Vamana, no doubt, starts the question and offers to solve it by declaring that the diction or Rīti is this 'soul' ; but we have seen that in Vämana's view, the Riti is not the expression of poetic individuality but the objective beauty of representation called forth by a definite adjust- ment of certain fixed literary excellences. The older writers, therefore, put the greatest emphasis on the alamkāra (or poetic figure), or on the riti (or diction in the objective sense), the advantages of which were considered sufficient for poetry;

17 ed. Kāvyamālā p. 5. Samudrabandha agrees with this view. 18 Nami-sadhu, explaining Rudrata xii. 2, states that in the opinion of his author, sabda and artha constitute the 'body' of poetry, the poetic figures take the place of artificial ornaments, while Rasa resembles natural qualities like prowess and beauty (rasās tu saundaryādaya iva sahajā gunāh). But there is nothing in Rudrata which will support this description of his standpoint, especially as Rudrata can scarcely be supposed to look upon poetic figures, which are of great importance in his view of poetry, as mere artificial embellishments of poetry. See above pp. 59-60, 61.

LOLLAȚA AND OTHERS 117

and, cognisant as they were of that aesthetic delectableness which must be present in all poetry and which in Sanskrit goes by the name of rasa, they could not harmonise it well with their theory of externals and treated it more or less as an embellishment of the language by including it in poetic figures or by allowing it to form an element of one of the excellences of diction. This was the only way in which they could recognise Rasa. It is partly for this reason that the Dhvanikāra (iii. 52) condemns earlier theories as crude and insufficient for the purpose of explaining the nature of poetry, and expounds his own system in which the suggestion of Rasa (rasa-dhvani) plays such an important part.

(2) The reason why Rasa was, even thus perfunctorily, admit- ted into the older systems appears to have been the fact that Bharata's treatment of Rasa in the drama had already esta- blished itself, having been further elaborated by a number of commentators and writers on the subject, and it naturally influenced, to a limited extent, the enquiry of early thinkers. We have already noted that Bharata's famous sūtra on Rasa19 by its ambiguity taxed the ingenuity of his followers and led to a great deal of controversy regarding its true interpreta- tion ; and as each writer tried to explain it in his own way, it gave rise to a number of theories on Rasa. There are four such theories associated with the names of Lollata, Śankuka Bhațta Nāyaka and Abhinavagupta; but Jagannātha (Rasa- gangādhara p. 28) speaks of eight different interpretations. Bhatta Lollata appears to have been one of the earliest formulators of such an explanation. His work is unfortuna- tely lost, and very little can be gathered from the brief review

19 Viz. vibhāvānubhāva-vyabhicāri-samyogād rasa-nispattih (see above p. 20), the different theories starting with the different explana- tions given of the terms samyoga and nispatti. Besides the vagueness of these two terms, it is noteworthy that term sthāyin or sthāyi-bhāva does not occur in this dictum.

118 SANSKRIT POETICS

of his opinion in Abhinavagupta's commentary on Bharata20, which is copied more or less by all subsequent writers who deal with Lollata's views21. But it is clear even from this summary exposition by an adverse critic that Lollata, in explaining Bharata's sūtra, took the vibhāva as the direct cause (kāraņa) of Rasa, which therefore is an effect (anukārya or utpādya), and the term nispatti of Bharata should be explained as utpatti or pusti. The Rasa, found in characters like Rāma, is attributed to the actor, who imitates the characters in form, dress and actien, and thereby charms the spectator. Mammata and his followers make this inter- pretation of Lollata's view more clear by saying that the per- manent mood or sthāyin is directly connected (mukhyataya vrttyā=sākşāt sambandhena) with the hero like Rāma, but it is recognised as existing in the actor through a clever imitation of the original character, this imitation being apparently the source of the charm to the spectator. The Rasa, therefore, re- sides in the hero; but the objection is that it is not clear how a mental state which belongs to the hero can be transferred to the acter, and how the spectator can be charmed by a feeling which does not exist in him. The spectator's mere apprehen- sion of the feeling imitated by the actor cannot produce even a semblance of the original feeling and consequent delight in the mind of the spectator ; otherwise such a delight would be brought about even on witnessing a love-affair in the ordi- nary world, as distinguished from the world of poetry22. It is

20 Abhinava's review of the opinions of Lollata, Śankuka and Bhatta Niyaka, which is followed by Mammata and all later writers, is extensively reproduced by Hemacandra (pp. 57-66) and Māņikya- candra (pp. 40f, ed. Anandasrama). This portion of Abhinava's commentary will be found in the article already mentioned above on the Theory of Rasa (reprinted in S. K. De, Some Problems of Sanskrit Poetics, Calcutta 1959). Detailed references, therefore, are not as a rule given here. 21 See for instance, Mammata ch. iv, Hemacandra p. 57, Mallinātha on Vidyādhara p. 85, Govinda on Mammața p. 63 etc. 22 Govinda criticises Lollata's view thus: tad apesalam, sāmājikeşu

LOLLAȚA AND OTHERS 119

also argued that this cause-and-effect theory of Lollata can- not satisfactorily explain the relation of the vibhāvas etc. to the Rasa. An effect may exist even when its efficient cause is destroyed ; but as the life of the Rasa is circumscribed by the exhibition of the vibhavas, it disappears when the latter disappear, a fact which goes to prove that the Rasa must not be taken as an ordinary laukika effect23. Again, the cause and the effect cannot be contemporaneous ; if the Rasa is supposed to be an effect, its relish cannot be, as it actually is, contem- poraneous with the appearance of the vibhavas. Hence Vis- vanātha remarks (p. 86) that if the Rasa is an effect, having for its cause the perception of the vibhāvas, then at the time of the relish of Rasa the vibhavas would not be perceived ; for we do not find the simultaneous perception of a cause and its effect. The perception of the touch of the sandalwood unguent and the perception of the pleasure pro- duced thereby cannot take place simultaneously, however rapidly the one may succeed the other. Śankuka, the next important writer on this subject mentioned by Abhinavagupta and others24, therefore rejects this interpretation of the utpatti-vādins, who are said to follow, in their peculiar theory, the Mimamsa school of philosophers. Lollata's view does not deal with Rasa as a matter of the spectator's feeling. Sankuka, on the other hand, thinks that the Rasa is not produced as an effect, but inferred by the spectator, and the inferred feeling is relished by him as Rasa. The permanent mood of the hero is inferred to exist in the actor (though not actually existing in him) by means of the vibhavas etc., cleverly exhibited by him in acting, so as to produce an illusion of identity with the

tadabhāve tatra camatkārānubhava-virodhāt, na ca tajjnānam eva camatkāra-hetuḥ, laukika-śrngārādi-daršanenāpi camatkāra-prasangāt (ed. Kāvyamālā 1912, p. 63). 23 Govinda p. 69, Mallinātha pp. 87, 93-4. 24 The reference to Śankuka's views is to be found in Mammata and others, as cited above in fn 21.

120 SANSKRIT POETICS

feelings of the hero25; and the mood thus inferred, being sensed by the spectator through its exquisite beauty, adds to itself a peculiar charm26, and thus develops into a relishable condition of his own mind which is called Rasa. The realisa- tion of Rasa, therefore, is a process of logical inference, and the nispatti of Bharata's sūtra is explained as anumiti, the vibhavas standing to Rasa in the relation of anumāpaka, or gamaka to anumāpya or gamya. But the mood itself, though inferred in this way from the relation of logical major and middle terms, is yet cognised as different from the objects. of ordinary inference, being inferred, as it were, by force of its connexion with the vibhavas, which factors, though artificial in themselves, are not then recognised as such. This cognition or knowledge is characterised as being based on what is called citra-turaga-nyaya (or the analogy by which a horse in a picture is called a horse), and should be differentia- ted from the true ('he is Räma'), the false (he is Räma' with a following negation 'he is not Rama'), the doubtful ('he may or may not be Rāma') knowledge, as well as from the knowledge of similarity ('he is like Rama')27. The theory, however, has been discredited by later schools on the ground (as Govinda concisely puts it28) that it disregards the well- recognised fact that the inference of a thing can never produce the same charm as direct cognition. It has also been pointed

25 rāmādyabheda-bhāvitena națe ·tat-prakāitair eva vibhāvādibhir- anumitah, Mallinatha p. 85. 26 vastu-saundarya-balād rasanīyatvena sthāyinām anyānumeya- vailakşanyāt Govinda p. 65, practically paraphrasing Mammața. 27 As interpreted by Mammata as well as Abhinava p. 241. Hema- candra expanding the exposition of Abhinavapupta and Mammata puts it in this way: na cātra nartaka eva sukhīti pratipattih, nāpyayam eva rāma iti, na cāpyayam na sukhīti, nāpi rāmah syād vā na vāyam iti, na cāpi tat-sadrsam iti, kim tu samyan-mithyā-samsaya-sādrśya-pratītibhyo vilakşaņā citra-turagādi-nyāyena yah sukhī rāma asāvayam iti pratītir asti (p. 59). 28 pratyakşam eva jnānam. sa-camatkāram nānumityādir iti loka- prasiddhim avadhūyānyathā-kalpane mānābhāvaḥ, p. 65.

LOLLAȚA AND OTHERS 121

out that the Rasa is not capable of being cognised by the ordinary means of arriving at knowledge, for the feeling of a hero like Rāma, being past, cannot be cognised directly by the organs of sense belonging to the present.29 The anumana-theory is criticised elaborately in connexion with the theory of 'suggestion', coming topically within the province of the suggestion of Rasa (rasa-dhvani). We shall have occasion to deal with this aspect of the theory in its proper place; but the general argument with which it is sought to be discarded is that the vibhāvas cannot be taken as the middle term in proving the sthäyin, because the former do not stand in the same relation to the latter as the middle term (sādhana) does to the major term (sādhya), but are simply its suggestors (vyañjaka). The vibhāvas, therefore, do not constitute either the efficient cause (kāraka-hetu) or the logical cause (jñāpaka- hetu) of Rasa, as held respectively by Lollata and Śańkuka, Apart from technicalities, Lollata's view appears to be that the spectator ascribes to the well-trained actor the same mental state as belonged to the hero, and his apprehension of this imparted feeling produces a similar feeling in his mind, causing delight. Sankuka thinks that the well-trained actor so cleverly simulates the action of the hero that the spectator apprehends the actor to be identical with the hero, and infers from this illusion the actual feeling of the hero in his own mind, being moved by the extraordinary beauty of the represented action. In both these theories, however, the difficulty remains, viz. that if Rasa is an objective entity, produced or inferred, how can it bring about a subjective feeling of relish in the audience in whom these factors (vibhāvas etc.) are presumably absent ? If, on the other hand, it is supposed that the Rasa exists in the audience also, the question still remains as to how the particular feeling of a particular hero (like Räma, who is different from or superior

29 Vidyādhara p. 94.

122 SANSKRIT POETICS

to the spectator himself) can be relished or realised as his own by the spectator ?. These objections are thus ably set forth by Bhatta Nayaka30, as interpreted by Abhinavagupta in his °Locana (pp. 67-8): "If the rasa is perceived as belonging to another person, then it is a case of tātasthya i. e. one would not himself be personally affected by it. It is also not perceived as belonging to oneself out of poetry dealing with the deeds of heroes like Rāma. If it is perceived as belonging to oneself, then origin of rasa in self is admitted. But this is not reasonable, for there is nothing there which can operate as a vibhava for the audi- ence. If it is objected that the generalised idea of the beloved (kāntātva), which lies dormant and awakens in us germs of latent impressions, operates in the capacity of a vibhāva, then how can it be applied to the description of a deity and the like ? The recollection of one's own beloved does not intervene in one's consciousness. How can vibhāvas, like the construction of a bridge over the sea, which form the attributes of an extraordinary hero like Rāma, become generalised (in the mind of an individual spectator) ? One does not recollect only Rāma's energy, because it has no resemblance (to one's own energy). The rasa is not perceived when one learns it from a verbal composition, because a man would learn it in the same way from the direct observation of a pair of lovers. If it is assumed that rasa is produced, then a man would feel disinclined to tragedy, inasmuch as he finds only pain following upon the production of the pathetic mood (karuņa) "31 30 Another objector to Śankuka's view appears to be Bhatta Tauta, whose opinions are summarised by Abhinava (who refers to him simply as asmad-upādhyāya) in a passage which is substantially reproduced by Hemacandra at p. 59 under Bhatta Tauta's name. See also Māņikya- candra p. 43 who draws also upon Abhinava's exposition of Tauta's view. 31 Raso yadì para-gatatayā pratīyate, tarhi tāțasthyam eva syāt. Na ca sva-gatatvena rāmādi-caritamayāt kāvyād asau pratīyate. Svātma-

LOLLAȚA AND OTHERS 123

Bhatta Nāyaka, therefore, attempts to refute these earlier theories and set up a peculiar doctrire of aesthetic enjoy- ment (bhoga) which makes the relish of Rasa possible. He argues (as interpreted by Mammata and others) that (i) Rasa cannot be produced as an effect, because the causes (namely, the vibhavas), being non-realities, cannot bring about a real effect ; (ii) it cannot be inferred, because the real character (e. g. Räma), not being before the audience, his feeling does not exist, and what does not exist cannot be inferred (na tattvato rāmasya smrtih, anupalabdhatvāt). Nor is it a case of revelation (abhivyakti) of something potentially existing (śakti-rūpa); for in that case, the potential emotions, once awakened, would occupy their field of action in diverse degrees, thus contradicting the nature of Rasa as one. Moreover, there would be the same difficulty as to whether the Rasa is revealed in oneself or in another person. To solve these and other difficulties, Bhatta Nāyaka (as interpreted by Abhinavagupta) maintains that Rasa is enjoyed in connexion with the vibhavas through the relation of the enjoyer (bhojaka) and the enjoyed (bhojya). This school32 postulates three different functions of a word, namely, abhidhā (already admitted by the Mīmāmsakas and gram- marians), bhāvakatva and bhojakatva, and thus ascribes to a poem threefold potency of its own, namely, the powers

gatatvena ca prafītau svātmani rasasyotpattir evābhyupagatā syāt, sā cāyuktā, sāmājikam praty avibhāvatvāt. Kāntātvam sādhāraņam vāsanā-vikāsa-hetur vibhāvanāyām prayojakam cet, devatā-varņanādau tad api katham? Na ca sva-kāntā-smaraņam madhye samvedyate. Aloka - sāmānyānām ca ramadīnam ye samudra-setu-bandhādayo vibhāvās te katham sādhāranyam bhajeyuḥ? Na cotsāhādi-mātram smaryate, ananurūpatvāt. Sabdād api tat-pratipattau na rasopajanah, pratyakşād iva nāyaka-mithuna-pratipattau. Utpatti-pakşe ca karuņa- syotpādād duhkhitve karuņa-prekşāsu punar apravrttih syāt, tan na. 32 Bhatta Näyaka's views are set forth and criticised by Abh. on Bh. p. 244, °Locana p. 68, Mammata ch. iv, Hemacandra pp. 61 f., Govinda p. 66.

124 SANSKRIT POETICS

of denotation, of generalisation and of enjoyment.33 The abhidha is not merely the actual Denotation of a word, but is given an extended meaning so as to include lakşaņā or Indication in its scope (abhidhā laksaņaiva),34 thus embracing the two functions already analysed by previous speculation. lt is meant probably that the Denotation (as postulated by Bhatta Nāyaka) also gives to the expressed sense a metaphorical significance as the basis of Rasa. The bhāvakatva (or rasa-bhāvanā), which, as Abhinavagupta suggests, is apparently derived from Bharata's general definition of bhäva, is described as the power of generali- sation which makes the vibhavas as well as the sthāyi- bhava, sensed in their general character without any reference to their specific properties. The vibhāva, Sītā, for instance, is understood through this power not as a particular individual but in the general character of a woman, and the sthāyi-bhava (here Rama's love towards her) is taken as love in general without any reference to the agent or the object. In this way the audience can appropriate the vibhāvas, as well as the sthāyi-bhāva, as universal. After the Rasa is thus generalised, comes its enjoyment. By the third function of bhojakatva, the sthāyin is enjoyed in this general form, accompanied by the vibhāvas, sensed also in a general form; and this enjoy- ment is described as a process of delectation similar to the enlightened, self-sufficient and blissful knowledge, arising (in the language of the Samkhya philosophers, which is borrowed by these theorists) from the prominence 33 T. R. Cintamani would prefer the terms signification, idealisa- tion and illumination. 34 T. R. Chintamani (JOR i, 1927, p. 275 fn) would read abhidhā- vilakşanaiva. He thinks that having been a Mīmāmsaka, Bhatta Nāyaka would not include Laksaņā in Abhidhā. But K. P. Trivedi (Ekāvalī p. 425) reads as we do: abhidhā lakşaņaiva. V. Raghavan appears to agree (JOR vi, 1932, p. 211 fn) that Abhidha is "here used not in its restricted sense of Sakti but in its larger and more general meaning, viz., the poet's expression as a whole."

LOLLAȚA AND OTHERS 125

of the attribute of goodness (sattva) in a man, and different from what is known as wordly happiness, being divested of personal relations or interests. It is differentiated from the two kinds of knowledge, anubhava and smarana ; and con- sisting of the qualities of melting, pervading and expanding the mind, it is compared to the indescribable bliss of divine contemplation (brahmāsvāda-sacivah). According to Bhatta Nāyaka, therefore, Rasa consists in the sthāyi-bhāva or the permanent mood, experienced in a generalised form in poetry and drama through the powers of abhidhā and bhāvakatva, and enjoyed by a blissful process, known as bhoga, till it is raised to a state of pleasurable relish, which is not wordly (a-laukika) but disinterested and which is akin to the philo- sophic meditation of Brahma. It will be noticed that these different theories about Rasa, though applied to drama and poetry, are yet generally tinged with the doctrines of the various schools of Indian philosophy. Lollata, it is clear, is a Mīmämsaka who believes in the far- reaching function of the Denotation of a word and thinks that it is capable of expressing all other implied or suggested sense in the shape of the Rasa. Sankuka, on the other hand, is a Naiyāyika or logician who would demonstrate the Rasa by means of syllogistic reasoning. He believes that the im- plied Rasa can be reached by the logical process of inference from the expressed sense, although he has to admit that the inferred mood is cognised differently from the objects of ordinary inference, being sensed by the spectator through the force of its exquisite charm. In Bhatta Nāyaka we mark a further development. In his theory there is not only a transition from what may be called the objective to the sub- jective view of Rasa, and an understanding that the whole phenomenon should be explained in terms of the spectator's inward experience, but also the fact that Bhatta Nāyaka in his peculiar theory of aesthetic enjoyment (bhoga) is sub- stantially following the teachings of the Sāmkhya philo- sophers.

126 SANSKRIT POETICS

We need not enter here into the details of Sāmkhya psychology or metaphysics, but we may indicate briefly the application of its main teachings to the conception of poetry and the artistic delight resulting from it35. The purpose of evolution in Sämkhya is the attainment of bhoga (experience of pleasure and pain) and apavarga (spiritual emancipation through right knowledge). The enjoyer of Rasa in poetry is like the knower of Brahma, but the aesthetic attitude is. different from the philosophic. The aesthetic attitude is indeed one of samvit (or cit-svabhāvā), i.e., pure contemplation dis- sociated from all personal interests, and results in viśrānti or composure; but in the philosophical attitude there is. . complete detachment or aloofness from pleasure and pain and egoistic impulses, for the knower becomes impersonal by transcending his buddhi. This comes about through the predominance of the sättvika guna in both cases ; but in the spiritual attitude the buddhi, which contains in it vāsanās: or acquired impulses, modifying its intrinsic sāttvika charac- tor, is purged of all its egoistic tendencies, and the true knower, realising the intrinsic disparateness of prakrti and purusa, transcends the empirical plane. Such complete detachment is not possible in the aesthetic attitude. The world of poetry, being idealised, is different from the natural world and does not evoke egoistic impulses ; for the objects. contemplated in poetry have no reference to any one in parti- cular but are entirely impersonal. These impersonalised forms, therefore, afford to the enjoyer of poetry escape from the ills which arise from personal relations, but they are the means only of temporary release from the natural world, for he cannot, like the ordinary man, transcend his buddhi al- together. The three stages in the appreciation of poetry which lead ultimately to the aesthetic experience of Rasa indicate that the apprehension of the meanings of words

35 M. Hiriyanna, Indian Aesthetics, in Proceedings and Transactions of the First Oriental Conference, Poona, vol, ii, has dealt with the subject at some length.

LOLLAȚA AND OTHERS 127

(abhidha) is not important in itself but only as a means of apprehending the generalised conceptions which are unrelated to any one in particular, the bhāvakatva being the process of such generalising, by which the factors of the feelings, as well as the feeling itself, become impersonalised. These idealised creations of poetry lead to enjoyment or bhoga, which im- plies that the condition produced is one of pleasure, as distinguished from the case of the natural attitude which is not always pleasurable, as well as from the spiritual attitude which is neither pleasurable nor painful.

( 3 ) If we may judge from the somewhat elaborate criticism levelled against Bhatta Nayaka's theory, it seems to have produced a greater impression than earlier theories, and paved the way, no doubt, for the later theory of Abhinavagupta to whom belongs the credit of explaining the new aesthetic system of the Dhvanikāra and Anandavardhana. The Dhvanikära, however, in his exposition of rasa-dhvani and rasa, seems to have been greatly influenced by the Dramaturgic Rasa school. Bharata had declared that the business of the drama was to evolve one or more of the eight Rasas ; and therefore a more or less elaborate psychology of human sentiments had been analysed in the service of the dramatic art even before poetic theories began to be seriously discussed. Bharata's ideas on these psychological processes. and on Rasa, which is the final internal experience consisting in the consciousness of a certain condition of the ego, were elaborated by his commentators and followers, until the Dhvanikāra, followed by Anandavardhana and Abhinava- gupta, came into the field. From the earlier drama and dramatic theory, the idea of Rasa was naturally taken over to poetry and poetic theory; and as the transition from naïve to sentimental poetry was accomplished, the theorists went a step further and erected Rasa into one of its essential. foundations. Anandavardhana is quite explicit on this point

128 SANSKRIT POETICS

when he says (p. 181): etac ca rasādi-tātparyeņa kāvya- nibandhanam bharatādāvapi suprasiddham eva. In other words, what was already well established in the drama by Bharata and others thus found its way into poetry, profoundly modifying, as it did, the entire conception of the Kāvya36. From his extensive literary and philosophical studies as well as from his interest in the work of Bharata and his followers, Abhinavagupta goes further and lays down: nātyāt samudaya- rūpād rasaḥ, rasa-samudayo hi nātyam; na nātya eva ca rasaḥ kāvye'pi nāțyāmāna eva rasaḥ kāvyārathaḥ. These theorists realised that no system of Poetics, as no system of Dramaturgy, can ever ignore the feelings, moods and sentiments, and must find an important place for Rasa, the manifestation of which is as much the business of poetry as of the drama. Gradually stress came to be laid on the emotional mood, as well as on the imaginative thought, which the poet succeeds in communi- cating to us; the outward expression, on which the older writers pinned their faith so much, being regarded only as a means of suggesting or pointing to the implicit significance of such a mood in poetry. The insufficiencies of the earlier theories on Rasa are obvious and are therefore rightly criticised by Abhinavagupta ; but it was a happy idea to elaborate the theory in such a way as not only to supply these deficiencies but also to fit it well into the theory of 'suggestion' or dhvani formulated by the new school. It is not necessary for us here to enter into the details of the Dhvani-theory, which will be treated in its proper place ; but we may for convenience and continuity of treatment indicate here generally how the idea of Rasa was worked up into them, The Dhvani school, in its analysis of the essentials of poetry, found that the contents of a good poem may be generally distinguished into two parts. The one

36 Rudrabhatta states (i. 5) in the same way that Bharata and others have already discussed Rasa in connexion with the drama, while his own object is to apply it to the case of poetry. Cf M, Lindenau, Rasalehre, p. 2.

LOLLAȚA AND OTHERS 129

is that which is expressed and includes what is given in so many words ; the other content is not expressed but must be added to it by the imagination of the reader or listener. The unexpressed or suggested part, which is distinctly linked up with the expressed and which is developed by a peculiar process of suggestion (vyañjana), is taken to be the 'soul' or essence of poetry. To the grammarians and learned writers, it perhaps seemed paradoxical to state that the very essence of poetry was that which was not even expressed. On the other hand, some form of symbolical speech, in which wisdom demands that one should express oneself more in hints and suggestions than in actual words, was always in vogue, and the poets had been more or less partial to the method of speaking in metaphor or wrapping up their ideas in transpa- rent allegory. But suggestive poetry is something different from the merely metaphorical, which Vamana had already amply recognised and on which the Alamkāra and the Rīti schools had put so much emphasis. The metaphorical or the allegoric, however veiled it may be, is still in a sense express d and must be taken as such; but the suggestive is always unexpressed, and is therefore a source of greater charm by its capacity of concealment. This unexpressed or inexpressible is called into being by a particular function of suggestion, appertaining to words and their meanings, which this school postulates. Now the unexpressed, through the suggestive power of sound or sense, may be an unexpressed thought or matter (vastu), or an unexpressed figure of speech (alamkāra), but in most cases it is a mood or feeling (rasa) which is directly inexpressible. The Dhvani school, therefore, took up the moods and feelings as an element of the unexpressed and tried to harmonise the idea of rasa with the theory of dhvani.37 It was realised that poetry was not, as Dandin thought, the

37 Anandavardhana himself says (Dhv. p. 163) that his object is not merely to establish Dhvani but also to harmonise it with Rasa. 32

130 SANSKRIT POETICS

mere clothing of agreeable ideas in agreeable language; the feelings and moods play an important part in it. But the feelings and moods are in themselves inexpressible. We can give a name to them, but naming a mood or feeling is not equivalent to expressing or developing it. At best, therefore, we can suggest it. What the poet can directly express or des- cribe are the vibhavas etc .; but with the help of these express- ed elements which must be generalised and conceived, not as they appear in the natural (laukika) world, but as they may be imagined in the world of poetry, the poet can awaken in us, through the power of suggestion inherent in words and their meanings, a particular alaukika (dissociated) condition of the soul in which the relish of the feeling is possible. It is true that the poet cannot rouse the same mood or feeling as, for instance, Räma whom he describes felt, but he can call up a reflection of it, which is similar in some respects; and the condition of the reader's soul in the enjoyment of such feeling is in poetry and drama the relish of Rasa, which can be brought into consciousness only by the power of sugges- tion inherent in words or ideas. Here comes in the new colour given to the Rasa-theory by the exponents of the Dhvani school. They interpret Bharata's central dictum to mean that the Rasa is suggested by the union of the sthayin with the vibhavas through the relation of the suggested (vyangya) and the suggestor (vyañjaka) ; the nispatti of Bharata, therefore, should mean abhivyakti. The elaboration of the Rasa-theory, however, by this school in the direct tradition of Lollata, Sankuka and Bhatta Nāyaka, is associated by Mammata and others with the name of Abhinavagupta38. Commenting on Bhatta Nayaka's theory, Abhinava points out that there is no need, as there is no authority, for assuming the two powers of bhāvakatva and

38 There is some difference in the general theoretical positions of Änandavardhana and Abhinavagupta which will be noticed later.

LOLLAȚA AND OTHERS 131

bhogikarana; for they are implicitly included in the idea of rasa-vyañjanā and its ultimate āsvāda. Bharata's dictum kāvyārthān bhāvayantīti bhāvaḥ implies that bhāvakatva is an inherent capacity of all bhävas, as the means of bringing into consciousness the sense of poetry, the term sense indicat- ing here the principal sense consisting of the relish of Rasa. Hence the sthayin, together with the vyabhicārin, being bhāvas themselves, bring into existence through this inherent power the extraordinary relishable sense of poetry, cognised in a general or impersonalised form. In this way, the sthāyin may be regarded as the bhāvaka or nişpādaka of Rasa ; and this so-called bhavakatva, according to Abhinava, consists in noth- ing more than a suitable use of Guņa and Alamkāra (samu- cita-gunālamkāra-parigrahātmakam) for the ultimate purpose of awakening Rasa through the suggestive power of word and sense. Thus, partially admitting bhāvanā or bhāvakatva but explaining it somewhat differently, Abhinava turns to the other power assumed as bhoga or bhogikarana by Bhatta Nāyaka. He remarks that beyond pratiti or perception of Rasa, he is not aware of any other process called bhoga. If it is relish or enjoyment, it is already admitted as the essence of Rasa, and nothing is gained by giving it a new name. Abhinava thinks that the bhoga supposed by Bhatta Nayaka is nothing more than the āsvada or relish of Rasa, based on permanent moods like rati etc., and made possible by the suggestive power of poetry. It falls naturally, therefore, within the domain of suggestion and need not be taken as a separate function (bhogīkaraņa-vyāpāraś ca kāvyātmaka- rasa-vişayo dhvananātmaiva). This prātiti of Rasa, Abhinavagupta maintains, results from its abhivyakti or manifestation by the power of sugges- tion, and consists of a state of relish known as rasanā, āsvāda or carvana. What is manifested is not the Rasa itself, but its relish ; not the mood itself but its reflection in the form of a subjective condition of aesthetic enjoyment in the reader. This taste or relish partakes, no doubt, of the nature of cogni-

132 SANSKRIT POETICS

tion ; it is nevertheless different from the ordinary laukika forms of the process, because its means (viz. the vibhāvas) are not to be taken as ordinary or laukika cause89. Although Rasa requires these three factors for its manifesta- tion and cannot exist without them, it cannot yet be regarded as an ordinary effect, and the cause-and-effect theory is in- applicable ; for in the transcendental sphere of poetry, it is

39 This will make it clear why the Rasas like karuna, bibhatsa and bhayanaka, which cause pity, disgust or horror, can be termed Rasas in which enjoyment is essential. The relish of Rasa is supposed to be an extraordinary bliss, dissociated from personal interests, and not to be likened to ordinary pleasure and pain in which personal or egoistic impulses predominate. The mind is so entirely lost in its contemplation that even when the sentiment of grief or horror relished in such a state, pain is never felt, and even when felt it is a pleasurable pain. This fact is borne out by the common experience that when grief is represen- ted on the stage, the spectator says 'I have enjoyed it'. Hence Abhinava- gupta says: samājikānām harşaika-phalam nātyam na śokādi-phalam. Viśvanātha similarly remarks (iii. 6-7 and Vrtti) that those very things which are called causes of pleasure and pain in the world (e.g. banish- ment of Sita in the forest), when consigned to poetry and drama, possess the right to be called, in consequence of their assuming such an imper- sonalised form, alaukika vibhavas etc., and from them only pleasure ensues, as it does from bites and the like in amorous dalliance. If pain were really felt, no one would have been inclined to poetry and drama (kim ca teşu yadi duhkham na ko'pi syāt tad-unmukhaḥ). It is also maintained that tears constitute no proof that anything but pleasure is felt in poetry ; for the tears that are shed by the reader are not those of pain but those of sentiment. Jagannatha's remarks in this connexion are interesting. He says (p. 26) that the shedding of tears and the like are due to the nature of the experience of particular pleasures, and not to pain. Hence in a devotee tears arise on listening to a description of the deity; in this case there is not the slightest feeling of pain. Such is the power of detachment which poetry produces that even unpleasant things like sorrow generate dissociated pleasure ; and this pleasant relish of impersonalised or idealised artistic creations should be distinguished from the ordinary experiences of life .- The Nātya-darpana, however, sets forth a theory (pp. 158-59) that Rasa has a touch of sorrow in it, and Bhoja in his Srngāra-prakāśa says: rasā hi sukha-duhkhāvasthā- rūpāh. See V. Rāghavan. Number of Rasas (Adyar 1940), p. 155.

LOLLAȚA AND OTHERS 133

said, the connexion between cause and effect gives place to an imaginative system of relations, which has the power of stirring the reader's soul into Rasa, The resulting Rasa cannot be identified with the constituent vibhāvas, for the latter are not experienced separately, but the whole appears as Rasa, which is thus simple and indivisible. At the time of relish nothing else but the Rasa itself is raised to our consciousness. The writers on Poetics are fond of explaining this phenomenon under the analogy of a beverage which, made up of black pepper, candied sugar, camphor and other ingredients, gives us yet a taste different from that of its constituents. The result, therefore, is an indissoluble unity of taste from which every trace of the constituent elements is obliterated. Abhinavagupta goes a step further also in maintaining that the permanent mood (sthāyin) inferred from its laukika causes (e.g. women, garden etc.) remains in the hearts of the appreciating audience in the subtle form of latent impressions, the idea of vāsanā or latent impression having been already admitted by the philosophers. On reading a poem or witness- ing a drama, this permanent mood, remaining in the form of latent impression, is suggested by the depicted vibhāvas etc., which cease to be called laukika causes but go by the name of vibhavas etc. in poetry and drama, and which are taken in their general form without specific connexions. The vibhavas, therefore, are generalised or impersonalised in the minds of the reader, and do not refer to particularities, not through the power of bhāvakatva, as supposed by Bhatta Näyaka, but generally through the suggestive power of sound and sense and specifically through a skilful use of Guna and Alamkāra in poetry, and clever representation in the drama. In the same way, the sthayi-bhava, which is the source of Rasa40, is also generalised, because the germ of it is already

40 The sthäyin is so called because, in spite of its being transient like all feelings, its impression in the form of vāsanā or samskāra is

134 SANSKRIT POETICS

existent in the reader's mind in the form of latent impressions ; and this, together with the beauty of the generalised re- presentation of the vibhavas ete., removes all temporal and . spatial limitations. The mood is generalised also in the sense that it refers not to any particular reader but to readers in general, so that the particular individual, while relishing it, does not think that it is relished by him alone, but by all persons of poetic sensibility. This subjective relish in the mind of the spectator or reader is known as Rasa in poetry and drama. To state it briefly and without any technicality, there is in the mind a latent impression of feelings which we once went through (or which we acquired from previous births), and this is roused when we read a poem which describes similar things. By universal sympathy or community of feeling we become part and parcel of the same feeling and imagine ourselves in that condition. Thus the feeling is raised to a state of relish, called rasa, in which lies the essence of poetic enjoyment. It will be noticed that these theorists presuppose latent impression of experience (vāsanā) and univeral sym- pathy (sādhāraņya ar sādhāraņīkarana). Those who have not experienced the feeling of love, for instance, and have there- fore no impression of experience left in them, as well as . those who have no sense of community of human feelings, can never relish Rasa. The vāsanā, we are told, is natural (naisargikī) and may have been left in our mind through the samskāra of previous births, but it may also be acquired by

more or less permanent, being called up when the Rasa is cognised. Cf Prabhā p. 61: antahkaraņa-pravrtti-rūpasya ratyāder āśu-vināśatve'pi samskārātmanā cirakāla-sthāyitvād yāvad-rasa-pratīti-kālam anusaņdhā- nāc ca sthayitvam. But possibly it was originally called sthāyin because it constituted the permanent mood or sentiment in the composition, which nothing akin to it or opposed to it could overcome, but which could only be strengthened by other bhāvas. But the sthāyin itself is not rasa ; it must be vyakti-viśişta and vibhāvādi-melaka, and thus made carvanopayogī or relishable, Govinda p. 62. The ultimate relish of rasa is free from the contact of the sthayin, as it is of the vibhāvas.

LOLLAȚA AND OTHERS 135

study and experience. The writers on Poetics, therefore, are merciless in their satire on dull grammarians and old Mīmāmsakas, to whom such relish of Rasa is denied, and they declare unanimously that the rasika alone is capable of realising the rasa; for Rasa is not an objective entity which can reside in the hero or the actor, but a subjective condition realised by the reader's own capacity of aesthetic enjoyment. Thus, a degree of culture, experience and aesthetic instinct is demanded in the critic, the rasika or sahrdaya, in conformity with this subtle conception of poetry. As Abhinavagupta puts it, adhikārī cātra vimala-pratibhāna- sali-hrdayah, and describes such a sahrdaya (°Locana p. 11) as yeşām kāvyānuśīlanābhyāsa-vaśād viśadībhūte mano-mukure varņanīya-tanmayībhavana-yogyatā te hrdaya-samvādabhājaḥ sahṛdayāḥ. It may be pointed out here that this subtle conception of Rasa makes it difficult to express the notion properly in Wes- tern critical terminology. The word has been translated etymologically by the terms 'flavour,' 'relish,' 'gustation,' 'taste,' 'Geschmack' or 'saveur'; but none of these renderings seems to be adequate. The simpler word 'mood', or the term 'Stimmung' used by Jacobi may be the nearest approach to it, but the concept has hardly any analogy in European critical theories. Most of the terms employed have associa- tion of subtle meanings of their own, and are therefore not strictly applicable. For instance, the word 'taste' or 'relish' though literally correct, must not be understood to imply aesthetic judgment, 'good or bad taste,' but must be taken to indicate an idea similar to what we mean when we speak of tasting food. At the same time, this realistic description must not lead us to drag it down to the level of a bodily pleasure ; for this artistic pleasure is given as almost equiva- lent to the philosophic bliss, known as ananda, being lifted above worldly joy. This peculiar condition of the mind, the rasa, is realised, according to Abhinavagupta, through the characteristic func-

136 SANSKRIT POETICS

tion of vyañjana or suggestion inherent in word and sense, The idea is elaborated by later theorists who take pains. to shew that it does not not come under the province of Denotation (abhidhā), nor of Import (tātparyā), nor of Indication (lakşaņā), nor of Perception (pratyakşa) nor of Inference (anumāna), nor of Reminiscence (smarana),. which means of knowledge are admitted by philosophers. and grammarians. Into these technicalities which properly come under the discussion of the vyañjana-vrtti, we neeď not enter; but it may be noted here that Abhinava describes this abhivyakti, which is taken as synonymous with carvaņā, as vītavighna-pratīti or realisation freed from obsta- cles. Jagannatha and the author of the °Prabha commen- tary on Kāvya-pradīpa describe vyakti as bhagnāvaraņā cit. Both these terms constitute a link connecting the present theory with the teachings of the Vedanta. The dismissal of the avidya and the elimination of kāma and karman (interest and activity) lead us to a point of detachment where we realise the intrinsic identity of self with Brahma and appre- hend the bliss or änanda resulting from such a realisation. The idea of Vecantin's Moksa, which consists of a condition, not to be produced but to be made manifest by the removal enveloping obstacles, finds an analogy in the idea of the manifestation of Rasa, implied in its abhivyakti, which con- sists not in the expression of anything new but in the reveal- ing of something already existing. The brahmāsvāda is likened to the rasasvāda because in both cases the intimate , realisation comes after the limitations of the ego-centric attitude are transcended, and all separate existence is merged in the unity or harmony realised. This happens in the case of rasāsvāda when the poetic sentiment, which remains in his heart in the form of latent impression, is made to shine forth, and the spectator's mind is purged of all egoistic impulses by the force of the idealised or generalised creations of poetry, consisting of the vibhavas etc, which are therefore termed xighnāpasārakas or removers of obstacles. It is, therefore,

LOLLAȚA AND OTHERS 137

alaukika, being unlike the taste of interested worldly happi- ness and being incompassable by the ordinary processes of knowledge. Its essence consists in its relish or taste, āsvāda, carvaņā, or rasanā ; but it is a relish in which the Rasa alone, apart from its constituent elements, is raised to consciousness. It is, therefore, described as a relish in which the contempla- tion of anything else but Rasa is lost (vigalita-vedyāntara), or which is free from the contact of aught else perceived (vedyāntara-sparśa-sūnya), like the state of mind lost in the philosophic contemplation of Brahma. It is not capable of proof or designation and cannot be made known, because its perception is inseparable from its existence; or in other words, it is identical with the knowledge of itself. The only proof of its existence is its relish itself by the sahrdaya or the man of taste (sakala-sahrdaya-hrdaya-samvedana-sākşika) ; and the sahrdaya to whom alone this bliss is vouchsafed, is like the yogin or devotee who deserves this preference through his accumulated merits (punyavantah praminvanti yogivad rasa- samtatim)41. This, in its general outline, is the Rasa-theory as finally fixed by the Dhvani school; and all later writers, from

41 The artistic attitude is, therefore, different from the natural, and more akin to the philosophic. But art affords only a temporary release from the ills of life by enabling one to transcend his personal relations or practical interests, and restores equanimity of mind (viśrānti) by leading him away from the common world and offering him another in its place. This is an attitude of pure pleasure, of disinterested contem- plation (samvit), but not of true enlightenment which comes to the knower who, no longer on the empirical plane, transcends completely the sphere of pleasure as well as of pain. The attitude is variously termed camatkāra-nirveśa (awakening of poetic charm), rasanā (relish), āsvāda (taste), bhoga (fruition), samāpatti (accomplishment), laya (fusion) and viśrānti (repose), which terms indicate the philosophical colouring given to the doctrine. For the idea of camatkāra involved in Rasa, see S. K. De, introd. to Vakrokti-jīvita (2nd ed. 1928) p. xxxvi, fn 33 and below under Jagannätha (ch. vii). See also V. Raghavan, Some Concepts, pp. 268-71.

138 SANSKRIT POETICS

Dhanañjaya to Jagannātha, accept, more or less, this new interpretation and attempt to work it out in detail. Even Mahimabhatta, who tried to demolish the Dhvani-theory, acknowledges the importance of Rasa and declares that on this point there is no difference of opinion between himself42 and the Dhvanikära, the only difference existing with regard to the function par excellence which manifests the Rasa. Thus, an endeavour was made by the Dhvani-theorists not only to explain the concept of Rasa in terms of inward ex- perience, but also to absorb this idea of aesthetic delectation into the new theory of Dhvani, and make it applicable to poetry as well as to the drama; and the Rasa school, properly so called, began to merge from this time onwards into the dominant Dhvani school, to the consideration of which we now turn, leaving the treatment of the later development of the Rasa-theory to a subsequent chapter.

42 kāvyasyātmani sangini (angini ?) rasādi-rūpe na kasyacid vima- tiḥ, p. 22.

CHAPTER V

THE DHVANIKĀRA AND ĀNANDAVARDHANA

(The Dhvani System)

The origin of the Dhvani school, like that of other schools of Poetics, is lost in obscurity; but the first clear formulation of its theory of dhvani as a whole is to be found in the memorial verses of the Dhvanikāra, whose date is unknown but who could not have been very far removed from the time of his commentator Anandavardhana. It is possible, however, that the Dhvanikära himself is following a much older tradi- tion. The fact that he shows himself conversant with some theory of rasa, alamkara and rīti need not be cited to the credit or discredit of this conjecture ; for these systems them- selves cannot be traced back to any definite period of time, and there is also no conclusive evidence that the Dhvanikāra was aware of the particular views of Bharata, Bhämaha or Dandin, with whom we begin the historic period of growth of these systems. But the very first line of the first verse of the Dhvanyäloka itself states that the theory that dhvani is the essence of poetry was traditionally maintained by earlier thinkers (kāvyasyātmā dhvanir iti budhair yaḥ samāmnāta- pürvaḥ). Accepting this statement of the Dhvanikāra, it is difficult, however, to explain why the dhvani-theory did not in the least, as the rasa-theory did to a certain extent, influence such early writers on Poetics as Bhāmaha. Dandin or Vämana. It is easy to maintain, on the other hand, that the vyañjana as a function must have been evolved by the school which set up a theory of vyangya artha or dhvani, as this function is not traceable in philosophers or philosophical grammarians before the time of Anandavardhana. Bhamaha1,

1 See above ch. ii, pp. 52f.

140 SANSKRIT POETICS

Vāmana (iv. 3. 8) and other early theorists discuss or show themselves cognisant of such a 'suggested sense' in general ; but they never use the terms vyañjanā, or vyangya artha or dhvani, as they probably would have done if it had been so universally known or accepted as the Dhvanikāra's statement would apparently imply. But this non-recognition by other schools need not be taken as a serious argument, as it admits of several reasonable explanations. It is well known that the philosophers refuse to recognise the vyanjanā as a Vyāpāra sui generis2, and even after it was put forward by the Alamkārikas, they would explain it by another recognised function as anyatha-siddha ; it is not surprising, therefore, that orthodox grammarians or philosophers should entirely ignore it. The absence of any direct reference to dhvani in early writers on Poetics may be explained by the not unlikely supposition that probably the Dhvanikāra him- self, who summed up and uttered the theory in a definite form, was contemporaneous with these writers, as we cannot put him much later if we are to leave sufficient margin bet- ween him and his commentator, as well as make room for intermediate scholastic activity evidenced by the recapitula- tion-stanzas cited by Anandavardhana in his Vrtti3. Even leaving aside this conjecture, the cases of such non-recogni- tion are actually explained by the Dhvanikāra himself, in the verse we have already cited, as constituting really cases of half-recognition ; for he seems to indicate that these early writers were cognisant of dhvani, but not having understood its nature they naïvely and uncritically looked at it from other points of view, some comprehending it in other elements of poetry, some thinking it incomprehensible, and others (like the poet Manoratha cited by Anandavardhana) going to the extreme of denying its existence altogether. One of the objects of the Dhvanikāra in this statement was, no doubt, to in dicate that he was not putting forward something entirely 2 See Jacobi in ZDMG, lvi, 1902, p. 397 fn 2, and p. 398 fn 1. 3 See vol. i. p. 108, and BSOS i, 4, 1920, pp. 7-8.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 141

new, and to find an authority for his procedure in the implied attitude (real or imaginary) of certain older writers ; but, apart from this, it is clear that although there is nothing explicit in the older writers, one can never affirm that some kind of 'suggested sense' was not known to them. It would be extraordinary indeed that a work like Dhvani- kara's could have sprung into existence without having had a previous history, although such earlier forms of the theory .as might have enabled us to trace directly its origin and growth were either not committed to writing or had disappear- ed in course of time ; for at its first appearance as we have it in these Karikas, we find the theory in a relatively complete shape, the outlines of which, definitely settled; may require considerable filling up but no important or substantial modifi- cation. This is probably implied by Abhinavagupta's gloss on the word parampara in Anandavardhana's explanation of the phrase samamnata-pūrvah used by the Dhvanikāra with reference to the previous existence of the theory. Abhinava explains (p. 3) that the theory was stated in unbroken tradi- tion by previous thinkers without its being discussed in particular books (avicchinnena pravāheņa tair etad uktam, vināpi viśista-pustakeşu vivecanāt). It is true that Mukula refers (p. 21) to a theory of dhvani being newly described by some men of taste (sahrdayair4 nūtana-tayopavarņitasya) as something not comprehensible by the recognised function of lakşanā, and dces not discuss it for its over-subtlety (etac .ca vidvadbhiḥ kuśāgrayā buddhyā nirūpaņīyam ... ityalam

4 The word sahrdaya here cannot be taken (see vol. i, p. 105f) as a proper name referring to the Dhvanikāra ; nor is it to be taken as a title of the propounder of the dhvani-theory. As in most of the places, it refers in general to the critics or men of taste who established the new theory, or in particular to the Dhvanikara or Anandavardhana ; and there is no need to go beyond this ordinary meaning of the term in Alamkāra literature. Possibly the reference is directly to Ānanda- vardhana who was a contemporary of Mukula's father Kallata (see vol. i, p. 74).

142 SANSKRIT POETICS

ati-prasangena) ; but he may in this passage be directly refer- ring to the Dhvanikära, who for the first time probably summed up in his memorial verses the floating traditions, or to Anandavardhana to whom belonged the credit of fixing the theory into a new and complete shape. This conjecture about the traditional existence of the dhvani-theory in some form or other even before the Dhvanikāra receives support from the fact that the theory in its essence derived its inspiration from the works of early grammarians and their semi-philosophical speculations on speech. Originating as a theory of expression, the theory of vyañjanā, no doubt, received no recognition from orthodox grammarians ; but not choosing to appear as an entirely novel theory, it sought the protection of the grammarian's authority by pretending that it was founded on the analogy of their an- cient sphota-theory. We have already noted5 the great in- fluence of the older science of grammar on poetics, and Änandavardhana himself is careful in noting that the system demonstrated by him is built on the system of the grammari- ans, who were the earliest theorists to apply the term dhvani to the spoken letter which reveals the sphota6. Abhinavagupta commenting on this passage, perhaps goes too far in follow- ing up, after the authority of the Vākyapadīya, all the details of the sphota-theory, but there is hardly any doubt that the writers on Poetics had this theory before them when they

5 See vol. i, pp. 6-7. 6 prathame hi vidvāmso vaiyākaraņāh, vyākaraņa-mūlatvāt sarva- vidyānām. Te ca śrūyamāņeşu varņeşu dhvanir iti vyaharanti. Tathai- vānyais tan-matānusāribhih sūribhiḥ kāvya-tattvārtha-darśibhir vācya- vācaka-sammiśraḥ śabdātmā kāvyam iti vyapadeśyo vyañjakatva-sāmyād dhvanir ity uktah (pp. 47-8). See °Locana on this. Cf also : pari- niścita-nirapabhramśa-śabda-brahmaņām vipaścitām matam āśrityaiva pravrtto'yam dhvani-vyavahāra iti taih saha kim virodhāvirodhau cintyete (p, 199). In this last passage, the reference is not to Vedänta, as some would think, but to the grammatico-philosophical theory of sabda-brahma. See Jacobi's note on this passage in ZDMG Ivii, 1903, p. 56 fn 1.

THE DHVANIKÄRA AND ANANDAVARDHANA 143

elaborated their own system of dhvani. The sphota, which has been likened to the neo-platonic logos, is often translated by the terms 'expression,' 'concept' or 'idea'; but none of these terms brings out its essential nature. Some philosophers propounded and the grammarians took it for granted that a word has intrinsically a word-prototype corresponding to it. The sphota is not exactly this word-prototype, but it may be explained as the sound of a word as a whole, and as convey- ing a meaning apart from its component letters (varnas). The sphota does not contain exactly the sounds of the word in the order peculiar to the letters, but the sounds or something corresponding to them are blended indistinguishably into a uniform whole. When a word is pronounced, its individual sounds become reflected in some degree in the order of the sphota in which the particular sounds are comprised ; and as soon as the last sound dies away, the sphota, in which the idea corresponding to all these sounds is comprised, becomes manifest and raises to our consciousness the idea thus asso- ciated. The sounds of a word as a whole, therefore, and apart from those of the constituent letters, reveal the sphota. Taking their cue from this somewhat mystical conception, the Alamkarikas developed the idea of dhvani by analogy. The several expressed parts of a poem, they held, reveal the unex- pressed deeper sense, which is something singular and different from the denotative and indicative elements both in order and in essence, and which is termed the dhvani (lit. 'sound', 'echo", . 'tone') or vyangya artha (suggested sense) in poetry. The word dhvani itself, as Anandavardhana pointed out, is some- times used by the grammarians for the word or letters which reveal the sphota. Mammata's remarks in this connexion are pertinent. In his Vrtti on the definition given by him of dhvani (i. 4), he says that the dhvani is, according to the grammarians, that word which reveals the all-important sphota, inasmuch as through it arises the knowledge of the word's meaning. Others, by whom he signifies the writers on the dhvani-theory in Poetics, carry this doctrine of the gram-

144 SANSKRIT POETICS

marians a step further and apply the term dhvani to the meaning, as well as to the word which is capable of suggest- ing a meaning superseding the one which is directly express- ed. Intrinsically the two theories have scarcely any mutual connexion; but what the Alamkarikas really wanted was an authority for their assumption of the power of vyañjanā, which the great grammarians did not acknowledge. The sphota-theory of the grammarians, however, presupposed something similar, for the varnas of a word reveal, as it were, the ideal word. Hence it afforded an analogy which could at least boast of the authority of the Vaiyakaranas, the prathame vidvamsah, and which could therefore be seized upon by the Alamkārikas as the foundation of their own theory of sugges- tion. It may also be pointed out that the sense of 'manifesta- tion' which exists in the vyanjana is an idea which, we have already noted7, is not unfamiliar to Indian philosophical speculation. The vyañjana does not consist in the utterance of something new, but in the manifestation of something al- ready existing ; it is, to use a familiar illustration from Indian philosophical systems, like the revealing of the already exist- ing jar by the lamp. Although the general concept of dhvani connects itself with such half-mystical currents of thought, Anandavardhana yet takes care to point out (pp. 232-4) that this dhvani is not, as often supposed, something mystical but it is something that can be properly defined and grasped; and he has no sympathy with those schools which would dismiss it, as Kapila has dismissed the sphota from the philosophical realm, on the ground that it is something inexplicable (anākhyeya). Although it accepted, with some modifications, the grammarian's analysis of the nature and function of speech and based its theory of dhvani on the analogy of the theory of sphota, the school really started indepen- dently with a distinct theory of expression of its

7 See vol. i, p. 9.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 145

own, which demonstrated a function of vyāñjanā and vyangyārtha untraceable in earlier speculative literature. But the influence of other schools of Poetics on the com- posite work on the Dhvanikara and Anandavardhana cannot be ignored. The latter, if not the former, appears to be perfectly familiar with the views of Bharata, Bhamaha, Ud -. bhata and Vamana, most of whom are cited directly by name; but even the Dhvanikāra must have known the theories of the Rasa, Alamkara and Rīti schools in some form or other. For, the Dhvanyāloka has two professed objects in view, viz., (1) the establishment of the theory of Dhvani and demon- stration that this idea cannot be comprehended by the theories of earlier or contemporaneous schools of Poetics, and (2) an examination of the existing ideas of rasa, alamkāra, rīti, guņa and dosa with a view to correlate them with the idea of dhva- ni, and thus by synthesis to evolve a complete and systematic scheme of Poetics. It succeeded so far in realising both these objects that not only was the concept of dhvani accepted inplicitly by almost all later writers, but the systems, which emerged after Anandavardhana and of which Mammata may be taken as the first and foremost representative, cannot be regarded strictly as constituting independent schools, nor can they be affiliated readily and entirely with the older Rasa, Alamkāra or Rīti schools. They constitute in substance a new aesthetic scheme in which the ideas of all these schools are worked and harmonised into a comprehensive doctrine, the outlines of this new adjustment being first clearly marked and the foundations firmly laid by Anandavardhana. Starting with a theory of expression, the Dhvani school concerns itself, first of all, with the grammatico-philosophical problem about the function of words and their meaning, or in other words, about the relation of a word to that which is expressed by it. The grammarians, logicians and the Mīmām - sakas had already laid down that the function by which the primary or intrinsic meaning (mukhya or śakya artha) of a word is known as abhidha, generally translated by the term 33

146 SANSKRIT POETICS

Denotation, which gives it its conventional significance (samketita artha). Thus, the concept of the cow is given by the word 'cow' by its power of Denotation. It has been defined as that power of a word which conveys to the under- standing the meaning attached to it by convention, without the intervention of any power. This convention (samketa) consists in a particular word conveying a particular meaning (asmāc chabdād ayam artho boddhavya ityākāraḥ śakti-grāhakaḥ samayah), which is comprehended by observing what takes place in the world (vyavahāra). We need not concern ourselves with the question whether this śakti is iśvarecchā: or icchāmātra (divine or human will); but there are several theories as to where this convention is to be understood, held respectively by the grammarians, logicians, Saugatas and Mīmamsakas. The writers on Poetics maintain, after the grammarians8, that it has reference either to genus (jāti), individual (dravya), quality (guņa) or action (kriyā). When this abhidheyārtha or the primary meaning of a word is incompatible, another power called laksaņā or Indi- cation (i. e. transference of sense) is communicated, whereby another meaning connected therewith is apprehended, either through usage (rūdhi) or from some special motive (prayo- jana). Thus, one can say 'the country rejoices', but since the country itself cannot rejoice, it is indicated that the people of the country rejoice. This power really belongs to the sense (artha-vyāpāra), as later analysis points out, but it is attributed to words and is thus an āropita-abdavyāpāra. That is to say (as other writers explain it) we have first śabda or the word, then its vācyārtha or direct denoted meaning, after which or in connexion with which comes the laksyārtha or indicated meaning through the power or Indication. It is thus santara (and not nirantara like abhidha), having the

8 Both Mukula and Mammata (Šabda-vyāpāra°, p. 2) point out that this view of the Alamkārikas is based on the dictum catustayi sabdānām pravrttih, occurring in the Mahābhāsya (ed. Kielhorn p. 19, 1. 20).

THE DHVANIKĀRA AND ĀNANDAVARDHANA 147

vācyārtha coming in between ; for the lakșunā is resorted to when the primary sense is incompatible (bādhita) and is so far artha-nistha as based on the expressed sense9. Hence the three essential requisites of the laksanā are the incom- patibility (or exhaustion) of the primary sense, the connexion of the indicated sense with the primary sense, and the reason or motive (prayojana) for resorting to it. As the Denotation is dependent on worldly convention (vyavahārıka samketa). so is the Indication (as Mammata points out) upon the special convention based on these three requisites; and as there can hardly be any indicated or transferred sense without the primary sense, the Indication is sometimes called the tail, as it were, of Denotation (abhidhā-pucchabhūta). In fact, writers like Bhatta Nāyaka, as we have seen10, would include lakşanā under abhidhā, of which it is supposed to be an extension. The lakșaņā being thus of a derivative nattre, its relations to abhidhā have been summarised differently in different works. The Nyaya-sūtra gives an exhaustive list of the relations on account of which a word is used in a secondary or transferred sense for another (ii. 2. 63), corresponding to the lakşanā of the Ālamkārikas; but Mukula quotes the authority of Bhartrmitra11 who summarises them in a verse

9 śakya-vyavahita-lakşyārtha-vişayatvāc chabde āropita eva sa vyāpārah, vastuto'rtha-nistha evety arthah, tad uktam-'sāntarārtha- nisthah' iti, °Pradīpa, ed. N.S.P., 1912, p. 27. 10 See above ch. iv, p. 124. 11 Abhidhā-vṛtti-mātrkā p. 17. The verse is also quoted anony- mously in Mammata's Šabda-vyāpāra p. 8, in Kāmadhenu p. 133 and in many other works. Abhinavagugta (°Locana p. 56) alludes to it, and discusses these five categories .- Mukula Bhatta's work consists of 15. Kārikās with prose Vrtti. Its object is to examine the principle which: should regulate words in their meanings. It includes Laksaņā in Abhidha, for it says that the functions of Abhidha are twofold, direct and indirect, both of which lead to the understanding of the import of words. Mukula discusses Abhidha only, but Mammata on his Sabda- vyāpāra-paricaya establishes three distinct functions of words, of which the last is Dhvani.

148 SANSKRIT POETICS

into five categories, viz. sambandha (connexion), sādrśya (similarity), samavāya (inherence), vaiparītya (contrariety) and kriya-yoga (association through action). 'The fat Devadatta does not eat in the daytime' (pīno devadatto divā na bhunkte), 'the lad is a lion' (simho mānavakaḥ), 'the herd- station on the Ganges' (gangāyām ghoșaḥ); 'this fool is a Brhaspati' (brhaspatir ayam mürkhaḥ) and 'in a great war thou art a Satrughna' (mahati samure śatrughnas tvam) are given as respective instances of the usage. We need not further dilate upon these niceties of analysis, nor enter into the elaborate classifications of laksanā, but we may note here that the laksana or transferred expression lies at the root of figures like metaphor and of metaphorical mode generally, which consist of the fancied transference of the qualities or action of one object to another. It has been pointed out that the transferred expression, resolving into the metaphorical, is the source of a particular beauty, because the special motive (prayojana) with wbich the poet chooses the trans- ferred expression becomes realised along with it, without being directly or at all expressed. When we say, for instance, 'youth is the springtime of life', we mean to imply at once, without directly expressing it, the beauty, vigour or enjoy- ments of spring-time. The prayojana or motive, though unexpressed, is yet apprehended. This is supposed to be one of the reasons, as we shall see, why we should admit, besides Denotation and Indication, a third function of vyañjanā or Suggestion, by which something not expressed is revealed. But there is a limited class of writers who postulate another function, called tatparya or Purport, which leads us to apprehend the connexion among the meanings of the constituent words in the form of the import of the whole sentence. This function conveys the connected meaning of the several words and therefore differs from abhidhā and lakanā which convey the meaning of a particular word, the tātparyārtha being manifested, not by word, but by a whole sentence, and therefore remaining distinct from the meanings

THE DHVANIKĀRA AND ĀNANDAVARDHANA 149

denoted or indicated by individual words The words have, according to this view, the power of denoting or indicating things and not the connexion (anvaya) among things, which is known not from the import of words as such, but from their relations of compatitility (yogyatā), proximity (samnidhi) and expectancy (ākānkșā). When the logical connexion or anvaya is thus known, a special sense arises which is called tātparya. Mammața explains (ii. 1, Vrtti) the position of these Abhihitā- nvaya-vādins, as they are called, thus12: "When the meanings of the words, to be hereafter explained, are connected in accordance with expectancy, compatibility and proximity, another sense arises, called purport, which has a distinct form and which, though not constituting the sense of words is yet the sense of the sentence-this is the view of the Abhihitan- vaya-vādins". The theory of this school is rejected by another school of Mīmāmsakas, called the Anvitābhidhāna- vādins, who deny the necessity of postulating a special func- tion like tätparya; for they hold that words have a power to denote not only things but also their purport or connexion along with them. To put it in another way, words do not express their sense generally but connectedly. In ordinary life, for instance, we first understand meanings from sentences, and words convey ideas not absolutely but relatively, ie., as having a connexion with one another. Mutatis mutandis, the theory would remind one of Berkeley's denial of abstract ideas. The formulators of the dhvani-theory do not enter into these minute discussions but appear to recognise them implicitly, although most writers from the time of Mammata (who deals with these questions in his Kāvya-prakāśa as well as separately in his Śabda-vyāpāra-paricaya) start with a preliminary analysis of word-function ; and some later works like Appayya's Vrtti-vārttika are devoted specially to the 12 ākānkşā-yogyatā-samnidhi-vaśād vakşyamāņa-svarūpāņām padār- thānām samanvaye tātparyārtho višeşa-vapur a-padārtho'pi vākyārthah samullasatīty abhihitānvaya-vādinām matam.

150 SANSKRIT POETICS

subject. All writers from Anandavardhana's time accept as a rule the abhidhā and lakșaņā, but they are not unanimo us with regard to the tātparya as a separate function, which they take as included in the vyañjana vrtti, this being the third and most important function established by the Dhvani school as the theoretical foundation of dhvani or the 'suggested sense' in poetry. The vyanjanā or power of suggestion is generally defined as that function of a word or its sense by which a further meaning comes into being, when the other functions, viz. abhidhā and lakşaņā, are ex- hausted in their scope. Ideas or notions are what are conveyed by words through their powers of Denotation and Indication ; these, put together in a sentence, convey a complete thought through the supposed power of the sentence, styled Purport. Now, another power is postulated by which a deeper sense, the vyangya artha, is revealed, consequent upon but distinct from the simple thought13. All good poetry, called par ex- cellence the dhvani-kāvya14, must have such a sense implicit in it, a sense which can only be realised by the vyañjanā- vrtti or power of suggestion postulated by this school. Now the question has been animatedly discussed as to whether it is necessary to postulate this separate function of vyañjana, or whether it may not be comprehended in other recognised functions like abhidha or lakşaņā, and in other

13 A word (or its sense), in virtue of these three powers, is called respectively the expressive (vācaka), the indicative (laksaka) and the suggestive (vyañjaka) ; and the sense which arises is termed respectively an expressed (vācya), indicated (laksya) and suggested (vyangya) sense. 14 The word dhvani (lit. 'echo' or 'tone') is used almost synony- mously (cf. Hemacandra p. 26) with the word vyangyartha (suggested sense), and sometimes wrongly as co-extensive with vyañjana, which term properly designates the process manifesting it. The dhvani-kāvya is so-called because the vyangyärtha, which predominates in it over the vacyartha, is 'echoed' par excellence in this class of poetry. Viśva- nātha (p. 198) explains the term etymologically thus: vācyād adhika- camatkāriņi vyangyārthe dhvanyate'sminn iti vyutpattyā dhvanir nāmo- ttama-kāvyam.

THE DHVANIKÄRA AND ĀNANDAVARDHANA 151

intellectual processes like anumāna or inference. Jayaratha cites (p. 9) a verse which enumerates twelve different ways in which the problem of Suggestion may be and perhaps was explained away ; but broadly speaking, we need notice, as Anandavardhana and his followers have done, only the principal attempted explanations. The first verse of the Dhvanyāloka summarises these antagonistic views into three groups. One sceptical school entirely denies the suggested sense in poetry. A second school, which is agnostic in this respect, holds that it is beyond the province of words (kecid vācām sthitam avişaye tattvam ūcus tadīyam), and can only be perceived by a man of refined discernment (sahrdaya- hrdaya-samvedyam, Ananda p. 10). A third school would try to trace it back to the recognised functions like abhidhā, lakşaņā and tātparya, or to some such means of knowledge as anumana or syllogistic reasoning. These three schools naturally divide themselves into two distinct standpoints: the one absolutely denies or ignores the concept of dhvani and thus does away with the necessity of vyañjana ; the other admitting the dhvani, attempts to explain away the necessity of vyañjanā, as it is sufficiently accounted for by the ordinary recognised functions. Against the attack of the systems which deny the existence of the suggested sense, the old argument that nothing can be denied which is not apprehended is applied ; but apart from such purely scholastic objections, the real grounds for pos- tulating the suggested sense are, the consideration, in the first place, that being a profound verity, it can positively be established by an examination of aesthetic facts as well as facts of experience ; and in the second place, that there are some elements of poetry (e. g. the Rasa) which cannot be satisfactorily explained as revealed by abhidhā, lakşaņā, anumāna or similar other means. This brings us to the consideration of the views of those who accept the concept of vyangya artha but dispense with the necessity of such a separate and unauthorised Vrtti as

152 SANSKRIT POETICS

vyañjana, which is regarded as included in other functions of sound and sense, Some Mīmāmsakas hold, for instance, that th e so-called suggested sense is conveyed by the abhi- dha-vrtti or the denotative power of a word. In this con- nexion, the dīrgha-vyāpāra-vādins15 are said to have main- tained that as a single arrow, discharged by a strong man, destroys by a single movement, called velocity, the armour of the enemy, pierces through his body and takes away his life, so a single word, used by a good poet, brings before us by a single power, called abhidha, the sense of the word, teaches us its logical connexion (anvaya) and makes us apprehend the suggested sense. The substance of this graphic des cription appears to be that such is the more and more exp anding function of abhidha that it is not to be measured in the balance and confined to the single business of making us understand the samketita artha, but it is competent to express whatever sense is apprehended after a word is heard. But it is urged in reply that the abhidha has not the power to give us the perception of a matter (vastu), an imaginative fact (alamkāra) or an emotional mood (rasa), because it ceases, in the orthodox opinion, after conveying the conven- tional (i. e. literal) sense, and the Rasa etc. are not matters of mere convention. Nor is the denoting, for instance, of component vibhvas, which give rise to Rasa, a denota- tion of the Rasa itself ; for it is acknowledged that the Rasa is not realised by a mere naming thereof but partakes of the nature of a self-manifested joy, the development of which can at most be suggested. These facts cannot be satisfactorily explained unless we assume with the dīrgha-vyāpāra-vādins an all-comprehensive power for the abhidha, for which there is hardly any authority. Moreover, if we assign such ex- tensive powers to Denotation, why even admit the power of Indication, since the sense conveyed by the latter might be understood from the Denotation itself ? 15 This view is sometimes ascribed to Lollata; but see on this. question vol. i, p. 36-7.

THE DHVANIKARA AND ANANDAVARDHANA 153

Admitting the denoted sense, therefore, strictly as that conveyed by convention, it cannot be said to be manifold, for it exhausts itself after conveying the particular conven- tional concept; the suggested sense, on the other hand, becomes varied in accordance with the diversity of the occasion, the speaker and similar other factors. The denoted meaning is shown to differ from the suggested (1) in point of form, for the suggested sense may sometimes be quite the opposite of the negative or positive (as the case may be) expressed sense, (2) in point of location, for the expressed sense resides in words alone, while the suggested sense may be found in the words, in their position, in their denoted meaning itself, in the affixes or suffixes, in the arrangement of letters and so forth, (3) in its effect, for the denoted sense brings a mere cognition, the suggested a surprise, (4) according to the nature of the speaker, the addressee, or the perceiver. The attempt to maintain that the suggested sense is conveyed by the tātparya or Purport, as some Naiyāyikas hold, is similarly shown to be insufficient; for the function of the Purport is exhausted by simply making us apprehend the logical connexion of the ideas in the sentence itself and cannot, therefore, take us to the vyangyartha, which arises after the sentence is understood. Nor is the laksanā-vrtti sufficient to explain the subtle power of Suggestion. Those who maintain, however, that the suggested sense is no other than the laksya or indicated sense are asked (°Locana p. 51) whether they consider the non- difference of Suggestion and Indication to mean (1) that the two functions are identical (tādātmya or tādrūpya), (2) that they consider the Indication to be the constant differentiating property (lakşaņa or vyāvartaka-dharma) of Suggestion, Or (3) that the Indication is an occasional differentiating mark (upalakşana or tațastha lakșana) of Suggestion in special cases. The Dhvanikāra and Anandavardhana discuss these views generally (pp. 50-9), but Abhinavagupta deals with them some- what elaborately.

154 SANSKRIT POETICS

With regard to the first of these views, viz. the tādrūpya or identity of dhvani and bhakti (by which term laksanā is meant), the Dhvanikara lays down that Suggestion cannot be identical with Indication, because both have properties peculiar to themselves (i. 17). Indication is based upon the consideration of the barring of the expressed sense, and consists merely in upacāra (upacāra-mātram tu bhaktiḥ, Ānanda p. 51), or, as Abhinavagupta expresses it, in the secondary application of a word (guna-vrtti). The suggested sense, on the other hand, though essentially distinct in character, does not yet cancel the expressed sense altogether. 'The later writers16 explain further that it is not a mere secondary application of a word through usage or special motive. For, if you say that in such a sentence as 'a herd- station on the Ganges', the supposed motive, viz. the coolness and purity of the site, is not suggested but indicated, then the notion of 'the bank', which is the real indicated sense, would become the primary meaning of the word 'Ganges' (for the motive and the secondary sense of 'bank' cannot both be indicated), and consequently would be cancelled, since there can be no Indication without the primary sense being cancelled. We must, therefore, acknowledge another indicated motive for the indication of the first motive (for there can be no indication without the supposition of an indicated motive), and a third motive again to this second indication, and so on ad infinitum. In fact, as already noted before, the prayojana or special motive is not expressed at all; if it is left un- expressed, how is it then apprehended, unless we suppose that it is suggested? It has also been demonstrated that Suggestion is based on the peculiarity of the speaker, the addressee and various other circumstances; and there is a difference as well in location, the Indication resides in a word only, the Suggestion in a word, its parts, its sense and in the style. Mammata adds that Suggestion cannot be said to be co-extensive with Indication and Denotation combined ; 16 e.g. Viśvanātha in his Sāhitya-darpana pp. 247-48.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 155

for it is seen to come into existence from mere letters without any specific Denotation. The second view that Indication is the laksana or the constant differentiating characteristic of Suggestion is shewn by the Dhvanikāra to be vitiated by the logical fallacies of too wide (ativyāpti) or too narrow definition (avyāpti). This is more or less a scholastic objection, and is based on the characteristic notion of Suggestion defined by its champions ; for both Anandavardhana and his commentator shew that Indication sometimes coverst a much wider, sometimes a much more limited, field than Suggestion. The vyañjanā, for instance, is not accepted when the prayojana of the Indication is not charming ; on the other hand, in cases of vivakşitānya- para-vācya dhvani, there is scope for Indication, for the Suggestion here is expressly based on Indication. The third view that Indication may be an occasional distinguishing mark (upalakşana) of Suggestion is not denied by the Dhvani- kāra, for Suggestion may sometimes rest ultimately on Indication, e. g. those cases which are admitted by Dhvani- theorists as based on lakşaņā (lakşaņā-mūla dhvani); but this does not prove the opponent's position that Indication is identical with Suggestion. Some of the oldest and most aggressive objectors to the admission of the vyañjana-vrtti are the adherent of the anu- mana-theory, whose views are refuted at some length by Anandavardhana himself. They are represented to us in later literature by Mahimabhatta in his Vyakti-viveka, a work which was written with the avowed object of establishing that the suggested sense can be arrived at by the process of syllogistic reasoning. Most of these controversies belong to the realm of scholastic speculation and are far removed from actual Poetics. We shall deal with Mahimabhatta's theory in its proper place ; it will suffice here to set forth the theory in its general outline as it obtained in Anandavar- dhana's time and notice the arguments with which it is sought to be disproved.

156 SANSKRIT POETICS

From Anandavardhana's repudiation (pp. 201 f) of the views of this school, it appears that its essential position consisted in establishing that the cognition of the unexpressed or suggested sense is nothing more than the cognition of the object of a logical conclusion, so that the relation of the suggestor and the suggested is that of the syllogistic middle and major terms (vyangya-pratitir linga-prafitir eveti linga- lingi-bhāva eva teşām, vyangya-vyañjaka-bhāvo nāparaķ kaścit). One of the alleged reasons for this assumption is that the Dhvani school itself admits suggestivity as depending upon the intention of the speaker, which intention is always an object of logical conclusion. Anandavardhana, however, demonstrates that this does not affect the general position of his school. He shews that words have two different aspects, the one inferable (anumeya) and the other communicable (pratipādya). The first, consisting of intention (vivakşā), may either be the wish to utter a sound or the wish to express an idea by a word ; the former, being a common characteristic of all animals, does not come within the sphere of speech. The communicable is something different from this, and consists of the idea itself which forms the object of the speaker's need of communication (pratipādyas tu prayoktur artha-prati- pādana-samīhā-vişayīkrtaḥ). It may be either expressed (vācya) or suggested (vyangya) ; for the speaker sometimes. wishes to communicate the idea directly by its Denotation, or sometimes he wishes to do so in such a way that it is not conveyed directly in words. This last-named inner content, Anandavardhana maintains, cannot be recognised in the form of a syllogistic conclusion, but can be by some other artificial or natural relation; for words, in the form of a logical middle term, can convey that an unexpressed idea is the object of intention, but cannot convey the unexpressed idea itself (vivakşā-vişayatvam hi tasyārthasya śabdair lingatayā prafiyate, na tu svarūpam). If the contrary is maintained, then, as every idea could be logically established, there would be no dispute about the correctness or falsity of an idea, any

THE DHVANIKÄRA AND ANANDAVARDHANA 157

more than about any other conclusion from a logical syllogism (yadi hi lingatayā śabdānām vyavahāraḥ syāt, tac-chabdārthe samyan-mithyātvādi-vivādā na pravarteran). It is only when the unexpressed takes the form of the intention of the speaker that it may be a matter of ordinary inference; but the inner content of the idea itself, when unexpressed, can be com- municated only by the supposition of another power like Suggestion ; for the natural mode of direct expression, as well as inference, is out of the question, With the establishment, against such hostile views, of the suggested sense and the function of Suggestion in poetry, which is variously termed vyañjanā (revealing), dhvanana (echoing), gamana (implication) or pratyayana (acquainting), we are introduced to the special doctrine of the system. The un- expressed or the suggested sense (vyangya artha), to which the name dhvani is appled when it is predominant, is definitely posed as the 'soul' or essence of poetry17, and poetry is classi- fied into three kinds in relation to the suggested sense. The best kind, specifically called dhvani-kāvya, is supposed to be that in which the suggested sense predominates and supersedes the expressed. It is thus defined by the Dhvanikāra (i. 13):

17 But the verse i. 2, in which this view is set forth appears, when literally taken, to state that "the sense which is praised by men of taste and which has been established as the soul of poesy, has two subdivi- sions, viz. vacya or the expressed, and pratiyamana or the suggested", implying thereby that the artha itself is the 'soul' or essence of poetry and that it includes the vacya as well, as one of its varieties. The Dhvanikāra, therefore, apparently declares that the expressed sense is also the essence of poetry, although this, as Viśvanātha objects, is opposed to his own statement in the first line of his work, which speaks of the suggested sense alone as the essence of poetry in accordance with the tradition of ancient thinkers. Abhinavagupta tries to reconcile these two apparently conflicting dicta by supposing that the real object of the Dhvanikāra in i. 2 is to distinguish between the väcya and the prafiyamāna sense, and not to establish both as the 'soul' of poetry. The objection is really over-fastidious ; for it can be easily shown that in the elaboration of the theory, the suggested sense alone is throughout taken as the ātman.

158 SANSKRIT POETICS

"The learned call that particular kind of poetry dhvani in which the (expressed) word and sense, subordinating them- selves, manifest that (other suggested) sense"18. This is par excellence Suggestive Poetry, and therefore pointedly called dhvani19. The second class of poetry, in which the suggested sense is not predominant but subordinate, is called guņībhūta- vyangya kāvya or Poetry of Subordinated Suggestion20. This Subordination consists in the suggested sense being either of equal or inferior prominence. It has been classified elaborate- ly, if not logically, on the hint furnished by the Dhvanyāloka itself, into eight varieties, according as the suggested sense is. (1) ancillary, (2) hinted by tone or gesture, (3) subservient to the completion of the expressed sense, (4) of doubtful promi- nence, (5) of equal prominence, (6) obscure, (7) unconcealed, or (8) not charming. That poetry, which is without any suggested element, is reckoned as the third and lowest kind, being merely 'pictorial in word' or 'pictorial in sense', and is. called citra or Pictorial Poetry21. In it could be included all verse which, on account of sound or magnificence of pictorial representation, or some such mechanical means, flatter the ear and is considered worthy of admiration. Under it also comes the whole body of expressed poetic figures (alamkāra- nibandho yah sa citra-vişayo matah, cited p. 221), which, containing no suggestive element, appeal by their turns of

18 yatrarthah śabdo vā tam artham upasarjanīkṛta-svārthau/ vyanktah, kavya-visesah sa dhvanir iti suribhih kathitah//. Here tam artham refers to artha defined in one of the pravious verses, e.g. in i. 4. 19 For the etymology of the word, see above footnote no. 14. 20 Dhvanyāloka, iii. 35. 21 Ānandavardhana describes citra-kāvya thus : rasa-bhāvādi- tātparya-rahitam vyangyārtha-visesa-prakāśana-sakti-sūnyam ca kāvyam kevala-vācya-vācaka-vaicitrya-mātrāśrayeņopanibaddham ālekhya-pra- khyam yad avabhāsate tac citram (p. 220). Abhinavagupta derives the word in different ways: vismayakrd-vrttādi-vaśāt ... kāvyānukāritvād vā citram, ālekhya-mātratvād vā, kalā-mātratvād vā (p. 34). In deference to Anandavardhana, Mammata speaks of citra as the third and lowest. kind of poetry; but Viśvanātha altogether rejects its claim as poetry.

THE DHVANIKÄRA AND ANANDAVARDHANA 159.

expression alone (vaicitrya) and which are characterised by Ānandavardhana as mere vāg-vikalpa. Ānandavardhana makes it clear that the citra-kāvya is not really fit to be called poetry, it is an imitation or copy thereof (kāvyānukāraḥ); for, strictly speaking, there can be no poetry in which there is no suggestion. It is admitted by him, however, to the category of poetry, because the poets, who are unfettered in their mode of expression, have, as a matter of fact, been found producing poetry of this kind, in which there is no intention of developing a suggested sense, but which is wholly taken up with the object of bringing about a strikingness of sound and sense. These three types of poetry are then elaborated and classified with somewhat minute and subtle ingenuity. In this treatment, we find the characteristic passion for reducing: everything to a formula and the scholastic delight in indulging in fastidious refinements; but at the same time there is a sincere effort to do justice to all the aesthetic facts, so far as. they have been recognised, and to unify the various currents of ideas obtaining in different schools by synthesising them with the central principle of suggestion in poetry. We see throughout the speculations of this school an anxiety to protect itself from the reproach of being too theoretical, of ignoring or doing violence to facts ; and this anxiety made the theorists evolve a scheme which should not overlook the inherited stock of notions but find a place for them in a comprehensive system. We need not take the Dhvani-theory here in all its minute details, and go through the five thousand, three hundred and fifty-five subdivisions22 of suggestive poetry, the object of which was possibly to mark out not

22 Viśvanātha gives this number. Vidyānātha in his Pratāparudrīya gives 1326 as the number of Suddha varieties of Dhvani, which with Miśra varieties mentioned by him, comes up to a total of 5304. Abhi- navagugta works out the possible number as 7420, and indicates that this number will increase infinitely if infinite varieties of Alamkāra are taken into account.

160 SANSKRIT POETICS

distinct classes, but distinct properties or circumstances. But we shall attempt to trace here briefly the effort made by this school to dispose of the already accumulated matter of Poetics, represented by the notions of rasa, rīti, guna, dosa and alamkāra, into the dhvani-system itself, by means of different arrangements or classification of the idea of suggestion. The true poetry, the dhvani-kāvya, is divided into two broad classes, viz., avivakşita-vācya and vivakşitānyapara- vācya, which two designations, clumsy as they are23, respec- tively indicate their nature. In the first case, the expressed ' sense is not meant ; in the second case, it is certainly meant but ultimately amounts to something else, viz., the un- expressed, The first is obviously based on lakşaņā or Indication, which the poet employs with the conscious purpose of bringing the unexpressed into comprehension ; and the question involved is about words and expressions which are taken not in a literal but in a transferred sense. This poetic transference, as we have already noted before, is at the root of metaphorical expression generally, the importance of which both the Alamkāra and Rīti schools amply recognised and industriously examined, and which Dandin specifically included in the samadhi-guna, and Vämana treated under the special figure vakrokti. As such, therefore, it could not be very well ignored, and by including it, as the Dhvani-theorists did, in one of the principal divisions of good poetry, they rightly assigned to it a prominent place in the new system. The second division of suggestive poetry, the vivakşitānya- para-vācya, in which the expressed is meant but is made to resolve itself into the unexpressed, is obviously based on abhidha or Denotation, and embraces the more important matter of Rasa, which has already been worked out by the

23 Mahimabhatta criticises both these terms, holding that the former is nothing more than a case of bhakti or laksana, and the latter contains an inherent contradiction (i.e., if a thing is vivakșita or pradhāna, it cannot be anyapara).

THE DHVANIKĀRA AND ĀNANDAVARDHANA 161

Rasa school in the sphere of the drama. Two possible cases of this division are enumerated, viz. (1) that in which the suggested is of imperceptible process (avamlaksya-krama), i.e., where the expressed denotation brings the suggested sense imperceptibly into consciousness, and (2) that in which the suggested is of perceptible process (samlaksya-krama). Under the first group comes the suggestion of rasa and bhāva, for it is made clear that these emotional states can be suggested only in this way. Under the second group are included the suggestions of matter (vastu) and of figure (alamkāra) by matter and figure in turns, based respectively on the power of word, or its sense, or both. Thus, the unexpressed, which is raised to comprehension by the suggestive power of a word, or its sense or both, can be an unexpressed fact or matter, an unexpressed imaginative mood which may be put into the shape of a poetic figure ; but in most cases-and these cases are of primary importance in poetry-it is an unexpressed emotional mood (rasa) or feeling (bhava), which is directly inexpressible, but which can only be suggested by an expressive word or its sense. We have already seen24 that the poet can at best directly express the three factors which bring about the Rasa, viz., the vibhāva, the anubhāva, and the vyabhicāri-bhava, but not the Rasa itself as a mood which is inexpressible in its nature. At the most, we can give a name to it, e.g. we can call it love, sorrow or anger, but the mere naming of the Rasa in poetry is not capable of awakening the mood itself in the reader which consists of a self-manifested state of the mind. Therefore, with the denotation or description of these factors, the poet can only suggest the Rasa ; in other words, he can call up a reflection of the mood which the reader realises as a particular condi- tion of his own mind25. The expressed factors, the vibhāvas

24 See above ch. iv, p. 130. 25 Abhinava explains (see above ch. iv, p. 132f) that the reader realises the feeling depicted because the artistic creations are generalised, and in this generalised form the reader realises them as his own, through 34

162 SANSKRIT POETICS

etc., are thus the suggestor or vyañjaka of the Rasa, which is the suggested or vyangya. The suggested, no doubt, depends for its manifestation on the expressed (vācyārthāpeksa), which consists of a denotation of the factors which suggest it, but it is in no way produced from it as an effect and differs entirely in essence. This suggestion is said to be 'of an imperceptible process', because the perception of the suggested Rasa by means of the various factors necessarily involves a process, but from its quickness the process is not perceived, like the process, as one writer graphically puts it, of the apparently simultaneous piercing of a hundred lotus- leaves placed one upon another. At the moment of relishing a poetic mood or feeling we are so absorbed in it that we do not perceive the process which suggests it, and this subtle suggestion may fittingly be described as one of 'imperceptible process'. By the side of the dhvani-kävya, the true poetry, in which the suggested sense is predominant, we have poetry of second- rate excellence, designated guņībhūta-vyangya kāvya, in which the unexpressed plays a subordinate part, in so far as it serves to emphasise or embellish the expressed. Here was an opportunity of including some of the results of earlier investigations of the Alamkāra and Rīti schools, which indirectly recognised a suggested sense but comprehended it, consciously or unconsciously, in some expressed poetic figures. Thus, in samāsokti was admitted the apprehension of a suggested matter, in dipaka of another suggested figure, in rasavat of suggested Rasa. But in all these cases the express- ed sense is meant to predominate and constitute the charm of the particular figure, the suggested sense being there only to emphasise or embellish it. Thus, in the much discussed figure rasavat, which was recognised by old Poetics and which helped to smuggle in, as it were, the idea of Rasa into their systems,

a certain community of human feelings, and because the germs of the feelings already remain in a latent form in his mind.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 163

the moods and feelings were supposed to have been roused, not for their own sake, but only to embellish the expressed thought. But this was not doing full justice to the fundamental significance of Rasa, and the point was bound to be re- examined. The Dhvani-theorists did not reject but justified this kind of poetry, in which the Rasa is suggested not direct- ly but secondarily, and included it in their second class of poetry. The other important case of this kind, known to earlier writers, in which something remains unexpressed but is understood, occurs in very many poetic figures which depend for their charm upon another analogous figure in- volved in themselves. Thus, Vamana thought that the upamā or comparison was involved in all figures, and Bhāmaha stated (in which Dandin substantially concurred) that all figures, in order to be charming, presupposed an atiśayokti, which he took as being involved necessarily in what he calls vakrokti (in the sense of a poetic figure). Udbhata assigned an apparently similar function to śleşa involved in some figures. Since the upamā, atiśayokti. and śleşa26 are them- selves independent figures, they can be involved in other figures as something unexpressed or suggested by the latter. But as the expressed figure is here in each case prominent, and the unexpressed merely helps to bring out its charm, these cases, in the opinion of the Dhvani-theorists, may also be fittingly relegated to this second class of poetry. In the third class are included those cases where there is no borrowed charm of a suggested sense at all, and where the appeal consists in some striking mode of direct expression, as in those figures of speech, for example, which delight us by their turns of expression alone. Thus, the suggested sense, or the unexpressed, has three different aspects; it may either be (1) a matter or an idea (vastu-dhvani), (2) a poetic figure (alamkāra-dhvani), or (3) a 26 Udbhata, however, is said to have held that when the ślesa is. involved in another figure, it predominates and dispels the apprehension of the figure itself.

164 SANSKRIT POETICS

mood or feeling (rasa-dhvani). The first occurs when a distinct subject or thought (a matter of fact) is suggested ; the second, where the suggested sense constitutes something imaginative (not a matter of fact) which, if expressed in so many words, would assume the form of a poetic figure; and the last, where a mood or feeling, which is directly inexpres- sible but which can be suggested, is the principal element. The Dhvani-theory, therefore, comprehends three kinds of poetry which deal with the communication of a fact (or a thought), or of an imaginative, or of an emotional mood. Abhinavagupta points out27 that this doctrine is not expressly taught in the Kārikas, but is clear from Anandavardhana's treatment in his Vrtti28. It appears, however, that both the Dhvanikara and Anandavardhana put a special stress upon rasa-dhvani ; and in spite of the fact that the citra-kavya or the lowest class of poetry is entirely devoid of it, it seems to afford the most weighty criterion by which a poem is to be judged. In a complete scheme, no doubt, the alamkāra-dhvani and vastu- dhvani, tacitly recognised by older writers and practised by the poets, must also be justified ; but the central question, which is carefully examined, is as to how a composition should help the Rasa to expression, for it is repeatedly laid down that neither the alamkāra nor mere narrative (p. 148) but the suggestion of rasa should be the guiding principle of the poet in his composition of word and sense29. In other words, the rasa appears to be the centre of gravity towards

27 yas tu vyācaste-'vyangyānām vastavlamkāra-rasānām mukhena' iti, sa evam prastavyaḥ-etat tāvat tri-bhedatvam na kārikākāreņa kṛtam, vṛttikāreņa tu daršitam, Locana p. 123. 28 e. g. sa hy artho vācya-sāmarthyākşiptam vastu-mātram alamkārā rasādayaś cety aneka-prakāra-prabheda-prabhinno darśayişyate, p. 15. 29 ayam eva hi mahākaver mukhyo vyāpāro yad rasādīn eva mukhyatayā kavyārthīkrtya tad-vaktyanuguņatvena śabdānām arthānām copanibandhanam, p. 181; paripākavatām kavīnām rasādi-tātparya- virahe vyāpāra eva na śobhate, p. 221.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 165

which everything else in a poem-rīti, guņa, doșa and alamkāra-should move; and stress coming to be laid on emotion in poetry, the suggestion of Rasa came to prevail over other kinds of suggestion. No doubt, it is laid down in ii. 7 that the unexpressed, apparently in all its three forms, is the angin or the principal element, and the Gunas and Alamkāras are to be esteemed in so far as they rest upon it. But this all-important angin is explained by Anandavardhana practically with special reference to Rasa (rasādi-lakşaņam)30, and the Dhvanikāra himself elesewhere discusses the merits of diction and the adjustment of words, letters and sentences with regard to their capacity of awakening the Rasa, a theme from which a theory of aucitya or propriety was evolved. Again, the Dhvanikara lays down that the gunībhūta-vyangya class of poetry can become true poetry (dhvani-kāvya) from the consideration of its tendency, if any, of developing a Rasa (iii. 41). In several places, Anandavardhana is so much carried away by his enthusiasm for Rasa that he goes almost near stating expressly that the Rasa is in fact the essence of poetry, as it is of the drama31. This borrowing from the Rasa-system-for the idea of Rasa, as Anandavardhana himself says, was already well established in the drama by Bharata and others-fills the outlines of the Dhvani-theory with a fundamentally important aesthetic content, which was not yet fully recognised in the poetic art as it had been in the dramatic. And, in this sense, the Dhvani-theory has been characterised as an extension of the Rasa-theory. But in reality it was not an extension so much as a rearrangement ; for the Dhvani-theorists accept

30 The term rasādi should be interpreted to mean the rasa, bhāva etc. as the angin; but the word adi might in every case be taken to imply strictly the other two kinds of vyangya artha, viz. vastu and alamkāra, which would be as much of an angin as the rasa, although such an interpretation is doubtful from the context. 31 rasadāyo hi dvayor api tayoḥ (=kāvya-nātyayoh) jīvabhūtāh, p. 182. See also the citations in fn 29 above.

166 SANSKRIT POETICS

the Rasa (despite the emphasis they put upon it) as only one of the aspects of the unexpressed in poetry. Neither the Dhvanikāra nor Ānandavardhana could, at least from the standpoint of theoretic consistency, explicitly make the sug- gestion of Rasa the exclusive end of poetry, inasmuch as the unexpressed may in some cases be a matter or an imaginative mood, although it can be shewn that their views practically tend to such a proposition and probably inspire later theorists to work out the thesis that the Rasa alone is the essence of poetry. The essentiality thus implicitly, if not explicitly, ascribed to Rasa by the formulators of the Dhvani-theory, is, however, expressed more definitely by Abhinavagupta, who appears to have attached little weight to mere theoreti- cal considerations. The point will be dealt with later; it will suffice here to indicate that Abhinavagupta in many places expresses himself unambiguously that the Rasa is in fact the essence of poetry; and, admitting that the unex- pressed may also take the form of Vastu or Alamkāra, he thinks that these two forms of suggestion terminate ul- timately in the suggestion of Rasa32. We shall see that this opinion probably inspired the somewhat extreme theory of Viśvanātha that the Rasa alone constitutes the essence of poetry ; but the considerations, which had wisely restrained the authors of the Dhvanyaloka from expressing it in clear terms, could not, as Jagannātha's criticism of Viśvanātha's view shews, be easily put out of the way, and recognition was refused to any further development of the theory out of itself. The Dhvanikara's idea was probably to make his con- ception of poetry wide enough to cover those varieties of

32 rasa eva vastuta ātmā, vastvalamkāra-dhvanī tu sarvathā rasam prati paryavasyete, p. 27. An almost similar view is expressed in his comment on the word ucita in Anandavardhana's exposition of the Dhvanikāra's remark on the essence of poetry: ucita-sabdena rasa- vişayam eva aucityam bhavafiti darsayan rasa-dhvaner jīvitatvam sūcayati, p. 13.

THE DHVANIKARA AND ĀNANDAVARDHANA 167

poetry which develop no Rasa, or, which develop it im- perfectly, although his real leaning to Rasa possibly betrayed itself in a different end, from which theorists like Viśvanatha drew the inevitable logical conclusion. Nevertheless, we have here an honest attempt to do justice to facts; not only to set forth what poetry ought to be but to establish the actual facts of poetry as they appeared to these theorists. They could not ignore the fact that the matter (vastu) or the imagination (alamkāra) played an important part in some kinds of poetry, although they were alive to the consideration that the emotion (rasa) was in most cases the important criterion. This attitude towards empirical analysis is also exemplified by the anxiety which made them never spare themselves the trouble of going so far as to classify the cases of the unexpressed into more than five thousand different aspects, taking into consideration all conceivable facts and circumstances, which can be made out by a careful analysis of the forms of poetic speech. This fidelity to facts did not also allow them to ignore the aesthetic ideas of earlier spe- culation ; for though these were found insufficient for explain- ing the whole problem, the concepts of rasa, rīti, guna, dosa and alamkāra had to be examined and their place properly defined in the new system before it could be established as a complete scheme. One of the triumphs of this school was, no doubt, the admission of the old idea of Rasa to its full importance in the art of poetry, as in the cognate art of the drama ; but the school did not forget at the same time to harmonise the other important elements into its comprehen- sive theory. The justification of the Riti is shown to consist in its relation to the suggestion of Rasa, and it is recognised in so far as it serves as a means to that end. The Dhvani-theorists, however, dispense with the somewhat useless classification of the varieties of Rīti (iii. 52, Vrtti)33, the nature of which 33 The numbering of this verse is wrong in the printed text: it should have been iii. 47. It is correctly given in the 4th ed. (1935).

168 SANSKRIT POETICS

is not discussed by Anandavardhana, but which, Abhinava points out, is explained by the position assigned to the Gunas (rīter hi guņeșveva prayavasāyitā)34. The function of the Gunas is justified only by their part in the development of the Rasa in the theme ; and from this standpoint, as we shall see presently, their minute classification is needless, Anandavar- dhana admitting only three Gunas corresponding roughly

34 Abhinava says (p. 231): yad āha-'viśeșo guņātmā' (Vāmana i. 2. 3) guņāś ca rasa-paryavasāyina eveti hy uktam prāg guņa-nirūpane 'śrngāra eva madhurah' (Dhva. ii. 8, p. 79) ity atreti. Vāmana has laid down that the riti is nothing more than a particular arrangement of words (visista-pada-racanā) and that the essence of this particularity of arrangement consists in the gunas. The nature and scope of the gunas, therefore, determine those of the riti. Now the Dhvanikara has pointed out in ii. 8f how the three Guņas, viz. mādhurya (in śrngāra), ojas (in raudra) and prasāda (in all the rasas) contribute to the development of the Rasas; and his remarks regarding the Gunas apply to the Rīti, which need not be taken separately. Roughly speaking, his three Guņas correspond, therefore, to the three Rītis of Vamana. Anandavardhana speaks of the Gunas as having samghatanā-dharmatva (p. 5), but this is probably only giving an exposition of the view of Udbhata who, accord- ing to Abhinavagupta (p. 134), had held that the Guņas are samghatanā- dharmāh. He might mean, as Mammata does, that particular combina- tions of letters or compounding of words produce particular Rasas. (see ii. 8f). No doubt, in iii. 5f, both the Dhvanikāra and Ānanda- vardhana speak of samghatanā in connexion with the Gunas; but they define samghatanā as depending on the length or shortness of compounds (which would correspond to the definition of Rīti given by Rudrata). The appropriateness of the samghatana depends on the ultimate object of manifesting the Rasa, as well as on the speaker and the theme. The question, therefore, resolves itself into a theory of suitability or propriety (aucitya) with regard to the disposition of words, letters and sentences, having a special reference to the Rasa (rasa-niyama), as well as to the theme in hand (vācya- or vişaya-niyama) and to the temper and character of the speaker (vaktr-niyama). Anandavardhana expressly lays down (p. 135) that the Gunas are not equivalent to samghatanā (na gunāh samghatanā-svarūpāh), nor do they depend on samghațanā (na ca samghatanāśrayā gunāh) ; on the other hand, the samghatanā depends on the Gunas. See S. K. De, Some Problems, pp. 91-94.

THE DHVANIKĀRA AND ĀNANDAVARDHANA 169

to the three Ritis of Vamana35. The relation of the Gunas to the Rasa is further made clear by drawing a sharp line of distinction between them and the Alamkaras, which also serve to embellish poetic expression. Expanding the dictum of Anandavardhana in his Vrtti on ii. 7, the later writers explain that the Gunas are the inseparable attributes of the Rasa36 without which they cannot exist, and are defined in terms of their having rasa-dharmatva, rasāvyabhicāri-sthititva and rasopakārakatva. If we sometimes speak of them as belonging to a word and its sense, it is said in a secondary or figurative way (upacara), and the old distinction between śabda-guna and artha-guna must be regarded in this light. The Alamkaras, on the other hand, belong essentially to sabda and artha, and through these means indirectly embellish the Rasa. Mammata describes their nature thus: "Poetic figures, like alliteration, simile and the rest, are those which sometimes help the existing (rasa), through the parts or members (i.e. śabda and artha), just as a necklace and the like (do to the human soul)"37. The Vrtti explains38 : "Poetic figures are those which help the principal existing rasa, through the excellence of the parts, consisting of the expressor

35 A similar function is assigned to the vrttis recognised by Udbhața. See p. 142, and also Abhinava's remarks on pp. 5-6. 36 Anandavardhana says (ii. 7 Vrtti): "The gunas depend on that sense which is the principal existing content (angin) in the form of rasa etc. Those, again, which rest upon the parts or members (anga), namely the expressed word and sense, are considered to be alamkāras. The former may be compared to qualities like bravery, and the latter to ornaments like bracelet". Mammata uses the term acala-sthiti (interpreted by Govinda as aprthak-sthiti) to indicate the relation of Guņa to Rasa. 37 upakurvanti tam santam ye'nga-dvāreņa jātucit | hārādivad alam- kārās te'nuprāsopamādayah / / 38 ye vācya-vācaka-lakşaņāngātiśaya-mukhena mukhyam rasam sambhavinam upakurvanți te kaņthādyangānām utkarşādhāna-dvāreņa śarīriņo'py upakārakā hārādaya ivālamkārāḥ. Yatra nāsti raso tatrokti- vaicitrya-mātra-paryavasāyinah, kvacit tu santam api nopakurvanti.

170 SANSKRIT POETICS

(word) and the expressed (sense), just as a necklace and the like add to the excellence of the soul through the adornment of a part of the body like the neck. Where there is no rasa, these end in mere strikingness of expression (vaicitrya) ; and sometimes when the rasa is existing, they do not help it". The Alamkäras, therefore, have only an indirect relation to the Rasa through their capacity of embellishing the expressed śabda and artha, and add to its excellence only secondarily. They can exist without the Rasa in the form of mere striking- ness of expression (ukti-vaicitrya) ; and even when the Rasa is present, the poetic figures are not invariably necessary39. As to how the Alamkara may sometimes help the Rasa, the question is discussed by the Dhvanikāra in ii. 19-20, and four possible circumstances are said to occur (1) when the poet, not dealing with it as the main point, intends its subordination to the main theme, e.g. the Rasa (tatparatvena, nāngitvena), (2) when he accepts or rejects it as suiting the occasion (kāle graha-tyāgayoh), (3) when he does not want to carry it out effectively to the end (nāti-nirvāhe), and (4) when accomplished effectively, it is still made subservient (nirvāhe'- pyangatve)40. The comparatively subsidiary position thus assigned to the Alamkāra41 must not, however, be taken to indicate any

39 This is explained by the following commentary: gunā rasam vinā nāvatisthante: guņā rasam avaśyam upakurvanti alamkārās tvavaśyam nopakurvanti ; guņā rasa-dharmā atah sākşād rase tișthanti, alamkārās tu na rase sākşāt tisthanti kiņ tu tvanga-dvāreņa. 40 Cf Hemacandra p. 17. 41 In Mammata's much criticised definition of poetry, therefore, the Alamkara is taken as an accident, not as an essential ; and though technically the phrase analamkrtī punah kvāpi is open to the objections brought forward by Viśvanātha and Jagannātha, the views of the latter on the point under discussion does not differ substantially from those of Mammata. In Mammata's definition there is no direct mention of vākyārthībhūta rasa or of the vyangya sense other than the Rasa (which are there by implication), but the Gunas and Dosas are expressly men- tioned. The explanation of these peculiarities of the definition must

THE DHVANIKARA AND ANANDAVARDHANA 171

tendency to minimise its importance, for Anandavardhana himself admits that poetry depends on it for its operation (kāvya-vṛttes tadāśrayāt). But the Alaņkāra is accepted only in connexion with the angin or the principal element in poetry, which in most cases takes the form of Rasa; and Alamkāras, other than such, which are devoid of or unconnected with the suggestion of Rasa and therefore un- poetic, are in Anandavardhana's opinion, mere vāg-vikalpas, and should be included in the citra-kavya, which is no poetry but an imitation thereof. The authors of the Dhvanyāloka ignored these because their system had no place for them ; but the poet may sometimes intend not to awaken Rasa or anything else unexpressed, but to produce mere strikingness of expression in the form of a poetic figure. Such cases, therefore, should be acknowledged and analysed. We shall see that followers of the Dhvani-system like Ruyyaka realised this deficiency in the treatment of the Dhvanikāra and tried to supply it by admitting the significance of such figures for poetry and analysing their content after the indication given by Kuntaka. The view indicated above regarding the nature of the Gunas necessarily dispenses with their endless multiplication and differentiation. Mammata and his followers, accepting the standpoint of the Dhvanyāloka in this respect, admit only three Guņas, viz., mādhurya (sweetness), ojas (energy) and prasāda (lucidity), out of the ten recognised since Bharata's time. They shew elaborately that these ten are either included in the three mentioned above, or else constitute mere absence of defects, while some of them are even positive defects. In fact, these three Gunas are defined broadly enough to include most of the ten Gunas of Bharata, Dandin and Vamana. Thus, the madhurya, found chiefly in the Erotic, the Pathetic and the Quietistic moods, is described

be sought in the historical development of these ideas in the earlier schools, and not in any attempt to invent an original definition. See below ch. vii.

172 SANSKRIT POETICS

generally as that excellence which brings delight (ahlāda) to the mind and makes it melt, as it were (druti-kāraņa); the ojas, arising in the Heroic, the Furious and the Disgustful moods, is that property by which the mind is brilliantly expanded (vistāra-kāraņa) ; while the prasāda, found in all poetic moods, causes them to pervade the mind (vyāpti- kāraņa), like fire pervading dry fuel, or water pervading a pure piece of cloth. As they are related to the main poetic mood Rasa in the composition and made suitable to its particular kind, the classification, as given here, naturally proceeds on a psychological basis having reference to their influence on the reader's mind (so as to lead up to the particular mood), and supersedes the old differentiation resting on an adjustment of sound and sense. It will be also seen from the somewhat comprehensive definitions of the three Guņas that the śleșa, samādhi and audārya of older writers may be included in ojas, and the artha-vyakti in prasāda ; while saukumārya and kānti are essentially the opposites of the defects of harshness (pāruşya) and vulgarity (grāmyatva) respectively, and samatā or uniformity of diction may sometimes be a positive defect. Consistently with this view of the Gunas, the Dosas or defects of a composition are recognised in so far as they are the repressors of the Rasa, as well as of the expressed sense. The Dosas, therefore, convey a positive significance, like the Gunas, in relation to the Rasa, in spite of the admitted fact that some Dosas approach guņābhāva (negation of Guņas) and some Guņas approach doşābhāva (negation of Doșas). The punarukta or tautology, for instance, is generally a fault, but it may sometimes be an excellence if there is an apprehension of the charm of the suggested Rasa through it. The justifi- cation of the distinction between invariable (nitya) and non- invariable (anitya) fault lies in the fact that in the case of some poetic moods, we can generalise the avoidance of par- ticular combinations as being always damaging for the effect. Thus, the Dhvanikara points out that when love or śrngāra

THE DHVANIKĀRA AND ÃNANDAVARDHANA 173

is the principal suggested mood, one should always avoid faults like unmelodiousness (śruti-dusta), although it is not a fault in the case of the Heroic or raudra-rasa. The attempt, therefore, to estimate the worth of a poem by analysing two kinds of meaning the one explicit and the other implicit, and judging it by a reference to the latter rather than to the former, explains in a new light the nature and function of the Gunas and Dosas, as well as of the Alamkaras which were admitted by previous speculation, but over which there had been so much controversy. The explicit, or express- ed word and sense, in which poetry is clothed constitutes its mere vesture, but this external or accidental feature alone appealed to earlier thinkers, whose attention was practically confined to the expressed sabda and artha. The Gunas and Doşas (along with the so-called Rīti), as well as the Alam- kāras, are only certain forms of these, being merely turns given to śabda and artha in expression, and are justified as such. They cannot, therefore, be taken as essential, for they do not touch the essence of poetry which consists of the im- plicit or unexpressed meaning. But at the same time, they cannot be ignored because they are the means by which the unexpressed is suggested, the expressed word and sense being the vyañjaka of the deeper vyangya sense. In classify- ing the implicit or the unexpressed, again, into communi- cation of a fact (vastu-dhvani), or suggestion of an ima- ginative mood (alamkāra-dhvani), or manifestation of an emotional state (rasa-dhvani), the theorists recognised the truth that the essence of poetry may consist of fact, imagina- tion or feeling as the predominant implicit factor, the out- ward expression being important as a means of pointing to this implicit significance. But it is also perceived that the emotional mood, which the poet succeeds in communicating to us, is of the highest importance in poetry ; and stress came to be laid on this emotional mood to the extent even of ignor- ing the imaginative or the realistic, and poetry came to have a deeper significance as a means of emotional realisa-

174 SANSKRIT POETICS

tion. This the Dhvani-theorists did by emphasising the rasa- dhvani in poetry. This, in brief, is an outline of the new system which at- tempts to take into consideration all the known facts and dogmas and build a compact theory of poetry on their basis. But its chief merit consists in its elaboration of the most necessary and fundamental principle of all higher poetry, viz ,. the art of suggestion, which should lead the reader through diverse routes from that which is distinctly expressed to that which is left unexpressed. With the arrival at this point, one. discovers the real significance of a poem and appreciates the taste or relish of the underlying poetic sentiment, which is in reality inexpressible. The ornamental fitting out of thought or word, as well as the literary excellences of structure or style, everything contributes towards this end, In this con- nexion, we must not mistake this suggestion to be a form of quiet hinting, or of absolute silence, such as we find in some modern poetic mystics, or that particular train of thought which holds that all things have their being in the unexpressed and resolve themselves into the indeterminable. Sanskrit poetry does not aim at leaving the unexpressed to be darkly gathered, nor does the theory of Poetics regard it as indeterminate. The unexpressed is bound up by means. of definite links with the expressed, without which it cannot exist; but it is wrapped up in such a manner as to make it possible only for the initiated in the poetic. hieroglyphics to comprehend it in its subtlety. The unexpressed is not understood by those who know grammar and lexicon, but only by men of taste and literary instinct who know the essence of poetry. It is the province of the sahrdaya, the connoisseur, who is expert in discerning through the intricate meshes of veiled word and sense into the aesthe- tic relish of deeper significance, in which the pleasure of the beautiful is mixed up with the pleasure arising from the fineness of the problem itself. This general scheme of Poetics outlined by the Dhvani

THE DHVANIKĀRA AND ĀNANDAVARDHANA 175

school, in spite of the loopholes that may be detected in the doctrinal edifice, is accepted as canonical by all important writers coming after Anandavardhana. Here and there an isolated theorist arose who dared to question the general creed, but he was at once put down as a heretic and condemn- ed to neglect and oblivion. The immediately following systems of the Vakroktijīvita-kāra and the Vyaktiviveka-kāra were, in spite of their able and ingenious efforts, unable to sup- plant the Dhvani-theory ; and, finding no strong adherents, themselves languished and died out. These views are taken notice of by later writers only for the purpose of refuting them. Bhatta Nayaka, judging from the long quotations from his lost work in Abhinavagupta and others, seems to have made a greater impression ; but even he does not appear to have been very successful. All these writers, no doubt, accept the concept of a suggested sense, but when they en- deavour to explain it in a different way, they could hardly find a patient hearing. Even Viśvanātha's attempt to push the theory to its logical extreme did not meet with universal. approval. The labours, therefore, of all later writers, typified by Mammata, consisted generally in working out the details of the Dhvani-theory and the scheme of Poetics standardised on its basis ; and they spent all their fine scholastic powers in refining and explaining but hardly in adding anything of abiding interest. No other work on Sanskrit Poetics has indeed exerted so much influence as the Dhvanyāloka, which brought to a focus the tentative efforts of earlier thinkers, and by its thoroughness and masterly exposition eclipsed all its predecessors, dominating, as it did, thoughts of generations of theorists even down to the present time.

CHAPTER VI

ABHINAVAGUPTA AND THE REACTIONARY SYSTEMS

( 1 )

Abhinavagupta The importance of Abhinavagupta as a writer on Sanskrit Poetics lies in his learned exposition of the Dhvani-theory in his well known commentary on the text of Anandavar- dhana ; and his erudition, reputation and influence as a great scholar and philosophical writer of his generation, no doubt, added weight to his championship of the theory, and contributed a great deal to its ultimate exclusive acceptance in later Poetics. His theoretical standpoint, however, does not differ, except in one material point which will be dealt with presently, from that of the formulators of the Dhvani- system ; and he may be fairly regarded as belonging to that group of faithful commentators who are more anxious to interpret than to incorporate new ideas into the system they comment upon. On the other hand, Abhinavagupta was also greatly interested in the dramaturgic work of Bharata and wrote an elaborate and stupendous commentary on this encyclopaedic text. From this interest in dramaturgy, we have seen1, he came to be deeply interested in the various theories about the origin and function of Rasa, not only in the drama but also in poetry; and one of the latest and most important theory on Rasa is directly associated with his name by Mammata, Hemacandra and others. In expound- ing this theory, he tried to explain clearly how the vyakti or vyañjanā of the Dhvani-theorists could be applied to the case of the manifestation of Rasa, thus correlating the Rasa-

1 See ch. iv, p. 128.

ABHINAVAGUPTA AND THE REACTIONARIES 177

doctrine with the Dhvani-theory. He defined the concept of Rasa and its place in poetic theory, and furnished a brilliant aesthetic explanation of a phenomenon which had already taxed the ingenuity of many a previous thinker on the subject. Having realised the importance of Rasa in poetry, Abhinava, however, went a step further than the Dhvanikāra and Anandavardhana in boldly setting it up as the only essence or aesthetic foundation of poetry, a view which has greatly influenced all later speculation on the subject. From the earlier drama and dramatic theory the authors of the Dhvanyāloka had admittedly worked up the idea of Rasa into poetry and poetic theory ; but as the emotional mood in poetry, which the fact of Rasa emphasises, came to be more and more prominent, the Rasa stood out more and more in relief as its essential aesthetic basis. We have seen2 that Abhinava's predecessors in the Dhvani school consider Rasa only as one of the elements of the unexpressed, which may take other forms in the shape of an unexpressed matter (vastu) or an unexpressed imaginative mood (alamkāra). No doubt, their theory puts great emphasis on the rasa-dhvani or suggestion of Rasa in poetry ; but both the Dhvanikāra and Anandavardhana are yet careful in taking into account other kinds of suggestion and do not, as they could not, erect the Rasa into the very 'soul' of poetry. No doubt, it may be thought that they show a decided partiality to rasa, which would practically lead to a conclusion of its essentiality ; but they could not, having regard to theoretical consistency give exclusive preference to it ; for in their complete scheme of Poetics the rasa-dhvani, which is only one of the three forms of the unexpressed, plays as much part as the vastu- and. alamkāra-dhvani. They had to recognise that the centre of gravity in a poem may lie in its material and its imagination, as much as in its emotional element. Abhinavagupta appears to have attached little weight to these theoretical considera- 2 See ch. v, p. 166. 35

178 SANSKRIT POETICS

tions, which had restrained his predecessors from explicitly stating what they practically implied; and brushing them aside, he carries their theory to its utmost logical consequence by declaring the essentiality of Rasa (rasenaiva sarvam jivati kāvyam), without which, in his opinion, there could be no poetry (na hi tac chūnyam, i. e. rasa-śūnyam, kāvyam kimcid asti, p. 65). He attempts, however, to explain the theoretical discrepancy by saying that the two other aspects of sugges- tion, concerned respectively with vastu and alamkāra, resolve themselves ultimately into the suggestion of rasa, which is in fact the essence of poetry (rasa eva vastuta ātmā, vastva- lamkāra-dhvanī tu sarvathā rasam prati paryavasyete, p. 27). This opinion, no doubt, influenced the view of later thinkers to a great extent ; for, although Mammata carefully follows the cautious attitude of the Dhvanikara and Anandavardhana, Viśvanātha, developing their theory (after Abhinavagupta) further out of itself, pushes it to its extreme limit and builds up his own scheme of Poetics on the basis of the theory that poetry consists of a sentence of which the 'soul' is Rasa (vākyam rasātmakam kāvyam). But we shall see that the considerations which led the Dhyanikära and his commentator to leave their view on this point wisely unstated could not be easily put out of the way, and they are repeated substantially by Jagannatha in his criticism of Viśvanātha's view. All later writers, however, agree in thinking that the rasa-dhvani is certainly the most important point for consideration in poetry ; and even if they do not explicitly state with Abhinava that the vastu- and alamkāra-dhvani resolve ultimately into rasa-dhvani, they yet show a decided partiality to the latter element. This, in brief, is the general position of Abhinavagupta as a champion of the new system established by Anandavardhana. The final dominance of this system in later speculation is due not only to the intrinsic worth of the theory itself and its. masterly formulation by Anandavardhana, but also probably to the authority which Abhinava's exposition as well as his

ABHINAVAGUPTA AND THE REACTIONARIES 179

reputation lent to it. We find in the immediate followers of the system, however, not the extreme position of Abhinava- gupta, but the theory and the scheme as finally outlined by Anandavardhana. With Anandavardhana, the Dhvani-theory, which was itself ancient, came to prevail; but with him also was evolved a more or less complete scheme of Poetics in which the divergent gleams of earlier thought and the accumulated stock of recognised ideas meet and are rationally adjusted. This scheme, with the concept of dhvani (especially rasa-dhvani) at its centre, was summed up and uttered in the concise form of a systematic text-book by Mammata, another Kashmirian, whose influence perhaps was not less potent than that of Abhinavagupta in raising it to almost exclusive authority in later times. This system, which for convenience we have called the Dhvani-system, absorbed and overshadow- ed all previous schools and systems, and came to reign supreme, only to be improved in detail by the large crowd of its followers who form the bulk of post-dhvani writers on Poetics. Jagannatha, one of the latest writers of this group, very aptly remarks, therefore, that the authors of Dhvanyāloka settled the path to be followed by later writers on Poetics (dhvanikrtām ālamkārika-saraņi-vyavasthāpakatvāt, p. 425). But it must not be supposed that the theory or system of Dhvani could obtain universal acceptance without some vigorous opposition. Before we take up the post-dhvani followers of Anandavardhana, it will be necessary to consider here some of the reactionary writers who either follow and develop other traditions of thought, or who refuse to acknow- ledge the new theory. Adherents of other schools, such as Pratīhārendurāja (pp. 79f) who commented on Udbhata, or Gopendra Tippa Bhūpāla (p. 72) who commented on Vāmana, carry on the older tradition and do not fail to criticise the new theory. Says Mukula, Pratībārendurāja's Guru: lakşaņā- mārgāvagāhitvam tu dhvaneḥ sahrdayair nūtanatayopavar- ņitasya vidyata iti ... etac ca vidvadbhiḥ kuśāgrīyayā buddhyā nirūpaņīyam, na tu jhagity evāsūyitavyam ity alam ati-

180 SANSKRIT POETICS

prasangena (p. 21). But more hostile opposition or attack came from some really thoughtful writers who urged new systems, or new explanations of the Dhvani-theory in terms of old ideas. Most of these theorists lived near enough in time to Abhinavagupta ; and coming later than the authors of the Dhvanyāloka, they accept or show themselves cognisant of the general concept of dhvani, but attempt to formulate other explanations of it. All of them, however, agree that the vyañjana vrtti need not be postulated or proved for explaining the suggested sense of poetry, and conservatively maintain that the suggested sense can be reached from the expressed sense by some of the recognised means or processes of knowledge (e. g. anumāna). None of these writers, there- fore, is what the Dhvanikāra would call an abhāva-vādin, i.e., none of them would deny the existence of Dhvani, but they would try to explain it in terms of already recognised concepts or processes. These theorists are: Bhatta Näyaka who probably preceded Abhinavagupta, Kuntaka who was probably the latter's contemporary, and Mahimabhatta who was either a younger contemporary or lived immediately after Abhinavagupta. It will also be convenient to take up in this connexion the school of opinion represented by the writer on Poetics in the Agni-purana and by Bhoja, which stands in many respects apart from the Kashmirian school of Anandavardhana and which appears to have been entirely untouched by the implications of the Dhvani-theory.

( 2 )

Bhatta Nāyaka It is unfortunate that Bhatta Nayaka's Hrdaya-darpaņa is now lost. From the citations of Abhinavagupta and others, the conjecture is likely that it was not a commentary on Bharata's Nātya-śāstra3 but an independent work written in prose and verse (i.e., with verse-kārikā and prose-vrtti) and 3 See vol. i, pp. 40f.

ABHINAVAGUPTA AND IHE REACTIONARIES 181

resembling Mahimabhatta's later Vyakti-viveka written in the same style and with the same object. Like the latter work, it was composed, if not for establishing a new theory of Poetics, at least for controverting the position of the Dhvanyāloka and formulating a different explanation of Dhvani, especially of rasa-dhvani. When Mahimabhatta later on took upon himself the task of "demolishing" the Dhvani-theory, he boasted at the outset of his elaborate attack that he had composed his Vyakti-viveka without look- ing into the Darpana4 (presumably Hrdaya-darpana, as explained by his commentator), which was therefore obviously written with the same object of dhvani-dhvamsa. No doubt, Bhatta Nāyaka was one of the four writers (mentioned by Abhinava, Mammata and others) who formulated explana- tions of Bharata's original sūtra on Rasa; but this in itself is no reason to take him as a commentator on Bharata's text5.

4 It is curious that Mahimabhatta says that in composing his own work he has not also looked into the Candrika, which was apparently an adverse commentary on the Dhvanyaloka. It is probably the same work as is referred to and criticised frequently by Abhinavagupta in his °Locana and as, he says, was composed by one of his ancestors. Abhinava's references and criticism also confirm the idea that it criti- cised the text of the Dhvanyāloka adversely on many points. This Candrikā is also apparently cited by Māņikyacandra and Someśvara in their commentaries on Mammata. See vol. i, p. 101. 5 The question has been already discussed by us in vol. i p. 40. There is a passage, already referred to by us, in the Abhinava-bhārafi ch. i, which appears (see Sovani's article on the Pre-dhvani Schools in Bhandarkar Commemoration Volume, p. 390; contra in JRAS, 1909, pp. 450-52) to indicate that the Hrdaya-darpana was a commentary on the Nātya-sastra. The passage runs thus (commenting on brahmana yad udāhrtam in Bharata i. 1): bhatta-nāyakas tu brahmaņā paramātmanā yad udāhrtam krta-nidarśanam ......... tad anena pāramārthikam prayo- janam uktam iti vyākhyānam hrdaya-darpaņe paryagrahīt. This passage is indeed important, for the relevancy of any comment on Bharata i, 1. is difficult to explain in a work which ex hypothesi is not a commentary on the text. But it appears to militate against those references to and passages from Bhatta Näyaka's work (prose as well as verse) which Abhinava cites and criticises in his °Locana (pp. 11, 12, 15, 19, 21, 27,

182 SANSKRIT POETICS

On the other hand, Abhinava's references in °Locana make it reasonably clear that the Hrdaya-darpana, like the Vyakti- viveka, had the special object of criticising in detail the text of the Dhvanyāloka as well as its theory ; and its discussion of Rasa might have come in topically in connexion with Bhatta Nayaka's general views regarding poetry and poetic expression. The question, however, cannot be definitely settled so long as we get only glimpses of Bhatta Nāyaka's views set forth in the brief exposition and adverse criticism of Abhinava and others. We have already considered at some length Bhatta Nayaka's views regarding the origin and function of Rasa in poetry6. We have seen that Bhatta Nayaka regards rasa-carvanā as the essence of poetry, but he is apparently not prepared to accept the function of vyañjanā as its means of manifestation7. It is possible that he admits a suggested sense, as he accepts

28, 29, 33, 63, 67-68) and which consist mostly of direct criticism of the text of the Dhvanyaloka. Either of two explanations is possible: (1) that the Hrdaya-darpana was in fact a commentary on Bharata's Nātya-sāstra, and Bhatta Nāyaka's criticism of the Dhvanyāloka might have constituted incidental discussions in it. But this does not explain the presence of verses in it, which later writers, including Abhinavagupta, quote from Bhatta Nayaka in their exposition of his views ; or (2) that it was an independent work in prose and verse, consisting of Bhatta Nayaka's propounding of his own views in opposition to those of the Dhvanyaloka; and the discussions of Rasa-theory and of Bharata's text are not altogether inexplicable, as they might have been topical in connexion with his general theory. This latter explanation, which we have already discussed vol. i, seems to be more likely. 6 See ch. iv, pp. 123f. 7 Bhatta Nāyaka's objection to the abhivyakti-th eory is thus sum- marised by Abhinavagupta (°Locana p. 68): "If the potentially existing śrngāra is supposed to be manifested by abhivyakti, then it would occupy its field of action in diverse degrees (vişayārjana-tāratamya-pravțttiḥ), i.e., thus contradicting the nature of Rasa as one. There would also be the difficulty mentioned before, namely, whether the Rasa is manifested as existing in oneself or in another person,"

ABHINAVAGUPTA AND THE REACTIONARIES 183

the suggestion of Rasa as the essence of poetry (rasa-dhvanis tu tenaivātmatayāngīkrtah, °Locana p.15) ; but from Abhi- nava's twitting him on this score it is probable that he denied vastu-dhvani (kim tu vastu-dhvanim dūşayatā rasa-dhvanis tad anugrāhakaḥ samarthyata iti suşthutarām dhvani-dhvam- so'yam, p. 20). Bhatta Nāyaka maintains in a verse attribut- ed to him by Abhinava (p. 27), Hemacandra (p. 4), Māņikyacandra (p. 4) and Jayaratha (p. 9) that the distinction between various kinds of literary composition lies in the fact that in the Sastra sabda predominates, artha in the Ākhyāna (=probably itihāsa), while in the Kāvya, both śabda and artha are subordinated (guņībhūta or nyagbhāvita). Elsewhere he is represented by Abhinava (p. 68) as saying that the verbal composition (śabda) which makes up poetry is different from other species of verbal composition by the fact that it possesses three elements. Of these elements, abhidha or Denotation belongs to the province of expressed meaning, bhāvakatva or power of generalisation to that of Rasa, and the bhojakatva or the power of enjoyment to the appreciating audience ; thus we have three functions attribut- ed to the three elements of poetry. If Denotation, among these, is taken by itself (i.e. without the other two), then what is the essential difference, he asks, between the poetic figures and the dogmas which form the method of Sastras? Or, if this manifold distinction of functions is without importance (metaphorically as well as intrinsically), then why avoid faults like unmelodiousness (śruti-dusta)? These considerations, Bhatta Nayaka thinks, would give us the second function, viz. bhāvakatva, by which generalisation is accomplished of poetry as well as of its factors (vibhavas). It is on account of this function that abhidha or Denotation is also Indication (lakșaņā), i.e., the Denotation can give to the expressed sense a secondary or metaphorical significance as the basis of Rasa. After the Rasa is thus generalised (bhāvita), comes its enjoy- ment or bhoga which, we have seen, Bhatta Nayaka represents, after the Sämkhya philosophers, as a process of distinterested

184 SANSKRIT POETICS

contemplation akin to the philosophic contemplation of Brahma. Bhatta Nayaka thus postulates a function of bhoga, beyond those of abhidha and bhāvakatva, inherent in poetry, in order to explain the working of Rasa. He seems to imply that the Rasa, which the Dhvani-theorists would take as the suggested emotional sense of poetry, is, in his opinion, purely sva-samvedya and therefore transcending definition. In other words, he belongs to that class of objectors to the Dhvani-theory regarding whom the Dhvanikāra says that they do not deny dhvani but think that its essence lies beyond the province of words (i. 1c). In a verse attributed Bhatta Nāyaka by Abbinava (pp. 15, 11) and Jayaratha (p. 9), he speaks of kāvyāngatva and not kāvya-rūpatā8 of what is known as dhvani; a statement which would indicate that having assumed the concept, Bhatta Nāyaka's object was to establish an explanation different from that of the Dhvani-theorists. Ruyyaka thinks that Bhatta Nāyaka would regard what is called vyangya-vyāpāra to be an element (kāvyāmśatva) and not an essence of poetry, being reached by the bold utterance of the poet (praudhokti). In this sense,. the kavi-karman or act of imagination on the part of the poet (indicated by Bhatta Nayaka, as Jayaratha points out, by the word vyāpāra), which makes śabda and artha subservient to itself, is the most important thing in poetry ; a view which approximates Bhatta Nayaka's theory to that of Kuntaka, who makes kavi-karman the source of what he calls vakrokti in poetry.

( 3 )

Kuntaka Kuntaka, anthor of the Vakrokti-jivita, on the other hand, had no direct intention of attacking or disproving the Dhvani-

8 Read in the verse kāvyāngatvam na rūpatā, as given by the reading ef MS ga indicated in °Locana p. 15.

ABHINAVAGUPTA AND THE REACTIONARIES 185

theory. He appears to have accepted the fact of a suggested sense in poetry but, following the tradition of Bhamaha's vakrokti, he develops a system of vakrokti of his own, in some aspects of which he includes all ideas of dhvani and rasa. Nearly the whole of his long lost work has been recently recovered, and a part of it (chs. i, ii and a part of iii) has been published by the present writer; it is now possible, therefore, to depend no longer on the references to Kuntaka in later literature for an account of his views, but gather it independently from his own statements10. The central idea in Kuntaka is that the vakrokti is the essence (jivita) of poetry ; and by Vakrokti he understands a certain striking or charming (vicitra) mode of expression (vinyāsa-krama), which is different from or excels the common or matter-of-fact expression of words and ideas in the Sāstras and the like (sāstrādi-prasiddha-śabdārthopanibandha- vyatireki). It is, therefore, a deviation from the established mode of speech for the purpose of attaining a certain strikingness (vaicitrya or vicchitti), or an imaginative turn of words and ideas (bhangī-bhaniti or bhaņiti-prakāra) peculiar to poetry, and abhorrent of common speech in which facts are more or less simply stated. This is the vakratva or vakra- bhava underlying all poetic speech. A distinction, therefore, is implied not only between the method of the sciences and the scriptures (sastradi), on the one hand, and that of poetry, on the other, but also between what may be called the "naturalistic" and the "artistic" mode of expression11. Kuntaka, therefore, holds that sālamkrta śabda and artha or embellished word and sense alone constitute poetry, and this embellishment consists of Vakrokti. The so-called embellishments, which go by the name of poetic figures

9 In the Calcutta Oriental Series, 2nd revised and enlarged ed. 1928. The work consisted probably of four chapters. See vol. i. p. 128 above. 10 An account of Kuntaka's theory of poetry is given in the introduction to the above edition, which see for detailed references. 11 See above ch. ii, pp. 48-49.

186 SANSKRIT POETICS

{alamkāras) in orthodox Poetics, are merely aspects of this Vakrokti, and can be properly included in its comprehensive scope. So can also the ideas of dhvani and rasa. This Vakrokti being the only possible alamkāra and being essential as such, Kuntaka finds fault with the common statement that the alamkara belongs to poetry ; for such a statement would imply that poetry may exist without it (i. 7, 11). Kuntaka then explains that the Vakrokti charms us by the skill of the poet, and is therefore called vaidagdhya-bhangi- bhaniti12. It rests ultimately on the conception (pratibha) of the poet, or on his skill (kauśala), or on an act of imagina- tion on his part, which is termed kavi-vyāpāra or kavi- karman. Kuntaka does not exactly define this kavi- vyāpāra, which is the ultimate source of poetry, perhaps because he is conscious of the fact that it is in its nature undefinable ; but he analyses it elaborately, and distinguishes and classifies its function in six different spheres, namely, in the arrangements of letters (varna), of the substantive and terminal parts of a word (pada-pūrvārdha and pada-parārdha), of a sentence (vākya), of a particular topic (prakarana), and of the composition as a whole (prabandha). He devotes

12 The word vidagdha is used in opposition to the word vidvat to signify a man versed in belles-lettres as distinguished from a scholar ; and the Dhvanyāloka often speaks of an appeal to vidagdha-vidvat- parișad (pp. 201, 239). Avantisundarī is cited in Kāvya-mīmāmsā p. 46 as saying vidagdha-bhaniti-bhangī-nivedyam vastuno rūpam na niyata- svabhavam. The vaicitrya is discussed by Anandavardhana at p. 243, in which connexion he uses the term bhaniti-krtam vaicitrya-matram. Abhinava speaks of infinite variety of upamā-vicchitti (upamā-vicchitti- prakārāņām asamkhyatvāt, °Locana p.5), and uses it also synonymously with carutva (p. 8). It would appear from the verse quoted by Ananda- vardhana at p. 130, the word vicchitti, used in this sense in poetic theories, is applied analogically from the same word used to signify a certain feminine charm or elegance derived from carelessness in dress and decoration (Bharata xxii. 16). See on this point Haricand Sastri, L'art poétiqne de l'Inde pp. 64-65. The word bhangi in the sense of a turn of expression is used in Dhva° pp. 139, 241, Etymologically it appears to have the same meaning as vicchitti.

ABHINAVAGLA AND THE REACTIONARIES 187

nearly the whole of his work, with the exception of the introductory portion of the first chapter, to the definition, classification and illustration of these varieties of kavivyāpāra- vakratā, which thus form the different categories into which poetic speech may be analysed. It is clear from this brief exposition that Kuntaka cannot admit as poetry a composition involving mere svabhāvokti, which he takes to be plain description without the requisite strikingness; and he consequently develops Bhamaha's indication that a kind of atisaya is involved in vakrokti- vaicitrya. This atisaya, if it is taken in the sense of the lokātikrānta-gocaratā of Bhāmaha's atiśayokti, would imply a kind of heightened charm of expression which is lokottara or dissociated from personal interests and relations. The dissociation, therefore, which is supposed in the artistic attitude involved in the relish of Rasa13, is also implied in Vakrokti ; and on this point Kuntaka appears to agree with the main position of the Rasa-theorists. Kuntaka also thinks that the ultimate test of this lokottara vaicitrya is tadvidāhlāda or pleasure of the appreciating sahrdaya, who plays here apparently the same part as he does in the Rasa-theory or in poetic theories generally. It seems, therefore, that the exponents of the different theories approach ultimately the same standard, albeit through different avenues of thought, and agree in holding that vaicitrya or camatkāra (in Alamkāra or Rasa) must be finally subjected to the taste of the sahrdaya. Thus a new turn was given to the Alamkāra-system of Bhamaha ; or rather, what was implicit or naïvely expressed in it was developed to its logical consequence by Kuntaka's systematic analysis of its implications.14 In spite of the obviously extreme nature of his central theory and his some-

13 See above ch, iv. 14 The Vakrokti-system of Kuntaka may properly be regarded as an off-shoot of the older Alamkāra-system (ch, 1i).

188 SANSKRIT POETICS

what quaint nomenclature, his work is of great value as presenting a unique system, or rather as systematising the Alamkāra-theory of earlier writers in a refreshingly original way. The Dhvani-theorists had either dismissed the poetic figures (alamkāras) as mere vāg-vikalpas, or considered them only as heightening the charm of the unexpressed element in poetry. They speak of the relation of the Alamkara to the principal suggested element of poetry (e.g. in the shape of Rasa) ; but there might arise cases where the poet's obvious intention is not to awaken Rasa or anything else unexpressed, but simply to produce a strikingness in the form of an expressed poetic figure. In these cases, the authors of the Dhvanyaloka think that all such figures which, in connexion with an involved unexpressed element, possess a peculiar charm, belong to the class of poetry called by the gunībhūta- vyangya; if there is no such unexpressed element involved, the figures have a pictorial effect merely, and may be included in the lowest class of poetry, called by them citra and describ- ed by them as no poetry but an imitation thereof. In other words, they take into consideration such poetic figures as being connected with the unexpressed possess a peculiar charm, and thus justify their position in poetry ; the figures unaccom- panied by the unexpressed or in no way connected with it are condemned to the level of no-poetry, as phases of speech which is of infinite variety. As Ānandavardhana says : anantā hi vāg- vikalpās tat-prakārā eva cālamkārāh. Kuntaka, on the other hand, justifies the significance of such figures in poetry as figures, and shews that this significance is independent of all considerations of their connexion with the unexpressed; for it consists in the very vaicitrya or strikingness involved in them, which is sufficient in itself, and does not borrow its power of appeal from elsewhere. But he justifies the alamkara as such only when it involves the vaicitrya, vicchitti or vakratva and becomes a phase of Vakrokti. He admits that the poetic figures are particular forms of speech, aspects of the expressed denotation

ABHINAVAGUPTA AND THE REACTIONARIES 189

{abhidhā-prakāra-viseşāh), in which there need not be any connexion with the unexpressed; but he supposes at the same time a specific differentia in them which consists in a peculiar turn of expression resulting in a characteristic strik- ingness (vaicitrya or vicchitti) and depending upon an act of imagination on the part of the poet (kavi-pratibhā- nirvartita). Thus, the so-called alamkāras of orthodox Poetics are admissible when they are found on analysis to possess these characteristics of peculiar charm imparted to them by the fertile imagination of the poet; and these, therefore, cons- titute the elements which go to make up the being of a poetic figure15. Kuntaka in this way not only supplies a remarkable deficiency in the teaching of the Dhvanikāra and Ānanda- vardhana, but also justifies the existence and fixes the conception of a poetic figure as distinguished from a mere speech-figure. It is no wonder, therefore, that later followers of the Dhvani school, who form the bulk of post-dhvani writers on Poetics, readily accept Kuntaka's analysis, and regard the two characteristics, viz. vicchitti and kavi-vyāpāra, as forming the ultimate test of a poetic figure. Mammata lays down that where there is no suggested Rasa, the poetic figures simply result in ukti-vaicitrya or charmingness of expression, and states generally that Alamkāra is charmingness itself (vaicitryam alamkārah). We shall see that Ruyyaka was the first writer who accepts Kuntaka's test of a poetic figure and applies it systematically to a detailed examination and classification of individual poetic figures. On Rīti Kuntaka puts greater stress than Bhāmaha, and gives a more elaborate classification of Gunas. He is aware of the classification of Marga or Rīti made by Dandin and

15 Hence, the word "poetic figure" or kāvyālamkāra, instead of "figure of speech." See above ch. ii. pp. 74. In a formal scheme of Poetics they no doubt correspond, but this point of view of its involving poetic charm would be entirely omitted in a treatise of rhetoric. It is, therefore, misleading to translate Sanskrit Alamkāra as Sanskrit Rhetoric. See ZDMG, lvi, 392 fn.

190 SANSKRIT POETICS

Vämana, but he does not accept it. He does not also believe that a particular Rīti is determined by Deśa-dharma (regional characteristics) or that it should be named after a particular locality ; for in that case one has to admit infinite varieties of Rīti, as there is infinite number of countries. The classification of Rītis into good (Uttama), bad (Adhama and indifferent (Madhyama) is also futile, for the best kind of mode alone is acceptable, and there is no point in admit- ting or framing rules for the so-called Adhama or Madhyama. mode. In Kuntaka's opinion, it is Kavi-svabhāva alone which. furnishes the criterion, and Rītis (Kuntaka employs the term Märga) should be classified according to the essential difference in the power (Sakti), culture (Vyutpatti) and practice. (Abhyāsa) of particular types of poets. One class of poets has special fitness for composition characterised by what he calls Saukumärya, while others prefer Vaicitrya, these being the two extreme modes of composition admitted by him. But there may still be other poets who would prefer to steer a middle course, thus favouring a mixed mode. In the Sukumāra Märga the natural powers of the poet find an unfettered scope in describing the Svabhäva of things, and. consequently whatever ornamentation is required is effected with the least effort ; while in the Vicitra Mārga, favoured. by all good poets, the art is chiefly decorative, and the Kavi- Kauśala is Āhārya, being characterised by more deliberate and greater skill. Each of these Märgas, according to Kuntaka, should contain four sets of excellences or Gunas,. which are designated by the same name but defined different- ly. In the Vicitra-marga, we have Madhurya=compactness of skilful structure avoiding laxity of form ; Prasada=lucidity due to the use of expressive words and easy syntax; Lavanya=beauty due to the arrangement of short and long syllables; and Abhijātya=elevatedness which is neither too soft nor too hard. In the Sukumara Marga, there should be Madhurya=sweetness due to the fewness of compounds ; Prasāda=perspicuity ; Lāvanya=beauty arising out of proper

ABHINAVAGUPTA AND THE REACTIONARIES 191

arrangement of letters and words; and Abhijatya=smooth- ness, The Madhyma Marga, which stands midway, combines the excellences of both (ubhāyatmaka). To these character- istics Kuntaka adds Aucitya (i. 53-54) and Saubhāgya (i. 55-56) as excellences common to the three Märgas. The Aucitya emphasises fitness of words and ideas, which Sau- bhagya arises out of the realisation of all the resources of a composition16. It follows from the prominence given by Kuntaka to Vakrokti in poetry that all ideas of Dhvani and Rasa should be comprehended in certain aspects of vakratā, just as the Vrttis of Udbhata, connected with anuprāsa, as well as anu- prasa itself and yamaka of orthodox writers, are taken as. kinds of varņa-vinyāsa-varkratā or vakratā depending upon the peculiar arrangement of letters. The idea of Dhvani is included partly in rūdhi-vaicitrya-vakratā, where Ānanda- vardhana's own verse tāla jaamti guna, as well as the verse snigdha-śyāmala-kānti° cited by Ānanda as an example of arthāntara-samkramita-vācya dhvani (i. e. suggestion where the expressed sense passes into another sense), is given as. instances. Other aspects of Dhvani are ackowledged in upacāra-vakratā, where the verse gaanam ca mattameham, cited by Ānanda as an instance of atyanta-tiraskrta-vācya dhvani (i. e. suggestion where the expressed sense disappears. entirely), is given as an example. From Kuntaka's treatment. it appears that he takes upacara in the sense of a supposed or fancied identification of two objects, however distinct, on. the basis even of the slightest resemblance. As such, therefore, it is admittedly involved in figures like metaphor (rūpaka) and forms the basis of metaphorical expression generally. This would come under the comprehensive domain of transferred. expression known as laksanā, and would be included by Dhvani-theorists under laksanā-mūla-dhvani, i. e. suggestion

16 See Har Dutt Sharma, Kuntaka's Conception of Gunas in Proc .. A.I.O.C, Patna 1933, pp. 581-91.

192 SANSKRIT POETICS

based on transference or Indication17. Kuntaka would thus belong to that group of writers regarding whom the Dhvanikära says that they do not deny the existence of dhvani but regard it as bhākta (bhāktam āhus tam anye), i. e. de- pending on a transference of sense or Indication. Regarding rasa-dhvani which comes under asamlaksya- krama-vyangya (i. e, suggestion of an imperceptible process), it is clear that Kuntaka, who admits not the essentiality of Rasa but that of Vakrokti, can comprehend Rasa only as an element in some aspects of Vakrokti. In the third chapter of his work, dealing with vākya-vakrata, he discusses how poetry may be made charming by delineating appropriate Rasas. In this connexion he examines in some detail such figures as rasavat, preyas etc., in which Rasa was admitted as an element by early theorists, whose system, maintaining the importance of Alamkāra in poetry, could not otherwise recognise Rasa independently. The special poetic figures like rasavat etc., constituted the back-gate, as it were, for the admission of the idea of Rasa in the Alamkāra-systems. When, however, the theory of Rasa assumed its proper importance in the schools, the necessity naturally arose of explaining how Rasa, which is essential and therefore fit to be embellished (alamkārya or upakārya) can itself be regarded as a means of embellishment (alamkāra or upakāraka) in figures like rasavat. We find accordingly in the Dhvanikāra and Ānanda- vardhana an attempt to comprehend the rasavat etc. under the class of poetry called by them gunībhūta-vyangya, in which the suggested sense (in this case the suggested Rasa) is subordinated to the expressed sense. The theory was put into shape by distinguishing the sphere of asamlakşya-krama dhvani from that of figures like rasavat on the ground that when the Rasa is predominant and forms the essence of the poem in question, it constitutes the principal suggested 17 It is for this reason that Ruyyaka thinks that the Vakrokti- jīvita-kāra comprehenđed all ideas of dhvani in upacāra-vakratā and the like (p. 8 with Jayaratha thereon).

ABHINAVAGUPTA AND THE REACTIONARIES 193

element, and as such it is alamkārya; but when it is subordi- nate to the expressed sense, it constitutes mere alamkāra or embellishment (Dhva° ii. 4 f). The Pradipa puts this con- cisely by saying: yatra pradhānam rasādis tatra dhvaniḥ, yatra tvapradhānam tatrālamkāra iti bhāvaḥ. Logically following this view, Mammata does not regard the cases of rasavat etc. as poetic figures at all, but only as a variety of guņībhūta-vyangya poetry18.

18 Later writers and commentators, however, unwilling to depart from the authority of the "ancients," attempt to explain the problem of rasavat in various ways by a method of ingenious interpretation, which keeps to the letter but changes the spirit of the old dictum. Most of these views are discussed by Visvanatha. One school holds that the designation alamkāra, given to figures like rasavat merely because they help the development of Rasa, is a purely secondary application of the term (bhakta) ; for they are not really alamkāras but should be accepted as such in deference to the practice of ancient writers (rasā- dyupakāra-mātreņehālamkrti-vyapadeśo bhāktaś cirantana-prasiddhy- angīkārya eva). These theorists admit a difference between rasavat, on the one hand, and alamkaras properly so called (such as upama), on the other ; for in the one case the Rasa directly embellishes another Rasa, while in the other case, the Rasa is indirectly embellished through the form of word and sense. But they maintain at the same time that there is one thing in common between the two kinds, viz. that both of them embellish the Rasa, either directly or indirectly, by being subservient to it. On account of this similarity of function, the designation alamkāra, which is properly applicable to such figures as upama, is applied to the rasavat by an extension of the sense (bhakti) ; and this usage has the sanction of ancient and respectable authority to which we must bow. But this explanation is rejected by others as being too fine. The differ- ence between alamkaras like upama, on the one hand, and the rasavat, on the other, which is supposed to be due to the fact of direct and indirect embellishment, is admitted to be true, but is explained away as purely accidental and immaterial; and, strictly speaking, we should designate both as alamkāras instead of indulging in fine distinctions. A third view, which altogether rejects this distinction between direct and indirect embellishment, maintains that the general definition of alam- kāra as that which embellishes the Rasa through word and sense is applicable as much to rasavat as to regular figures like upamā. 36

194 SANSKRIT POETICS

Kuntaka takes up the rasavat topically under vastu- vakratā, which may relate to both sahaja and āhārya vastu, the delineation of Rasa coming apparently under the latter head, which is described as kavi-śakti-vyutpatti-paripāka- praudha. He criticises the definitions of rasavat given by Bhämaha, Dandin and others, and holds that it is neither darśita-spaşta-śrngārādi-rasam, nor rasa-samśrayam, nor again rasa-peśalam, but rasena tulyam vartamānam; and consequently it is not an alamkāra but an alamkārya. In other words, the Rasa is awakened in these cases for its own sake, and not for the purpose of embelishing the ex- pressed word and sense. If not theoretically invulnerable, this view is interesting as indicating that the importance of Rasa, first advocated in poetic theories by the Dhvanikāra, appears to have influenced thinkers belonging to other traditions of thought. The Dhvanikära attempts to reconcile the older idea of rasavat as involving the idea of Rasa secon- darily, by admitting it in his second division of poetry ; but Kuntaka brushes aside even the view of his predecessor Bhamaha in this respect, and thinks that this case should be regarded as one in which the poet has an opportunity of creating a kind of vakrokti in which the Rasa supplies the principal charm. But he allows Rasa to play the greatest part in what he calls prabandha-vakratā, i. e. in vakratā occurring in the composition as a whole which, he thinks, must be accomplished chiefly by the aid of pleasing Rasas (rasāntareņa ramyeņa yatra nirvahanam bhavet). It is not the mere matter or plot, but the beauty imparted to it by the continuous sense of Rasa in it which can make the words of a poet live (nirantara-rasoddhara-garbha-saundarya-nir- varāḥ / girah kavīnām jīvanti na kathāmātram aśritāḥ). Kuntaka even accepts the Dhvanyāloka's judgment that in the Mahābhārata, the santa-rasa is the angin or predominant Rasa and constitutes its principal charm, although he thinks that it is ultimately the kavi-pratibha which is the all- important thing in poetry.

ABHINAVAGUPTA AND THE REACTIONARIES 195

( 4 )

Mahimabhatta Mahimabhatta begins his Vyakti-viveka, whose very name implies that it is a consideration of the theory of vyakti or vyañjana established by the authors of the Dhvanyāloka, with the proposition that his object is to comprehend all ideas of dhvani in the process of anumana or syllogistic reasoning (anumāne'ntarbhāvam sarvasyaiva dhvaneḥ pra- kāśayitum). He proceeds, therefore, to consider in detail the text as well as the theory of the Dhvanikara and Anandavar- dhana. He criticises minutely the definition of dhvani given in Dhva° i. 13 which, if properly considered, applies, he thinks, to anumāna. He considers (especially in the third chapter) most of the examples given in the Dhvanyāloka and tries to demonstrate that they are really cases of anumāna. Indeed, throughout his work he proceeds by an elaborate process of destructive criticism and makes the definition of dhvani, propounded by its advocates, conform to his definition of what he calls kävyānumiti as the process through which another sense is revealed by the expressed sense, or by a sense inferred from it connectedly (vācyas tad-anumito vā yatrārtho'rthāntaram prakāśayati/, sambandhataḥ kutaścit sā kāvyānumitir ity uktā, p. 22). This being his main position, he accepts only two senses of śabda, namely, the actually expressed (vācya) and the inferable (anumeya), including under the latter both lakşya and vyangya senses, whose independent existence he does not admit. He says (p. 7); "Meanings are of two kinds, the expressed and the inferable. Of these, the expressed belongs to the function of a word, and is alone called the primary sense of a word ..... From it, or being inferred from it, as from a logical hetu or middle term in a syllogism, another sense which is inferred, is called the inferable sense. This again is threefold, consisting of mere matter (vastu), the poetic figures (alamkāra) or the mood and sentiments (rasa). The first

196 SANSKRIT POETICS

two of these varieties can also become the expressed, the last is always inferable". It is clear enough from this that Mahimabhatta apparently accepts the recognised concept of a suggested sense in the shape of a vastu, alamkāra and rasa, but maintains that these are not revealed by vyakti or suggestion but by anumäna19; for the expressed sense and the so-called suggested sense stand in the relation of linga and lingin, the middle and the major terms of a syllogism (p. 12). Mahimabhatta maintains, by analysing many examples taken from the Dhvanyāloka, that the expressed sense does not really suggest the unexpressed sense, but that between the two, inferences are possible and do occur. The vyakti, as Anandavardhana himself admits (p. 192), is the manifestation of that which is desired to be manifested, and which becomes manifest along with that which manifests it, just as a jar in a dark room becomes visible along with the light which makes it visible. The vastu, alamkāra and rasa, which are the three suggested elements in the opinion of the Dhvani-theorists, are not manifested in this way ; for they are not comprehended along with the expressed which suggests them, but only afterwards. The interval between the perception of the expressed vibhāvas and the suggested rasa, for instance, is indeed very short, and is therefore called by the Dhvani- theorists themselves a process 'of imperceptible sequence (asamlakşya-krama) ; but this very nomenclature shows that the existence of a krama or sequence cannot be denied, and that the expressed and the unexpressed, therefore, are sequential. Being such, they must bear the relationship of a logical premise and its conclusion (pp. 11 f). Even in the case of the indicated sense, as in the phrase gaur bāhīkah, what one first understands is that the two (go and bāhīka) are not identical, and from this the conclusion arises that

19 This is the only important point of his disagreement with the Dhvanyāloka; in other respects, he says, there is hardly any dis- agreement (prānabhūtā dhvaner vyaktir iti saiva vivecitā/yat tvanyat tatra vimatiḥ prāyo nāstīty upekșitam/ /).

ABHINAVAGUPTA AND THE REACTIONARIES 197

they are meant as possessing similar qualities ; the indicated sense here is ultimately reached by anumāna (p. 24). Thus, artha is merely a ground of inference and not a vyañjaka. The process of Anumāna or inference is very wide in its scope, much wider than that of Dhvani which is naturally included in it (tasya, i.e. anumānasya, ca tad-apekşayā mahāvişayatvāt p. 12). With regard to śabda, it cannot be taken as the vyañjaka or suggestor of anything else but its literal mean- ing. As it exhausts itself after expressing its literal or primary sense, even the secondary indicated meaning (lakşya artha) has admittedly to be inferred, not from itself but from the latter ; how can it be supposed to suggest any deeper sense? But such words, through their expressed sense, can well become the ground or source of inference (anumāpaka), pp. 27 f. The process of inference in poetry by which the unex- pressed may be thus reached is presumably the ordinary process of syllogistic reasoning, which consists in the invari- able concomitance (vyapti) of the middle (linga) and the major (lingin) terms. The Dhvani or suggested sense is the lingin, and its suggestors (viz. word and sense) are apparently its linga. The invariable concomitance is ascertained in three ways, viz. by anupalabdhi (non-cognition), tādātmya (identity) and tad-utpatti (causation). In reply to Mahimabhatta's posi- tion, it has been shewn that none of these means of proving a syllogism is applicable to establishing the invariable con- comitance between the linga sabdārthau and the lingin dhvani. The non-cognition of word and sense does not prove the exist- ence of dhvani ; for non-cognition only proves that lingin which consists of the absence of something. That a jar is absent can be proved from its non-cognition. But here the lingin dhvani does not consist of the absence of anything. Therefore the hetu is vitiated, and the non-cognition of word and sense can only prove their absence, but not that of Dhvani. There can be no identity (tādātmya), again, between the sug- gested sense (dhvani) and that which suggests it (śabda and artha); for the suggested meaning is essentially different from

198 SANSKRIT POETICS

the expressed, and comes out prominently by keeping the latter in subordinate position. Similarly, the test of tad-utpatti or causation does not apply, for here the word and sense cannot be regarded as being caused by the suggested sense, in the same way as the smoke, which proves the existence of fire, can be taken as being produced from the fire itself. Viśvanātha puts the objections in another way. Inference is the knowledge of the lingin by means of the linga, quali- fied by its existence in the subject (paksa-sattva), its existence in similar instances (sapaksa-sattva) and its exclusion from opposite instances (vipakşa-vyāvartatva). For example, we conclude in the subject, e. g. a smoky hill, the existence of the lingin fire by the linga smoke, which we see existing in it, as well as in similar instances (such as in the culinary hearth, where there is no doubt as to the existence of fire), and which we see absent from opposite instances (e. g. such places where the absence of fire is certain). But this syllogistic method is not strictly applicable to establishing the suggested sense from the expressed ; for logical inference, Viśavnātha points out, has nothing to do with works of imagination. Take, for instance, the following verse :

drstim he prativeśini kşaņam ihāpy asmad-grhe dāsyasi prāyeņāsya śišoḥ pitā na virasāḥ kaupīr apa pāsyati / ekākiny api yāmi satvaram itaḥ srotas tamālākulam nīrandhrās tanum ālikhantu jaratha-ccheda nala-granthayaḥ//.

"O neighbour, will you cast your eyes for a moment here on our house? The father of this child will scarcely drink the tasteless water of a well. Though alone, I go quickly hence to the river whose banks are covered with tamala-trees. Let the densely swarming knots of reeds with their hard projec- tions scratch my body". Here the reed-knot's scratching the woman's body and her going alone to the quiet river-side may be taken as the linga of her enjoyment with a lover, which is the suggested sense (lingin) here. But these alleged reasons, "though they help to reveal the unexpressed sense, are not

ABHINAVAGUPTA AND THE REACTIONARIES 199

invariable; for dalliance with a gallant is not, from the logical point of view, universally predicable of a woman going alone to a river-side or from her being scratched by the reed-k nots. It is noteworthy that Mahimabhatta relies (p. 26) also upon the arguments of those objectors of the Dhvani-theory who think that dhvani is identical with bhakti20; but he opposes alike the views of those Mimamsakas who believe in the single pervasive power of the expressed sense21, as well as the view of the Vakroktijīvita-kara. His objection to the latter system is naturally based on his own idea of the importance of Rasa and unimportance of Dhvani. In his opinion, any deviation from common usage involving charmingness of expression, such as Kuntaka upholds, may take either of two forms, viz. (1) it may resolve itself more or less into a theory of propriety (aucitya)22, or (2) it may mean the manifestation of an implied sense other than the expressed sense. If the first alternative is meant, it is super- fluous to one who admits Rasa in poetry, as no theory of Rasa can dispense with a theory of propriety or suitability with regard to the adjustment of its factors. To admit the other alternative is to bring in the idea of dhvani in a more or less disguised form. Mahimabhatta's work is undoubtedly a masterpiece of

20 See above ch. v, pp. 152f. 21 See above ch. v, 168 fn. 22 Mahimabhatta treats the question of aucitya (already dwelt upon in the Dhvanyāloka) in the second Vimarsa of his work. He divides the subject of impropriety, which may be sabda-vişaya and artha-vişaya, into two heads according as it concerns the matter or form of poetry. The former, called antaranga anaucitya, consists in improper employ- ment of the vibhavas etc. in the manifestation of Rasa. It has already been dealt with in the Dhvanyaloka. The formal impropriety, called bahiranga anaucitya, is chiefly concerned with the occurrence of five defects, viz. vidheyāvimarśa (pp. 37-58), prakrama-bheda (pp. 58-66), krama-bheda (pp. 66-69), paunaruktya (pp. 69-84) and vācyāvacana (pp. 84-109). The question of aucitya will be dealt with in the next chapter.

200 SANSKRIT POETICS

scholastic argumentation, exhibiting much fastidious criticism and great learning of a miscellaneous kind ; but its avowed object is polemical and it does not pretend to set up a new system. Mahimabhatta possesses all the qualifications of a subtle controversialist and enters into his task with a decided animus, which constitutes the source at once of his weakness as well as of his strength. A fine product of a scholastic age, he cannot yet look beyond the pettiness of immediate issues ; and whatever might be the value of his peculiar proposition, he hardly ever adds to its limited interest any independent treatment of the larger problems of Poetics. This is perhaps one of the reasons why even his logical acumen and his erudition failed to keep the interest of his work alive; but the chief reason why his book was forgotten in later times and was cited only to be condemned -a fate which it shared with the Vakrokti-jīvita of Kun- taka-was that it pitted itself against the more formidable theory of the Dhvanikara and Anandavardhana which was destined to supersede it by attracting away the best thinkers of later times. The anumāna-theory of Mahimabhatta, like the vakrokti-theory of Kuntaka, appears to have never receiv- ed any liberal recognition in the hands of later theorists, nearly all of whom, since Mammata's time, accepted without question the system of the Dhvanyaloka. In attempting to explain away the new theory of Vyañjana in terms of the already recognised idea of Anumāna, or reviving Bhāmaha's old position in the face of the more widely received theory of the new aesthetic school, Mahimabhatta and Kuntaka were apparently fighting on behalf of a cause already doomed. ( 4 )

Bhoja and the Agni-purāna The school of opinion, represented in Poetics by the alamkāra-portion of the Agni-purana apparently follows a tradition which departs in many respects from the orthodox

ABHINAVAGUPTA AND THE REACTIONARIES 201

systems, and which we find developed by Bhoja in his own way in his Sarasvati-kanthābharana23. This apocryphal Purāna of uncertain date is ambitiously cyclopaedic. There is hardly any doubt, however, that the alamkāra-portion of this work, as we have already remarked, is chiefly a compilation, in a somewhat eclectic fashion, by a writer who was himself no theorist but who probably wanted to collect together and present a workable epitome, conform- ing in essentials to the teachings of no particular orthodox school, but gathering its material from all sources. This will be borne out not only by its independent, if somewhat loosely joined and uncritical treatment, but also from the presence of verses culled from various old writers. Taking the Alamkara-section of the Agni-purana and the Sarasvatī-kanthābharaņa side by side, one is struck at once by some fundamental characteristics which are common to both. The most peculiar feature of the Agni-purana theory is the absence of the doctrine of Dhvani, although the concept of dhvani is included casually, after the manner of ancient authors, in the figure āksepa (sa āksepo dhvaniḥ syāc ca dhvaninā vyajyate yatah, 344. 14). The word dhvani is also used in the opening verse (336. 1=Bhoja i. 1), which says generally that speech consists of dhvani, varna, pada and vākya (dhvanir varņāḥ padam vākyam ity etad vānmayam matam) ; but apparently this alludes to the grammatical word which reveals the sphota, and which is indicated by the same term in the Vākyapadīya. The work, however, recognises abhidhā and lakșanā, the ideas of which were already elaborated by philosophers and philosophical grammarians. At the same time, apart from obvious borrowings or copyings from Bharata, Bhämaha and Dandin, this work cannot be taken as substantially following the views of any one of the schools represented by these names. 23 Much of this section was printed originally as an article contri- buted to JRAS, 1923, pp. 537-49. On Visnu-dharmottara Puraņa see vol. i, p. 95f and on Agni-purāna i, p. 97f, where an account of their contents is given.

202 SANSKRIT POETICS

There is no doubt that in one verse, which is conveniently cited by Viśvanātha in support of his own extreme view, the Agni-purana speaks of Rasa as the "soul" of poetry in contrast with mere verbal ingenuity (vag-vaidagdhya-pra- dhane'pi rasa evātra jīvitam, 336. 33). It devotes a somewhat lengthy chapter to the description, after Bharata, of rasa and. bhva ; yet there is nowhere any central theory of Rasa or any elaboration of a system of Poetics on its basis. As to the origin of Rasa, it propounds a peculiar view that from infinite bliss (ananda) proceeds self-consciousness (ahamkāra), from self-conciousness proceeds conceit (abhi- māna), from conceit pleasure (rati), of which śrngāra (love), hāsya (laughter) and other rasas are modifications (338. 2-4). It admits with Bharata four fundamental Rasas, from which are derived five others. Although partiality is thus shown to Rasa in poetry and drama, the Agni-purana cannot be affiliated to the Rasa school; for it does not make any attempt to correlate with this central priciple the other factors of poetry, viz. rīti, guna and alamkāra, which are also recognised as of great, if not of equal, importance. One fact, however, worth noticing in this connexion is that al- though the Agni-purana recognises nine Rasas, adding śānta to the orthodox eight, it extols and gives prominence to śrngāra: a trait which is unique and which is found fully developed in Bhoja who, as we shall see presently, accepts no other Rasa than śrngāra in his Śrngāra-prakāśa and gives almost exclusive attention to this important Rasa in his Sarasvatī-kaņthābharaņa. On the other hand, although treatment is accorded to riti (ch. 339) and guna (ch. 345), the Agni-purāna does not follow the tenets of the Rīti school, as represented by Dandin and Vāmana. Daņdin classifies Rīti, which he calls Mārga, into two extreme types, vaidarbhi and gaudi, to which Vāmana adds pāñcālī as an intermediate type; but the classification, according to both, depends upon the presence or absence of certain fixed excellences of diction, known as Gunas. To

ABHINAVAGUPTA AND THE REACTIONARIES 203

this enumeration Rudrata adds lāfi, but by Rīti he means a definite arrangement of sentences with reference to the use of compound words of variable length. The Agni-purāna accepts this four-fold classification, but the distinction is supposed to lie not only in the length or shortness of com- pound words, but also in the qualities of softness or smooth- ness, as well as in the prominence or otherwise of metaphori- cal expression (upacara). The Gunas, again, are regarded as fundamental characteristics both by Dandin (i. 42) and Vämana (i. 2. 6-8), who take them as forming the essence of Rīti, and distinguish them carefully from Alamkāras, which, in the opinion of Dandin, form the general characteristics of both the Margas, and, in the opinion of Vamana, are merely accidental characteristics enhancing the charm of poetry already brought out by the Gunas. The Agni-purāna. however, defines the Gunas, which are nowhere connected directly with Riti, simply as those characteristics which reflect great beauty on poetry (yah kāvye mahatīm chāyām anu- grhņāty asau guņah, 345. 3), a definition which hardly distinguishes them from Alamkäras, the definition of which is here almost the same: kāvya-śobhākarān dharmān alam- kārān pracakșate (341. 17)24, and is merely copied uncritically from Dandin ii. 1. The classification of Gunas themselves, again, in this work is peculiar to itself, Ordinarily, the Guņas are classified as either śabda-gunas or artha-guņas, and this procedure is sanctioned by Vämana. The Agni- purāna brings in finer distinctions (345. 3 f.). The Gunas are here said to be of two kinds, speciflc (vaiseşika) and general (sāmānya), the former apparently confining itself to any specific part or feature of a composition, the latter existing as common to its several component parts. The sāmānya guna, again, is in its turn classified into three subdivisions,

24 This verse is also cited by Bhoja (ch. v, p. 355) ; but he remarks: tatra kāvya-śobhākarān ity anena śleşopamādivad guņa-rasa-bhāva- tadābhāsa-praśamanādīn apy upagrhnāti, apparently as a commentary on Dandin's view !

204 SANSKRIT POETICS

according as it appertains to sabda, artha, or both; the Agni-purāna (and Bhoja) admitting for the first time, so far as we know, this threefold classification. An altogether different scheme of enumeration of these Gunas then follows. Vāmana mentions in all ten Gunas, making each of these a śabda-guna as well as an artha-guna. The śabda-guņas, according to the Agni-purāna, are seven in number, viz., śleşa, lālitya, gāmbhīrya, saukumārya, udāratā, satyā, and yaugikī; the artha-guņas are six, viz. mādhurya, sam- vidhāna, komalatva, udāratā, praudhī, and sāmayikatā; the śabdārtha-guņas are again six, viz. prasāda, saubhāgya, yathāsamkhya, praśastyatā, pāka, and rāga. The character- istics of some of these Gunas are not very clearly marked25 ; and in Guņas like samvidhāna and yathāsamkhya are includ- ed ideas which are credited by other writers to Alamkāras. Although not enumerated as such, Dandin's ojas is re- produced (345. 10=Dandin i. 80) in the course of the treatment of individual śabda-guņas. In the same way, it can be easily shown that the influence of the Alamkāra school, as represented by Bhāmaha and Udbhata, is not very marked in this work. The śabdālam- kāras are, with some modification, developed, no doubt, on the general lines of Dandin's treatment26, but the arthālam- kāras do not strictly conform to the orthodox classification or definition. The Agni-purana gives eight varieties of the latter, viz. svarūpa (or svabhāva), sādrśya, utprekşā, ati- śaya, vibhāvanā, virodha, hetu and sama (343. 2-3); the figures upamā, rūpaka, sahokti, and arthāntara-nyāsa being included separately under sādrśya (343. 5), and mention being made of eighteen kinds of upama embracing most of

25 See V. Raghavan, Rīti and Guna in the Agni-purana in IHQ, x (1934) pp. 776-79. The printed text in the Ānandāśrama ed. appears to be corrupt; Raghavan suggests corrections and interpretations. 26 It recognises nine classes of Šabdālaņkāras, viz. Chāyā, Mudrā, Ukti, Yukti, Gumphanā. Vākovākyam, Anuprāsa (including Yamaka), Citra and Duşkara (including Prahelikā).

ABHINAVAGUPTA AND THE REACTIONARIES 205

Dandin's numerous subvarieties of that figures (343. 9 f). The Agni-purana is also one of the earliest known works which adds a separate chapter on the ubhayālamkāras (not recognised by earlier writer), and this includes six varieties, viz. praśasti, kānti, aucitya, samkşepa, yāvad-arthatā and abhivyakti (344. 2), some of which would come under Gunas of other writers27. Indeed, the classification and definition of the Gunas and Alamkaras, which are not differentiated very clearly, would appear crude and unsystematic, when compared to the elaborate critical treatment of the Rīti and the Alamkāra schools. From this brief outline, it will be clear enough that the Agni-purana follows, in its general standpoint, none of the orthodox schools of Poetics, so far as they are known to us, although with regard to its material it attempts to cull, in its cyclopaedic spirit, notions, expressions and even whole verses from the authors of the different schools, without, however, connecting them with a central theory. It borrows, for instance, Daņdin's definition of the kāvya-śarīra (iştārtha- vyavacchinnā padāvalī), but the attempt to supplement it by addıng kāvyam sphutad-alamkāram guņavad doşa-var- jitam (336. 6-7), is merely eclectic and hardly constitutes an improvement. The same remarks apply to its definitions of fundamental notions like guna or alamkāra, which are merely copied or paraphrased uncritically from earlier writers. At the same time, mere eclecticism is not enough

27 It is noteworthy in this connexion that Dandin's samādhi-guna is treated here under the context of laksana with a hint apparently of identifying them .- The borrowings from Bhamaha and Dandin by the Purāņa are extensive. For instance, the definitions of the figures rūpaka, ākşepa, aprastuta-praśamsā, samāsokti and paryāyokta given by the Purāņa· (343. 22 ; 344. 15, 16, 18, 17) are almost the same as those of Bhāmaha (ii. 21, 68 ; iii. 29, 8 ; ii. 79); while the definitions of rūpaka, utprekşā, višeşokti, vibhāvanā, apahnuti and samādhi (343. 23, 24-25, 26-27, 27-28 ; 355. 18, 13) appear to have been repeated from Dandin (ii. 66, 221, 323, 199, 304 ; i. 93) respectively.

206 SANSKRIT POETICS

to explain certain features of this work; the peculiar treat- ment and arrangement, for instance, of the gunas and alamkāras which depart very strikingly from orthodox views of the matter. In order to explain this novelty, we should, having regard to the essentially derivative nature of the work itself, admit the probable existence of an al- together different line of speculation, of which unfortunately no other early traces are preserved. This tradition of opinion we find fully developed in Bhoja. The prominence given to rasa and the absence of the dhvani-theory in Bhoja, therefore, need not surprise us ; nor should the peculiar arrangement of the gunas and alamkāras appear unintelligible. The same reverence to Bharata and Dandin is shown throughout; and in fact, Dandin is estimated to have supplied Bhoja with more than two hundred unacknowledged quotations28. At the same time, Bhoja very freely incorporates definite verses and illustrative stanzas from most of his well-known predecessors, especially from Bhamaha, Vāmana, Rudrața and Dhanika. He even appropriates Kārikās from the Dhvanyāloka29, although he does not accept its theory. His huge compilation, like its prototype the Agni-purāna, in more or less cyclo- paedic in scope and eclectic in spirit, and represents apparent- ly one of the several forms of arranging the teachings (with the exception of ignoring the dhvani-theory) of earlier schools in the light of a different tradition, of which another form is perhaps preserved, to a certain extent, in the two Jaina Vāgbhațas. But in some of the main points, similarity of his treatment to the Agni-purana is obvious, and here the teachings of the orthodox schools are of no avail. The verbal borrowings are numerous. Thus Agni 341. 18 f has much in common with Sarasvatī-kanthabharana ii, some

28 While the Agni-purana takes no less than 160 passages from Daņdin. 29 As the Agni-purāņa appropriates six of its Kārīkās.

ABHINAVAGUPTA AND THE REACTIONARIES 207

verses of the former being literally adopted by the latter. The Agni 341. 18-19, says : ye vyutpattyādinā śabdam alamkartum iha kşamāḥ/ śabdālamkāram āhus tān kāvya-mīmāmsā-kovidāḥ30// This definition of a sabdālamkāra is adopted by Bhoja, with the only verbal change of the defective last line into śabdālamkāra-samjñās te jñeyā jātyādayo budhaiḥ (ii. 2). Such instances can be easily multiplied, and we may cite for comparison Agni 341. 21 and Bhoja ii. 39; Agni 342. 10 and Bhoja ii. 79; Agni 338, 11 and Bhoja v. 3, etc. Apart from this. fact of literal similarity, which, however, is not conclusive, there is a striking coincidence, as we shall see presently, of treatment, as well as agreement of views on fundamental points, which is more than merely accidental. It is not suggested that Bhoja is directly copying from the Agni-purāna. or the Purana copying directly from Bhoja; it is quite possible that they exploit in common an unknown source. But there is hardly any doubt that they follow a common tradition which is different in many respects from that of the Kashmirian writers.31

30 Instead of kāvya-mīmāmsakā vidaḥ in the text. 31 With regard to the relation between the Agni-purana-compiler and Bhoja, our views do not appear to have been clear to P. V. Kane- and V. Raghavan. We have explicitly stated them in Poona Orientalist ii, p. 15-17 ; we repeat them here. A comparative study of Bhoja's Sarasvatī-k. and the Alamkāra-section of the Purana would, in our opinion, indicate that (1) Both the works are more or less compilations, (2) As compilations both are eclectic, but not very well assorted and critical, (3) Both follow a tradition of opinion which is distinctive and which stands apart from that of the orthodox Kashmirian writers, (4) Bhoja is more systematic and certainly more elaborate, and the distinctive topics are found in Bhoja in a more developed form. These considerations led us to believe that there might be no question of direct mutual borrowing, but both were drawing upon a common source, and that the more elaborate and systematic Bhoja was probably chronologi- cally later. If the Purana-compiler was later and took from Bhoja, it would be strange indeed that he should present as undeveloped and

208 SANSKRIT POETICS

Bhoja develops the definition of poetry given by the Agni- purāna by adding expressly Rasa among its essential characteristics, which, as the commentator Ratneśvara points out, indicates the influence of the "Kāśmīrakas": nirdoşam guņavat kāvyam alamkārair alamkṛtam/ rasānvitam ........ .. (i. 2). In conformity to this definition, which mentions rather uncritically all the requisite elements, Bhoja deals in the first chapter with the Dosas and Gunas and devotes the next three chapters respectively to the consideration of poetic figures (Alamkāras) of śabda, of artha, and of both śabda and artha. In the last chapter is given a detailed treatment of Rasa, for Bhoja thinks that rasokti is essential in poetry (v. 8). But like the author of the Agni-purana, Bhoja is not explicit with regard to the question of correlating this aesthetic element with other elements of poetry, and his conception of Rasa bears resemblance to that of the utpatti-vadins whose causal theory, as Abhinavagupta points out, is accepted by earlier authors like Dandin. No doubt, in one verse (i. 158) Bhoja is apparently of opinion that a poem is relished only if it contains the Gunas, even though it may possess various kinds of poetic figures; for even excellent poetic figures in a composition without the Gunas present an ugly aspect, as the form of a woman, destitute of youth, looks ugly even though she wears excellent ornaments. But this verse is only an unacknowledged quotation from Vāmana (iii. 1. 2, Vrtti), and must be taken as an instance of eulogistic statements, not unusual in Sanskrit writers, made for the purpose of simply emphasising a point, or as a characteristic of the uncritical and confused nature of the work itself ; for other- wise we cannot reconcile this dictum with others of a similar nature made in connexion with Rasa or Alamkāra.

unsystematic what was already developed and systematic in his pre- sumed source. It would hardly make any difference if the Purāna- compiler is proved later than Bhoja ; but as our available evidence is at best uncertain, it would be better to leave the question open.

ABHINAVAGUPTA AND THE REACTIONARIES 209

Although Bhoja puts a great deal of emphasis on Rasa, probably in accordance with the views of the new school of Ānandavardhana and Abhinavagupta, he cannot yet be taken as an adherent of the Dhvani school, nor of the older Rasa school. Bhoja mentions here as many as twelve Rasas, adding the śānta, preyas, udātta and uddhata, to the eight orthodox Rasas mentioned by Bharata (vi. 15)32 ; but in his treatment he follows the Agni-purana tradition in singling out the Srngāra for almost exclusive attention. This trait is also noticeable in his other work, Śrngāra-prakāśa, in which he accepts only one Rasa, the Erotic33, thus justifying its title. Bhoja modifies the Agni-purāna's classification of the Gunas by dividing the general (sāmānya) Gunas, according as they relate to śabda and artha, into external (bāhya) and internal (abhyantara), on the one hand, with specific (vaiśeşika) Guņas on the other. By the last he understands those which under special circumstances are Gunas in spite of their being essentially Dosas or faults (i. 60f). He carries the differentia- tion and multiplication of Gunas still further, and enumerates twenty-four śabda-gunas and as many artha-gunas, again, of identical names. Like the Agni-purāna, Bhoja is not very precise or critical in his definitions of individual Gunas, and he assigns to some Gunas properties which are ascribed to Alam- käras by other writers. It is curious to note that the artha- guņa kānti is defined, after Vāmana, as dīpta-rasatvam (i. 81), including Rasa therein ; and in the sabda-guna gāmbhīrya (i. 73) is incorporated the concept of Dhvani. At the same

32 Of these Santa and Preyas are already recognised. The four additional Rasas (to orthodox eight) are meant to be associated with the four kinds of heroes, namely, Dhīra-santa, Dhīra-lalita, Dhīrodatta and Dhiroddhata respectively. See V. Raghavan, Number of Rasas, pp. 121-22. 33 So says Vidyādhara, p. 98 ; also Kumārasvāmin, p. 221, and the author of Mandara-maranda-campū ix, p. 107. See Vol i pp. 136-38, for a detailed account of Bhoja's conception of Rasa in Sarasvati-k. and Śrngāra-pr. See V. Raghavan, Śrngāra-prakāsa pp. 418-542. 37

210 SANSKRIT POETICS

time, Rasa is taken elsewhere as a fundamental aesthetic concept, and the idea of Dhvani is omitted from his treat- ment. Bhoja, however, does not pay any homage to Vamana's classification of Riti, the elaboration of which he carries still further. He adds two more types of Rītis, viz. āvantikā and māgadhī (ii. 32), to the four mentioned by the Agni-purāna, the former of these being an intermediate kind between vaidarbhī and pāñcalī, and the latter forming only a Khanda-rīti, i.e. defective or incomplete type, It is also noteworthy that some of the upamā-doșas, such as hīnatva and adhikatva, are included in the general discussion of Dosas as hīnopamā and adhikopamā, and not mentioned, in the usual manner, in connexion with the figure upama itself. In the treatment of Alamkāras, Bhoja is one of the earliest writers who, in common with the Agni-purāna, classifies them into three groups, viz. śabdālamkāra, arthālamkāra, and ubhayālamkāra. Without entering into details here, we may state that Bhoja's treatment is much fuller34. He enumerates, for instance, and defines the largest number of sabdālamkāras mentioned by any author, namely twenty-four, and develops. further the treatments of Dandin, the Agni-purana and Rudrața. The number of arthālamkāras, however, is surprisingly limited, and a love of symmetry probably leads him to enumerate them also as twenty-four in number, which is also the number of the ubhayālamkāras. The most curious. chapter is that which deals with the last-named class of poetic figures, which includes figures like upamā, rūpaka, utprekşā, dīpaka, atiśaya and other well-recognised arthālamkāras. Mammata later on admits this three-fold classification of poetic figures, which is not recognised by all, but unlike

34 His treatment also is sometimes very curious. He makes poetic figures, for instance, out of the six pramānas of Jaimini (cf. Māņikya- candra on this point at p. 304). One of the results of this is that he has to admit the philosophical idea of upamāna (as a means of knowledge) in a poetic figure of that name, and distinguish it as a figure from the well-known figure upamā.

ABHINAVAGUPTA AND THE REACTIONARIES 211

Bhoja, he includes a very limited number in the mixed third class of Ubhayālamkāra, such as punaruktavad-ābhāsa, in which stress is laid equally on śabda and artha. This novel and somewhat unorthodox standpoint, which follows a peculiar line of speculation different in some respects from the accepted views of the various established schools, makes Bhoja's work an interesting study ; but its theoretic importance has been exaggerated. The work, no doubt, possesses a certain importance for this unique treat- ment in the history of Sanskrit Poetics ; but its value consists, not in its theories, nor in its discussion of general principles, but in its being a very elaborate, if somewhat diffuse, manual and an exhaustive store-house of definitions and illustrations, for which not only the works of Ālamkārikas but also of almost all the well-known poets have been laid under contribution. The later writers, in spite of the fascination which the magic name of Bhojarāja carries with it, cite this work chiefly for its abundant wealth of illustration, or for the purpose of supporting some unorthodox view to which Bhoja might have lent the authority of his name. The learning which this work parades, though extensive, is ill-assorted and uncritical, its ideas lacking in system and its expression in preciseness. The school of opinion which Bhoja represents does not appear to have received any support or following in later times35.

35 Apart from occasional citations from Bhoja by later authors, Vidyānātha (as well as Prakāśavarşa in his Rasārņavālamkāra) appears to be the one writer who goes to the length of following Bhoja's elaborate classification of the Gunas (see below, ch. vii) .- Bhoja's truly "mammoth" work, the Srngara-prakasa, has not yet been published, but a detailed account of its contents will be found in V. Raghavan's thesis on the same. It has the same eclectic and encyclopaedic character of an all-comprehending type (but on a much more extended scale) as his presumably earlier and smaller work, the Sarasvatī-kanthābharana. With regard to subject-matter and essential ideas, however, it adds nothing substantially new which is not contained in a brief form in Sarasvatī-k. In spite of its name the Srngāra-pr. comprehends in its

212 SANSKRIT POETICS

36 chapters most of the important topics of Poetics and some of Dramaturgy. Thus, ch. i-vi. deal with Sabda and Artha; ch. vii-xi with grammatical and poetical aspects of Sahitya of Sabda and Artha, including treatment of Dosa, Guņa and Alamkāra; ch. xii. mainly with Drama and its general features; ch. xiii-xiv with a preliminary treat- ment of Rasa; ch. xv-xvii with Vibhavas and Anubhāvas of Rati; ch. xviii-xxi with four Srngāras of four Purusārthas-viz. Śrngāras of Dharma, Artha, Kāma and Moksa; ch. xxii-xxxvi with elaboration of the lower Srngara Rasa (apart from the higher Srngāra of Abhimāna explained in ch. xi) of Rati between man and woman. Thus, after 'dealing with Dosa-hana, Guņopādāna, Alamkāra-yoga and Rasa-viyoga (which last should be avoided) he gives an exposition of his theory of Ahamkāra-Abhimāna-Srngāra Rasa. With reference t.) the general features of the Drama he devotes a large part of his work to the two phases of Vipralambha and Sambhoga Śrngāra viewed as a relation between man and woman.

CHAPTER VII

MAMMATA AND THE NEW SCHOOL

( 1 ) The foregoing sketch of the progress of the principal schools and systems, terminating in the dominance of the Dhvani school, will make it clear that the history of Sanskrit Poetics is marked by two or three well-defined stages.1 The dim beginnings of the science are indeed hidden from us, but we enter upon the first historic stage of its formulation, in a more or less developed form, in the works of Bhämaha and Dandin. This is followed by a fruitful and creative stage, ending with Abhinavagupta, in which the theories of the different schools or systems were settled in their general outlines, giving rise to four distinct schools of opinion, respectively represented by the Rasa-, Alamkāra-, Rīti- and Dhvani-systems. It covers more than three centuries, and includes some of the great names in the history of the discipline, like those of Bhamaha, Udbhata and Rudrata, of Lollața, Sankuka and Bhatța Nāyaka, of Daņdin and Vāmana, of the Dhvanikāra, Ānandavardhana and Abhinavagupta, of Kuntaka, Mahimabhatta and Bhoja: all of whom helped, in a constructive or destructive way, to shape the different currents of thought which ultimately ran into one stream in the standard text-book of Mammata. If we attempt to discriminate between these different schools of thought and roughly indicate the broad steps taken in the progress of the discipline, we can state generally that the Alamkāra system proposed to confine itself to a theory of embellishment (alamkāra) of expression consistently with what was probably the original tradition of the discipline ;

1 See vol. i, pp. 322 f.

214 SANSKRIT POETICS

the Rasa system, starting with the consideration of the drama, was responsible for introducing into poetic theory the subjective element of rasa, represented by the feelings, moods and sentiments ; the Rīti system laid stress on the objective beauty of representation realised by means of diction (rīti) and its constituent excellences (gunas) ; while the Dhvani system, admitting the underlying truth of all these doctrines, elaborated a peculiar theory of suggestion in poetry (dhvani), including the suggestion of rasa, to which everything else was correlated. It will be seen from this that a relative emphasis was laid on the elements of alamkāra, rasa, rīti (including guna and dosa) and dhvani by each of these systems ; and although the soft hand of Indian dialectics drew lines of fantastic ideas, consisting of odd and abstruse schemes, it is on these essential points that the theories centred themselves, and the main currents flowed thereof in different directions. It was, however, realised in the end that all these gleams of thought must be gathered into a focus, and all these currents must be made to flow into one stream. The purely normative character of the discipline began to disappear, and it was understood that, however much importance was attached to the fact of externalisation, to the consideration of embellishment or diction, it was far outbalanced by the most necessary and important principle of higher poetry, viz, the art of suggestion, especially connected with the art of suggesting a peculiar mental condition of enjoyment, technically known as rasa, of which the charm lies in a disinterested and impersonal pleasure in the mind, the attitude proper to contemplation of the beautiful2. 2 The problem, therefore, does not coneern rhetoric merely, and the ideal of beauty (if the expression is allowable) is no longer conceived from the outside, being associated with a peculiar condition of artistic enjoyment, the suggestion of which is taken as the chief function of poetry. As explained by an able critic of Sanskrit literature (Oldenberg, Die Literatur des alten Indien, pp. 207f), the Indian theorists permit intellectual vigour and subtlety, the masculine beauty, to stand behind that of purely feminine enjoyment born of the finest sensibility. Both

MAMMAȚA AND THE NEW SCHOOL 215

This period ends with the ultimate standardisation of a more or less complete scheme of Poetics, outlined in the Dhvanyāloka, in which an attempt is made to bring into a definite focus the scattered ideas of previous speculation. The period which followed this and with which we are concerned in this and following chapters, is necessarily a stage of critical elaboration, the chief work of which consists in summarising and setting forth, in the concise form of text- books, the results of earlier speculations. The stage is marked by great scholastic acumen, if not by remarkable originality or creative genius, but it denotes also a progressive deteriora- tion of the study itself. It covers the age of numberless commentaries, which may be characterised, like the scholia of European classical literature, as consisting mostly of "comments on comments of annotated annotations". They busy themselves with the explanation, expansion or restriction of the already established rules. We have also the rise of a number of popular writers and textbook-makers who wanted to simplify the science for general enlightenment, the lowest stage being reached when we come to manuals and school- books of comparatively recent times. It is difficult to classify some of these writers. Here and

these traits are found in the literature from the earliest times ; the idea of ecstatic rapture side by side with a strong inclination towards saga- city and subtlety. It is true that the dogmatic formalism of a scholastic discipline naturally sank to the level of a cold and monotonously inflated rhetoric; but at the same time it must be admitted that the theorists were not blind to finer issues, nor were they indifferent to the supreme excellence of real poetry and the aesthetic pleasure resulting from it. They always take care to add that despite dogmas the poetic imagination must show itself; and the ultimate test of poetry is the appreciation of the sahrdaya, the man of taste, whose technical know- ledge must be equal to his finer capacity of aesthetic enjoyment, born out of wide culture and identification with the feelings and sentiments of the poet. As this capacity, which is likened to the bliss of divine contemplation, is vouchsafed only to the fit and few, the critic as well as the poet is born, and not made.

216 SANSKRIT POETICS

there we find isolated and straggling followers of the older schools. Some are frankly uncritical, some merely eclectic ; while others are characterised by the very modest ambition of producing nothing more than a popular text-book. But the majority of the writers of this period, which covers more than five or six centuries, accept, with some reservations, the Dhvani-theory and the scheme of Poetics as finally determined by Mammata. There are small groups of writers who devote themselves to special topics, like kavi-śiksā or the subject of rasa (especially śrngāra-rasa), but this apparent branching off from the main stem of the finally authoritative Dhvani system, is to be explained as due rather to the following of older traditions, or perhaps to the refining or analytic spirit of the times, than to any real split in the- domain of general theory. With regard to matters of general theory and the main problems, the decadent Post-dhvani. writers as a rule thought that there was nothing new to set forth ; they consequently fell back on matters of detail which helped to satisfy their growing speculative passion: for fine distinctions and their scholastic bent for controversy. It would be tedious, as well as useless, therefore, to treat them here at any great length, for they repeat more or less the same idea in their own way, sometimes in the same stock manner and phraseology, and differ from each other only in matters of no great theoretic importance. The- only subject worth studying in them is their minute analysis and elaboration of numberless poetic figures, which are not treated, as not coming perhaps within the scope of their general exposition, by the Dhvanikāra and Ānandavardhana, but which occupy a very considerable position in later- literature. Here was room enough to supplement, as. Ruyyaka expressly states, the treatment of their predecessors ; and this portion of their work is beaten out with such extreme nicety and elaborateness that the Alamkāra-śāstra, judging from these works alone, would be, as it often has been, designated as a study of Rhetoric merely. But even from

MAMMATA AND THE NEW SCHOOL 217

Ruyyaka's time, the scope and nature of most of the in- dividual figures appear to have been fairly fixed, only to be criticised and improved upon here and there by such later writers as Jagannatha. It is important, however, to note that although this new school (navyāh, aravācīnh) accepts in the main the general position of the Dhvani school, it is yet not entirely free from the influence of older schools. It betrays a lurking regard for older writers and brings back, rightly or wrongly, some of the old ideas into the elaboration of its own theory of poetry. It is difficult, for this reason, to take these writers in a lump and affiliate them directly to the Dhvani school. Mammata's definition of poetry, for instance, is not altogether free from the influence of the views of such older writers as Vāmana; Ruyyaka follows Udbhata and Kuntaka extensively in his detailed analysis of poetic figures ; Viśvanātha clearly be- trays the influence of the Rasa school on his own system; while Jagannatha revives in a new form the old definition of poetry given by Dandin. It is remarkable that most of these writers attempt to arrive at a precise definition of poetry, a task which was wisely left alone by the Dhvani- kāra ; but in doing so, they probably meant to find out a comprehensive formula to cover the old ideas as well as the new, although it must be said that they succeed less often than they involve themselves in hopeless inconsistencies. This reactionary tendency, however, is interesting as indicating that they were not unconscious of the importance of earlier views as they were not entirely content with the clear-cut scheme of the Dhvanyāloka; a fact which would go to demonstrate, to some extent, that want of originality is a charge which cannot be brought in its entirety against these followers of the finally dominant Dhvani system.

( 2 ) Mammața The first and foremost writer of this group is Mammata,

218 SANSKRIT POETICS

whose Kāvya-prakāśa must have helped a great deal, judging from its popularity and influence, in finally establishing the authority of the Kashmirian school of Anandavardhana. This work, combining as it does the merit of fulness with that of conciseness, not only summed up previous speculations in Poetics in the succinct form of a text-book, but it became in its turn the starting point of endless text-books and exegesis. Mammata's general standpoint will be obvious at a glance by examining his well known definition of poetry. Although he adheres in the main to the teachings of the Dhvani school and accepts Rasa as an important element of poetry, his definition tad a-doşau śabdārthau saguņāvanalamkrtī punaḥ kvāpi ("poetry consists in word and sense, devoid of the defects and possessing the excellences, and sometimes devoid also of poetic figures") follows the time-honoured custom of starting with word and sense (śabda and artha) and mention- ing the guna, dosa and alamkāra; but it does not expressly include any reference to dhvani and rasa, which are apparent- ly comprehended by implication. For, following up this definition, Mammata begins with the discussion of the functions of sabda and artha, incidentally establishing the function of suggestion (vyañjana) and the superiority of the suggested sense (vyangya artha or dhvani), and divides poetry into three classes (viz. dhvani, guņībhūta-vyangya and citra) in relation to the suggested sense. This leads him to enu- merate and exemplify the various subdivisions of these three classes of poetry, and in this connexion dilate upon the nature and theory of Rasa, which is included in the scope of "suggestion of imperceptible process" (asamlakşya-krama vyangya). In this context, he examines and rejects the views of Lollata, Śankuka and Bhatța Nāyaka, and accepts the vyakti-vāda which he ascribes to Abhinavagupta. Mention is made of eight orthodox dramatic Rasas, (astau nātye rasāḥ smrtāh), but the ninth Rasa, the santa, is added, apparently as relevant to poetry.

MAMMATA AND THE NEW SCHOOL 219

Mammata then proceeds to discuss the Guņa and Doșa, not in relation to poetry in general as his definition would imply, but in relation to their subserviency or otherwise to the awakening of Rasa. The Gunas as excellences of com- position are interpreted in a new sense (after Anandavardhana) and brought into effective relation with the underlying sentiment in a work, as qualities which serve to heighten its charm. The verbal form of a work cannot be said to possess the qualities of energy or sweetness (except by way of analogy), unless we mean by it that the underlying sentiment is vigorous or sweet. The Gunas, therefore, are related to the Rasa, as virtues like heroism are related to the soul of a man. The verbal form, the mere sound, produces the excellences only as a means or instrument ; the real cause is the Rasa, even as the soul is the true cause of virtues like heroism in a man. The same consideration applies also to the case of poetic figures (Alamkāras), and their place in poetry is justified by tbeir relation to Rasa. They are compared to ornaments on a man's body; and as such, they adorn words and meanings which constitute the 'body' of poetry. They thus serve to embellish indirectly (through sound and sense) the underlying soul of sentiment, but not invariably. If the Rasa is absent, they produce mere variety of expression. It should also be noted that the Gunas are accepted, after Anandavardhana, as three in number3, and it is maintained 3 Mammata demonstrates with some care that it is not necessary to accept the ten Gunas of Vamana, but that it is quite enough if we postulate three comprehensive excellences, viz. ojas (energy), prasāda (lucidity) and mādhurya (sweetness). If we examine the Gunas of Vamana critically, we find that some of them can very well be included in these three; some constitute mere absense of defects; while others are sometimes positive defects. Thus, Vamana's ślesa, samādhi and udāratā are comprehended by ojas ; artha-vyakti is merely an aspect of prasāda; samatā, consisting of a certain uniformity of diction,is sometimes a fault; while saukumārya and kānti, defined respec- tively as freedom from harshness (or inauspiciousness) and vulgarity, are simply the reverse of the defects śruti-kasta and grāmyatva. These consi-

2Z0 SANSKRIT POETICS

that combination of particular letters signify particular Gunas, so that the three Vrttis of Udbhata (and roughly the three Rītis of Vamana) are equalised to the three Gunas defined by himself4. Mammata admits Doşas of pada, vākya

derations simplify the classification of the Gunas and put a limit to their useless multiplication or differentiation (witness, e.g. Bhoja's elaborate scheme of 24 Guņas). Mammata, therefore, thinks that the distinction between sabda-guna and artha-guna is meaningless, for the latter need not be separately considered. The mental activity, involved in the enjoy- ment of Rasa, is made to justify only three (and not ten) Gunas which are thus brought into effective relation with the principal sentiment of a composition. Thus, the ojas is supposed to cause a brilliant expansion (vistara) of the mind and resides in the moods of heroism, horror and fury ; the prasāda, proper to all the moods, is taken as the cause of a quick apprehension of the sense, extending over the mind at once (vyāpti or vikāsa), like a stream of water over a cloth, or like fire among dry fuel (cf. Bharata vii. 7) ; while the madhurya, residing nor- mally in the erotic mood of love-in-union, but also appropriate to and rising successively in degree in pathos, love-in-separation and calm, is regarded as causing a softening or melting of the heart (druti). The three conditions of the mind, viz. expanding pervading and melting, which accompany the poetic sentiments, are thus made the basis of the three Gunas; and though these mental states are sometimes mixed up and lead to various other mental conditions, these latter effects are too many and too indistinct to be taken as the basis of new Guņas. This exposition follows and expands Dhvanyāloka ii. 8-11 (see above pp. 171f) ; but it is possible that the original hint of associating these effects on the mental condition of the reader with the three Gunas was supplied by Bhatta Nāyaka (°Locana p. 68) who speaks of the enjoy- ment (bhoga) of Rasa as being characterised by the mental conditions of expanding (vistāra), prevading (vikāsa) and melting (druti). Viśva- natha substantially agrees with this view of Mammata; but he takes the technical objection that the ojas etc. are not the causes of, but identical with, the process of expansion etc. 4 See above p. 104. We have seen above that Mammata explains away the so-called excellences of sense (artha-gunas), so that the Guņas are confined to the sphere of sound. They are produced by a particular arrangement of letters (varna), compounds (samāsa) and mode of com- position (racana). Thus, the madhurya or sweetness results from the employment of (i) all sparśa letters or mutes (from k to m), excepting

MAMMAȚA AND THE NEW SCHOOL 221

and artha, as well as Dosas of Rasa, a mode of treatment which is followed by most later writers. Although the poetic figures are not, in his opinion, always necessary to poetry5, he rounds off his treatise with an elaborate analysis and illustration of figures of sound and sense (including a limited number of figures which are of a mixed kind), enumerating as many as sixty-seven independent figures. From this brief summary of the topics of Mammata's work, it will be clear that its value consists not in its originali- ty but in its orderly and concise discussion of the main issues (excepting those of Dramaturgy, which Hemacandra, Vidyā- nātha and Viśvanātha include in an attempt to supplement). His definitions as well as general treatment attempt to cover almost all fields of thought traversed by his predecessors.

the cerebrals (t, th, d, dh) coming with the last letters (nasals) of their respective class, (ii) r and n with short vowels, as well as from (iii) complete absence of compounds or presence of short compounds. The ojas or energy is produced by (i) compound letters formed by the com- bination of the first and third letters of a class with the letters immediate- ly following them (i.e. with the second and fourth letters respectively) (ii) conjunct consonants of which r forms a part, (iii) cerebrals other than n (which letter is indicative of madhurya), (iv) doubled letters, i. e. combinations of the same letters, (v) palatal and cerebral sibilants (ś, s), (vi) long compounds, and (vii) a formidable or bombastic style. For obvious reasons there are no rules for prasada. The letters mentioned here are mostly the same as those given by Udbhata as suggestors of upanāgarikā, purusā, and komalā (or grāmyā) Vrttis respectively. Mammata, therefore, thinks that Udbhata's three Vrttis, which Udbhata himself comprehends under vrttyanuprasa, are really equivalent to the three Ritis of Vamana and to his own three Guņas. 5 From the new standpoint, Vamana's view that the Gunas produce the beauty of a poem, while the Alamkaras merely heighten the beauty thus produced, is clearly inadequate. Mammata argues in this way. If the doctrine is taken to mean that the possession of all the excellences constitutes poetry, the gaudi and the pancali which are not marked by all the Gunas, would not be poetical; if, however, the presence of a single excellence is enough to dignify a composition to the rank of poetry, then we are driven to accept even a perfectly unpoetical passage as poetry, if it contains, say, the quality of energy.

222 SANSKRIT POETICS

The great popularity and authority which the Kāvya-prakāśa has always enjoyed and which is indicated by the large number of commentaries on it, must be explained as due not to any remarkable novelty of treatment, but to the clear and lucid (albeit the obscurities due to its brevity of exposition, necessitating commentaries) working out of the already accumulated stock of ideas in the light of the new scheme put forward in the Dhvanyāloka. But from the theoretical point of view, Mammata's definition of poetry has been subjected to much vigorous criticism. Viśvanātha, for instance, undoubtedly takes Mammata's work as the basis (upajīvya) of his own, but begins his Sahitya-darpana with a trenchant attack on his definition. He points out, in the first place, that the Gunas being merely properties of Rasa, should not have been logically included in the definition of essentials. He shows. next that if only faultless (adoșa) compositions be called poetry, some of the best poems will have to be given up, inasmuch as it is almost impossible to keep clear of all blemishes. Nor could it be said that faults mar only those particular words or their meanings in which they occur; for if they are faults at all, bearing relation to the underlying sentiment in the composition, they must mar the whole poem. Lastly, he maintains that no reference to poetic figures ought to have been made in the definition, as they are admittedly non-essential. Jagannatha's criticism is more fundamental, although he agrees with Viśvanātha as to the impropriety of including a reference to guņa, doşa, and alamkāra in the definition. He objects that a word and its sense are not what is denoted by the term 'poetry'; for the universal use of such expressions as 'a poem is read but its meaning is not under- stood' shows clearly that a particular kind of words only is. what is meant. If it is said that the essence of poetry lies in its capability of producing a mood (rasa), and that inasmuch as a word and its sense have this capability, both constitute- poetry, then it can be replied that, according to this too wide

MAMMAȚA AND THE NEW SCHOOL 223

view, musical tones and theatrical gestures will have to be called poetry. Some of these and other arguments appear, no doubt, fastidious and pedantic, and have been met with equal ingen- uity by the commentators and followers of Mammata; but the whole controversy indicates the futility of arriving at a precise logical definition of poetry and the difficulty of com- bining all the conventional elements in such a definition. The earlier theorists probably realised this and carefully avoided the task, for even the Dhvanikāra contents himself with des- cribing its general nature and its divisions, an omission on which Mahimabhatta ridicules him by saying: kim ca kāvya- sya svarūpam vyutpādayitukāmena matimatā tallakşaņam eva sāmānyenākhyātavyam. From the historical point of view, however, the definition is interesting, its apparent inconsistency and obscurity being a curious fact which can reasonably be explained by a reference to the views of the older schools and systems. The term rasa does not occur in the definition, and the fact that Mammata accepts the citra-kāvya, which is grudgingly admitted by Anandavardhana as a division of poetry, would indicate, as. Viśvanātha points out, that Mammata does not consider Rasa to be essential. Yet he defines Guņa and Doșa in terms of their relation to Rasa, a procedure which is not justifiable if the essentiality of Rasa is not admitted. On the other hand, if it is maintained that all reference to Rasa is omitted. in the definition because it is such a well known and establi- shed fact in the poetical and critical world, then the pro- minence of threefold suggestion and the division of poetry on its basis are hardly explicable. The mention of Guna and Dosa should in that case be omitted, as done by Jagannātha, from the definition, which corresponds more to the definition of Vāmana (i. 1. 1-3); and these two elements must be un- derstood in the sense in which Vāmana takes them, viz. as properties of sabda and artha.6 This and other discrepancies 6 This is the modified view of Jagannätha who realises the difficulty

224 SANSKRIT POETICS

make it probable that Mammata, belonging as he does to the new school, is influenced to a great extent by the views of the older schools. He accepts, no doubt, the general scheme and theory of the Dhvani school, but in trying to reconcile them with those of earlier theorists, he lands himself in object- ionable inconsistencies.

( 3 )

Viśvanātha

Even the definition of Viśvanātha, who took upon himself the task of criticising Mammata, is open to similar objections, and has been criticised in its turn by Govinda and Jagannātha. In declaring that poetry consists of a sentence of which the 'soul' is rasa (vākyam rasātmakam kāvyam), he is indeed betraying the unmistakable influence, of the Rasa school ; but he is also putting into a shape, in a way clearer than Abhi- navagupta does, the essentiality of rasa-dhvani, wisely left unstated but practically meant by the authors of the Dhvanyā- loka.7 In taking up this extreme position, Visvanatha involves himself, however, in the somewhat clumsy subterfuge of admitting a distant touch of Rasa (rasa-sparśa) in all poetry, even in what is professedly descriptive or ornamental. Even if it is ideally correct to say that a poem ought always to manifest the Rasa alone, it does not, as a matter of fact, always do so. Jagannātha rightly objects8 that Viśvanātha's definition would exclude poems in which the centre of gravity lies only in the matter (vastu-dhvani) or in the imaginative mood (alamkara-dhvani). The opponent cannot reply that this is exactly his own opinion, for thereby he goes against the view of ancient authorities, as well as against the establi-

and does not agree with Mammata (see Rasa-gangādhara p. 55). Similarly Vidyānātha, though belonging to the new school, would accept (p. 334) Udbhata's dictum samghatanā-dharmā guņāh. 7 See above ch. v, p. 166 ; ch. vi, p. 178f. 8 p. 7-8. Cf Prabhā, ed. N. S. P. 1912, p. 11.

MAMMATA AND THE NEW SCHOOL 225

shed practice of great poets, who have admitted the scope of fact and imagination, and described subjects like a flood or narrated a travel, in which there is hardly any touch of Rasa. It would not do, therefore, to accept the rasa-dhvani alone; for a complete definition must also take into account vastu- dhvani and alamkāra-dhvani. Viśvanātha anticipates this objection by saying that in cases other than those admitted expressly by himself there is always a semblance of Rasa (rasābhāsa) ; and the verse given in the Dhvanyāloka as an instance of vastu-dhvani is, in his opinion, admissible only because there is such a touch of Rasa in it, and not be- cause mere vyangya vastu can constitute the essence of poetry. Jagannätha replies that nothing is gained by such a supposi- tion of an indirect reference to Rasa, because such a reference may also be construed in phrases like 'the cow moves' or 'the deer leaps.' This cannot be taken as the sole criterion, because thereby any and every content of poetry would be reduced to the position of an excitant, an ensuant or an accessory of Rasa. Apart from this technicality and the emphasis put on the essentiality of Rasa, which however is not reconciled to other elements of poetry, Viśvanātha's scheme does not substantially differ from that of Mammata, on whose work he also appears to have written a commentary. In one passage, indeed, he pays an elegant tribute to his predecessor's work by admitting his own indebtedness to it. After defining poetry as a sentence the 'soul' of which is Rasa, he proceeds in the usual way to analyse the 'sentence' (vakya) and the different functions of its constituent word and sense, establishing suggestion or vyañjana as the function necessary and impor- tant for the purpose of conveying the suggested Rasa (bodhe rasādīnām). He accepts only two divisions of poetry, viz., dhvani and guņībhūta-vyangya kāvya, and rejects the third, the citra-kāvya (which is suffered by Anandavardhana and accepted by Mammata) on the ground that it is entirely devoid of Rasa and therefore inconsistent with his own 38

226 SANSKRIT POETICS

definition of poetry. It is curious, however, that Viśvanātha, following Anandavardhana, partially admits the suggestion of vastu and alamkāra under suggestion of perceptible process (kramoddyota-vyangya), based on the power of word or sense or both. The case of poetry of subordinate suggestion (gunībhūta-vyangya), where the Rasa involved is secondary, is justified by holding that the mere circumstance of the Rasa being collaterally suggested does not destroy the claim of such poetry ; for the relish of Rasa alone, whether circumstantial or essential, is the true criterion. It is this partiality for Rasa which makes him include a treatment, omitted by most writers on Poetics, of dramatic composition, in which the delineation of the Rasas, the moods and senti- ments, is already established as fundamental by both poets and theorists. Consistently with the same idea, a Dosa or blemish is defined as the detractor of the Rasa (rasāpakar- şaka), while a Guņa is explained as a particular mode or quality of the Rasa depending on śabda and artha and enhancing the charm of the Rasa when Rasa is principal. The Gunas are really attributes of Rasa, but they are secondarily spoken of as belonging to a word and its sense: which secondary use also explains the old distinction between sabda-guna and artha-guna. The Guņas are accepted as three in number, viz. mādhurya, ojas and prasāda, depending upon a combination of particular letters and justified by the attributes of expanding, pervading and melting the mind in its enjoyment of Rasa; and the ten Gunas of older writers are mentioned and criticised after the manner of Mammata. It is curious, however, that Viśvanātha admits the Rītis separately, instead of comprehending them, as Mammata does, under the Vrttis or considering them redundant after the enumeration of the three Gunas. He defines the Rīti as pada-samghatana9 or particular arrangement of words (and letters) which helps the Rasas (upakartrī rasādīnām). The Rīti, however, relates

9 By the term samghatanā stress in laid on its technical sense of Samāsa-vrttti, but this is not the only criterion.

MAMMAȚA AND THE NEW SCHOOL 227

entirely to the external framework of poetry, and is likened to the conformation of the body in relation to the soul10. Finally, the poetic figures (alamkāras), which are treated substantially after the manner of Mammata and Ruyyaka, are defined as those non-permanent attributes of word and its sense which add to their beauty and thus embellish the Rasa indirectly. The term 'non-permanent' (asthira) is explained, after Mammata, by the statement that the presence of the Alamkaras is not necessary but accidental, as compared to the Gunas which are necessary attributes, The above sketch of Viśvanätha's general position will sufficiently indicate that he is more or less a compiler and not an original writer, although he shows some constructive ability in elaborating a full and compact system of his own on the basis of rasa-dhvani. His borrowings from Ananda- vardhana, Mammata and Ruyyaka are very extensive ; and sometimes his judgment forsakes him, making him copy his originals rather slavishly. He is not always happy in his innovations, and sometimes (though rarely) he is wrong or inconsistent in his interpretation. In spite of these and other defects his work is interesting in the history of Sanskrit Poetics as an attempt at a further development of the Dhvani- theory out of itself, an attempt to which recognition was not

10 Viśvanātha speaks of four Rītis as follow: (i) vaidarbhī or sweet style (marked by letters indicative of madhurya, and by short compounds or absence of compounds). (ii) gaudi or bombastic style (marked by letters indicative of ojas and possessing a large number of compounds). (iii) pāñcāli (marked by letters other than those mentioned above and containing five or six compound words). (iv) lāti or the style intermediate between vaidarbhī and pāñcāli. This is only a variation on the conventional enumeration; but strictly speaking, Mammata is right in not considering the Rītis separately, as they are comprehended by the three Vrttis or even by the three Gunas accepted by the new school. Visvanātha alludes to Vrttis under vrttyanuprāsa and simply says (after Ruyyaka): rasa-vişaya-vyāpāravatī varņa-racanā vṛttih, tad-anugatatvena prakarşeņa nyasanād vrttyanuprāsaḥ.

228 SANSKRIT POETICS

universally accorded by other strict followers of the theory. The Sāhitya-darpaņa, written like the Kāvya-prakāśa in the form of Kārikā with Vrtti, has also the great merit of being written in a more simple and less controversial style than the treatises of Mammata and Jagannātha respectively ; and as a suitable and complete manual of Poetics, including a treat- ment of the dramatic art, it has always held its popularity as one of the most convenient text-books on the subject as a whole.

( 4 )

Ruyyaka One of the most important writers of this group is Ruyyaka, who comes immediately after Mammata and who also appears to have written a commentary on Mammata's work. In his treatment of the poetic figures with which his work (as its name Alamkāra-sarvasva implies) is directly con- cerned, he shows, however, a remarkable degree of insight and independence of judgment which distinguishes him from his predecessor. The value of his contribution in this respect may be judged from the fact that his Alamkāra-sarvasva not only helped to define and fix the conception of an alamkāra, of which the first indication was given by Kuntaka but which was left untouched by the authors of the Dhvanyāloka, but it had also a great influence in establishing by its careful analysis the scope and nature of individual poetic figures, so that his views on this matter have been accepted as autho- ritative by such important later writers as Viśvanātha, Vidyādhara, Vidyānātha and Appayya Dīkșita. The plan of Ruyyaka's work, written in the form Sūtra with Vrtti, is stated by himself in the introductory portion of his Vrtti. He starts, in common with other followers of the Dhvani school, with the suggested sense (pratīyamāna artha) and demonstrates by a rapid survey of the views of older writers that it was directly or indirectly recognised by

MAMMAȚA AND THE NEW SCHOOL 229

all. But he thinks that, in the opinion of the authorities who came before the Dhvani school, the chief function of the suggested sense consisted in embellishing the expressed meaning (vācyopaskāra), and therefore it was naturally com- prehended in the sphere of poetic figures in which the expressed sense prevailed11. This is generally the view of Bhamaha, Dandin, Udbhața, Vamana and Rudrața. The Vakroktijīvita-kāra, who came after Ānandavardhana, in- cludes all ideas of Dhvani in a variety of Vakrokti based on upacara or metaphorical expression. Bhatta Nāyaka maintains that the suggested sense, established chiefly through the forcible utterance of the poet (kavi-praudhokti) is only a secondary element in poetry, the essential thing being the relish of Rasa realised through a function called bhoga or enjoyment, which is distinct from and which transcends the denotative or generalising functions of words. The Vyakti- viveka-kāra takes the relation of the expressed and the suggested in terms of the logical linga and lingin, and regards the process of suggestion as a process of inference. None of these views comes up to that of the Dhvanikara, which is, therefore, accepted indisputably by Ruyyaka, who lays down sententiously at the end; asti tāvad vyangya-nistho vyañjanā- vyāpāraḥ. The three divisions of poetry into dhvani, guņī- bhūta-vyangya and citra are also recognised; but as the first two are already discussed in the Alamkāra-mañjarī12 (p. 15) and the Dhvanyāloka respectively, Ruyyaka proposes in this treatise to take up the remaining citra-kāvya, which, including in its scope all poetic figures devoid of suggestion13,

11 vācyopaskārakatvam hy alamkārānām ātma-bhūtatvam, Jaya- ratha p. 3. 12 Presumably this work was composed by himself. But Jayaratha does not expressly say so. In the Trivandrum edition, the reading is different. It reads kālidāsādi-prabandheșu (instead of alamkāra-mañį- aryām) darsitah. The Alamkāra-mañjarī appears to have dealt parti- cularly with rasa-dhvani, apparently laying stress on śrngāra-rasa. 13 See above ch. v. p. 171.

230 SANSKRIT POETICS

naturally covers an extensive field. As all detailed con- sideration of this topic is omitted in the Dhvanyāloka (as coming not properly within the sphere of its theory), here was an opportunity of supplementing the work of his pre- decessors. But the point had already been taken up and discussed in his own way by Kuntaka who recognised that the poet's intention need not always be to awaken the Rasa or anything else unexpressed but may be directed simply to producing a certain strikingness of expression in the form of an express- ed poetic figure. He analysed poetic expression and found that the elements which went to make up the being of such a figure consisted of a peculiar turn of expression, which produced a certain charm (called vaicitrya or vicchitti-viśesa) and which ultimately depended on the conception of the poet (kavi-pratibhā-nirvartitatva)14. Both these terms are not new, the kavi-pratibha having been acknowledged as essential in a poetic composition by older theorists, and the saundarya of Vämana reappearing under the name of vaicitrya or vicchitti. The ukti-vaicitrya is discussed in another connexion by Anandavardhana (p. 243) ; and Mammata (probably under the influence of Kuntaka) lays down that the alamkāra is nothing but vaicitrya itself. Abhinavagupta speaks of endless varieties of upama-vicchitti (p. 5), and in another passage (p. 8) uses the term as almost synonymous with kāmanīyaka or cārutva-hetu. Ruyyaka does not elaborate a doctrine on this point but he accepts Kuntaka's analysis implicitly and applies it to the detailed examination of individual poetic figures, a procedure which is followed by Viśvanātha, Appayya Dīkșita and Jagannatha. That he derived this idea from Kuntaka is indicated by Jayaratha in a passage in which the commen- tator, while rejecting on this ground the claim of the yatha-samkhya to be counted as a poetic figure, says: etac ca

14 See above ch. vi, p. 188f.

MAMMAȚA AND THE NEW SCHOOL 231

vakroktijīvita-krtā saprapañcam uktam ity asmābhir nāyastam (p. 149). In the Alamkāra-sarvasva Ruyyaka does not define the term vicchitti, but in the commentary on the Vyakti- viveka attributed to him, he says (p. 44): tathā ca śabdār- thayor vicchittir alamkārah, vicchittiś ca kavi-pratibhollāsa- rūpatvāt kavi-pratibhollāsasya ānantyād anantatvam bhaja- māno na paricchettum śakyate (Then again, an alamkāra consists of the charm or vicchitti of sound and sense ; and it is not possible to define vicchitti exactly, inasmuch as it is of infinite variety, being identical with the play of the poetic imagination, which itself is infinite in its scope'), the boundlessness or infinite scope of poetic conception having been already admitted by Anandavardhana himself (ch. iv), as well as by Kuntaka. Ruyyaka, however, takes this vicchitti, brought out by the productive imagination of the poet, to be the test of a poetic figure; or, in other words, a form of expression or a mere speech-figure (if the phrase is allowable) becomes a poetic figure when a certain charm is lent to it by the peculiar conception of the poet, Thus, a form of expression involving the logical anumana would not prima facie constitute the figure anumāna, unless this special charm is involved in it ; or, the doubt involved in the figure samdeha must be brought into being by the imagination of the poet, for it should not be an ordinary doubt but a 'poetic' doubt. Jayaratha makes this doctrine more explicit than his author in many places in his commentary. He lays down repeatedly that a special charm (vicchitti-viśeşa) depending on the conception of the poet (kavi-karma or ° pratibhā) is to be taken as the essential factor of an Alamkāra (pp. 144, 149-50, 183), and all so called figures are to be accepted or rejected accordingly15.

15 The question has been dealt with in some detail by Jacobi in his Ueber Begriff und Wesen der poetischen Figuren in der indischen Poetik in GN, 1908, and also in the present writer's introduction to Vakrokti-jīvita, 2nd ed. 1928, pp. xlvi-lviii.

232 SANSKRIT POETICS

In later writings this doctrine is admitted as settled beyond question. Appayya Dīksita explains it at the beginning of his Citra-mīmāmsa (p. 6), and Jagannatha repeatedly states:j alamkārāņām bhaniti-viśesa-rūpatvam. In addition to the terms bhaņiti-viśesa, vaicitrya and vicchitti, Viśvanātha and Jagannātha use the terms cārutva, hrdyatva, camatkāritva and saundarya almost synonymously, while the latter attempts to define it (p. 466, 470) more precisely as the poetic imagination with reference to the power of poetic production ; or rather, as the charm which is thereby brought into being, upon which the poetic figures distinguish themselves in their special peculiarities. Ruyyaka's work is also important for its acute analysis of the scope and nature of individual poetic figures, of which nearly eighty independent varieties are dealt with. At first sight one would be inclined to classify him as belonging to the Alamkara school. There is no doubt that Ruyyaka was a great admirer of Udbhata, on whose work his father Tilaka (as Jayaratha informs us) wrote a Viveka or Vicāra. Ruyyaka himself tells us (and he is borne out by Jayaratha and Samudrabandha)16 that he is a follower of the views of the "ancients" (e. g. cirantana-matānusrtiḥ, p. 205), by which he means apparently the older Alamkāra school of Bhāmaha and Udbhata ; but of course, he corrects, modifies or expands. older authoritative opinions in the light of the progressive study of the subject. Ruyyaka's development of Udbhata's idea of ślesa may be taken as a typical instance. The contro- versy regarding the divisions of ślesa and its relation to other figures in cases of combination, started, as Ruyyaka himself and his commentators point out, from Udbhata's time.

16 Jayaratha refers to Ruyyaka's following of cirantana-mata at: pp. 72, 83, 103, 172 etc., and of Udbhata at pp. 10, 20, 34, 87, 93, 97, 98, 125, 126, 150 etc. Samudrabandha's references are at pp. 4, 7, 10, 11, 14, 21, 74, 82, 103 etc. Ruyyaka's own references to Udbhata's views will be found in numerous places, at pp. 3, 7, 23, 59, 82, 86, 92, 123, 126, 148, 174, 191 etc.

MAMMATA AND THE NEW SCHOOL 233

Ruyyaka accepts the division of this figure into sabda-ślesa and artha-śleșa (adding ubhaya-śleşa), and holds that the principle of this distinction consists in the dictum yo'lamkāro yad-āśritaḥ sa tad-alamkārah. He rejects Mammața's view that the distinction is based on the ground that the sabda- ślesa is incapable of enduring a change of synonym (parivrtti- asaha), while the artha-śleșa is capable of doing so; for Mammata holds that it is not āśrayāśrayi-bhāva (mutual dependence or inherence) but anvaya (connexion) and vyatireka (contrast) which must be taken as the test for determining whether a figure is of sabda or of artha. Accor- ding to Ruyyaka, however, a śabda-ślesa occurs when the expression, being differently split up, yields two different meanings. Here the words are really different, as is indicated by a difference of accent as well as by the effort required in pronouncing them. They present the appearance of sameness or coalescence (ślesa), just in the same way as the lacquered wood appears to be one single object, though really lac is put on the wood. The artha-ślesa occurs where the expression is the same and has the same accent and effort, but possesses two meanings, just as two fruits hang down from a single stem. The ubhaya-śleşa is the case where both these circum- stances exist17. Rergarding the implication of ślesa in other poetic figures, the question has been raised whether it should be regarded (i) as stronger than and thus dispelling the notion of the accompanying figure, (ii) as being equally powerful and therefore entering into combination with them, or (ii) as being weaker and therefore not prominent where other figures occur18. Udbhata takes the first position, and thinks that where the ślesa is present (e. g. along with upamā) there is only the appearance (pratibha) of the other figure,

17 Viśvanātha follows Mammata, but Vidyādhara agrees with Ruyyaka's interpretation in this matter. 18 Jagannātha, p. 393, sums up the views thus: ayam cālamkārah: prāyeņālamkārāntarasya vişayam abhiniviśate, tatra kim asya bhādha- katvam syād āhosvit samkīrņatvam utāho bādhyatvam iti.

234 SANSKRIT POETICS

the real figure being in such a case the ślesa (and not upamā). Ruyyaka demurs to this view, and agrees with Mammata in pointing out that in such cases of conflict the possession of common attributes (sādharmya) inherent in upamā is alone sufficient to constitute the latter figure ; for the unqualified definition of sādharmya as community of attributes or cir- cumstances is not exclusive of the verbal sameness conveyed by the accompanying ślesa. The upamā, therefore, is pre- dominant and the subordinate ślesa only helps it ; for in such cases, the common property is not arrived at without the ślesa, and without the common property there can be no upama. If the two figures are thus found together, one helping the other, we have samkīrņatva of ślesa and upamā. From these and other instances which we need not multi- ply, the influence of the Alamkara school on Ruyyaka will be obvious ; but it will be also obvious that the views of the older school never receive unqualified acceptance from him. His following of ancient opinions, a trait which he shares with Mammata and Viśvanatha in their following of Vamana and the Rasa-writers respectively, should be explained as an earnest attempt on his part to reconcile the views of later theorists with those of the earlier, of which he was a professed admirer. It is this impulse probably which made him take up the Vakroktijīvita-kāra's conception of a poetic figure, and apply it to his own detailed analysis of individual figures, as this topic of Poetics was not sufficiently dealt with in the Dhvanyāloka. It cannot be said19, however, that Ruyyaka was a follower of the Vakroktijīvita-kāra, for Ruyyaka himself declares his own adherence to the Dhvani-theory; and, in spite of his borrowing from Udbhata and Kuntaka, he cannot by any means be directly affiliated with the Alamkāra school.

19 As suggested by Harichand Sastri p. 108.

MAMMATA AND THE NEW SCHOOL 235

( 5 ) Vidyādhara and Vidyānātha To most of the writers who followed in the footsteps of Mammata and Ruyyaka, there was hardly any original work that seemed left or unaccomplished. The details of the new system having already been established, there was apparently no occasion for any creative work, and even the task of critical elaboration had well-nigh run its course. Nor did any of the writers possess the genius of making an entirely new departure. This was also the period of early Muham- madan incursions, and was marked, as it was to be expected, by a general decadence of all investigations, reflecting a cor- responding ebb in the tide of intellectual, as well as social and political, activity. In the centuries that follow there arose a host of commentators, Mammata alone claiming no less than seventy, who busied themselves in interpreting the already established rules and in adding here and there minor points of detail, not clearly made out by their predecessors. The task of remodelling and presenting the new theory in an easier style was also undertaken, giving birth to works like the Ekāvalī of Vidyādhara and the Pratāpa-rudra-yaśo-bhūşaņa of Vidyänātha, the chief merit of which consisted in systematic compilation and arrangement. On the main problems, these treatises and even the later works of Jayadeva and Appayya Dīksita, which carry in particular the process of analysing the poetic figures to its utmost limits, throw little valuable light. Vidyādhara, for instance, models his text (consisting of Kārikā and Vrtti) on the Kāvya-prakāśa of Mammața, and in the treatment of poetic figures follows Ruyyaka in the main.20 After characterising the Kāvya as dhvani-pradhāna21, and 20 In this he agrees with Viśvanātha, Vidyanatha and others. See his definition of figures like vicitra, vikalpa or ullekha which are passed over by Mammata. 21 In the first chapter of his work Vidyadhara follows the Dhvanyā- loka very closely, and some of his Kārikās, e. g. i. 6, 13, are mere para- phrases of the Karikas of the older work.

236 SANSKRIT POETICS

setting forth its purpose as well as the qualifications necess- ary for the poet, he devotes the first chapter to the establish- ment of the dhvani or suggested sense in poetry. In this connexion, he refutes at some length the views of those schools which maintain the non-existence of Dhvani or its inclusion in other processes and draws mostly on the Dhvanyā- loka and the Kāvya-prakāśa. The second chapter deals with the three functions of word and sense, viz. abhidhā, lakşaņā and vyañjanā, while the third chapter classifes the dhvani- kāvya, in which the suggested sense excels the expressed, explaining incidentally the different theories of Rasa, which constitutes the province of one of the eight varieties of the imperceptible process of suggestion (asamlakşya-krama vyangya). The second class of poetry, the gunībhūta-vyangya kāvya, is taken up in the next fourth chapter. The fifth chapter defines the gunas, distinguishing them from the alamkāras, and concluding with the treatment of the rīti, with an incidental attack of older views and general support of Mammata's position. The next chapter is concerned with the doşas, while the last two chapters deal with the poetic figures, the śabdālamkāras and arthālamkāras respectively, adhering in general to the treatment of Ruyyaka. This brief enumera- tion of the contents of Vidyadhara's work will sufficiently indicate the scope and nature of these subsequent treatises, as well as the fact that they embrace the same topics as are dealt with in the Kāvya-prakāśa, even the different chapters some- times corresponding, in regard to their subject-matter, to the different ullāsas of the latter. The scope of Vidynatha's work, written also in the form of Kārikā with Vrtti, is much wider, and its plan somewhat different, but from the theoretical point of view it is perhaps less interesting. Its nine prakaranas cover almost the same ground as the ten paricchedas of Viśvanatha's Sāhitya-darpaņa. Like Vidyādhara, Vidyānātha follows Mammața in general, but prefers Bhoja in the matter of Gunas and Ruyyaka in the matter of poetic figures. The author justifies the pro-

MAMMATA AND. THE NEW SCHOOL 237

duction of his work by stating that although the older writers have already dealt with the different branches of the subject, none of them has described a näyaka or hero in it; but as the greatness of a composition depends on the representation of the merits of the hero described in it, the first prakarana, entitled nāyaka-prakarana, deals with the attributes of a hero, as well as of the heroine, and their necessary adjuncts. Then follows the kāvya-prakarana, which describes in the usual way the nature of a Kavya and its constituents, the vrttis and ritis suitable to the development of different sentiments, the śayya or repose of words in their mutual favourableness, the pāka or maturity of sense, and the divisions of Kāvya. It is curious that Vidyanātha's definition of poetry (gunālamkāra- sahitau śabdārthau doşa-varjitau | gadya-padyobhayamayam kāvyam kāvya vido viduh) follows closely Mammața's known definition which is quoted in a slightly modified form immedia- tely afterwards. He speaks of sabda and artha as the 'body' of poetry, vyangya as the 'soul', the gunas and alamkāras in the usual manner being likened to natural qualities like hero- ism and to outward ornaments like bracelets respectively. The ritis are described as natural dispositions which lead to the excellence of the soul (ātmotkarşāvahāḥ svabhāvāḥ). After dealing with the three functions of word and sense, he goes on to the consideration of the vyaňjanā-vrtti (pp. 52 f) and mentions (pp. 77 f) in passing 5304 varieties of Dhvani. And yet he defines the excellence, called gāmbhīrya, as dhvani- mattā, after Bhoja! He lays down racanāyā api rasa-vyañja- katvam prasiddham, which leads him to a separate considera- tion of the suggestion of Rasa. The third chapter, styled the Nataka-prakarana, deals with the subject of Rupaka or dramatic composition, a theme generally omitted by most writers, taking up the Nātaka as the most important variety and analysing its plot into five samdhis. Although based avowedly on Dhanañjaya's Dasa-rūpaka, this chapter is one of the important later contributions to the subject of Drama- turgy, and a great interest attaches to its inclusion of a model

238 SANSKRIT POETICS

drama illustrative of all its characteristics and eulogistic of the author's patron Prataparudra. Next comes the rasa- prakarana dealing with the nature and theories of Rasa. The next two chapters are the dosa-prakarana and the guna- prakarana, while the last two chapters are devoted to the topic of śabdālamkāra, arthālamkāra and miśrālamkāra. It is curious that Vidyanatha follows Bhoja in mentioning as many as twenty-four Gunas. The definitions are almost identical in the two authors. The Gunas are: (i) śleșa, coale- scence of words (owing to the imperceptibility of samdhi, when it is not harsh to the ear and when the letters belong to the same sthāna or organ of pronunciation). (ii) prasāda, lucidity arising from carefully selected words which lead to the inten- ded sense at once. (iii) samatā, uniformity of diction (rejected by Mammata as being often a defect). (iv) mādhurya, dis- tinctness of words (prthak-padatva) on account of the absence of samdhi. (v) saukumārya, softness of expression due to the use of soft-sounding letters. (vi) artha-vyakti, clearness of sense due to the completion of a sentence in all its parts, (vii) kānti, gracefulness of diction, explained as follows by the commenta- tor Ratneśvara: apratihata-padair ārambhaḥ samdarbhasyaiva kāntiḥ ...... 'kusumasya dhanur' iti prahatam, ‘kausumam' ity aprahatam; 'jalanidhau' iti prahatam, 'adhijalam' ity aprahatam; 'gurutvam' ity prahatam, 'gauravam' iti aprahatam ityādi ...... asti tu tulye'pi vācakatve padānām kaścid ābhyantaro viśeşo yam adhikrtya kimcid eva prayuñjate mahākavayah, na tu sarvam. (viii) audārya, where the sentence is so arranged with formidable letters (vikatāksara) or hard vocables (vikata, explained as kathina-varņa-samghațanā-rūpa by Jagannātha) that the words proceed as if they were dancing (nrtyadbhir iva padair yad vākya-racanā). (ix) udātta, the use of praiseworthy epithets (cf Agni-purāna 345. 9) ; Kumārasvāmin notes that it is the absence of the defect known as anucitārtha. (x) ojas, strength due to the presence of compounds. (xi) sauśabdya, elegance in the use of nominal and verbal forms (cf Bhāmaha i. 14-15 ; Rājaśekhara p. 20). (xii) preyas, statement of agree-

MAMMATA AND THE NEW SCHOOL 239

able or flattering things (Bhämaha, Dandin and some other writers regard this as an Alamkāra and not a Guņa) ; the commentator notes that this is the absence of the defect parușa (see above p. 14, fn 38). (xiii) aurjitya, compactness of structure22. (xiv) samādhi, attribution of the properties of one object to another (an echo of Dandin's samādhi), e.g., attribu- tion of the properties of an inanimate object to an animate object (what Kuntaka would comprehend under upacāra- vakratā and other writers under the figure rūpaka). (xv) vistara, detailed corroboration of what is said. (xvi) sammitatva, use of the absolutely necessary words, neither more nor less, to convey the intended sense, i.e, balance of sound and sense (arthasya padānām ca tulā-vidhrtatvāt tulyatvena sammitat- vam). (xvii) gāmbhīrya, presence of the suggested sense or dhvani (dhvani-mattā). (xviii) samksepa, brevity of statement. (xix) sauksmya, minuteness or subtlety of sense. (xx) praudhi, maturity of sense (this comes under pāka detailed below). (xxi) ukti, cleverness of speech. (xxii) riti, homogeneity of manner (corresponding to Vāmana's samatā), consisting of the com- pletion of a sentence or theme in the manner in which it was begun. (xiii) bhāvika, conduct of a sentence according to its underlying emotion or sentiment (bhāvatah). (xxiv) gati, a pleasing effect produced by long and short vowels (suramyat- vam svararohavarohayoh, in which svarāroha is explained as dīrghākşara-prāyatva, and svarāvaroha as its reverse)23.

22 This is said to be the absence of the defect called visamdhi. This defect, mentioned also by Bharata and Bhamaha, is explained as visamhito virūpo vā yasya samdhih, Kumārasvāmin explaining visam- hitaḥ as vigatā samhitā varnānām paraspara-samnikarso yatra, and virūpaḥ as simply karņa-kathorah, The samhitā therefore, means close proximity of letters which leads to euphonic, combinations sanctioned by grammar. The fault occurs (a) when there is no samdhi (viślesa) and (b) when the samdhi is harsh to the ear (kasta). Mammata (pp. 331f) adds a third case of its occurrence when the samdhi gives rise to the implication of an indecent (aslila) idea. See Trivedi's note to Pratāparudra pp. 73-75. 23 The Agni-purana, while admitting most of these excellences,

240 SANSKRIT POETICS

The theory of pāka and śayyā, mentioned by Vidyādhara and Vidyanatha follows from the stress laid on felicity of expression, which depends on poetic genius and which lies at the root of all discussions on style, poetic figure and kindred topics. The word śayya is old, having been used, apparently in this sense, by Banabhatta in one of the introductory verses of his Kadambarī; while the Agni-purāna uses the word mudra with a similar connotation. Vidyadhara and Vidya- nātha develop it further as a special excellence of expression. The śayya is defined as the repose of words in their mutual favourableness like the repose of the body in a bed, the similitude explaining the etymology of the term. This mutual friendship (maitrī) of words is so close that they cannot, as Mallinātha explains, be replaced even by synonyms: a theory of the immutability of words which mutatis mutandis would remind one of Flaubert's half-platonic view, developed by Walter Pater, that each idea has its fixed word-counterpart. The theory of paka, is very closely allied with this. The word pāka, meaning literally 'ripeness', 'maturity' or 'fruition', is as old as Vāmana. He speaks of pāka (i. 2. 21 Vrtti), resulting from the vaidarbhi riti in a delightful effect on the connoisseur, as "that attaining which the excellence of a word quickens and in which the unreal appears as real". Elsewhere he says (i. 3. 15) that śabda-pāka occurs when the words are so carefully chosen that they cannot bear an exchange of synonym. Later theorists elaborate the doctrine as consisting of (1) śabda-pāka, which may be explained, after Vāmana, as maturity of expression due to the perfect fitness of a word and its sense, and (2) artha-paka or depth of sense which is of various kinds brought about by the different tastes of different poetic Rasas. Mangala, according to Rājaśekhara, regards pāka only as sauśabdya (excellence of words) or tinām supām ca vyutpattih (proficiency in the use of nouns and verbs. cf.

classifies and defines them somewhat differently. See ch. 345, and above p. 204.

MAMMATA AND THE NEW SCHOOL 241

Bhāmaha i. 14-15). Vidyādhara admits only what is called artha-påka above ; but he alludes to the other theories which say that pāka consists of pada-vyutpatti (Mangala) or of pada-parivrtti-vaimukhya (Vāmana). Vidyānātha calls this last śayyā, and accepts and defines pāka as depth of sense, Bhoja would call it praudhi and enumerate it as a śabda-guņa. Rājaśekhara's discussion of earlier views on this point (p. 20) is interesting and deserves quotation. "The ācāryas ask: 'what is paka?' Mangala says: 'it is maturity (parinama)'. 'What again, is maturity?' ask the ācāryas. Mangala replies: 'it is the skill in the use of nouns and verbs'. Hence it is sauśabdya or excellence of language. 'The pāka is fixedness in the application of words' say the acāryas. It is said [by Vāmana i. 3. 15]: The insertion and deletion of words occur so long as there is uncertainty in the mind ; when the fixity of words is established, the composition is success- ful'. So the followers of Vamana say: the pāka is aversion of words to alteration by means of synonyms'. Therefore it is said [Vāmana, loc, cit.] : The specialists in the propriety of words have called that śabda-pāka in which the words abandon the capability of being exchanged (by synonyms)'. But Avantisundarī thinks that this want of capability is not paka. Since the varied expressions of great poets, with regard to one and the same object, all attain maturity, the pāka consists in the composition of word and sense proper to the development of Rasa. So it is said: 'That is vākya-pāka to me by which the mode of stringing together word and sense, according to guna, alamkāra, rīti and ukti, is relished'. And again: 'There being the speaker, there being the word, there being the rasa, there is still not that by which the nectar of words flows'. Hence the Yāyāvarīyas say: 'Since the pāka, which is capable of being communicated by śabda (word) through its inferrability from its effect, is in a high degree the province of Denotation (abhidhā), still it is subject to usage of what is established by the sanction of the sahrdaya'." 39

242 SANSKRIT POETICS

From this passage it would seem that Rājaśekhara admits that the paka is conveyed chiefly through words ; and taken as sauśabdya or śabda-vyutpatti, it comes primarily under the province of abhidha; but it finds its scope only in the artha which is established by the taste of the sahrdaya. In this connexion it is proper to note that the term paka, like the word rasa, has a reference to its etymological meaning of physical taste which has been fancifully likened to that resulting from the ripeness of fruits. As such ripeness of fruits bears different tastes, some theorists carry the analogy into distinguishing and naming pākas after various kinds of fruits. Thus, Vamana quotes two old verses. (under iii. 2. 15) which speak of vrntāka-pāka ; while Vidyā- nātha speaks of two kinds of pāka (1) drākşā-pāka, or the maturity of grapes in which the taste flows both in and out, and (2) nārikela-pāka, the ripeness of cocoa-nut which is rough outside but tasty inside. Bhoja similarly distinguishes between. mrdvīkā-pāka and nārikela-pāka ; but Ratneśvara in his com- mentary alludes to various kinds of pāka, named after saha- kāra (mango), vārtāka (egg-plant) and nīlakapittha (feronia. elephantum). Rājaśekhara goes to the length of mentioning. nine such cases of pāka named after the following nine fruits. (pp. 20-21):, picumanda (nimba, azadirachta indica), badara (jujube), mrdvīkā (grapes), vārtāka (egg-plant fruit), tintidī (tamarind), sahakāra (mango), kramuka (betel-nut), trapusa (cucumber) and nārikela (cocoa-nut).

CHAPTER VIII

SOME LATER WRITERS OF THE NEW SCHOOL

( 1 )

Hemacandra and the Vāgbhatas The group of three Jaina writers, Hemacandra and the older and the younger Vägbhatas, may be conveniently mentioned here, but they do not call for any special remark. Hemacandra's Kāvyānuśāsana, written in the form of Sūtra with Vrtti, and its commentary, called Viveka1, composed by himself, indicate extensive learning and constitute a compact manual of Poetics in eight chapters; but there is hardly anywhere any striking trait of originality2 or even indepen-

1 The Sūtra-portion is called Kāvyānuśāsana, the Vrtti is styled the Alamkāra-cūdā-mani, while the brief commentary which explains the Vrtti may be called Viveka from its mangala-verse. 2 Hemacandra's treatment of poetic figures, however, is somewhat peculiar. He speaks of six sabdālamkāras, viz. anuprāsa, yamaka, citra, śleşa, vakrokti and punaruktavad-ābhāsa. The arthālamkāras are much reduced in number and limited to twenty-nine (viz. upamā, utpreksā, rūpaka, nidarśana, dīpaka, anyokti. paryāyokta, atisayokti, ākşepa, virodha, sahokti, samāsokti, jāti, vyāja-stuti, śleșa, vyatireka, arthāntara-nyāsa, sasamdeha, apahnuti, parāvṛtti, anumāna, smṛti, bhrānti, vişama, sama, samuccaya, parisakhyā, kāraņamālā and samkara). He includes samsrsti under samkara, and treats ananvaya and upameyopama as varieties of upamā. The aprastuta-praśamsā similarly goes under anyokti. All figures like rasavat, preyas, ūrjasvin and samahita that have a touch of Rasa and Bhava are omitted as being comprehended (so also Mammata thinks) in the class of poetry called guņībhūta-vyangya. Hemacandra does not deal with parikara, yathā- samkhya, bhāvika, udātta, āsīh and pratyanīka for reasons explained by himself at pp. 292-4. Hemacandra, however, defines some figures somewhat broadly so as to include other recognised figures in them, e. g. his dipaka would include tulyayogitā, his parāvrtti would contain. the paryaya and parivrtti of Mammata, his nidarsana would comprehend. prativastūpamā, drstānta and nidarśanā of other writers.

244 SANSKRIT POETICS

dent thinking out of the main problems. This work is chiefly a compilation3. Hemacandra not only paraphrases literally most of the standardised definitions, and reproduces almost unhesitatingly the illustrative quotations of Mammata; but his acknowledged and unacknowledged borrowings from the Dhvanyāloka and Locana, from the Abhinava-bhāratī, from the Vakrokti-jīvita, from Rājaśekhara's Kāvya-mīmāmsā, as well as from other well known works are indeed numerous. No doubt, Hemacandra adds a chapter on Dram aturgy, chiefly compiled from Bharata and others, but the account of the doctrines of dhvani, rasa, guņa, doşa and alamkāra is closely and somewhat uncritically copied from Mammata, supplemen- ted, however, by excerpts, in the commentary, of other views on the subject. In trying to improve upon Mammata's imperfect definition of poetry by substituting sālamkārau ca in the place of analamkrtī punah kvāpi, he puts himself open to greater technical objection, although he adds the gloss: ca-kāro niralamkārayor api śabdārthayoḥ kvacit kāvyatva-sthāpa- nārthaḥ. . . The older and the younger Vägbhatas, on the other hand, though making considerable use of Mammata's text (the latter especially borrowing from Hemacandra's version too), do not admit dhvani, and are allied in their sympathies with the Pre-dhvani schools. The authority of Dandin, for instance, carries great weight with them; and the younger Vāgbhata admits some of Rudrata's peculiar poetic figures. At the same time, the unmistakable influence of the new

3 The eight chapters of Hemacandra's Kāvyānuśāsana comprehends the following topics: i. The purpose (proyojana) of poetry, its causes (hetu) viz. pratibhā to which are added vyutpatti and abhyāsa ; the defini- tion of poetry ; the nature of sabda and artha ; the denoted, indicated and suggested meanings. ii. The rasa and its factors. iii. The dosas of pada, vākya, artha and rasa. iv. The gunas, accepted as three after Mammata, and the letters which produce them. v. Six figures of sound. vi. Twenty-nine figures of sense. vii. The nāyaka and nāyikā. viii. Division of poetry into preksya and śravya, and their characteristics and subdivisions.

SOME LATER WRITERS OF THE NEW SCHOOL 245

school on them precludes us from affiliating them directly with the older Alamkara and Rīti schools. The older Vāgbhata defines poetry as: sādhu-śabdārtha-samdarbham guņālamkāra-bhūşitam/. sphuta rīti-rasopetam kāvyam kurvīta kīrtaye/ / while the younger Vagbhata, whose work is written in the sutra- and vrtti-form like Hemacandra's, adopts literally the latter's modification of Mammata's definition. The criterion of poetry, according to them, is that it must contain, through its word and sense, the guna, alamkāra, riti and rasa, but these elements are mentioned rather in an eclectic than critical spirit. The older Vägbhata accepts without question the ten Gunas of older writers, but the younger Vägbhata follows Mammata in limiting them to three, with the pointed remark: iti daņdi-vāmana-vāgbhațādi-praņītā daśa kāvya-guņāh, vayam tu mādhuryauja-prasāda-laksaņān trīn eva guņān manyāmahe. The younger Vagbhata speaks of Rasa as the 'soul' of poetry4 ; but beyond a description, after Hemacandra and others, of the different Rasas, he does not touch upon the theoretical aspect of the question, nor does he indicate the mutual relation of the different elements of poetry with reference to the Rasa. Indeed, both of them do not appear to accept the reconciliation proposed by the Dhvani-theorists ; and the younger Vägbhata specifically includes dhvani, after Bhämaha and Udbhata, in the figure paryāyokta with the remark: evamadi-bhedair dhvanitoktir bhavati, param grantha-gaurava-bhayād asmābhir nodāhriyate, sa prapañcas tvānandavardhanād avagantavyah (p. 37). The object of these Jaina manuals (though there is nothing specifically Jaina in them) appears to have been the presenta- tion of a popular and easy epitome of the subject, allaying themselves to no particular school or system, but following

4 doşa-muktam guna-yuktam alamkāra-bhūşitam sabdārtha-rūpam uktam kāvya-śarīram, param tat tvaprāņi-śarīram iva nirātmakam na pratibhāsate, atah kāvyasya prāņa-bhūtam rasam āha, ch. v, p. 53).

246 SANSKRIT POETICS

the traditional notions in a spirit of eclecticism, without critically systematising them in the light of a central theory. In this respect, they bear a close resemblance to the alamkāra- section in the Agni-purāna and the Sarasvatī-kanthābharana of Bhoja, whose definition of poetry is forcibly recalled by that of the older Vägbhata quoted above. The topics dealt with in the five paricchedas of the Vāgbhațālamkāra are as follow: (i) The definition of Kāvya ; pratibhā as the source of Kāvya, aided by vyutpatti and abhyāsa ; the circumstances favourable to poetry and the conventions observed by poets. (ii) The language of poetry (Sanskrit, Prakrit, Apabhramśa and Bhūta-bhāșā); forms of poetry (metrical chandonibaddha and non-metrical) ; its divisions into padya (verse), gadya (prose) and miśra (mixed verse and prose) ; eight dosas of pada and of vākya respecti- vely, and the dosas of artha. (iii) The ten gunas. (iv) Four poetic figures of sound, viz. citra, vakrokti, anuprāsa and yamaka, and thirty-five figures of sense ; the two rītis (vaidarbhī and gaudīyā). (v) Nine rasas; kinds of nāyaka and nāyikā and kindred topics. The Kāvyānuśāsana of the younger Vāgbhata is, unlike the Vāgbhatālamkāra (which is written in the metrical form, generally in the anustubh with only one prose passage at iii. 14), composed in the Sūtra- and Vrtti-style of Hemacandra's Kāvyānuśāsana. It is also divided into five chapters with topics as follow: (i) The prayojana, and the hetu (pratibha aided by vyutpatti and abhyāsa) of Kāvya ; its division into padya, gadya and miśra ; the classi- fication of poetical composition into mahākāvya, ākhyāyikā, kathā, campū and rūpaka. (ii) Sixteen dosas of pada, fourteen of vākya, and fourteen of artha; the ten gunas of Vāmana and Dandin reduced to three, viz. mādhurya, ojas and prasāda ; three rītis, viz. vaidarbhī, gaudīyā and pāñcālī. (ii) Sixty-three figures of sense, in which some of Rudrata's old figures reappear. (iv) Six figures of sound, viz. citra, śleșa, anuprāsa, vakrokti, yamaka and punaruktavad-ābhāsa. (v) Nine rasas ; the topic of nāyaka-nāyikā ; and the doșas of rasa.

SOME LATER WRITERS OF THE NEW SCHOOL 247

( 2 )

Jayadeva, Appayya and Jagannātha

We have now practically closed our survey of the principal Post-dhvani writers who deserve mention and treatment. The school of kavi-śikşā and the erotic rasa-writers stand apart in many respects, and we propose to deal with them separately in the following chapters. But the above account of the Post- dhvani writers must not be supposed to exhaust the ex- traordinary wealth of scholastic activity of this period. The commentators and textbook-writers continued to multiply, and a glance at their names given in our preceding volume5 will show the extent to which their activity was carried; but hardly any of these later works, except perhaps Jagannātha's Rasa-gangadhara, with an account of which we shall close our survey, deserves separate or detailed mention. Even Keśava Miśra's Alamkāra-śekhara, or Acyuta Rāya's more modern Sāhitya-sāra6, convenient and well-written compen- diums as they are, and standing as they do much above the average, add hardly anything fresh to our knowledge. Keśava Miśra draws largely upon Mammata and the younger Vägbhata (besides older writers). He declares that his work is based upon the Kārikās of one alamkāra-vidyā- sūtrakāra bhagavān Sauddhodani (see vol. i, pp. 220f) ; but it does not present any theory nor set up any new system. There are, however, certain opinions which are peculiar to the work, the chief of which is that it sets up Rasa as the essence of poetry. The work is divided into eight ratnas, consisting of twenty-two maricis. The first ratna defines Kāvya as rasādimat vākya, and discusses pratibhā etc. as its sources. After an enunciation of three Rītis, vaidarbhi, gaudī and mägadhī (which are defined with reference to the em-

5 See vol. i, ch. x, pp. 262-315, Minor Writers. For commentators, see Bibliography given under each writer. 6 For a summary of the contents of this work, see vol. i, p. 264.

248 SANSKRIT POETICS

ployment of compounds), along with ukti (4 kinds) and mudrā (4 kinds), it goes on to discuss the three usual Vrttis, viz. śakti (=abhidhā), lakșaņā and vyañjanā. Then comes the doşa-ratna, in which are detailed two series of eight faults each of word and sense, and twelve defects of sentences. The next third section, called guna-ratna, deals with five excellences of śabda (viz. samksiptatva, udāttatva, prasāda, ukti and samādhi), and four excellences of sense (viz. bhāvikatva. suśabdatva, paryāyokta and sudharmitā). This is followed by a discussion of the cases where some of the above Dosas may sometimes become Gunas. The influence of Bhoja's opinions on this part of the work is obvious. Then comes the alamkāra-ratna, where mention is made of eight figures of sounds (citra, vakrokti, anuprāsa, gūdha, ślesa, prahelikā, praśnottara and yamaka) and only fourteen figures of sense (upamā, rūpaka, utprekşā, samāsokti, apahnuti, samāhita, svabhāva, virodha, sāra, dīpaka, sahokti, anyadeśatva= asamgati of Mammața, viśeșokti, and vibhāvanā). This is followed by a curious chapter, entitled varnaka-ratna, in which are detailed the upamanas appropriate for describing a damsel, her complexion, hair, forehead, eyebrows etc. It goes on to give practical hints as to how poets should describe the physical characteristics of the hero, mentions words which convey the idea of similarity, details the conventional usages. of poets (kavi-samaya), as well as the topics for description (such as the king, the queen, a town, a city, a river etc.) and the way of describing them, the colours of various objects in nature, words that convey numerals from one to one thousand, certain tricks of words such as bhāsā-sama (where a verse reads the same in Sanskrit as in Prakrit), samasyā-pūrana, the nine Rasas, the kinds of hero and heroine, the different Bhävas, the Dosas of Rasa, and lastly, the arrangement of letters favourable to each Rasa. Jayadeva's Candrāloka7 has been a deservedly popular

7 See vol. i, p. 199 for a résumé of its contents.

SOME LATER WRITERS OF THE NEW SCHOOL 249

manual, but in spite of its clearness and brevity of exposition and aptness of its illustrations, it is nothing more than a con- venient epitome, its most remarkable feature being its detailed treatment of poetic figures, which occupy nearly half its bulk. The Candraloka deals with ten gunas and one hundred alamkāras. The third chapter, curiously enough, is devoted to laksanas, which are not mentioned by later writers except in connexion with Dramaturgy (as by Viśvanātha). Instead of Bharata's thirty-six laksanas (ch, xvi 6-39 ; see above pp. 3-5), Jayadeva defines and illustrates only ten, viz. the economical combination of letters to convey a striking meaning (akşara- samhati), the prohibition of a fault by the indication of an excellence (śobhā). the deliberation resulting in a negation of what is said (abhimāna), the determination of a proposition by a rejection of other possible alternatives (hetu), the disregard of well established causes (pratisedha), the inter- pretation of a name both as true and false (nirukta), false attribution where both the major and middle terms of a pro- position are absent (mithyādhyavasāya), substantiation of the excellence of an object by emphasising its well-matched resemblance to a well known object (siddhi), establishment of a particular attribute through the drift of two different mean- ings (yukti), and the accomplishment of a purpose through some action or occurrence (phala). Viśvanātha, however, adds thirty-three dramatic embellishments (natyālamkāras) to his laksanas, which are enumerated after Bharata as thirty-six in number, but which do not correspond exactly to Bharata's laksanas, as some of the latter fall also under Viśvanātha's nātyālamkāras. The two classes cannot indeed be distingui- shed on any conceivable theory ; and though Viśvanātha adopts the conventional enumeration, he remarks in the end:) eşām ca laksaņānām nātyālamkārāņām ekarūpatve'pi bhedena vyapadeśaḥ gaddalikā-pravāheņa. Thus, we find included under laksanas the combina- tion of Gunas with Alamkaras; the economical grouping of letters to produce a charming import ; the use of double

250 SANSKRIT POETICS

entendre for the purpose of conveying a less known import along with one more well known ; the use of analogy and example; the brief citation of a reason for the intended meaning; the expression of doubt in the case of an object whose nature is not known; the surmise from a matter coincident with the course of nature; the fitting of expression to the sense ; the citation of admitted facts to refute inadmissible views; the supposition of a non-existent object or fact from resemblance ; the inference of an object from some of its peculiarities ; the deduction by reasoning of a fact which is not capable of sense-perception ; the descrip- tion of an object under the similitude of time and place ; the statement of agreeable views in accordance with the Sastras ; the indication of acts contrary to one's qualities ; the attribu- tion to an object of qualities in excess of its ordinary qualities ; the discrimination of a particular meaning out of other well known meanings by an allusion to the literal sense ; the repetition of a proposition already established : the mention of various objects in eulogy of the intended object ; the unconscious expression, under the influence of passion, of the contrary of what one means; the alteration of a conclusion through doubt; the compliance with other people's views by words or acts; the persuasion by means of affec- tionate words ; the statement in succession of several means to attain a desired object ; the suggestion and strengthening of one view by a different view ; the reproach ; the respectful enquiry ; the employment of names of well known persons or things in eulogy of the person or thing under description ; the mistaken resemblance of apparently similar things causing resentment ; the offer of oneself in the service of another; the flattering statement; the employment of a comparison to convey a sense which is not directly desired; the indirect expression of desire; the veiled compliment; and the expression of gratitude in pleasing terms. The nātyālamkāras are the benediction, the lamentation, the deception, the unforgiving attitude, the arrogant ex-

SOME LATER WRITERS OF THE NEW SCHOOL 251

pression, the expression of a resolution or of an excellent purpose, the raillery, the desire for a charming object, the agitation due to reproach, the repentance for missing an object through folly, the use of an argument, the longing for an object, the request, the commencement of an undesirable act, the mentioning of a purpose, the provocation, the reproach, the observance of the Sastras, the covert rebuke administered by citing a common opinion, the narration, the prayer, the apology, the reminding of a duty neglected, the recounting of previous history, the determination of an act by reasoning, the ecstacy and the instruction. It will be seen that the division is not only over- lapping, but both the laksanas and the nātyālamkāras refer largely to modes of exposition, to the use of what other writers would regard as specific figures or excellences of diction, or they may sometimes appertain to the feelings and emotions which come within the sphere of Rasa and Bhava. This fact is retognised very early by Daņdin who includes lakşuņas under alamkāras in the wider sense. Dhanañjaya does the same, but he recognises also that some of them come under Rasa and Bhava. Viśvanātha, therefore, includes them under the guņa, alamkāra, bhāva and samdhi, but deals with them only in connexion with the drama. There is practically no need for them in later. Poetics from which they ultimately disappear, their function having been assigned to other recognised elements of poetry. Appayya Dīksita's three well known manuals8 one of which is directly based on Jayadeva's work, and Viśveśvara's Alamkāra-kaustubha9 are indeed noteworthy for their ela- borate treatment of poetic figures and have merits of their own, but they are in reality nothing more than elementary text-books, excellent résumés which methodically register

8 See vol. i, pp. 223-25. 9 See vol. i, p. 303. The work, as its name implies, deals entirely with poetic figures in an elaborate way. The number of independent figures dealt with is about 76.

252 SANSKRIT POETICS

previous speculations on the subject. There is such a general sameness of characteristics, such a monotony of treatment, as well as repetition of conventional topics in conventional phraseology, that it is not worth while to linger over the activity of these lesser writers. The work of the great Kashmirian writers was over, and although Bengal and the Deccan had come into prominence as fields of later activity, the age of really original or thoughtful writers was long gone by. It was succeeded by an age of commentators, interpreters and critics (some of them were very able and painstaking) as long as there was the need of critical ela- boration, of understanding and explaining a great author. But in course of time, even this became superfluous, and there was nothing to be done but the writing of smaller and simpler manuals adapted to general comprehension. The declining age of most of the schools witnessed a host of such manuals and manuals of manuals ; but this was the period when the declining age of the Post-dhvani school, as represented by Mammata, went through the same process. Even this was not. enough. Out of the debris of these schools there grew up a spirit of eclecticism, of which we have already an early indication in the works of Bhoja and the Vägbhatas after the decline of the older Rasa, Alamkāra and Rīti schools ; and we meet with hand-books which depend upon no system but which are apparently written for the enlightenment of lay understanding. The different systems of Sanskrit Poetics may now be supposed to have well-nigh run their course and attained their natural termination.

( 3 )

Jagannātha Jagannātha's Rasa-gangādhara is the last remarkable work on Poetics. We do not, however, find in it a complete presentation of the subject, as the available text forms about two-fifths of what the work was originally designed to be by

SOME LATER WRITERS OF THE NEW SCHOOL 253

its author and is thus extant only in an unfortunately in- complete shape. Nearly three-fourths of this, again, and the whole of his Citra-mīmāmsā-khandana are taken up with the discussion and illustration of poetic figures, a topic which, as here set forth, forms indeed one of the most exhaustive and noteworthy presentation of later times, but is of little theoretical interest from the standpoint of general Poetics. Jagannätha's style is erudite and frightens the student by its involved language, its subtle reasoning and its unsparing criticism of earlier writers. The most criticised authors in this respect are Ruyyaka, his commentator Jayaratha and his follower Appayya Dīksita. But in spite of this tendency towards controversy, which is combined with an aptitude for hair-splitting refinements, Jagannatha's work displays an acute and independent treatment, or at least an attempt at a rethinking of the old problems. He shows himself con- versant with the poetic theories of older writers, which he does not ignore but which he endeavours to harmonise with the new currents of thought. Along with some other important writers of the new school, Jagannātha marks a reaction in this respect ; and the school of Mammata and Ruyyaka does not receive from him unqualified homage. Jagannātha defines poetry as a word or linguistic com- position which brings a charming idea into expression (ramaņīyārtha-pratipādakah śabdaḥ) : a definition which reminds us of Dandin's well known description of kāvya- śarīra as iştārtha-vyavacchinnā padāvalī, but which is further explained in this way. The charmingness belongs to an idea which causes unworldly or disinterested pleasure. This quality of disinterestedness is an essential characteristic, which is a fact of internal experience and which is an attribute of pleasure, being synonymous with camatkāra or strikingness. The cause of this pleasure is a conception or a species of representation, consisting of continued contemplation of something characterised by the pleasure itself. Thus, there is no disinterestedness in the pleasure conveyed by the ap-

254 SANSKRIT POETICS

prehension of the sense of a sentence like 'a son is born to you' or 'I shall give you money' ; in such a sentence, there- fore, there is no poetry. Hence poetry consists of words which express an idea that becomes the object of contem- plation causing such pleasure10. The beautiful (ramanīyata) in poetry, therefore, is that. which gives us disinterested or impersonal pleasure. This pleasure is specifically different from that which one finds in the actually pleasing, and depends upon taste formed by continued contemplation of beautiful objects. It will be noticed that this definition not only gives us a remarkable analysis of the beautiful but includes in its generality and comprehensiveness all the elements of poetry recognised by previous theorists, without specifically naming them. We have already noted that the poetic sentiment or Rasa, excited. in the reader's mind, is peculiar in its nature ; it is, no doubt, a fact of one's own consciousness but it is essentially universal. and impersonal in character, being common to all trained. readers and possessing no significance to their personal. relations or interests. A distinction is made between a natural emotion and a poetic sentiment ; the former is individual and immediately personal and therefore may be pleasurable or painful, but the latter is generic and disinterested and marked by impersonal joy. The poetic sentiment in this sense is. supernormal (alaukika), and those things which cause disgust,. fear or sorrow in ordinary life and those normal emotions .. which are far from pleasant in actual experience, being conveyed in poetry, become ideal and universal, and bring about this supernormal pleasure which is not to be compared.

10 ramaņīyatā ca lokottarāhlāda-janaka-jñāna-gocaratā; lokot- taram cāhlāda-gatas camatkāratvāpara-paryāyo'nubhava-sākşiko jāti- viśeşah ; kāraņam ca tad-avacchinne bhāvanā-viseşah punaḥ punar anusamdhānātmā ; putras te jātah, dhanam te dāsyāmīti vākyārtha- dhī-janyasyāhlādasya na lokottaratvam, ato na tasmin vākye kāvyatva- prasaktiḥ. Itham ca camatkāra-janaka-bhāvanā-vişayārtha-pratipādaka- sabdatvam. Cf Jacobi in Internat. Wochenschrift, 1910, ix. 821.

SOME LATER WRITERS OF THE NEW SCHOOL 255

to the very mixed pleasure experienced in ordinary life. This pleasure, dissociated from all personal interests, is the essence of the mental condition involved in the enjoyment of Rasa; it is also the essence of all poetry, as conceived by Jagannātha. In the same way, the definition includes the concept of the suggested sense (dhvani), and Jagannātha proceeds to divide poetry on this basis into four (viz. uttamottama, uttama, madhyama and adhama) classes, corresponding to the three classes recognised by his predecessors since the Dhvanikara's time. The first occurs where the sound and sense, subordinating themselves, suggest another charming sense ; it corresponds to the principal dhvani-kāvya of the Dhvanikāra. The second and third classes, the guņībhūta- vyangya and citra, mentioned by the Dhvanikāra, are split up11 into three cases, viz. (i) where the suggested sense, though not principal, is yet the cause of special charm, (ii) where the charm of the expressed sense is predicated equally with the charm of the suggested, and (iii) where the charm of sound, being embellished by the charm of sense, is principal. This lowest class of poetry, corresponding to the śabda-citra and artha-citra of Mammata (a distinction which is rejected by Jagannātha), apparently comprehends those cases where the artha-camatkrti is swallowed up or strengthened by śabda- camatkrti. Jagannatha adds that although it is possible to count a still lower fifth class of poetry, in which the charm of sound is altogether devoid of all charm of sense (e. g. cases of conundrums like the padma-bandha) and which is allowed by the practice of some poets, yet in view of the definition of poetry already given, as consisting of words expressing a charming sense, these instances have to be excluded or ignored. 11 The object of this splitting up is to dispense with the necessity of minutely subdividing the various cases of the gunībhūta-vyangya kāvya and also to include generally all poetry which is alamkāra- pradhāna.

256 SANSKRIT POETICS

After this classification, Jagannätha follows the conven- tional way of dividing Dhvani, infinite aspects as it may present (asamkhya-bheda), into two broad groups, based on Denotation (abhidhā-mūla) and Indication (lakşaņā- mūla) respectively. The former has a threefold aspect, according as it is a suggestion of rasa, alamkāra or vastu, while recognition is given to the two cases of the latter, viz., (i) where the expressed meaning passes over to another sense (arthāntara-samkramita-vācya) and (ii) where the expressed sense is made to disappear entirely (atyanta-tiraskta-vācya). This brings our author topically to a detailed consideration of Rasa-dhvani which is characterised as. parama-ramaņīya. and an elaborate discussion of the nature and theory of Rasa and Bhava and its constituent elements. He speaks in this connexion of bhava-dhvani (pp. 74-98) and takes into consi- deration different phases of Rasa and Bhāva, such as rasā- bhāsa (p. 99), bhāva-śānti (p. 102). bhāvotpatti, bhāva-samdhi and bhāva-sabalatā (p. 103f)12. The discussion of the Gunas come in this context, inasmuch as they are related to the Rasa. Jagannātha enumerates and discusses the ten śabda- and artha-gunas of Vämana and other older writers; but he appears to accept only three Guņas after Mammata, viz. mādhurya, ojas and prasäda, classifying them on the basis of their respective effects on the reader's mind, viz. druti (melting), dipti (brilliance, i. e. expansion) and vikāsa (pervasion). Jagannatha remarks in this connexion: guņānām caişām druti-dīpti-vikāsākhyās tisraś citta-vrttayah krameņa prayojya, tat-tad-guņa-višista-rasa-carvaņā-janyā iti yāvat, making it clear that the justification of this classifi- cation consists in the divergent nature of the mental activity involved in the relish of Rasa. He does not agree, however, with Mammata in the latter's statement that when we speak of a composition as madhura we use the word in a secondary

12 These topics are also dealt with by Mammata, Viśvanātha and . others.

SOME LATER WRITERS OF THE NEW SCHOOL 257

sense (as when we say "the appearance of this man is brave"), inasmuch as the Gunas being the properties of Rasa, we apply to the 'body' what appertains to the 'soul' of poetry by an extended use of the term. Jagannatha maintains that when we say the sentiment of love (śrngāra) is madhura in a particular case, we mean to imply its influence, such as druti etc. on the mind, and this must be taken to refer not only to Rasa but also to śabda and artha and the composition in general (śabdārtha-rasa-racanā-gatam eva grāhyam). The next chapter proceeds to discuss other varieties of suggestion, including suggestion based on laksaņā, which is dealt with in detail. It then takes up the poetic figures (to the number of about 70), to which the rest of the work, breaking off in the middle of the figure uttara, is devo- ted. The poetic figure or Alamkāra comes in as the source of the charm or ramanīyatā essential in the principal sugges- ted element of poetry already defined (prāg-abhihita-lakşaņa- sya kāvyātmano vyangyasya ramaņīyatā-prayojakā alam- kārāḥ, p. 156). The aesthetic pleasure (camatkāra or lokottaratva) into which this ramaņīyatā resolves itself is an essential element in the poetic figure ; Jagannātha thus har- monises his own conception of poetry with Ruyyaka's theory of the alamkāra (which he accepts and elaborates) as involving this camatkāra (also called hrdyatva, cārutva, saundarya, or denoted by the technical terms vaicitrya, vicchitti-viśeșa or bhaņiti-prakāra) imparted by the conception of the poet (kavi-pratibha). Kuntaka, from whom Ruyyaka appears to have derived his analysis, laid down that in every poetical production the activity of the poet, which consists in an act of the productive imagination (pratibha), is the principal point, and it should result in poetic expression. Jagannätha asserts that the pratibha alone is the source of poetry and therefore of poetic expression, and as such it fixes the nature of the alamkāra. The special charm (vicchitti- viśesa), which is thus imparted to poetic figure by the ima- gination of the poet, is taken (pp. 466, 470) as the basis upon 40

258 SANSKRIT POETICS

which the poetic figures distinguish themselves in their special peculiarities, and is explained as an act of imagination on the part of the poet in so far as it is produced in the poem, or as that aesthetic pleasure13 which is thereby brought into being. As to how this vicchitti is determined, Jagannātha settles the question by resting it not only upon established usage (sampradaya) but also upon one's own internal ex- perience (anubhava). On this fundamental principle, the various figures are minutely defined, differentiated, illustrated and classified ; and this portion of Jagannätha's work, in spite of its subtlety and polemic attitude, is one of the most acute, though unfortunately uncompleted, treatment of the subject.14

13 This translation of the term camatkāra is justified by Jagannātha's own definition of poetry. 14 For an elaborate account of Jagannātha as a literary critic see V. A. Ramaswami Sastri, Jagannātha Pandita (Annamalai Univ. Skt. Series 1942), pp. 78f (chs. iv-vi).

CHAPTER IX

LATER WRITERS ON RASA

( 1 ) The doctrine of Rasa, which is advocated, if not first enunciated, by Abhinavagupta, is finally adopted by almost all writers on general Poetics who accept rasa-dhvani as an important element of poetry. With the exception of Viśva- nātha and Kesava Miśra, they do not indeed go so far as to declare expressly with Abhinavagupta that Rasa alone is the essence of poetry, but they accept in reality the suggested sense in the form of Rasa as essentially the main element. The Rasa is viewed as a pleasant sentiment belonging to the reader whose dormant emotions, derived from experience or inherited instincts, are evoked by the reading of poems into an ideal and impersonalised form of joy ; an appreciation or enjoyment, consisting of a pleasant mental condition in which the reader identifies himself with the feelings of the hero and experiences them in a generic form, the fulness of the enjoy- ment depending upon the nature and experience of the particular reader. The sentiment thus evoked is essentially universal in character, and the aesthetic pleasure resulting from it is not individual (even though enjoyed as an in- timately personal feeling), but generic and disinterested, being such as would be common to all trained readers (samasta- bhāvaka-svasamvedya). It is, therefore, described as something supernormal (alaukika) and invariablv pleasant, not to be compared to the normal pleasures of life which have always a reference to one's personal relations or interests, and which may be pleasant or painful. Things, which would be called causes of an emotion in the normal sense and which may produce disgust, horror or pity in real life, awaken these feelings indeed in poetry and drama, but convey them in such an ideal and generic form that these emotions, which

260 SANSKRIT POETICS

are far from pleasant in ordinary life, are converted into an impersonal joy, which is ineffable and indivisible. One may be moved by disgust, horror or pity and shed real tears ; but the underlying sentiment is always one of exquisite joy1, which must be distinguished from an ordinary feeling. This is the general position of all later theorists with regard to the nature and function of Rasa in poetry. Dhanañjaya, for instance, gives us the same process of transformation of an ordinary emotion, dominant in a composition, into a poetic sentiment, as formally laid down by Bharata and interpreted by Abhinavagupta; and in this he is practically in agreement with Mammata, Vidyādhara Viśvanātha and others. The dominant emotion (sthāyi-bhāva), he says, becomes a sentiment (rasa) when it is brought into a relishable condition through the co-operation of the excitants, the ensuants and the accessories (including the sāttvika bhāvas). This statement is further amplified by the assertion that the enjoyer of Rasa (rasika) is the audience (sāmājika) on whose capacity of enjoyment it depends, and that the dominant feeling becomes a sentiment when it is so enjoyed. The Rasa, being a mental state, a subjective experience of the reader, in which enjoyment (āsvāda, carvaņā, rasanā or bhoga) is essential and in which the enjoyer and the object of enjoyment become identical, the reader receives the represented feeling into his own soul and thereby enjoys it2. The locus of the Rasa is not in the

1 But the Nātya-darpana, as noted above (p. 132 fn), as well as Bhoja (sukha-duhkhāvasthā-rūpa), believes that Rasa is sukha-dukhātmaka. The Rasa-kalikā (vol. i, p. 318) also holds this view. See the elaborate arguments set forth in Nātya-darpana (ed. GOS, p. 159) in support of this view. Siddhicandra (Kāvyaprakāsa-khandana pp 16-21) refers to this theory of the "Navyas" that all Rasas are not pleasurable, but some distinctly painful. They accordingly admit the four pleasurable Rasas, viz. Śṛngāra, Vīra, Hāsya and Adbhuta only, and not those which involve pain, viz. Karuņa, Raudra, Bībhatsa and Bhayānaka. See the question discussed by V. Raghavan, Number of Rasas, ch. viii. 2 Cf. Jacobi in GgA, 1913, pp. 308f.

LATER WRITERS ON RASA 261

represented hero who belongs to the past; nor is it in the poem itself, the task of which is merely to exhibit the excitants etc. by which the dominant emotion is brought into expression and the Rasa, on its part, becomes revealed to the reader. Nor does the Rasa consist of the reader's mere apprehension (prafiti) of the emotions exhibited in the poem or enacted by the actor ; for the reader would then apprehend not the Rasa but a feeling varying in different individuals, just as in real life the spectacle of a pair of lovers in union gives different spectators who witness it the varying emotions, according to their individual nature, of shame, envy, desire or aversion3. The vibhavas etc., therefore, bring the sthāyi- bhāva to the enjoyment of the rasika, the aesthetically receptive reader or spectator, and thereby convert it into rasa ; but they must be generalised and have no specific relation to a particular individual (parityakta-viśeşa). Thus, the vibhāva Sīta, Dhanika explains, must refer to woman in general, and not to the particular individual who was the daughter of Janaka. Hence things, which are the exciting, ensuing or accessory circumstances in ordinary life, act as vibhāvas etc. in poetry, and generalise the dominant feeling into Rasa. The spectator, say, of the deeds of Arjuna on the stage may be compared, therefore, to the child who, in playing with clay elephants, experiences the sensation of its own energy as pleasant. The enjoyment in the spectator's mind is a mani- festation of that joy which is ianate as the blissful nature of self, a circumstance which gives us the frequent comparison of rasāsvāda with brahmāsvāda. The mental activity involved in this enjoyment has got four aspects taken in connexion with the four primary sentiments of the erotic (śrngāra), the heroic (vira), the

3 These circumstances, Dhanika thinks, disprove the vyangyatva of Rasa. It seems that Dhanika does not accept the vyangya-vyañjaka relation of Rasa to Poetry, but holds some views similar to the bhāvya- bhāvaka theory of Bhatta Nāyaka (ed. Parab, 1917, p. 96).

262 SANSKRIT POETICS

horrible (bibhatsa) and the furious (raudra) admitted by Bharata4, and consists respectively of the conditions of un- folding (vikāsa), expansion (vistāra), agitation (kşobha) and distraction (vikşepa). We have seen that Bhatta Nayaka (along with Abhinavagupta) speaks of the bhoga (or āsvāda) of Rasa as involving only three mental conditions, named vikāsa (pervasion), vistāra (expansion) and druti (melting), which later theorists have taken as the basis and justification of the three Guņas of prasāda, ojas and mādhurya respectively. With regard to the ninth Rasa, the quietistic, which is not mentioned by Bharata but which is acknowledged by some theorists, Dhanañjaya forbids its delineation in the drama (iv. 35) ; for the sentiment of absolute peace is in its own nature undefinable, and consists of four states mentioned by philosophers5, viz. maitrī, karuņā, muditā and upekşā, which are not realisable by the Sahrdaya. If it exists at all as Rasa, it must comprehend the fourfold mental activity enunciated above, as corresponding to the fourfold states recognised by philosophers in śamaG. It is not necessary in this connexion to take up in detail the views of Mammata, Vidyädhara and other writers, for it - would be repeating substantially what has already been said regarding the final doctrine of Rasa. Viśvanātha is the only important writer, among later theorists, who boldly accepts Abhinavagupta's extreme view that the rasa-dhvani alone is the essence of poetry and builds up a system of Poetics on its basis7.

4 See above p. 23. The fourfold division is probably adopted as an ostensible rationale for the doctrine of four primary and four secondary Rasas recognised by Bharata. 5 E. g. Yoga-sūtra i. 33. 6 na ca tathābhūtasya śānta-rasasya sahrdayāh svādayitāraḥ santi, atha tad-upāyabhūto muditā-maitrī-karuņopekşādi-lakşaņas tasya ca vikāsa-vistāra-kşobha-vikșepa-rūpataiveti. 7 Bhanudatta, who substantially follows the doctrine of Rasa detailed here, is however singular in his classification of some aspects of Rasa.

LATER WRITERS ON RASA 263

Following up his own definition of poetry as "a sentence of which the soul is the Rasa", Viśvanātha gives us an elaborate analysis of Rasa in almost all its aspects. He sums up at the outset the characteristics of Rasa in two verses thus: "The Rasa, arising from an exaltation of the quality of sattva " or goodness, indivisible, self-manifested, made up of joy and thought in their identity, free from the contact of aught else perceived, akin to the realisation of Brahma, and having for its essence supernormal wonder (camatkāra), is enjoyed by those competent in its inseparableness (as an object of knowledge) from the knowledge of itself". He explains camatkāra as consisting of an expansion of the mind and as synonymous with vismaya. In this connexion, Viśvanātha quotes with approval an opinion of his ancestor Nārāyana who put a premium on the sentiment of the marvellous (adbhuta rasa) and maintained that it was essential in all Rasas. It is also explained clearly that the Rasa is identical with the enjoyment of itself, or, in other words, there is no distinction between the object and the operation in the apprehension of Rasa; so that when we say 'the Rasa is enjoyed', we only use a figurative expression. It follows from this that the enjoyment of Rasa is different in its nature from the ordinary processes of knowledge. Viśvanātha insists very strongly on the necessity of vāsanā

He speaks of Rasa as laukika and alaukika, subdividing the latter into svāpnika (enjoyed in a dream), mānorathika (fanciful like a castle in the air) and aupanāyika (as depicted in poetry). He again gives us (Rasa-taran. ch. viii, p. 65, ed. Regnaud) a three-fold arrangement of Rasa with reference to its manner of manifestation: (i) abhimukha, when it is manifested by means of the bhāva, vibhāva and anubhāva. (ii) vimukha, when these elements are not directly expressed; so called because it is comprehensible with difficulty. (iii) paramukha, which has again two aspects according as it is (a) alamkāra-mukha, i. e, where the alamkāra is principal and the rasa is secondary. This includes probably the cases of figures like rasavat, which are included in guņī- bhūta-vyangya kavya by the Dhvani-theorists, and (b) bhāva-mukha where the bhäva is in the same way principal.

264 SANSKRIT POETICS

in the spectator, which consists of experience (idānīntanī) or instincts acquired from previous births (prāktanī). If one is not endowed with these germs of the capacity of appreciation, one may develop them by study of poetry and experience of . life. In the case of the grammarian, the philosopher or one well-versed in the sacred lore, these susceptibilities are deadened. If it is sometimes found that an eager student of poetry is still deficient in the capacity of relishing Rasa, we must assume that it is the result of his accumulated demerit of a previous birth. Thus, Viśvanātha is anxious to show that experience and cultivation of the power of imagination are essential in one who seeks to enjoy Rasa. Viśvanātha also insists that the vibhāvas etc. as well as the dominant feeling (sthāyi-bhāva) must be felt as generic or impersonalised. The reader must not take the feeling as his own individual emotion; for it would then remain as his feeling (and never become Rasa) and would sometimes (e.g. in the case of the pathetic sentiment) cause pain, and not joy. Nor should the feeling be taken as pertaining solely to the hero ; for then it can not, as the feeling of another person, affect the reader and become Rasa. It is necessary, therefore, that the excitants etc. as well as the dominant feeling, should be generalised by a generic function (sādhāraņī krti) inherent in themselves, which corresponds to the generic power (bhāvakatva) postulated for poetry by Bhatta Nāyaka. This universalisation of the factors and the feelings enables the reader to identify himself with the personages depicted ; and this conceit of community removes all difficulty about accept- ing extraordinary episodes of exalted personages who may be superior in virtue or prowess to the average reader. The excitants etc. are indeed normally called causes, but in reality the Rasa is not an effect in the ordinary sense ; for in the case of Rasa there is the simultaneous presence of itself and its excitants, which is not true of an ordinary cause and effect. It is also pointed out that all the factors (vibhāvas etc.) need not be present at once, for the presence of one would revive

LATER WRITERS ON RASA 265

the others by association of ideas. In other words, what might seem wanting in the utterance of poetry is supplied, from the suggestive character of poetry itself, by force of association of ideas. It also follows from the character of Rasa described above that it is not necessarily found in the actor, who in assuming the rôle of the hero performs his part only mechanically by rule and rote ; he ranks as a spectator (and therefore as a recipient of Rasa) in so far as he is himself a man of taste and actually experiences the feelings he enacts.8

( 2 ) In spite of the unquestioned dominance of the Dhvani School, which no doubt recognised the importance of Rasa but regarded it as one of the phases of the unexpressed only, one class of writers, who still adhered to Rasa as the only element worth considering in poetry, continued to devote exclusive attention to it and built up a system, so to say, on the basis of the Rasa alone. Of all the Rasas, however, as śrngāra (or love) forms the absorbing theme of Sanskrit poetry and drama in general, and as this particular poetic sentiment has an almost universal appeal, these writers naturally work out this important Rasa in all its phases ; and we have in consequence a series of erotico-rhetorical treatises, of which the earliest known and the most remarkable is Rudrabhatta's Srngāra-tilaka9. Rudra states distinctly at the beginning of his work that although Bharata and others

8 This follows Dhanañjaya's dictum that the enjoyment of Rasa is not precluded in the actor, if he realises in himself the feelings depicted. 9 The topics dealt with in its three chapters are: I. The rasas, the sthayi-bhavas, the dramatic vrttis; srngara and its division; the Nãyaka, classified with illustrations ; his assistants ; classification of the Nāyikā, II. Characteristics of love-in-separation, of pūrva-rāga, the ten stages of love, the upāyas, etc. III. The other rasas, viz. hāsya, karuņa, raudra, vīra, bhayānaka, bībhatsa, adbhuta and šānta; the four vrttis appropriate to the rasas.

266 SANSKRIT POETICS

have spoken of Rasa in the drama, his object is to apply it to the case of poetry, and that a Kavya, in his opinion, must possess Rasa as its constant theme. Following upon this we have Bhoja's Śrngāra-prakāśā10, which deals with the subject in the usual elaborate cyclopaedic manner of its author, with profuse illustrations of every phase of the erotic senti- ment in no less than eighteen out of its thirty-six chapters. After this come innumerable works of a similar nature11, which take Rasa, especially śrngāra, as their principal theme and which were composed apparently with the object of guiding the poet in the composition of erotic pieces so popular and profuse in Sanskrit poetry. Of these, the Bhāva-prakāsa12 of Sāradātanaya, which reproduces the substance of most of the chapters of Bhoja's work, and the exhaustive Rasārnava- sudhākara of Singa Bhūpāla13, as well as the two well-known works of Bhanudatta14, deserve mention. But none of these later treatises adds anything new or original to a subject already thrashed out to its utmost. A new turn was given to the theory by Rūpa Gosvāmin's

10 See above p. 209. 11 See vol. i, pp. 238f, and chapter on Minor Writers. 12 See vol. i. p. 240. 13 See vol. i. p. 241f. The three vilasas of this extensive work deal with the following topics: i. The hero, his qualities and classification ; his adjuncts ; the heroine, her classification and qualities, her sāttvika excellences ; the uddipana-vibhavas ; the riti and the gunas ; the drama- tic vrttis ; the sāttvika bhāvas. ii. The vyabhicāri-bhāvas, the anu- bhavas, the eight sthayi-bhavas, the eight rasas. iii. The drama and its varieties, characteristics etc. 14 The eight tarangas of Rasa-tarangini are: i. Definition of bhava and subdivisions thereof ; the sthāyi-bhāvas. ii. The vibhāvas. iii. The anubhāvas. iv. The eight sāttvika bhāvas. v. The vya- bhicari-bhavas. vi. The rasas and detailed treatment of śrngāra. vii. The other rasas. viii. The sthāyi-bhāvajā and rasajā drsti. The Rasa-mañjari, a much smaller work, devotes more than half of itself to the nāyika and her companions, and applies the rest to the śrgāra- nāyaka, his assistants, the eight sāttvika gunas, the two aspects of śrngara and the ten stages of vipralambha-śrngāra.

LATER WRITERS ON RASA 267

Ujjvala-nīla-mani, which attempted to deal with Rasa in terms of the Vaisnava idea of ujjvala or madhura rasa, by which was meant the śrngāra rasa, the term ujjvala having been apparently suggested by Bharata's description of that Rasa15. The madhura rasa, however, is represented not in its secular aspect but primarily as a phase of bhakti-rasa (madhurākhyo bhakti-rasaḥ, i. 3) ; for according to Vaișņava theology there are five Rasas forming roughly the five degrees of the realisation of bhakti or faith, viz., śānta (tranquillity), dāsya (servitude or humility, also called prīti), sakhya (friend- ship or equality, also called preyas), vātsalya (parental affect- ion) and mādhurya (sweetness). The last, also called the ujjvala rasa, being the principal, is termed bhakti-rasa-rāj16 and constitutes the subject-matter of the present treatise. The krsna-rati or the love of Krsna forms the dominant feel- ing or sthāyi-bhāva of this sentiment, and the recipient here is not the literary sahrdaya but the bhakta or the faithful17. This sthāyi-bhāva, known as madhura rati, which is the source of the particular Rasa, is defined in terms of the love of Krşņa18 ; and the nature of nāyaka and nāyika is defined in the same manner and their feelings and emotions illustrated by adducing examples from poems dealing with the love- stories of Krsna and Radha. The work is, therefore, essentially a Vaisnava religious treatise presented in a literary garb, taking Krsna as the ideal hero, with the caution, however, that

15 yat kiñcil loke śuci medhyam ujjvalam darśanīyam vā tac chṛngāreņopamīyate, ed. Grosset, pp. 89-90. 16 i. 2, explained by Viśvanātha Cakravartin as śānta-prīti-preyo- vātsalyojjvala-nāmasu mukhyeşu ......... sa evojjvalāpara-paryāyo bhakti- rasānām rājā madhurākhyo rasaḥ. 17 svādyatvam hrdi bhaktānām ānītā śravaņādibhiḥ/esā krsņa- ratiḥ sthāyī bhāvo bhakti-raso bhavet, cited by Viśvanātha Cakravartin, p. 4. 18 madhurākhyāyā rater lakşaņam coktam-mitho harer mrgāksyās ca sambhogasyādi-kāraņam/ madhurāpara-paryāyā priyatākhyoditā ratih, ibid, loc. cit.

268 SANSKRIT POETICS

what is true of Krsna as the hero does not apply to the ordinary secular hero (i. 18-21)19. With the exception of the Ujjvala-nila-mani, which attempts to bring erotico-religious ideas to bear upon the general theme of Rasa, these specialised treatises have, however, very little importance from the speculative point of view; and as they belong properly to the province of Erotics rather than Poetics, treatment of them should be sought elsewhere. The simple idea, elaborated more or less in all these works is that the awakening of Rasa is all-important in poetry, and that the fundamental Rasa is śrngāra or the erotic, which is consequently treated in its various phases with copious illustrations. This is clearly expressed in the attitude of the author of Agni-purana and of Bhoja, who accept only one poetic Rasa, viz. the erotic20. In the same way, Rudrabhatta declares śrngāro nāyako rasaḥ (i. 20), and Bhānudatta appears to take it for granted that śrngāra occupies an honoured place among all the Rasas (tatra raseşu śrngārasyābhyarhi- tatvena etc, ed. Benares, p. 21).

( 3 ) It is unnecessary, as it is unprofitable, in the diseussion of general principles, to enter here into the elaborate defini- tions, distinctions and classifications of the amorous sentiment with all its varying emotional moods and situations, which these treatises industriously discuss and which have always possessed such attraction to mediaeval scholastic minds. The theorists delight in arranging into divisions and sub-

19 The orthodox theorists (cf Jagannatha pp. 47f) would regard bhakti (which being based on anuräga or attachment cannot be com- prehended by śānta rasa) as included in bhāva, being devādi-vişayā rati, and as inadmissible as a fully developed rasa. Cf. Bhānudatta, Rasa- tarangiņī ch. vi .- On Ujjvala-nīlamani and Vaisnava theory of Rasa see S. K. De, Vaisnava Faith and Movement, Calcutta 1942. 20 See vol. i, p. 137. Cf. also Mandāra-maranda-campū ix, p. 107 (ed. Kāvyamālā).

LATER WRITERS ON RASA 269

divisions, according to rank, character, circumstances and the like, all conceivable types of the hero, the heroine and their adjuncts, together with the different shades of their gestures and feelings, in conformity with the tradition which already obtained in the cognate sphere of dramaturgy.21 Thus Rudrabhatta, after a preliminary enumeration and definition of the Rasas and the Bhavas, proceeds to speak of two as- pects of Śrgāra, viz. sambhoga (love in union) and vipra- lambha (love in separation)22, and classify the hero (nāyaka) into the faithful (anukūla), the gallant whose attention is equally divided among many (daksiņa), the sly (śatha), and the saucy (dhrsta), according to his character as a lover. Later writers, however, subdivide each of these, again, into the best (uttama), the middling (madhyama) and the lowest (adhama), and arrange the whole classification under the four- fold division of the genus hero into four types viz., (i) the brave and the high-spirited (dhīrodātta) (ii) the brave and haughty (dhīroddhata) (iii) the brave and sportive (dhīra- lalita) and (iv) the brave and serene (dhīra-praśānta), thus giving us altogether forty-eight subdivisions of the hero23.

21 See Bharata ch. xxii-xxiv ; Daśa-rūpaka iv. 50f and iii. 22 This statement follows Bharata and is accepted by most theorists including Bhoja; but Dhanañjaya distinguishes three cases, privation (ayoga), sundering (viprayoga) and union (sambhoga): the first denoting the inability of lovers, through obstacles, to secure union, and the second arising from absence or resentment. The first case of love may pass through the well-known ten stages (longing, anxiety, recollection, praise of the beloved, distress, raving, insanity, fever, stupor and death; cf Singabhupāla ii. 178-201) ; while the second condition may be caused by a quarrel, due to discovery or inference of unfaithfulness (which may be counteracted by six upāyas, viz. conciliation, winning over her friends, gifts, humility, indifference or distracting her attention) or by absence arising from business, accident or a curse. 23 The good qualities of the hero are innumerable. For his characteristics, see Dhanañjaya ii. 1f; Viśvanātha iii. 30f; Singa- bhūpāla i. 61f ; etc. On the theme of Nāyaka-Nāyikā, as treated in Alamkāra works, see V. Raghavan, Introd. to his ed. Akbarshahi Brngāra-mañjārī pp. 14-90.

270 SANSKRIT POETICS

Then follows a brief description of the assistants of the hero24 in matters of love (narma-saciva), viz. the Comrade (pītha- marda), the Companion (vita) and the Buffoon (vidūşaka), some adding ceta (or the servant) in the enumeration. In the same way, the heroine is taken broadly in threefold aspects in her relation to the hero as his wife (svīyā), or belonging to another, (parakīya) and as common to all (sāmānyā). The Svīyā is subdivided again into the adolescent and artless (mugdha), the youthful (madhya), and the mature and audacious (pragalbha), i.e. the inexperienced, the partly experienced and the fully experienced. Later authors introduce greater fineness by subdividing each of these accord- ing to her temper, into the self-possessed (dhīrā), the not- self-possessed (adhīra), and the partially self-possessed (dhīrādhīra), or according to the rank, higher (jyestha) or lower (kanistha), each holds in the affection of the hero. The Parakīyā or Anyadīyā who, according to Vaișņava ideas, is the highest type of the heroine, is twofold, according as she is a maiden (kanyā) or married (ūdhā)25 ; while the sāmānyā heroine, who is sometimes extolled (Rudrabhatta) and sometimes deprecated (Rudrata), is only of one kind, the veśya or courtesan26. The sixteen types of heroine thus

24 The Pratinayaka or the rival of the hero is dhiroddhata, haughtiness being his essential characteristic ; but he is described also as stubborn and vicious (Dasa° ii 9; Sāhitya-darpaņa iii. 130, p. 136). The pithamarda of the hero possesses, in a lesser degree, the qualities of the hero (e. g. Makaranda in the Mālatī-mādhava). The term pīthamardikā in the feminine occurs in the Mālavikagnimitra in the sense of a trusty go-between, applied to the nun Kausiki. The vita, usually neglected in the serious drama, except in Carudatta and Mrcchakatika, appears in all his glory in the Bhana, for which he is prescribed as the hero. 25 An amour with a married woman cannot, according to Rudrata and Rudrabhatta form the subject of dominant Rasa in a play or poem ; but this is the central theme of Vaisnava yrics. 26 Bharata xxii. 197-206: Dhanañjaya ii. 21f: Viśvanātha iii. 67-70 ; Singabhūpāla i. 121-51. Rarely a heroine, she must be represented as in love when she is a heroine ; but she cannot be so when the hero

LATER WRITERS ON RASA 271

obtained are further arranged according to eightfold diversity in their condition or situation in relation to her lover, viz. the heroine who has the lover under absolute control (svādhīna-patikā), the heroine disappointed in her assignation through misadventure or involuntary absence (utka), the heroine in full dress expectant of her lover (vāsaka-sajjikā), the heroine deceived (vipra- labdhā), the heroine separated by a quarrel (kalahānataritā, also called abhisamdhitā), the heroine outraged by the discovery of marks of unfaithfulness in the lover (khanditā), the heroine who meets her lover by assignation27 (abhisārikā) and the heroine pining for the absence of her lover gone abroad (prosita-patikā). We arrive in this way at an elaborate classification of the heroine into three hundred and eighty- four types; and one of the later writers states characteristi- cally that there are other types also, but they cannot be specified for fear of prolixity (Viśvanātha iii. 88, p. 120). But here the theorists do not stop. The hero is endowed further by a set of eight special excellences, as springing from his character (sāttvika): e. g. brilliance (śobhā) including heroism, cleverness, truthfulness, emulation with superiors and compassion to inferiors; vivacity (vilāsa) indicated by his glance, step and laughing voice; grace (mādhurya) displayed in placid demeanour even in trying circumstances ; equanimity (gambhīrya) consisting of superiority to emotions ; steadfastness. (sthairya) in obtaining one's object ; sense of honour (tejas) manifested in his impatience of insult; gallantry (lalita) in his word, dress or deportment ; magnani- mity (audārya) exhibited in generosity, agreeable words and equal treatment to friend or foe. The heroine is allowed a

divine or royal. The exception occurs in a prahasana or farce (and incidentally in a bhāna or the erotic monologue) where she can be represented in her low and avaricious character for comic effect. 27 The usual meeting places are given as a ruined temple, a garden, the house of a go-between, a cemetery, the bank of a stream, or any dark place generally ..

272 SANSKRIT POETICS

more generous set of qualities. First we have the three physical (angaja) characteristics ; bhāva or first indication of emotion in a nature previously exempt, hāva or movement of eyes and brows indicating the awakening of emotion, helā or the decided manifestation of feeling. Then we have seven inherent qualities: e. g. brilliance of youth, beauty and passion, the touch of loveliness given by love, sweetness, courage, meekness, radiance and self-control. Then are enumerated her ten graces, to which Viśvanātha adds eight more. All her gestures, moods or different shades of emotion, e.g. giggling, trepidation, hysterical fluster of delight, involuntary expression of affection, self-suppression through bashfulness, affected repulse of endearments, as well as the deepest and tenderest display of sentiments, are minutely analysed and classified. To this is added a detailed descrip- tion of the modes in which the different types of heroines display their affection, the maidenly modest demeanour of the mugdha or the shameless boldness of the more experien- ced heroine. We should recognise the subtle power of analysis and insight which these attempts indicate; but speaking generally, the analysis is more of the form than of the spirit, based on what we should consider accidents rather than essentials. At the same time, marked as.it is by much of scholastic formalism, there is an unmistakable attempt to do justice to facts, not only as they appeared to the experience of these theorists but to the observation of general poetic usage ; and in the elaborate working out of the general thesis that the Rasa is evolved on the basis of one or other of what they call the permanent mental moods, with the help of the various emotional adjuncts, the writers on Poetics have proceeded a long way in the careful analysis of poetic emotions, the psychology of which bears an intimate relation to their theory and in itself deserves a separate study.

LATER WRITERS ON RASA 273

( 4 ) The discussion of this extensive topic of the nāyaka and nāyikā comes in topically under the theory of vibhāva and anubhava, which act as factors of Rasa. The mood, which is at the root of sentiment, is held to be the sthāyi-bhava, the dominant feeling, the main theme of the composition in question. These feelings, according to Bharata, who is accep- ted on this point by all writers, can be classified into eight categories, viz. Love (rati), Mirth (hāsa), Sorrow (śoka), Anger (krodha), Energy (utsāha), Fear (bhaya), Disgust (jugupsā) and Astonishment (vismaya), though some later writers add, as we shall see, Tranquility (sama or nirveda) to the number. These dominant feelings are worked up into a corresponding number of sentiments or Rasas through the means of the vibhavas etc28. The vibhāvas or Excitants are said to be of two

28 Theoretically the Rasa is one, a single ineffable and impersonal joy, but it can be subdivided, not according to its own nature but according to the emotions which form its basis. Bharata (ch. vi) and other theorists give a full description of the sthāyi-bhāvas, vibhāvas etc. in the case of each Rasa, into which space forbids us to enter. A summary of it will be found in Lindenau, Rasalehre Leipzig 1913, pp. 18f. Thus, in the case of the heroic sentiment (vira), the dominant feel- ing is energy (utsāha) ; the excitants (vibhāvas) are coolness (asam- moha), resolve (adhyavasāya), circumspection (naya), strength (bala) etc. ; the ensuants (anubhāvas) are firmness (sthairya), heroism (śaurya), sacrifice (tyaga) etc .; the vyabhicarins or accessory feelings are those of assurance, arrogance etc. Viśvanātha gives them somewhat different- ly. The essential excitant (alambana-vibhāva) of the heroic sentiment, according to him, consists of those to be vanquished, and their acts and gestures form the enhancing excitants (uddipana-vibhāvas) ; the ensuants comprehend the desire or seeking for assistants and adherents; while the accessory feelings are patience, intelligence, remembrance, cogita- tion etc. The sentiment may take three forms of courage (Bharata vi. 79=ed. Regnaud vi. 80), viz. in battle (yuddha-vīra), in virtuous deeds (dharma-vīra) and in liberality (dāna-vīra), to which later writers (e.g. Viśvanātha) add dayā-vīra. It should also be noted that a special colour and a presiding deity is attributed to each Rasa. Thus, red, black, white, dark (śyāma) and grey are associated, not unreasonably, with 41

274 SANSKRIT POETICS

kinds29, viz., (1) the Substantial or Essential (ālambana), which consists of such material and indispensable ingredients as the hero, the heroine, the rival hero and their adjuncts, and (2) the Enhancing (uddipana), viz, such conditions of time, place and circumstance as serve to foster the Rasa, e.g. the rising of the moon, the cry of the cuckoo etc. in the case of the erotic sentiment. The anubhavas or the Ensuants, which follow and strengthen a mood, comprise such outward manifestations of feeling as sidelong glances, a smile, a move- ment of the body, or such involuntary action of sympathetic realisation of the persons depicted (sāttvika)30 as fainting (pralaya), change of colour (vaivarnya), trembling (vepathu) etc., which are, again dogmatically classified into eight varieties. There are other feelings of a more or less transitory nature, which accompany or interrupt the permanent mood without, however, supplanting it ; and these are known, as we have noted, by the name of Accessories or vyabhicāri-bhāvas. These are likened to servants following a king or to waves of the sea, whereby the dominant mood is understood as the king and the sea respectively, and classified elaborately into thirty-three categories, first mentioned by Bharata (p. 23f above) and implicitly accepted by his followers. All these elements contribute towards developing the eight or nine sthyi-bhāvas into eight or nine different types of

the furious, terrible, comic, erotic and pathetic sentiments, although it is difficult to explain why horror is dark blue (nila), wonder is orange, and heroism is yellow. The respective deities are Visnu (erotic), Yama (pathetic), Pramatha (comic), Rudra (furious), Indra (heroic), Kāla (terrible), Mahākāla (disgustful), Brahmā (marvellous) .. Viśvanātha adds that Nārayana is the presiding deity of sānta rasa and the colour associated is that of jasmine (kunda). 29 These two divisions of vibhāva are not maintained by Bharata but distinguished by Dhanañjaya (iv. 2) and traditionally handed down by Viśvanātha. 30 See above p. 24, fn 55. The sāttvika bhāvas in later works form a special class of anubhāvas.

LATER WRITERS ON RASA 275

Rasa. We have the earliest and most orthodox mention in Bharata (p. 23 above) of eight sthāyi-bhavas and the result- ing eight Rasas corresponding to them, of which the Erotic (śrngāra), the Heroic (vīra), the Furious (raudra) and the Disgustful (bibhatsa) are the main, leading to four others, the Comic (hāsya), the Marvellous (adbhuta). the Pathetic (karuna) and the Terrible (bhayānaka). Dandin accepts this classifica- tion (ii. 280-87), but Udbhata (iv. 4) adds 31 the Quietistic (santa) as the ninth Rasa, although Bharata32 neither defines it nor mentions its corresponding vibhāvas. Rudrața is singular in postulating a tenth Rasa, called the Agreeable (preyas), which is accepted by Bhoja, with the addition of two new Rasas, Udātta and Uddhata, as well as Sānta. Rudrabhatța admits nine Rasas in poetry ; so do Hemacandra and the two Vägbhatas. The Agni-purana in the same way mentions nine Rasas (and eight sthāyi-bhāvas), but follows Bharata in regarding four as principal and lays special stress on the Šņngāra. Ānandavardhana admits Sānta (pp 138, 238). Those later authors who accept the ninth Rasa, the Quietistic, necessarily postulate nirveda or self-disparagement, arising out of the knowledge of reality (tattva-jñāna), as its sthāyi- bhava, which is called by some authorities śama, or repose resulting from freedom from mental excitement33. The Vaişņava writers (especially Kavikarņapūra add Dāsya, Sakhya, Vātsalya, Preman and Bhakti.34

31 If the verse is genuinely Udbhata's. See above p. 114, fn 15 .- On the Santa Rasa in Bharata and Dhanañjaya see S. K. De, Some Problems pp. 139-41. On the number and nomenclature of Rasas generally see V. Raghavan, Number of Rasas, Adyar 1940. 32 The Santa texts in Bharata, available in certain recensions, are interpolations. See Raghavan, op. cit. pp. 15f. Kālidāsa knew only eight Rasas, Vikramorvaśīya ii. 18, where Muni Bharata is also mentioned. 33 This sentiment is also closely related to the sentiment of disgust ; for it arises from an aversion to worldly things. 34 See S. K. De, Vaisnava Faith and Movement, p. 145.

276 SANSKRIT POETICS

The author of the Daśa-rūpaka, however, contends that there can be no such sthāyi-bhāva as nirveda or śama, for the development of that state (if it is at all possible to destroy utterly love, hatred and other human feelings) would tend to the absence of all moods; and in the drama, the object of which is to delineate and inspire passion, it is inadmissible. Others, again, hold that the Quietistic Rasa does exist, as it is experienced by those who have attained that blissful state, but it has no sthāyi-bhāva in dramatic composition; for nirveda, being the cessation of all worldly activity, or sama being freedom from all mental excitement, it is not fit to be represented. Hence Mammata takes eight Rasas in the drama (p. 98) and nine in poetry (p. 117). Bhoja, in accordance with the views of the school which lays special emphasis on the Śrngara, accepts only one Rasa, the Eroti in his Srngāra- prakāśa; and although he mentions as many as ten Rasas in his Sarasvatī-kanthabharana, including the santa and the preyas, he appears to devote almost exclusive attention to the Srngara in his treatment of the Rasas in this work. The views about the admissibility of the santa are discussed by the author of the Ekāvali (pp. 96-7) who maintains that Bharata has mentioned nirveda as a vyabhicāri-bhāva immediately in context after the enumeration of the sthāyi-bhāvas and at the beginning of the list of the vyabhicāri-bhāvas ; and this fact is interpreted as indicating that the sage meant it both as a sthāyi-bhāva and as a vyabhicāri-bhāva; but Hemacandra (p. 81) anticipates and rejects this quibble of verbal interpreta- tion, though agreeing in the general proposition as to the admissibility of Sānta as the ninth Rasa, Viśvanātha primarily admits s eight orthodox Rasas (iii, p. 160) but adds the ninth Sänta in deference to the views of these authorities, and a tenth Rasa, called vātsalya or parental affection, subscribing apparently to Vaisnava ideas (pp. 185-6)35. He quotes a verse to explain that the mood,

35 Bhanudatta counts (Rasa-taranginī) māyā under the Rasas.

LATER WRITERS ON RASA 277

called by the great sages the Quietistic, which has, among all sentiments, tranquillity (śama) as its basis, is that state in which there is neither pain nor pleasure, nor hatred, nor affection, nor any desire36. But the question arises how can the Quietistic, being of the nature described, arising only in a state of emancipation wherein there is an absence of all feelings like the Accessories etc., be Rasa, which implies a state of relishable enjoyment37. To this objection Viśvanāthā replies that the Quietistic is a Rasa because in that state the soul is only about to be emancipated (yukta-viyukta-daśā) and is not completely absorbed in the Divine, so that the presence of feelings, like the Accessories etc. in it is not incompatible. As for the statement that there is an absence of even pleasure in it, it is not contradictory, for it refers only to worldly pleasure88. Jagannätha, the latest writer on the subject, advocates nine Rasas and maintains (pp. 29-30) that like all other Rasas, the Santa is capable of being represented and appreciated by the audience. Since the clever performance of the actor, representing such a state of mind, free from disturbance and not affected by passions or desire, is found in actual experience to produce an impression on the mind of the audience, it is their state of mind, exhibited by their silent and rapt attention, which ought to settle the question. The representation of absolute indifference or the actor's power of representing it is not the point in issue: it is the capacity of

Rudrața mentioned preyas (friendship), which Rasa is accepted by Bhoja. Some writers add śraddhā, along with bhakti. See Bhānudatta, op. cit. p. 56, ll. 25f (ed. Regnaud), Śinga-bhūpāla admits only eight Rasas, but his treatment is from the standpoint of dramaturgy. 36 na yatra duhkham na sukham na cintā, na dveșa-rāgau na ca kācid icchā/ rasaḥ sa śāntaḥ kathito munīndraiḥ, sarvesu bhāveşu śama- pradhānah//, cited also in Daśa° iv. 49 (comm). 37 ity evam-rūpasya śāntasya mokşāvasthāyām evātma-svarūpāpatti- lakşaņāyām prādurbhūtatvāt tatra sañcāryādīnām abhāvāt katham rasatvam. 38 yaś cāsmin sukhābhāvo'pyuktas tasya vaişayika-sukha-paratvān na virodhaḥ.

278 SANSKRIT POETICS

the spectator who actually feels the sentiment. Jagannātha also adds that even those, who do not admit this Rasa in the drama, should accept it in poetry from the fact that poems like the Mahābhārata have for principal theme the delineation of Santa Rasa, which is thus èstablished by universal experience (akhila-lokānubhava-siddhatvāt). Nāgeśa remarks on this that the santa rasa should also be admitted in the drama on this ground, inasmuch as the Prabodha-candrodaya is universally acknowledged as a drama (p. 30). Coming to the essential basis of Rasa, viz. the bhāva, we have seen that Bharata defines it in general terms as that which manifests the sense of poetry through the three kinds of representation, vācika, āngika and sāttvika39 ; it is the emotion which ultimately becomes a sentiment, if it is dominant and therefore, serves as the basis of Rasa. But later writers arrive at a greater precision and apply the term technically to those cases where there is no proper or complete development of Rasa. Both Dhanañjaya and Bhānudatta expand the definition of Bharata, the latter defining it as a deviation from the natural mental state (vikāra) which is favourable to the development of Rasa (rasānukūla) and which may be either physical (śārīra) or mental (āntara). But Mammata fixes the conception of Bhāva as ratir devādi- vişayā vyabhicārī tathāñjitah ('love having for its object a deity or the like, and also the suggested Accessory'), on which he adds the gloss: ādi śabdān muni-guru-nrpa-putrādi-vişayā, kāntā-vişayā tu vyaktā śrngārah (by the term the like are meant sages, preceptor, the king, son etc., the one having a beloved woman for its object becomes the erotic') Govinda explains that the word rati here implies the sthāyi-bhāva which has not attained to the state of Rasa40. What is meant is that when the sthāyi-bhāvas, like rati, have for their objects

39 A fourth kind of abhinaya is sometimes added, viz. āhārya (extraneous) i.e. derived from dress, decoration etc. 40 ratir iti sthāyi-bhāvopalakşaņam, devādi-vişayety apy aprāpta- rasāvasthopalakșaņam, p. 206.

LATER WRITERS ON RASA 279

god, king, son and the like, or when the vyabhicāri-bhāvas are manifested as the principal sentiment in a composition, there is no rasa but bhāva ; and this definition is accepted by all writers after him. Thus, Viśvanātha explains the Bhava as follows: sañcāriņaḥ pradhānāni, devādi-vişayā ratiḥ/ udbuddha-mātraḥ sthāyī ca. bhāva ity abhidhīyate / / In other words, when the Accessories are principal, or when love etc. has a deity or the like for its object, or when a dominant feeling (sthāyi-bhāva) is merely awakened, we have Bhäva. His own gloss upon the above verse explains it in this way. Although they are always concomitants of Rasa in which they finally rest, such Accessories as are for the time being principally'developed, like a servant for the time being followed by his king in his marriage procession ; or love etc. having a deity, a sage, a spiritual guide, a king and the like for its object ; or such sthāyi-bhāvas as are merely awaken- ed or have not attained the state of a Rasa from their not "being fully developed, are denoted by the term bhāva. In all these cases apparently there is no complete or proper development of Rasa ; and a Bhäva, therefore, in later termi- nology, may be generally described as an incomplete Rasa. But this must be distinguished from the rasābhāsa or sem- blance of Rasa and the analogous bhāvābhāsa, which occur when the poetic sentiments and emotions are falsely attribut- ed (e.g. sentiments in animals such as described in Kumāra- sambhava, iii. 36-7), or when they are brought out improperly, i.e, when there is a lack of entireness in them as regards their ingredients41. The cases occur, according to Bhoja (v. 20) when the mood or emotion is developed in an inferior character (hīna-pātra), in animals (tiryac), in the rival hero (nāyaka-pratiyogin) or in any other subordinate object (gauna

41 anaucitya-pravṛttatve ābhāso rasa-bhāvayoh (Mammața) explain- ed as: anaucityam cātra rasānām bharatādi-pranīta-lakşanānām sāmagrī- rahitatve tveka-desa-yogitvopalakşana-param bodhyam.

280 SANSKRIT POETICS

padārtha)42, but Viśvanātha elaborately summarises various other cases (iii. 263-66), especially noting improprieties in connexion with particular Rasas. Thus, there is an impro- priety if the Terrible (bhayānaka) is made to reside in a noble personage, or the Comic (hāsya) in a spiritual guide. It must be noted, as Jagannātha explains, that if a mood or feeling is developed by impropriety, the impropriety, unless it acts as a bar, does not constitute a fault43. In the same way (1) when there is an excitement only (and not full development) of sentiments, (2) when two opposing sentiments, striving for mastery, are represented as being relished in one and the same place and at the same time, or (3) when a number of sentiments, of which each succeeding one puts down the preceding, they constitute res- pectively bhāvodaya, bhāva-samdhi and bhāva-śabalatā. Now, all these phases of sentiment are taken as Rasa topically, inasmuch as they are capable of being tasted (sarve'pi rasa- nād rasāḥ). These cases do not seem to have been formally recognised by Bharata, though hinted at by him in vi. 40, as we learn from Abhinava's commentary on ch. vi, which is partially reproduced also in his ° Locana, p. 66. They are first met with in Udbhata, who includes them under ūrjasvin (iv. 6) ; but in Rudrata (xii. 4) and the Dhvanikāra (ii. 3) we find them definitely established. This incomplete development of Rasa and its subordina- tion must be distinguished from the cases of the opposition (virodha) of simultaneously existing sentiments in the same theme. It is laid down formally that some Rasas are intrinsi- cally inconsistent with one another, e.g. the Erotic is opposed to the Disgustful, the Heroic to the Quietistic, and so forth44.

42 Śinga-bhüpāla (pp. 141-2) distinguishes two cases (i) where Rasa is ascribed to an inanimate object and (ii) where it is developed in an inferior character or in animals. 43 yāvatā tvanaucityena rasasya pastis tāvat tu na vāryate, rasa- pratikūlasyaiva tasya nişedhatvāt. 44 Some Rasas again are mutually consistent, e. g. karuņa and

LATER WRITERS ON RASA 281

The incongruity or opposition results in three ways, viz. (1) from identity of the exciting cause (ālambana-vibhāva) (2) from identity of the subject of emotion and (3) from immediacy of succession. The incongruity in the first two cases may be removed by representing the sentiments as having different exciting causes respectively, or as existing in different subjects (e.g. in the hero and the rival hero). The last case of conflict may be removed by placing, between the two immediately succeeding sentiments, a sentiment which is not opposed to them. These are cases where two or more Rasas stand in the relation of principal and subordinate; the term 'subordinate' being misleading, it is sometimes called a concomitant Rasa (sañcarin), which implies that it cannot terminate absolutely in itself and at the same time is distinct from a fully developed Rasa, as well as from a mere un- developed Bhäva45. There is also no incongruity where a conflicting Rasa is recalled or described under a comparison. All these questions properly come under the theory of Propriety or Aucitya in relation to Rasa, elaborated by Anandavardhana and his followers, and is ultimately based on the dictum attributed to the Dhvanikāra (p. 145, cf. ° Locana p. 138), which lays down in general terms that the secret of Rasa lies in conforming to the established rules of propriety.

bībhatsa go with vira; śrngāra goes with hāsya (cf Bharata vi. 40) etc. On this question see Lindenau, Rasalehre (pp. 71f). According to Viśvanātha, the Rasas hostile (i) to śrngāra are karuņa, bībhatsa, raudra, vīra and bhayānaka (ii) to hāsya-bhayānaka and karuņa (iii) to karuņa -hāsya and śrňgāra (iv) to raudra-hāsya, śrngāra and bhayānaka (v) to vīra-bhayānaka and śānta (vi) to bhayānaka -śrngāra, vīra, raudra, hāsya and śānta (vii) to śānta-vīra, śrngāra raudra, hāsya and bhayānaka (viii) to ->bībhatsa-śrngāra. Bhānudatta gives the antagonistic Rasas as follow: śrngāra->bībhatsa; vīra->bhayānaka; raudra->adbhuta ; hāsya->karuņa. 45 ata evātra pradhānetareşu raseșu svātantrya-viśrāma-rāhityāt, pūrņarasa-bhāva-mātrāc ca vilakşaņatayā, sañcāri-rasa-nāmnā vyapa- deśaḥ prācyānām, Viśvanātha, p. 420.

282 SANSKRIT POETICS

The doctrine of the Dhvanyāloka that in a composition in which the sentiment is awakened, proprieties of various kinds (e. g. with reference to the speaker, the theme, the employ- ment of the vibhāvas etc., the use of the alamkāras and other elements, pp. 134f, 144f) should be observed, and that certain items of conflict (virodha) with the dominant sentiment should be avoided, gave rise to a theory of Propriety, which is generally comprehended by later writers under the discussion of the Dosas of Rasa. Thus, in later treatises, the rasa-dosas occupy a separate and important place, in addition to the conventional doşas of pada, padārtha, vākya, vākyārtha recognised since Vämana's time. It is Ksemendra alone who emphasises the importance of the subject by making it the theme of his Aucitya-vicara-carca which will be noticed in its proper place. Mahimabhatta, in the second chapter of his work, considers the question of anaucitya in some detail. According to him, impropriety or incongruity has two aspects, according as it refers to śabda or to artha respectively. Then he speaks of propriety as external (bahiranga) or internal (antaranga), apparently as it is śabda-vişaya or artha-vişaya. The cases of internal propriety, which consists in the proper employment of the vibhāvas etc. have already been explained by previous writers (e.g. the Dhvanyāloka pp. 144f). Mahima- bhatta, therefore, takes up the question of external propriety, which he thinks falls under five faults of composition, viz. vidheyāvimarśa (non-discrimination of the predicate), prakrama-bheda (violation of uniformity in the expression), krama-bheda (syntactical irregularity), paunaruktya (tauto- logy) and vācyavacana (omission of what must be expressed), to the explanation and exemplification of which he devotes, amidst several digressions, the rest of the chapter (ch. ii). It is difficult to say why these faults of expression alone are singled out as defects resulting in a violation of Rasa (rasa- bhanga). Later writers would include them under general defects, reserving the cases of virodha or opposition of Rasas as specific instances of rasa-doşas.

CHAPTER X

WRITERS ON KAVI-ŚIKSĀ

The small group of writers who deal with the theme of kavi-śikșā ("education of the poet") does not, strictly speaking, come directly under general Poetics, but deserves notice, partly from the reputation and authority enjoyed by some of these authors but chiefly because it displays a peculiar tendency which emphasises one aspect of Poetics as a discipline, namely, its practical object which developed side by side with the theoretical consideration of general principles These treatises do not deal with the conventional topics of Poetics, with its theories, dogmas and definitions, but they are meant chiefly as manuals to guide the poet in his profession, their primary object being kavi-śikșā or instruction of the aspiring poet in the devices of the craft. It is difficult, in the absence of data, to determine the origin of this school, but the attitude adopted is significant, being almost co-extensive with what may be supposed to have been the original standpoint of Poetics itself as a more or less mechanical Ars Poetical. The ancient as well as the modern writers on general Poetics, no doubt, touch occasionally upon the question of the practical training of the poet2 ; and it is not improbable that this in course of time formed the object of a separate study and multiplied these convenient handbooks, of which necessarily we possess comparatively late specimens.

( 1 )

Kşemendra Kşemendra's two works, Aucitya-vicāra-carcā and Kavi- kanthabharana, which may be conveniently grouped here, are

1 See above pp. 33-34 and footnote 3. 2 See above pp. 42f.

284 SANSKRIT POETICS

curious and valuable in many respects. In his theory of Aucitya3 or propriety, he takes as his thesis, mainly, Ānan- davardhana's treatment of the same question with reference to Rasa which is crystallised in the oft-quoted verse from the Dhvanyāloka: "There is no other circumstance which leads to the violation of Rasa than impropriety ; the supreme secret of Rasa consists in observing the established rules of propriety"4. To depict Rasa, it is necessary to observe the rules of propriety ; and the subject, which is anticipated by Bharata (who, for instance, speaks of the proper employment of anubhāvas), may take various forms according as it relates to the subject-matter, the speaker, the nature of the sentiment evoked or the means employed in evoking it. We have al- ready noted that this theme has been discussed topically at some length by the formulators of the Dhvani-theory, by Mahimabhatta, as well as by most post-dhvani writers who consider it generally in connexion with rasa-doşas. Kşe- mendra develops and pushes the idea to its extreme, and speaks of Aucitya as the essence of Rasa (rasa-jīvitabhūta), and as having its foundation in the charm or aesthetic plea- sure (camatkāra) underlying the relish of Rasa. The Alam- kāra and Guņa in poetry are justified by and receive their respective significance from this element which may, therefore, be fittingly regarded as the 'soul' of poetry5. That which is suitable or conformable to another is called ucita in its relation to that object6. This Aucitya may have application with reference to various points in a poem, such as word (pada), a sentence (vākya), the sense of the composition as a whole (prabandhartha). its literary excellences (gunas) its

3 V. Raghavan (Some Concepts. pp. 194-257) gives an able and detailed account of the history of Aucitya. 4 anaucityād rte nānyad rasa-bhangasya kāraņam/ prasiddhaucitya-bandhas tu rasasyopanişat parā/ / 5 ed. Kāvyamālā, Gucchaka i, pp. 115-16. 6 ucitam prāhur acāryāḥ sadršam kila yasya yat, explained by the gloss as: yat kila yasyānurūpam tad ucitam ucyate .- On the theory of Pāka and Sayyā, allied to the theory of Aucitya, see above pp. 240-41.

WRITERS ON KAVI-ŚIKȘĀ 285

poetic figures (alamkāras), the Rasa or the sentiment in a poem, the employment of the verb (kriyā), the use of the case (kāraka), of the gender (linga), of the number (vacana), of preposition, adjective, particles (upasarga), or consider- ations of time and place (desa and kāla) etc .; and the cases of application are dogmatically summarised as twenty-seven in number (śl. 8-10). The treatment of each of these cases is accompanied by profuse illustrations of every point from the works of various poets, the favourite method being first to cite examples of verses which comply with a rule and then adduce one or two examples of verses which do not do so. There is hardly anything original in the theory itself ; for though it rightly insists upon a standard of propriety in poetic expression, it ultimately resolves itself into assuming a more or less variable criterion of taste or personal apprecia- tion, which Anandavardhana and others admit as sahrdayatva, but which is bound to be, as it is, rather vaguely defined and therefore incapable of exhaustive formal treatment. It is really the province of taste or criticism rather than of Poetics proper. At the same time, the fact must not be overlooked that works of this nature in Sanskrit, ostensibly meant as they are for the guidance of the aspiring poet, display, in their discussion of what is right and proper in poetry, a tendency towards genuine criticism, taken apart from the beaten paths of orthodox Poetics ; and they set up in their naïve way a standard, whatever it may be, of taste and critical judgment. No doubt, most writers on general Poetics, betray some critical acumen and give us a great deal of critical or semi- critical matter while considering the application of a rule or principle, especially in the chapters on Dosa and Guņa; but their outlook is often and necessarily limited by their confining themselves to rigid rules and specific definitions7.

7 The growth of artificial poetry, we have seen, made the technical analysis of rhetoric and instruction of it, a necessity ; but rhetoric involves (and sometimes becomes identical with) criticism ; and it is

286 SANSKRIT POETICS

Ksemendra's work, in this respect, possesses a unique value, and the part of his treatment which discusses the illustrative verses is extremely interesting as an evidence of "apprecia- tion" which is comparatively rare in Sanskrit. Kse- mendra deals out praise and censure, within his limits, as a true critic who is no respector of persons ; even the honoured names of Amaru, Kālidāsa or Bhavabhūti make no difference. In more than one instance, he illustrates two sides of a question, regarding both merit or defect laid down by a rule, by different verses from his own work ; and in some cases he does not hesitate to go against orthodox opinion8. Whatever may be the intrinsic value of his critical dicta, some of which may appear too trivial or crude to us, he shows a wide acquaintance with the whole range of classical Sanskrit Poetry and an undoubtedly cultured taste. If the common saying that a bad poet often turns out to be a good critic carries any wisdom in it, it is very apt in the case of Ksemendra whose critical powers cannot indeed be ignored.

almost impossible for Alamkārikas, who also theorised on principles, not to busy themselves with the forms and general phenomena of literature. Thus, most works, whether on general poetics or on rhetoric, do involve some amount of criticism which could not be avoided. It must also be borne in mind that our modern ideas of Aesthetics, Poetics or Rhetoric are not sharply distinguished in these old authors, nor was there any well-defined notion of the respective spheres of these studies. The theorists drew their ideas of poetry mainly from existing classical Sanskrit literature which, though magnificent in partial accom- plishment, was not fully equipped for purposes of general criticism. The absence of some other literature for comparison-for later Prakrit and allied specimens are mainly derivative-was a serious drawback. This will explain partially why their outlook is so limited, and their principles and definitions so stereotyped. 8 E. g. while discussing the question of propriety of the contents of a composition, he cites (p. 120) from Kumāra-sambhava viii-which canto he accepts as Kālidäsa's-and severely censures the poet's manner of describing the amours of Hara and Parvatī in terms of ordinary dalliance, against the authority of Anandavardhana who defends (p. 137) it against the imputation of vulgarity.

WRITERS ON KAVI-SIKSĀ 287

Kşemendra's other work, Kavi-kanthābharana9, though less interesting, is equally remarkable for its refreshingly novel treatment. Ksemendra postulates two impulses for the attainment of poetic capacity, viz. divine help (divya- prayatna) and individual effort (paurușa). The first includes prayer, incantation and other heavenly aids ; but from the latter standpoint, he classifies three groups of persons10 with whom instruction in the art of poetry is concerned, viz. those who require little effort (alpa-prayatna-sādhya), those who require great effort (krcchra-sädhya), and those in whom all effort is fruitless (asdhya), and sums up by saying that the capacity for poetry is vouchsafed only to the fit and few. The next chapter discusses with illustrations the question of borrowing or plagiarism11, a theme which is just touched upon in the fourth chapter of the Dhvanyaloka12 but which is dealt with extensively by Rājaśekhara13. Kșemendra divides

9 A sketch of this work will be found in Kşemendra's Kavikanthā- bharana by J. Schonberg (Wien 1884), pp. 9f. The five samdhis or sections of this work deal respectively with the following themes: (i) attainment of poetry by an unpoetical person (akaveh kavitvāptiḥ), (ii) instruction of the poet already gifted (śiksā prāpta-giraḥ kaveḥ), (iii) strikingness (camatkrti), and the faults and excellences of poetry, (iv) familiarity which a poet should possess with other arts and sciences as a source of charm to his poetry (paricaya-cārutva). 10 Cf Vāmana 1. 2. 1-5 ; Rājaśekhara iv. 11 Bāņa (Hara-c. i. 5-6) distinctly condemns poetasters and plagiarists. Vämana appears to be the first writer on Poetics who in his classification of Artha refers to the question of plagiarism. 12 In Anandavardhana's opinion, the province of poetry is unlimited, in spite of the fact that hundreds of poets have composed works for centuries; but the thoughts of two inspired poets may bear certain resemblance, which may be like that between an object and its reflection, between a thing and its picture, or between two human beings. The first two kinds of resemblance should be avoided, but the third is charming (iii. 12-13). 13 In ch. xi-xii. For a summary of his views see V. M. Kulkarni, Sanskrit Writers on Plagiarism in JOS, iii (1954), pp. 403-411. Rājaśekhara declares that "there is no poet that is not a thief, no merchant that does not steal, but he who knows how to hide his theft flourishes without

288 SANSKRIT POETICS

poets from this point of view into those who imitate the general colour of a poet's idea (chayopajīvin), those who borrow a word or a verse-line (padaka- and pāda-upajīvin), or an entire poem (sakalopajivin) and lastly, those who borrow from sources considered universally as legitimate (bhuvano- pajīvya, e.g. Vyāsa). Then he lays down elaborate rules for regulating the life, character and education of the poet. This is followed by a discussion of camatkāra or poetic charm, without which, we are told, no poetry is possible, and an illustration (by means of examples draw from the works of various poets) of its tenfold aspect, according as it appeals with or without much thought (avicārita-ramaņīya or vicārya- māna-ramanīya14), resides in a part or in the whole composi- tion, appertains to the sound, the sense or both, or relates to the poetic figure, to the sentiment, or to the well-known nature of the theme. Then we come to the treatment of the ex- cellences and defects with reference to the sense (artha), the verbal expression (śabda), or the poetic sentiment (rasa) involved; and the work is rounded off by indicating the extent of knowledge which a poet must possess and giving a long list of the arts and sciences in which he must be proficient, which is thus set forth: tatra tarka-vyākaraņa-bharata-cāņakya- vātsyāyana-bhārata-rāmāyaņa-mokşopāyātmajñāna-dhātuvāda- ratnaparīkşā-vaidyaka-jyotişa-dhanurveda-gaja-turaga-puruşa- lakşaņa-dyūtendrajāla-prakīrņeşu paricayaḥ kavi-sāmrājya- vyañjanah. This rapid summary of the contents of this work

reproach". He deals with two kinds of plagiarism, namely, that which should be avoided and that which should be adopted. In his opinion, a poet may be a creator (Utpādaka) or an adapter (Parivartaka), or a coverer up (Ācchādaka) or a collector (Samgrāhaka). He who sees something new in word and sense and writes up something old may be accounted a great poet. Rājaśekhara accordingly gives an elaborate classification of Artha so far as it is Anya-yoni, Nihnuta-yoni or A-yoni from the point of view of plagiarism. See below under Rājaśekhara. 14 Rājasekhara attributes a dictum to Udbhata which says that the sense may be vicārita-sustha or avicārita-ramaņīya, according as it is found in the Sastra or the Kāvya respectively. See above p. 59, fn 33.

WRITERS ON KAVI-ŚIKSĀ 289

will show that it hardly puts forward any special claim as a work of great theoretic importance, but that its value consists not in its substance but in its treatment of practical issues, its careful and minute illustration of every point by examples taken from various poets, with not a little amount of knowledge and critical discernment1 5.

( 2 )

Arisimha, Amaracandra and Deveśvara The Kāvya-kalpalatā-vrtti of Arisimha and Amaracandra and the Kavi-kalpalata of Deveśvara, written in imitation of that work, need not detain us long. They are essentially treatises on the composition of verses, including a practical treatment of prosody and rhetoric. They furnish elaborate hints on the construction of different metres, on the display of word-skill of various kinds, on jeu de mots and tricks of producing double meaning, conundrums, riddles, alliterative and rhyming verses, and various other devices of verbal ingenuity, concluding with a chapter on the construction of similes and enumeration of parallelisms for the purpose of ordinary comparisons. It gives also a list of kavi-samayas or conventions observed by the poets, and states in detail what to describe and how to describe it. These decadent treatises, therefore, offer such adventitious aids for ready- made poetry, as may-to take a particular point-be afforded, for instance, by a modern rhyming dictionary or works of similar nature. A summary of the main topics dealt with in the Kāvya- kalpalata and its Vrtti will make the standpoint clear and give an idea of the general scope and nature of such works. The first pratana of this work is called chandah-siddhi (prosody)

15 For Ksemendra's satiric and didactic works see S. K. De, Aspects of Sansk. Lit. Calcutta 1959, pp. 279-83 and Hist. of Sansk. Lit., Calcutta 1947, pp. 404-410. 42

290 SANSKRIT POETICS

and consists of five sections on (i) the construction of the anustubh metre (anustubh-śāsana), (ii) enumeration of the principal metres, shifts in grammatical forms of the verb. Prakrit loan-words (where Hemacandra seems to be super- ficially quoted), transmutation of one's own or another poet's ideas into the same or different metres, conversion of one metre into another, caesura (yati), the whole section being. generally entitled chandobhyāsa, (iii) use of expletive particles. and words for filling up the verse (chandah-pūrana), such as. śrī, sam, sat, drāk, vi, pra etc. (sāmānya-śabdaka), (iv) argu- mentation, pointed sayings, subjects of laudation or vituper- ation, interrogations, e.g. in kula-śāstrādi, sva-śāstrādhyayana- prathā etc. (vāda), (v) subjects for descriptive poetry,. how to describe the king, his ministers, the prince, the army,. battle and hunting, as well as a city, a village, a garden, a lake and so forth ; enumeration of the kavi-samayas (varnya- sthiti). The second chapter, called śabda-siddhi treats of etymology, derived meanings of compounds, alliteration and rhyme in the middle of a verse, with a list of words suitable for this purpose, enumeration of sambandhin expressions ; denoted, indicated and suggested meanings, showing the influence of the Dhvani school. The next chapter, entitled ślesa-siddhi, upon play on words of various kinds, consists of the following sections: (i) composing of words in such a way that they can be read differently according as they are divided, with a list of ślesopayogi words, (ii) a kind of śleşa occurring in the description of an object by analogies, in which the same quality or condition has to be traced in the same words or in synonyms, (iii) cases of double meaning, produced by homonyms capable of widely different inter- pretations, (iv) ambiguity produced by similarity of inflections of different origins, (v) surprises of different kinds, such as verses in which the same consonant or vowel is repeated in each syllable, verses spread out in diagrams to be read in different ways, e.g. according to the move of the knight in chess etc :- a subject which is treated in some detail, for

WRITERS ON KAVI-ŚIKSĀ 291

instance, in the Vidagdha-mukha-mandana of Dharmadāsa Süri16. The last chapter, called artha-siddhi, is devoted to the construction of similes, ellipsis and similar figures, and gives · long lists of parallelisms arising from like conditions or attri- butes of the objects compared, e. g. the lips may be compared to the coral, to the bimba fruit or fresh-blown twigs and so forth. A large part of this treatment is repeated in Keśava's Alamkāra-śekhara, as well as in Deveśvara's Kavi-kalpalatā, the latter work being directly modelled on the Kāvya-kalpa- latā with considerable plagiarism of passages in extenso These works, therefore, do not require any separate notice .. Keśava's text (see vol. i, pp. 220-21 above) deals, besides this, with the ordinary topics of Poetics, set forth as the views of his master Sauddhodani, but in substance and form it follows the views of Mammata, Hemacandra and the Vāgbhatas; it thus possesses hardly any claim to originality either in theory or in treatment. Most of the works of Jaina authors, even including those of Hemacandra and the Vägbhatas, are written apparently from the practical standpoint of composing a suitable text-book, and they always, in their discussion of general principles, incorporate hints on matters helpful for the practical working out of poetry17.

( 3 )

Rājaśekhara (Although written in a fanciful style and hardly presenting one systematic theory, Rājasekhara's Kāvya-mīmāmsā may be noticed here, inasmuch as it mixes up the topics of kavi-śiksā with those of Poetics proper, at the same time giving us a somewhat rambling treatment of various extraneous matters.

16 The subject is dealt with as early as Dandin, Rudrata, and the Agni-purāņa. 17 E. g. Hemacandra, pp. 5-15, 126-135 ; the younger Vägbhata pp. 38-68. Both borrow largely from Ksemendra and Rājaśekara.

292 SANSKRIT POETICS

The work is also remarkable for its varied collection of different opinions, as well as for the light it throws on the literary practices of a certain period. Its views cannot be directly connected with any particular school, but it is quite . possible that its author follows in the main a tradition of opinion inherited from his literary ancestors, whom he frequently quotes as the Yāyāvarīyas. The origin of Poetics is attributed by Rājasekhara to the Supreme Being and the celestials, and he claims a very high position for the discipline, which is regarded as the seventh anga without which the significance of Vedic texts cannot be grasped. ) The self-born Srīkantha taught this science to his sixty-four will-born disciples, among whom the most vener- able was the Kāvya-purusa, born of Sarasvatī, who figures as the nominal hero of this half-allegorical work. As Praja- pati set him to promulgate the science to the world, he imparted it to his seventeen divine pupils, Sahasrākșa and others, who embodied it in eighteen separate adhikaranas on the portions learnt by each18. 'Our author seeks to set forth in one book, consisting of eighteen adhikaranas, the substance of these teachings which were in his time, to some extent, lost. If we are to accept this plan of the author, only the first adhikarana on kavi-rahasya exists of this ambitious work. ) The Kavya-purusa, from whom metrical speech first began and who stands symbolically for the spirit of poetry, is the son born to the goddess of learning, Sarasvati, as the result of her long penance on the Himālayas. In order to keep the boy company, Sasasvatī creates Sāhitya-vidyā as his bride who follows him and wins him over. {On this slight conceit the book proceeds to set forth its peculiar doctrines, including in its desultory scope various literary remarks and dogmas, as well as topics like general geography, conventions observed by poets, a disquisition on the seasons, an account of kavi-gosthī and other relevant and irrelevant subjects.

18 See vol. i, pp. 1-2.

WRITERS ON KAVI-ŚIKSĀ 293

The work begins (ch. ii) by dividing literature (vānmaya) into śastra (both human and revealed) on the one hand, and kāvya, on the other. It enumerates the differnt Sastras and defines their nature and form, including under the revealed Sāstras the Vedas, the Upanisads, and the six angas (the Yāyāvarīyas taking Alamkāra-śāstra as the seventh), and comprehending under human Sāstras the Purāņas, Itihasa, Ānvikşikī, the two Mīmāmsās and the Smrtis. It then mentions fourteen (or eighteen) vidyā-sthānas, bringing under it several technical and philosophical disciplines. The mean- ings of the terms sūtra, vrtti, bhāşya, samīkşā, țīkā, pañjikā, kārikā and vārttika, which are the different forms or styles of the Sästras, are then explained, incidentally giving an ety- mological definition of sāhitya-vidyā19. Then, after a digression (ch. iii) on the fable of the Kāvya-purusa, the author goes on to deal (ch. iv) with the different kinds of pupils to whom a knowledge of the science can be imparted, viz. buddhimat and ähārya-buddhi, the latter of whom may be again anyathā-buddhi and durbuddhi, and discusses in this connexion the force of śakti (genius), pratibhā (poetic ima- gination), vyutpatti (culture) and abhyāsa (practice). The Yāyavarīyas think that sakti is the only source of poetry and it gives rise to pratibha and vyutpatti ; but others hold that the aid of concentration (samadhi) and practice (abhyāsa) is also required. The pratibha20 may have a twofold aspect, accord- ing as it is creative (kārayitrī) or discriminative (bhāvayitrī). The creative faculty may be natural (sahaja), adventitious (āhārya) or acquired by instruction (aupadeśika), and poets are accordingly classified as sārasvata, ābhāyasika and aupa- deśika. The discriminative faculty (bhāvakatva) is distinguish- ed from the poetic (kavitva). The bhāvaka may be either 'the dicontented' (arocakinah, i.e. those who possess the faculty but

19 See above p. 37, fn 5. 20 Defined as: yā śabda-grāmam artha-sārtham alamkāra-tantram mukti-mārgam anyad api tathāvidham adhihrdayam pratibhāsayati sā pratibha.

294 SANSKRIT POETICS

require to be guided), 'those feeding on grass' (satrnābhyava- hāriņaḥ, i.e., vulgar persons absolutely devoid of the faculty21), 'the envious' (matsarinah) and lastly, 'the really discerning' (tattvābhiniveśinaḥ) who are rare. In the next chapter (ch. v) we have elaborate classifications of the poet from different points of view. Poets may be group- ed generally into three classes, the śāstra-kavi, the kāvya kavi and the ubhaya-kavi. The śastra-kavi may either compose the sāstra, or produce kāvya-effect in the Sāstra or śāstra-effect in the Kāvya. The kāvya-kavi is classified elaborately, if not very logically, into eight groups, viz. racanā-kavi, śabda-kavi, artha-kavi, alamkāra-kavi, ukti-kavi, rasa-kavi, mārga-kavi, and sāstrartha-kavi. Then we have an enumeration of ten grades of apprenticeship through which a poet has to pass until he becomes a kavi-raja, which is indeed not the highest distinction but which, according to Rājaśekhara who was himself so designated, indicates a status even higher than that of a mahākavi. Elsewhere in ch. x, he gives an account of the test or literary examination of poets for such honour and recognition, in which the successful poet was conveyed in a special chariot and crowned with a fillet (patta-bandha). He speaks also of purity of body, speech and thought necessary for a poet, and describes the house of the poet, his attendants, his writing materials, the division of his whole day into eight parts and duties appropriate thereto. The chapter under discussion concludes with a reference to the theory of pāka22, of which as many as nine varieties, named after the taste of different fruits, are mentioned. The next chapter (ch. vi) deals with the word and the sen- tence, and their functions grammatical, logical or otherwise. In this connexion Rajasekhara states that a sentence possessing the literary excellences (gunas) and embellished by poetic figures (alamkāras) constitutes poetry (guņavad alamkrtam ca vākyam

21 Cf Vāmana 1. 2. 1-3. 22 See above pp. 240-42.

WRITERS ON KAVI-ŚIKSA 295

eva kāvyam, p. 24). If any definite conclusion can be drawn from this statement, Rājasekhara, in general theory, appears to recognise tacitly the position of the Rīti school ; for in this sentence he reproduces Vämana's well-known dictum (kāvya- sabdo'yam guņālamkāra-samskrtayoḥ śabdārthayor vartate, on i. 1.1). This is supported also by the apparent disfavour he shows towards the view of Udbhata and Rudrata, as well as by the marked partiality attached to the opinions of Mangala and Vāmana, whose classification of Rīti is accepted on p. 31. It is true that his school lays special stress also on Rasa23, and like most writers coming after Anandavardhana, Rājaśekhara does not fail to bring Rasa into prominence. This makes it difficult to take his work as framed definitely for any particular system. But it is clear that his sympathies ally him with the older Riti and Rasa schools, rather than with the new school of Anandavardhana who, though cited at p. 16, does not appear to have influenced his views greatly. It is probable that he is following some old tradition, which :stands apart from orthodox schools, but which has many things in common with the older currents of thought and opinion. y The rest of the work, devoted to topics of a similar character, does not throw any further light on his general view of Poetics. The seventh chapter, which comes next, analyses modes of speech on a novel basis, having reference to the promulgation of different religious doctrines, into brāhma, śaiva and vaisnava, with their sectarian subdivisions ; and after a brief mention of the three Rītis of Vamana24, we have some remarks on Käku and on the methods of reading or pronunciation of different peoples, incidentally discussing the question of appropriate language and style of gods,

23 E.g. kim tu rasavata eva nibandho yuktah, na nīrasasya p. 45. 24 Rajasekhara's account of the origin of Rītis is curious. He says that on account of the Sahitya-vidya's wanderings through various countries, different poetic forms evolved themselves, the important jamong them being the three Ritis mentioned by Vamana.

296 SANSKRIT POETICS

Apsarasas, Piśācas etc. The eighth chapter enumerates the- sources or auxiliaries of poetry (kāvya-yonayah), already referred to by Bhāmaha (i. 9) and Vāmana (i. 3), such as the- scriptures, the law-books, the epics, the Puranas etc., and gives a long list of arts and sciences, as well as philosophical systems, which contribute to the content of poetry25. The next chapter (ch. ix) is concerned with the possible themes of poetry, topically referred to by Ānandavardhana (p. 146), according as it deals with incidents and personages, human, divine, or pertaining to the lower world (pātāla), by them- selves or in different combinations. But he adds that the subject-matter must be rasavat. The tenth chapter speaks. of the conduct of a poet, his household and surroundings, his daily duties and routine of work ; it then proceeds to speak of the king who patronises him, and one of whose duties is to call assemblies of poets and scholars. Two very interesting chapters (xi-xii) follow on the elaborate26 classification of the different shades of borrowing or plagiarism (harana). with. reference respectively to borrowing of words and borrowing

25 These are: śruti, smrti, itihāsa, purāņa, pramāņa-vidyā, samaya -- vidyā, rāja-siddhānta-trayī (artha-śāstra, nāțya-śāstra and kāma-śāstra), loka, viracanā (= kavi-manīşā-nirmitam kathā-tantram artha-mātram vā), and prakīrņaka (miscellaneous, like hasti-śikşā, ratna-parīkşā, dhanur- veda etc.). In ch. x. he speaks of (i) kāvya-vidyās, viz. nāma-dhātu- pārāyaņa (=grammar), abhidhāna-kosa (lexicon), chando-viciti (prosody) and alamkāra (poetics), (ii) sixty-four kalās, called upavidyās (accessory studies) and (iii) kāvya-mātaraḥ, viz. kavi-samnidhi, deśa-vārttā, vidagdha-vāda, loka-yātrā, vidvad-gosthī, and purātana-kavi-nibandha. 26 Hemacandra (pp. 8f) and Vagbhata (pp. 12f) plagiarise and reproduce this portion of Rajasekhara's treatment and draw also partly upon Kşemendra (see above p. 287f). On these passages, see F. W. Thomas in Bhandarkar Commemoration Volume, pp. 379-383). To Anandavardhana's classification of three kinds of resemblance which may be found in two poets (see above p. 287 fn 12) these writers add a fourth kind, viz. "foreign-city-entrance" likeness (parapura-praveśa- pratimatā), i. e. where there is substantial identity, but the garnishing is widely different. And of these four kinds, the superiority is in the ascending order.

WRITERS ON KAVI-ŚIKSĀ 297

of ideas. A verse is cited towards the end which says that there is hardly any poet who does not 'steal' from others, but the best of stealing is cleverly concealing the fact27. But mere reflection or copying of ideas is condemned as unpoetical (so'yam kaver akavitva-dāyī sarvathā pratibimba kalpaḥ pariharanīyah, p. 68). The true poet is said to be one who discovers something novel in the expression of words and ideas, as well as restates what is old28. The next chapter (ch. xiii), therefore, details thirty-two different modes by which plagiarism or literary borrowing may be skilfully turned to advantage (a question which must have assumed some im- portance in Rājaśekhara's time), all the points in these interest- ing chapters being profusely illustrated by examples drawn from the works of various poets. This discussion is followed by three chapters (ch. xiv-xvi) on the established poetic conventions (kavi-samaya), with reference to countries, trees, plants, flowers etc., as well as about intangible things (e. g. a smile should always be described as white). There are two more chapters (ch. xvii-xviii) on geography (deśa-vibhāga) and the seasons (kāla-vibhāga) respectively, the former mentioning the countries, rivers, mountains etc. of India, the products peculiar to each, the colour and complexion of various peoples, and the latter describing the winds, flowers and birds, and actions appropriate to various seasons. This bare outline of the eighteen chapters of the Kāvya- mīmāmsa, so far as it is available and actually published, will make it clear that nearly the whole of its content falls, strictly speaking, outside the province of general Poetics, whose conventional topics have thus far been hardly touched upon. At the same time, some of the subjects dealt with by Rājasekhara have been referred to, if not elaborately dealt

27 nāsty acaurah kavi-jano nāsty acauro vanig-janaḥ/ sa nandati vinā vācyam yo jānāti nigūhitum/ / 28 śabdārthoktişu yah paśyed iha kimcana nūtanam/ ullikhet kimcana prācyam manyatām sa mahākavih / /

298 SANSKRIT POETICS

with, by even orthodox writers like Vamana; and the unique evidence of the comparatively early work of Rājaśekhara on this topic, written ostensibly in conformity with some old tradition, will go to support the hypothesis that shitya or the art of poetry originally included in its comprehensive scope all such varied literary topics, until there was a gradual branching off of kavi-śiksā as an allied but separate discipline, and a limiting of the Sastra itself to the discussion of more or less general principles. In themselves, however, these topics are extremely interesting and throw a great deal of light on some of the literary aspects of classical Sanskrit Poetry and its practice, They are made all the more delight- ful reading by Rājaśekhara's concise but easy and picturesque style, especially as it is enriched by judiciously selected and varied illustrations, very unlike the conventional illustrations one meets with in an ordinary text-book on Poetics.

SUBJECT-INDEX TO

Vol. II Entries are confined to those passages which contain a substantive reference to, and not mere citation of, the persons, works or subjects indicated. The figures denote pages, and references to footnotes are marked with an asterisk. Occasional references to vol. i are given with figure i. The following abbreviations are used: fig=poetic figure; Agp=Agnipurāņa; Anv=Ānandavardhana; Abh=Abhi- navagupta; Bh=Bharata; Bhã=Bhamaha; Bh N=Bhatta Nāyaka; Bhānu=Bhānudatta; D=Dandin; Dh=Dhanañ- jaya ; Dhk=Dhvanikāra; Hc=Hemacandra; Jg=Jagan- nātha; Knt=Kuntaka; Ks=Kşemendra; L=Lollata; Mmt=Mammața; Mkc=Māņikyacandra; Mbh=Mahima- bhatta; Rdt=Rudrata; Rk=Ruyyaka; Ud=Udbhata; Vg I=the older Vāgbhata ; Vg II=the younger Vāgbhata ; Vm=Vāmana ; Vis=Visvanātha ; Vid=Vidyādhara ; Vin= Vidyānātha; Snbh=Śingabhūpāla. abhańga-śleșa 64 in relation to vyañjanā abhāva-vādin, the school 150f ; Dīrghavyāpāra- which denies dhvani 53; vādins on 152f ; as basis of Bha and Ud are not of this viva- kşitānya-para - vācya school 53; reactionary dhvani 160, 321-2 ; postula- writers of this school 180 ted by Bh N 183 ; poetic abhidha, denotation of a word, figs as particular forms of 145-47; includes lakşaņā in 188 ; admitted by Agp 201 Bh N 124, 234 ; its far- abhidheya (or samketita) artha reaching function 125; as 152 ; distinguished from a means of generalisation vyangyārtha 153 127 ; cannot comprehend abhihitanvaya-vadins 149 Rasa 136 ; as defined by Abhinavagupta on, the object different schools 183-4; of poetry 40 ; on pratibhā

300 SANSKRIT POETICS

or śakti 42 ; on rītis and abhivyakti, a fig, in Agp 205 vrttis 105 ; comments on abhiyoga. See abhyāsa Bha's vakrokti 109 ; inter- abhyāsa, poetical exercise or prets Dandin's idea of practice 42-43, 293 ; same rasa 111-12 ; influence of as D's abhiyoga 42 Bh N on I27; maintains acala-sthiti or aprthak-sthiti essentiality of rasa in or asthiratva, of figs in poetry and in drama 128, relation to rasa 169, 227 166, 176, 224 ; criticises actor, when he is an appre- Bh N's theory of rasa ciator of rasa 265 130 ; his theory of rasa- Acyuta Raya and his work bhivyakti (vyakti-vāda) 21, Sāhitya sāra 247 131f, 135. 161*, 166, 177f ; adbhuta (rasa), implied in on previous existence of Bh's udāra 15 ; emphasis- dhvani-theory 144; on the ed by Nārāyaņa 263 ; its. relation of indicated and presiding deity 274 suggested senses 153-54 ; adhikopamā 210 defines citra-kāvya 158 ; adhīrā (nāyikā) 270 importance of his exposi- tion of dhvani-theory 176, adhyavasāya, complete super- imposition, as basis of 179; influence of his classification of figs 73 theories of rasa 166, 259 Aesthetics,involved in Poetics. abhinaya, vācika 17 ; āhārya 49, 285* 278 ; three kinds of 20, 278 aesthetic pleasure. See plea- abhineyartha, drama, 44. sure See rūpaka aesthetic attitude, distin- abhiplutärtha, a doșa in Bh 9 guished from the philoso- abhimāna, a lakșaņa, 249 phic and the natural, 126, abhimukha (rasa) 263 136f, 137* abhisaņdhitā (nāyikā). See Agni-purāņa, treatment of kalahāntaritā yamaka in 7 *; its . abhisārikā (nāyikā) 271 definition of udāra 15 *; abhivyakti (of rasa), Abh's its classification of rīti and (q. v) theory of; synony- guņa 104, 202-4 ; schook mous with carvanā (q. v.) of opinion represented by 136 180, 200-1; nature and

SUBJECT-INDEX 301

scope of the alamkāra chapters its position explained 33- in 201, 204f ; features common 34, 213f ; Bhä not its ori- with Bhoja's work 201f ; ig- ginator 38; Ud as its expo- nores dhvani 201 ; unorthodox nent 54f ; Rdt's relation to standpoint and spirit of it 59f; its decline connect- eclecticism in 201f, 205f ; its ed with the rise of rīti- treatment of rasa 202; its system 66; D's relation to classification of figs 204f ; it 66-67, 75-76, 78 ; its accepts nine rasas 275, but external theory of embel- emphasises śrngãra 202; its lishment not accepted treatment of rīti and guna fully 67; influence on later 202f ; first admits sabdārtha- systems 68 ; compared to guņas 204; its borrowings rīti-doctrine 102f; not from older sources 201f, 205 ; coextensive with rhetoric first admits ubhayālamkāra 74, 189 *; developed by 72, 205 Knt (q. v), Rk (q. v) and agrāmyatva, in D's mādhurya others 80, 110, 111 *. See alamkāra, as a term applied grāmyatva to the discipline i. 4f, āhārya, a kind of abhinaya 12f, 15 ; 32f, 67 278 alamkära, in the general sense āhārya-buddhi (pupil) 293 of beauty or embellish- akşara-dambara, of the Gau- ment 82-83, 99 das 75 alaņkāra, poetic figure, as a akşara-samhati, a lakșaņa, technical term unknown 249 in early lit. 2; its limited Alamkāra, doctrine or system sense and number in Bh of 25-26 ; no particular 2, 3, 5-7 ; not defined by theory in Bh 32; why so Bhã 26, but its prominence called 32-33 ; contempor- in his system 32, 46, 52; aneous with dramaturgic conclusions re develop- Rasa systems 33 ; earliest ment of the discipline known exponents of 38; drawn from the treatment coextensive with the origi- of Bh and Bhã 27f; nal standpoint of the Bhatti's treatment of 30- discipline itself 33, 213; 31 ; the discipline itself

302 SANSKRIT POETICS

takes its name from 33-34; 249, 251 ; in Agp 204; in in Ud 54f; in Rdt . Bhoja 206, 210; in Mmt 60, 61, 63f, 115, 116; in D 221; in Vis 227; in Hc 66-67, 78, 84f ; its wider 243 *; in Jayadeva 249 and sense in D 67, 82-83 ; not Appayya 251; in Jg 253,. distinguished from guņa 257-58 in Bha, D and Ud 58, 66-67, - 82-83f; sharply alamkāra-dhvani, 129, 163,.

distinguished from guna 173, 177f; not conveyed by

by Vm 99f ; its existence abhidhã 152 ; included in

justified in poetry by Vm samlakşya- krama- vyangya.

100 ; its place in dhvani- 161; in dīpaka 162 ; its.

system 67, 158, 169-171, place in a complete sche-

173, 187-88 ; Knt's theory me of dhvani 163-64 ; re-

of 185f; as an aspect of solves ultimately into rasa- dhvani 166, 177-78 ; can be and ccextensive with vakrokti (q. v.) 51-52, 49 ; vācya as well as anumeya 195-6 ; Vis on 224 significance of, justified by its vaicitrya (q. v.) due alamkāra-doșa 89 Alamkāra-kaustubha 251 to kavipratibhā (q. v.) 74, 185f, 188f, 217, 228, 230f, Alaņkāra-mañjarī 229

231, 257 ; included in citra- alamkāra-mukha (rasa) 263*

kāvya 158, 188; its relation Alamkāra-samgraha 54. See. Udbhața to rasa 169-70, 219 ; its importance recognised in Alamkāra-sarvasva 228. See

later systems 68f, 169f, Ruyyaka

216-17; process of its Alamkāra-śāstra. See Sanskrit multiplication 63, 68, 72- Poetics

73, illustrated by an exam- alamkāra-sūtras 2

ple 68f ; difference of opin- Alamkāra-śekhara 242, 291'

ion on the nature and alamkriyā, used in the sense scope of individual figs of alamkāra 83-84 70f, the subject not fully alaukika, supernormal or dis- treated 70 *; classification into subvarieties72f; works interested, an attribute of

entirely devoted to 71-73, the relish of rasa 125, 130, 132 and fn, 137, 254, 259;

SUBJECT-INDEX 303

essential to poetry 255 by Śańkuka 119f See lokottara anumeya artha, inferable Amaracandra. See Arisimha meaning, 156, 195f anaucitya, See aucitya anupalabdhi, non-cognition, anibaddha (kāvya) 76* as applied to the inference antaranga (or artha-vişaya) of dhvani 197 aucitya 199*, 282 anuprasa, alliteration, 7 ; antarbhāva-vādin 53 distinguished from yamaka anubaddha or anuvamśya 27 ; in Ud 55; in Rdt ślokas i. 25, 26-27 62, 64; in Bhoja 80 ; in anubhava, a kind of know- Vm 101 ledge 125 ; as a means of anvaya, logical connexion, establishing vicchitti 258 149,153 anubhava, factor of rasa, 17; anvaya-vyatireka, a principle explained by Bh 20f; re- of distinguishing figs of cognised by Ud 56, 114; sound and sense 72, 233 the term not used by Bhā anvitabhidhana-vadins 149 109; defined by later anyadīyā (nāyikā). See para- writers 274 kīyā anukārya (or utpādya), rasa anyārtha, a doșa, 10 as 118 anyathā-buddhi (pupil) 293 anukūla (nāyaka) 269 apabhraņśa (kāvya) 44, 77 anumana, logical inference, apahetu, a doșa, 82 cannot comprehend rasa apahnava, conceal ment, basis 136, 197f ; Mbh's theory of classifying figs 74 of 195f; how comprehends apahnuti, a fig, 69 ; defined. dhvani 196f by D, Vm and others 101f anumāna, a fig, includes fig apakrama, a doşa, 10 hetu in D 65 ; distinguish- apavarga (Sāmkhya) 126 ed from logical anumāna apārtha, a doșa, 9 231 Appayya Dīkşita, accepts anumāna-theory (or anumiti- Rk's doctrine of alamkāra vāda) re vyañjanā-vrtti, 232 ; his three works 251; repudiated by Anv 156f; criticised by Jg 253 as set forth by Mbh 156, aprasiddhi, an upamā-doșa, 65 180, 195f ; re rasa, as held aprastuta-praśamsā a fig, 52

304 SANSKRIT POETICS

:apratīta, a doșa, 88* arthālamkāra, distinguished .ardhabhrama, a fig, 85 from śabdālamkāra 7, 27*, Arisimha and his commentator 37, 61-62 ; reason of this Amaracandra, their Kāvya- distinction 72; in Rdt kalpalatā and Vrtti 289; 62 ; number of, subject to indebtedness of Deveśvara fluctuation 73; classifica- to 289, 291 tion 73f; in D 85f; in Vm arocakin 293 101; in Agp 204f; in artha, sense or idea, its differ- Bhoja 210 ; in Mmt 221 ent forms 150*, 157 ; as arthantara, a doșa, in Bh 8; a ground of inference 196 possibly same as vācyā- artha-dosa 88 vacana (q. v,) and includes artha-dușța, a doșa, in Bhā svaśabda-vācyatā of rasa 11, called aślīlatva by (q. v.) 8* later writers arthāntaranyāsa, a fig, in Bhā artha-guna, Bh's gunas mostly 28 ; in Agp 204 16; clearly distinguished arthāntara-samkramita-vācya from śabda-guņa by Vm (dhvani) 191, 256 15, 82, 93; implied by D artistic expression, theory of 82 ; Vm's scheme of 94f, 49 ; artistic attitude 137* criticised as useless by aślīlatva 11. See grāmyatva Mmt 169, 220 ; in Agp asambaddha, a doșa, 88 203 ; in Bhoja 209 asambhava, an upamā-doșa artha-hīna, a doșa, 9 65 artha-pāka. See pāka asamlakşya-krama vyangya, artha-rasa, in D 111* 161; its function in sug- artha-śleșa 38, 56, 64, 233 gesting rasa 161, 192, 218 ; arthavyakti, a guņa in Bh why so called 162; its 14; in D 80, 81, 82; in sphere distinguished from Vm 15*, 95; equivalent that of fig rasavat 192-93 ; to D's svabhāvokti 15*, Mbh's explanation of 196 95, 97 ; of older writers, asam artha, a dosa, 88 * comprehended by Mmt's asādhāraņopama 86 prasāda 172, 219 *; in atimātra, a doșa, 88 * Vin 238 atiśaya, involved in vakrokti artha-vyāpāra 146 50-51,187. See atiśayokti

SUBJECT-INDEX 305

atiśaya, elevatedness, a basis aupadeśika (poet) 293 of classifying figs in Rdt Aupakāyana 38 62, 73 ; a fig in Agp 204 aupamya, comparison, as a atiśayokti, a fig, involved in basis of classification of other figs 50-57, 85, 163 ; figs in Rdt 62, 73; in identified in substance Vm 62*, 73, 101; in Rk with vakrokti by Bhā and 73 Knt 50, 163 ; lokottaratva (q. v.) implied in it 51-2; aupanāyika (rasa) 263*

in Ud 55 ; in Bhoja 210 aurjitya, a guņa, 239 Avantisundarī on pāka 241 .atyukti, opposite of kānti-guņa in D 81, 87 avasara, a fig, 63* avācaka, a doșa, 10 atyanta-tiraskrta-vācya avicārita-ramaņīya (artha) 59*, (dhvani) 191, 256 288 aucitya, a fig, in Agp 205; avidyā (philosophical) 136 an excellence in Knt 191 -avivakşita-vācya dhvani, bas- aucitya, theory of, anticipated ed on lakşaņā (lakşaņā- by Bh 284; based on Dhk's mūla) and includes meta- dictum 199, 281, 284; in phorical expression 160 relation to doșa esp. rasa- ayoga (śrňgāra) 269* doșa 88*, 282 ; Rdt's use ayuktimat, a doşa, 11 of the term 115* ; in rela- Ābhīra 77* tion to samghatanā (q. v.) ābhyantara guņa 209 165; Mbh's treatment of ābhyāsika (poet) 263 199*, 282; in Ks 282, āgama-virodhi, a doșa, 10 283 ; aspects of 283-85; ākānkșā, expectancy of words defined by Ks 284 ; equiva- 9 fn 25, 149 lent to sahrdayatva and ākșepa, a fig, 28 ; difference of incapable of formal treat- opinion re its scope and ment 285 nature 70-71, 101 Aucitya-vicāra-carcā 282, 283. ākhyāna 76 fn 4 See Kşemendra ākhyāyikā, distinguished from Audbhata 33*, 98 kathā (q.v.) 44f, 76 audārya, a guņa. See udāra ānanda 40, 135, 136 audārya, an excellence of the Ānandavardhana, on yamaka hero 271 7 *; on rītis and vrttis 104, 43

306 SANSKRIT POETICS

105fn, 167-68 ; his familia- bhangī, bhangī-bhaņiti, in- rity with the views of Bh, volved in vakrokti (q.v.) Bhã and Vm 19f, 145; his 52, 185, 186 influence in later schools bhaņiti or °prakāra, or °viśeșa, 145, 175; on the relation involved in vicchitti or of lakşya to vyangya artha vakrokti 185, 186, 232 154 ; defines citra-kāvya Bharata, his views on Poetics 158-59, 171 ; exposition of in his Nātya-śāstra 1f ; on rasa 18f, 46*, 23, 174-75, kāvya 3 fn 5; on lakșaņa. 177f; on three kinds of 4f ; on figs 5-7; on doșa suggestion 157f; on guņa 7f ; on guna 11-16 ; on rasa and alamkāra in relation to 17f; his sūtra on rasa rasa 16, 168, 169 ; on figs differently interpreted and devoid of suggestion 171; giving rise to divergent. accepts three gunas 171; theories 20f, 21, 108, 117, on plagiarism 287* 118, 120, 130 ; on bhäva āropita-śabda-vyāpāra 146 20-21, 278; accepts eight āśīḥ 5* rasas 23, 275; on the āśrayāśrayi-bhāva, as a prin- function of the drama- ciple of distinguishing tic art 39 ; his ten. figures of sound and sense gunas ultimately reduced 72f, 233 to three 176 (see guņa); āsvāda, relish (of rasa), 21, his treatment of Poetics 131, 139 and fn, 260f compared to that of Bhã āvantikā (rīti) 104, 265 25f ; influence of 16, 19f, āvrtti, a fig, 86 127,260 bahirańga (or śabda-vişaya) Bhartrmitra 147 aucitya 199, 282 Bhatta-mata i. 37 bādhayat, a doșa, 88* Bhatta Nāyaka. See Nāyaka bāhya, guņa, 209 Bhațta Tauta. See Tauta bhadrā, a vrtti, 64 Bhatti 6*, 7*, 30-31 bhakti, used for lakşaņā (q.v.) bhayānaka as rasa 132* 154 ; the theory that dhvani bhākta (lākșaņika). See is bhākta 154f, 192, 199 bhakti bhakti, as rasa 276*, 267 ; as Bhāmaha, ignores drama- bhāva 279* turgy and Rasa 2; on āśĩḥ

SUBJECT-INDEX 307

5 ; on doșa 9-11, 46; on bhava, a fig, in Rdt 60, 63* ; rīti and guņa 16,45-6,'79 ; Ud's view on 60 on figs 27f, 46; practical bhava, feeling or emotion, as character of his treatment the basis of rasa 20, 21, of rhetorical categories 39, 24*, 278; recognised by 47 ; his idea of vakrokti Ud 56; in later writers (q. v.) 47f, 48f, 185 ; first 278-79 ; distinguished from definite scheme of Poetics rasābhāsa 279 in 32f ; his general idea of bhava, an excellence of the poetry 37; onfigs of sound heroine, 272 and sense 37; on the pur- bhāva-dhvani 256 pose of poetry 39f; on bhāvakatva (or rasa-bhāvanā), the sources of poetry and postulated by Bh N 123-24 equipment of the poet 183, 184 ; its derivation 41f; on pratibhā 41; 124 ; Abh on 130-31, 133 contents of his work 44f; Bhāva-prakāśa 266. See his classification of kāvya Śāradātanaya 44; his indifference to bhäva-mukha (rasa) 263* rīti and guņa (q. v.) 45-46 ; bhāvayitrī (pratibhā) 293 on vaidarbha and gauda bhāvika, a fig, 47 fn 22, 50* ; kāvya 46, 75, 79 ; on excluded by Vm 101 rasa 52, 108-9 ; on dhvani bhāvika, a guņa, 239 52f, 229 ; his relation to bhāvotpatti, bhāva-śabalatā Ud and Knt 54f, 187, to and bhāva-samdhi, aspects Rdt 65-66, to D 76 of bhava, defined 280; Bhamaha-vivaraņa 68 not expressly mentioned Bhāmahālamkāra or Kāvyā- by Bh 280; included in lamkāra of Bhāmaha 32f fig ūrjasvin by Ud 280; bhāna 270, 271* definitely established in Bhanudatta, his works on Rdt and Dhk 280 rasa and their contents bhinna-vrtta, a doșa, in Bhā 10 266 ; on bhāva 278 ; his bhinnārtha, a doșa, in Bh 9 peculiar classification of Bhoja on Upamāna i. 11 ; on rasa 262*, admits śrngāra yamaka 7 ; on the pur- as principal ; accepts māyā pose of poetry 39 ; on rīti as rasa 276 104, 210 ; school of opinion.

308 SANSKRIT POETICS

followed by 180, 200-01, buddhi (philosophical) 126 206-7; borrowings by buddhimat (pupil) 293 206f; nature and object Candrikā, on Dhvanyāloka of his work 206-7 ; deve- 181* lops Agp's definition of camatkāra, involved in rasa poetry 208 ; gives pro- 137*, 253, 263 ; involved minence to rasa, but in vakrokti 187, 232 ; as accepts srngara as the an element of alamkāra only rasa 202, 209, 276 ; 257 ; as explained by Jg his relation to utpatti- 253 ; as the foundation vādins 208 ; ignores dhvani of aucitya 284; its tenfold 206, mentions twelve rasas aspect as the basis of 209, 275, 276, 277* ; his poetry 288 classification of gunas carvaņā (of rasa) 21 . 131, 209-10; apparent emphasis 136, 137, 260 laid on guņa 208f; his caturvarga, as the aim of praudhi-guņa equivalent poetry 40, 115 to pāka 241 ; on rasā- cärutva, used for vicchitti bhāsa 279-80; value of (q.v) 42, 230, 232, 257 his work 211 cit-svabhāvā (samvit) 126, 136 bhojakatva (or bhoga), enjoy- citra or citra-bandha, a class ment of rasa, postulated of fig, in Rdt and D 61, by Bh N 123-5, 183; Abh 62, 84, in Agp 204 ; not on 131 ; a term denoting mentioned by Bh, Bhã or the aesthetic attitude 137, Ud 64 ; Māgha on 64, 85*, 260; mental activity Mmt on 64 ; discredited in involved in 220*, 262 ; idea later times in theory 85 *; of, admitted in Sāmkhya specialised works on 85 ; 126 its relation to citra-kāvya bhukti-vāda. See Bhațța (q.v.) 85 Nāyaka citra-kāvya 85 ; explained bībhatsa, as a rasa 132 158-59 ; includes figs devoid body · of poetry. See kāvya- of suggestion 158, 163, 171, śarīra 188; Rudrata on citra 61, brahmāsvāda, likened to rasā- 85; not poetry but an svāda 125, 136, 261 imitation thereof 159, 164;

SUBJECT-INDEX 309

suffered by Mmt but rejec- on the distinction between ted by Vis 158*, 225 ; why guņa and alaņkāra 83f; admitted by Anv 159 ; the on rasa 110f ; rasa in his exclusive province of Rk's mādhurya-guņa 110f ; rasa work 229 ; Jg on 255 as an element of certain Citra-mīmāmsā 232 figs 111f ; his idea of rasa Citramīmāņsā-khaņdana 253 similar to that of Lollata citra-turaga-nyāya 120 (q.v.) 112; indebtedness Citrāngada 38 of Agp and Bhoja to 202, Critic, the. See sahrdaya 203, 204, 205 ; his attitude Criticism, in Sanskrit, its towards dhvani 229 comparative rarity and its Daśa-rūpaka 2, 260 limitations 285 ; in Ks dāsya. See prīti 285-86 Denotation or denoted (con- cūrņa (prose) 13, 76* ventional) sense. See abhi- chāyā 204* dhã chekānuprāsa 55 deśa-bhāşā 77* dakşiņa (nāyaka) 269 deśa-virodhi, a doșa, 10 Dandin, ignores dramaturgy deśya, a doşa, 88* 2 ; on lakşaņas 4, 315 ; on Deveśvara, his Kavikalpalatā āśīḥ 4 ; on doșa and guņa 289 ; its scope and its 8, 9*, 13*, 14*, 15*, 25, 26, borrowings 291 27*, 28, 29, 30, 86f ; on Dhanañjaya and Dhanika, on kāvya-śarīra 34-5 ; on the lakşaņas 4, 251 ; Vin's purposes of poetry 39 ; on indebtedness to 237; on pratibhā 41; intermediate śanta as rasa 276f ; on the position between Rīti and nature and function of Alamkāra schools 66, 75- rasa 260-61 ; his classifica- 76, but allies himself with tion of śrňgāra 269 the latter 76 ; his agree- Dharmadāsa Sūri 85*, 291 ments and differences with dhvanana, a name for sugges- Bhämaha 76f ; his classifi- tion 157 cation of poetry 76; his dhvani, use for sphota 143f ; theory of rīti and guna Mmt's definition of 143, 78f ; his classification of 218; accepted by most gunas 79f ; on figs 82f, 84f ; post-dhvani writers 145;

310 SANSKRIT POETICS

used synonymously with on alamkāras as helping vyangyārtha (q. v.) 150 *; rasa 170; on rasavat 192f; etymology of 150 *; promi- followed by later writers nent in dhvani-kāvya (q.v.) 174-75, 179, 215, 217, 224, 157; Kuntaka on 191f; 228, 236. accepted by Vis 225f, by dhvani-kāvya 150, 157f, 255; Rk 229, by Vid 235f, by Jg classed into two types 160. 255f See dhvani and vyańgy- Dhvanikāra The, follows an ārtha earlier tradition 139-40; Dhvani-theory (or system or sahrdaya as his name or school), in relation to rasa- title 141 ; his knowledge and alamkāra-theories 67f; of some theories of rasa, harmonised with the rīti and alaņkāra 139, 145; theory of rasa 128, 137-38, the object of his work 145; 161f, 177 ; its origin 139f; influence of Bh on 127f; why it did not influence applies rasa-theory from early theorists of other drama to poetry 127-8; systems 139-40 ; early first formulation of dhvani- forms of, lost 141; its theory 139; on the relation inspiration from gramma- of indicated sense to rians 142f ; its relation to suggested sense 154 ; three sphota-theory 142-43 ; hos- kinds of suggestion not tile attacks on 151f; not taught in the kārikās of mystical or inexplicable 164 ; lays stress on rasa- 144 ; its analysis of expres- dhvani 127f, 164f, but does sion 145f ; suggested sense not speak of rasa as the set forth by it as the es- essence of poetry 166, 177f; sence of poetry 157 and as his comprehensive defini- not compassable by earlier tion of poetry 166-67; theories 145 ; its classi- accepts three gunas 168, fication of poetry 157f; 171; on the relation of its anxiety to do justice guņas to rasa 16, 169f; to facts 159, 167 ; how it justification of rīti 169f ; comprehends all tradition- on the distinction between al notions 160f, 167; on guņa and alaņkāra 169; metaphorical expression

SUBJECT-INDEX 311

(q. v.) 160 ; on suggestion dīrgha-vyāpāra-vādins i, 52, ii. of rasa, vastu and alam- 119, 152 kāra 161f, 164; stress on doșa, blemishes of composi- rasa-dhvani 164f ; whether tion, in Bh 7f; Bba's two an extension of rasa- lists compared with those theory 165-66 ; rasa accep- of Bh and D 8-11, 86-88 ; ted as an element of the when becomes guņa 11; unexpressed 129-30, 164f ; whether positive entities on guņa and alamkāra 67f, or negations of guņa 12, 168f, 171, 173 ; on rīti 167; 87f ; as subservient to rasa on doșa 173 ; its teachings 88*, 172, 226 ; of alaņkāra summarised 174f ; its chief 65; the doctrine of 86f; contribution 173-74, 214 ; four classes of in Vm 88; on the nature of sugges- Rdt's classification of 88 ; tion 174 ; its influence on in later schools 88, 172, later writers 175, 179 ; its 226 ; whether nitya or importance established by anitya 89*, 172; Mmt on Abh 176; oppositions to 220-21 179f; how explained by doşābhāva 12, 87, 89, 172. Bh N 181 ; criticism of drama. See rūpaka by Mbh 195f; ignored by Dramaturgy, as a separate Agp, Bhoja and Vgs 201f, discipline 2-3, 18 ; ignored 206, 244, 245 by Bhã 2, 32, and by D dhīrodātta, dhīroddhata, and Vm 2; by Mmt 221; dhīra-lalita, and dhīra- included by Hc, Vid, Vis praśānta (nāyaka) 269 3, 221, 237, 244 dhīrā, dhīrādhīrā (nāyikā) 270 dravya, individual, 146 dhrsta (nāyaka) 269 druti, associated with mā- dīpaka, a fig, in Bh 5, 6; in dhurya-guņa 172, 220*, Bhã 27; involves appre- 256, 262 hension of a suggested fig drstānta (logical) 10 162 ; in Bhoja 210 drstānta (or kāvya-drstānta), a dīpta, opposite of D's suku- fig 56, 65, 69 māratā 80, 87 durbuddhi (pupil) 293 dīpti, associated with ojas by duşkara, a fig, 204 Jg 259 ekārtha, a doșa, 8 fn 20, 9

312 SANSKRIT POETICS

embellishment, theory of. See 146 ; incapable of relishing: alamkāra rasa 135 ; the vyangyārtha empirical abalysis, 33-4, 91, not intelligible to 129; 159, 167 influence on Poetics and enjoyment involved in rasa, on the dhvani-theory 142 its nature 124, 126f, 132*, grammatical correctness (sau- 135-37, 183-84. See śabdya) 37, 38, 47*, 55 rasa grāmya, a doșa, 88* erotic rasa-writers 265f gramyatva, rejected by D gadya, prose. See padya 80; meaning of 110; gamana, a name for sugges- opposite of kāntiguņa 219* tion, 157 grāmyā, a vrtti (also called garbhita, a doșa, 88* komalā) 55, 64, 221* gatārtha, a doșa, 88* grāmyānuprāsa 55 gati, a guna, 239 grāmyāpabhramśa 77* gauda kāvya, Bhā's view on gūdha-śabdābhidhāna, a doșa 46, 75, 79* 11 gauda mārga in D 67, 75, 79f, gūdhārtha, a doșa, 8 ; differ- 202. See Vaidarbha entiated from ekārtha and gaudī (or gaudīyā) rīti, in Rdt paryāyokta 8* 60, 104, 203; in Vm 90, gūdhārtha-pratīti, a basis of 202 ; whether earlier than classifying figs 73 vaidarbhī 91* ; in Agp 104, 203 ; in Vis 227* guņa, excellence of diction or of poetry in B 11f, 15, 70 gāmbhīrya, a guņa, 204, 209, fn 9, 97f; whether mere 237, 239 ; an excellence of negation of dosa (q. v.) 12, the hero 271 87f ; differences in the geya (kāvya) 77* definition of individual 16, gomutrikā 85 46, 95f, 97; Bha's indiffer- Gopendra Tippa Bhūpāla 179 ence to 46, 79 ; D's view Govinda, on L's theory of on 36, 67, 79, 82; its rasa 118 *; on Śankuka's relation to rīti 90 ; how rīti. view 120 ; on Vis's defini- is characterised by 90-91 ; tion of poetry 224; ex- plains bhāva 278 essential to poetry in Vm's

Grammarians, on abhidhā view 90f, 93, 99 ; cogni- tion by sahrdaya a proof

SUBJECT-INDEX 313:

of its existence 93. its guņābhāva 12, 89*, 172. See classfication in relation to doşābhāva śabda and artha 93, 219- guņībhūta-vyangya (kāvya) 20 ; Vm's scheme of defined and its eight kinds. 94f ; protest against its of 158; includes figs which. multiplication 95-96, 107, involve a suggested sense 168, 171, 219 ; three gu- 162, 188 ; when it becomes. nas of Bha 46, 79*, of Anv true poetry 165 ; Mmt, and Mmt 96, 109, 168-69, Vis and Rk on 218, 225,. 171, 219, 226 ; how differ- 226, 229 entiated 171f, 219*, 220*, gumphanā, a fig, 204* 226, 262 ; relation to alam- haraņa. See plagiarism kāra 58, 67, 82f, 99-100, 'hāva ) 169f, 219; functions of helā J an e an excellence of the

alamkāra assigned to 97; heroine 272 relation to samghațanā 58- Hemacandra, includes treat- 59 and fn 32, 168 ; place ment of dramaturgy 2, 221,. of, in dhvani system 67, 244 ; reviews guņa-doctrine 168f, 219 ; related to rasa 97, 98; explains rasa in 106, 169f, 173, 219f ; Agp D's mādhurya 111; his. on 203-4 ; Bhoja on 206, work chiefly a compilation. 208-9 : Mmt on 219, criti- 244; its nature and cised by Jg 256f ; appro- content 243 ; on figs 243 ;. priate to particular rasas accepts nine rasas 275; 220 ; caused by combina- includes practical treat- tions of particular letters ment of Poetics 291 220-21 and fn; equalised hetu (logical) 10 with Vm's rītis and Ud's hetu, fig, 28 ; rejected by Bhā. vrttis 104, 168*, 278 ; con- 50,65 ; in D, Ud and Rdt fined to śabda by Mmt 63*, 65, 86 ; in Agp 204 220* ; Vis on 226 hetu, a lakșaņa, 249 guņa, quality, 146 hīna-pātra 279 guna-vrtti, subordination or hīnopamā 210 secondary application Hrdaya-darpana, by Bh N, 154 whether a commentary on guņa-viparyayas 87 Bharata 180-81

314 SANSKRIT POETICS

hrdyatva, synonymous with test of a fig 231; criticis- vicchitti (q. v.) 42, 232, ed by Jg 253 257 jāti, a fig, another name for Indication, or indicated or svabhāvokti (q. v.) transferred sense. See jāti, genus, 146 lakşaņā jāti-bhāșā 77* (fn 5) intention, of the speaker (vi- jāyā-sammita. See kāntā- vakşā) when an object of sammita logical inference 156 jñāpaka-hetu 121 Jagannātha on enjoyment jyesthā nāyikā 270 of rasa 132 ; defines vyakti kaņişthā nāyikā 270 136 ; criticises Vis's and kalahāntaritā (nāyikā), also Mmt's definitions of poet- abhisamdhitā, 271 ry 166, 178, 222, 223; kalā-virodhi, a doșa 10 on the influence of the kalpanā-dușța, a doșa 11 Dhk 179; revives D's kalpitopamā, the simile with definition of poetry 217 ; an imagined object 6 follows Knt's theory of karman (philosophical) 136 alamkāra and defines karuņa (rasa) in D 117 ; pain vicchitti 232, 257f; the following upon 122 ; how scope and nature of his enjoyable as rasa 132 Rasa-gangādhara 252f; karuņā (philosophical) 262 his Citra-mīmāmsā-khaņ- katha, varieties of 76 . See dana 253 ; his criticism ākhyāyikā of earlier authors 253; kathānikā 77 his reactionary tendencies kavi, classification of, by Ks 253 ; defines poetry 253- 287 ; by Rājaśekhara 293, 54; on suggestive poetry 294 ; his education etc, see 255f ; on guna 256 ; on the kavi-śikşā śānta rasa 277; on rasā- kavi-goșthī 292, 294 bhāsa 256; on bhakti as Kavi-kalpalatā 289, 291. See rasa 268* Deveśvara Jayadeva, his Candrāloka 248f ; on ten lakşaņas 249 Kavi-kanțhābharaņa 287f. See Kşemendra Jayaratha, on Rk's indebted- kavi-praudhokti 184, 229 ness to Kat 230-31 ; on the kavi-pratibhā (or °karman,

SUBJECT-INDEX 315

°vyāpāra, °kauśala) as the 96, 172 ; in Bhoja 209, 238 ; source of vakrokti (q.v.) in Vin 238 48, 184, 186, 189, 230, 257 ; kānti, a fig, 205 analysed 187 ; as the source kāraka-hetu 121 of poetry, see pratibhā kārayitrī (pratibhā) 293 kavi-rahasya, Rājaśekhara's Kāśmīrakas 208 treatment of 292 kāvya, in relation to nāțaka kavi-rāja 294 2; definitions of 37, kavi-samaya, poetic conven- 40, 78, 90, 157, 166, 185, tion 248, 289, 290, 292, 297 205, 208, 217, 218, 222, kavi-śiksā, theme of, in older 224, 237, 244, 253f, 295 ; authors 35*, 41, 216, 247, distinguished from scien- 283, 298 ; origin and scope ces and scriptures (śāstra) of 35*, 283; Ks on 40, 49* (fn 25), 102, 287f ; Rājaśekhara on 109, 185; function of 292f ; in Arisimba 289f; vakrokti (q. v.) in 50; in Deveśvara 289; in its relation to the poet Keśava 291; in Hc and 50* ; classification of 44, Vg II 291* 157f ; persons entitled to kāku, intonation, as the basis study 78, 287, 293f ; its of vakrokti in Rdt 63; not 'body' and 'soul' (see admitted by Hc and Rāja- kāvya-śarīra and °ātman) ; śekhara 64 its sources (see °hetu, kāla-virodhi, a doșa 10 °yoni); function of rīti kāma (philosophical) 136 (q.v.) in 90f, 102, 116; Kämadhenu, the comm. of place of guņa (q.v.), Gopendra Tippa Bhūpāla alamkāra (q.v.) and rasa (q. v.) 100 (q.v.) in ; earlier theories kāmanīyaka 230 of, insufficient 117 kāntā-sammita or jāyā-sam- kāvya-doșa. See doșa mita 40, 109 kāvya-guņa, See guņa kānti, a guņa, in Bh 15, 95, kāvya-hetu 39, 287, 293 113; in Vm 15*, 95, 97, kāvya-hetu, a fig. See kāvya- 98, 112, 219 *; in D 81, 82, lińga 97; as a mere doşābhāva Kāvya-kalpalatā 289. See (opposite of grāmyatva) Arisimha

316 SANSKRIT POETICS

kāvya-kavi, distinguished from on 35*, 36, 38, 76, 78, 90, śāstra-kavi 294 205; Vm on 35*, 90; kavya-lakşaņa. See lakșaņa Rdt's view on 36*, 115; kāvya-liňga (also called kāvya- Agp on 205 ; attention of hetu), a fig, 56, 65 earlier writers confined to kāvyālaņkāra, why so called 36*, 116f, 173 189 . See alamkāra kāvya-śobhā 83, 99, 102, 203 Kāvyālamkāra of Rudrața kāvyātman 34f (see kāvya- (q.v.) 60 śarīra) ; not explained by Kāvyālamkāra of Bhāmaha alaņkāra-system 67, 90 ; (q. v.). See Bhāmahālam- Vm's enquiry on 35, 90,. kāra 116 ; attitude of earlier Kāvyālaņkāra ot Vāmana 89* writers towards 34f, 116; Kāvyālaņkāra-vivțti. See the dhvani-theorists on 157 Bhāmaha-vivaraņa kāvya-yoni (kāvyānga or kāvya-mātaraḥ. See kāvya- kāvya-matarah) the arts. yoni and sciences comprised in Kāvya-mīmāmsā 291f. See 42f, 78, 288, 296 Rājaśekhara Keśava Miśra, professes to be kāvyānga, sources or auxilia- a follower of Śauddho- ries of poetry. See kāvya- dani, but really belongs to hetu and kāvya-yoni post-dhvani group 17*, kāvyānumiti 195. See anu- 247, 291 ; scope and con- māna-theory tent of his work 257f; Kāvya-prakāśa 40, 149, 218f. influence of earlier opi- See Mammața nions on 248, 291 ; accepts kāvya-prayojana, object of rasa as the essence of poetry 39f poetry 247, 259 kāvya-purușa, 35*, 292, 293 khaņda-rīti 104, 210 kāvya-rasa, in relation to khaņditā (nāyikā) 271 nātya-rasa 18*, 19; its kiņcit-sadrśī upamā, the meaning in Bhã 109; in simile based on partial D and in old classics 109 . similarity 6 See rasa kīrti, as an object of poetry 39 kāvya-sarīra body of poetry klişta, a doșa 10 34f, 35 ; Bhã ơn 36f; D komalā, a vītti. See grāmyā

SUBJFCT-INDEX 317

komalatva, a guņa 204 on the test of an alamkāra kośa (kāvya) 76* 188f, 230, 257; how he krama, a name for yathāsam- supplies a deficiency in khya, 29 fn 63 dhvani-theory 189, 228, krama-bheda, a doșa 199*, 282 230; on dhvani 191, ikramoddyota vyańgya 226 229 ; on upacāra 191 ; on kriyā, action, 146 rasa 192f ; criticised by .krodha, as a basis of raudra Mbh 199; his theory rasa 23, 112 ignored in later times 200 ; Krsņa-rati as sthāyi-bhāva 262 followed by Rk and others Ksemendra, his two works 189, 230-31 283f ; his theory of au- Kuvalayānanda 73 citya 284-85 ; critical acu- lakşaka 150* men displayed by 286 ; on lakşana, Bharata on 4-5; the means of poetry and included by D and Dh in classification of poets alamkāras and rasa 4, 287f ; on plagiarism 287; 249-50; disappears in on the sources of poetry Poetics but persists in 288 Dramaturgy 5, 16, 25, 249 ; ksobha, a mental state involv- Jayadeva on 249 ; Vis on ed in bībhatsa rasa 262 249, 251 ; not distinguish- Kubera 38 able from nātyālamkāras kulaka 76* 249, 251 Kuntaka, the vakrokti-jīvita- lakşaņa (=vyāvartaka dhar- kāra develops Bhā's idea ma) 153 of vakrokti 48, 54, 187 ; his lakşaņā-vrtti or upacāra, indi- system an off-shoot of the cation or transference of older alamkāra-system sense, as the basis of D's 187 ; on svabhāvokti 49*, samādhi (q.v.) and Vm's 187 ; on atiśaya involved in vakrokti (q v.) 97, 129; vakrokti 50f, 187f, on figs basis of metaphor or meta- which help rasa but result phorical exp. 147-48, 191 ; in mere vaicitrya 171, 188f an extension of abhidhā 290; on Rīti and Guņa 147, but included in it by 189-90 ; could not supplant Bh 124, 147, 183; as an dhvani-theory 184-5, 200 ; artha-vyāpāra 146 ; classi-

318 SANSKRIT POETICS

fication of 147-48; its logos 143 relation to vyañjana 150; loka-nyāya 73 whether it can convey the loka-virodhi, a doșa, 10 suggested sense or rasa lokātikrānta-gocaratā 49* ; 136, 153f; may be the involved in atiśayokti and basis of suggestion, e.g. vakrokti 187. See loko- avivakşitavācya 155, 160; ttara Knt on 191; Agp on 201 lokottara, disinterested or su- lakşaņā-mūla-dhvani, see lak- pernormal 42, 49* (ex- şaņā and avivakșita-vācya plained), 187, 257. See lakşyärtha, indicated or trans- alaukika ferred sense 146f, 150 ; Lollața, Bhațța on Bh's sūtra distinguished from vācyā- on rasa 19, 21, 112, 117f ; rtha and vyangyārtha 146, his lost work 117 ; review 153-55 of his opinion in later lalita, an excellence of the writings 118, 218; his hero 271 theory on rasa (utpatti- lalitā, a vrtti, 64 vāda) 118-19, 121. rejected laya 137 by Śańkuka 119f; said laukika 119, 130, 132, 133. to have been a Mīmām- See alaukika saka 119, 125; whether a lālitya, a guņa, 204 dīrghavyāpāra-vādin 152* lāțī or lāțīyā (rīti) mentioned madhura (rasa). See mādhu- by Rdt and Vg I 60, 104, rya 203 ; an intermediate style madhurā, a vrtii, 64 in Vis 227* madhurā rati, sthāyibhāva for lāțānuprāsa or lāțīyānuprāsa, ujjvala rasa (q. v.) 267 kind of alliteration 55, 80 madhyā nāyikā 270 fn 11 Mahimabhatta, the Vyaktivi- leśa, a fig, 28*, 50 veka-kāra, the scope and linga, lingin, syllogistic major object of his work 195f; and minor terms 156, 196, compared with that of Bh 197, 198, 229 N 181f ; his criticism of °Locana. See Abhinavagupta dhvani theory 195-96; logical correctness 10, 47, 87 acknowledges rasa 138; logicians 145, 146. See Nyāya comprehends dhvani in

SUBJECT-INDEX 319

anumāna 195-97, 229 ; 230 ; on rasavat 193 ; ex- his theory of kāvyānumiti cludes dramaturgy 221 ;. 195f ; his theory criticised scope and value of his 197f ; could not supplant work 221-22 ; reactionary dhvani-theory 175, 180; tendency in 223 ; defines. opposes dīrgha-vyāpāra- bhäva 278 vādins 199; on Knt's Mangala, belongs to the Rīti. theory 199 ; on anaucitya school 98 ; on pāka 240 199*, 282, 284; value of maņikulyā, a kind of kathā. his work 200; why his 76*

theory ignored in later Manoratha 140

times 200 mata, a fig, 63*

maitrī (philosophical) 262; matallikā, a kind of kathā.

or friendship of words 240 76*

Mammata, dismisses yamaka mādhurya, a guņa, in Bh 14;

7 fn 16; on kāvya-śarīra in Bhã 46, 79* ; in D 80, 36* ; on the aim of poetry 81, 82, 95, 110 ; in Vm 95, 40 ; on śakti or pratibhā 97 ; in Dhk 168* ; defined 41-42 ; on guņa 96, 169, by dhvani- and post-dhvani. 219f, 221*, 256 ; on rīti and writers 171, 219*, 220 ; in vrtti 104, 220, 226 ; on the Agp 204; as appropriate Abhihitānvaya-vādins 149; to certain rasa 220 ; pro-

on śabda-vrttis 149; duced by certain letters. his definition of poetry 220 ; in Bhoja and Vin 170, 218f, criticised 238 ; in Jg 256 222, explained histori- mādhurya or madhura, also. cally 223-4; establishes called ujjvala, a rasa in Anv's system and makes Vaișnava theology 267 older speculations con- mādhurya, an excellence of verge in his work 179, 213, the hero 271 218 ; the principal post- māgadhī (rīti) 104, 210 dhvani writer 216, 217f ; mānorathika (rasa) 263* contents of his work 218f ; mārga a name for rīti (q. v.) on rasa 19, 218, 262; on 46,67, 75, 78 śānta as rasa 276; on māyā, rasa 276* alamkāra 169, 219, 221, Medhvin 29, 38

320 SANSKRIT POETICS

metaphorical expression, not narma-saciva 270 equivalent to suggestive Navyas i. 177, ii. 217, 260* poetry 129 ; comprehended nața-sūtra 2 by D's samādhi 81, 82, Nāgeśa on śanta rasa 276 160; by Vm's vakrokti 13*, nāma-dvirukti (anuprāsa) 80 48, 101, 160; recognised fn 11 by earlier systems 129, Nārāyaņa 262 160; based on lakşaņā nāțaka 77 (q. v.) 148, 191 ; compre- nātya-rasa, Bh's doctrine of hended by avivakșitavācya 19f. See kāvya-rasa and dhvani 160 rasa mitra-sammita 40* Nāțya-śāstra. See Bharata, miśrālaņkāra. See ubhayā- also vol. i lamkāra nātyālaņkāra, mentioned by mithyādhyavasāya, a laksaņa Vis 249 ; not distinguish- 249 able from dramatic lakşa- Mimāmsā school or Mīmam- ņas 249-51 sakas 9f; influence of, on nāyaka, the hero, classifica- Lollața 119, 125 ; on abhi- tion of 269; excellences dhã 123, 146, 152 ; imper- of 271 ; comes topically in vious to rasa 135 the discussion of alam- mokşa, as the aim of poetry bana-vibhāva 273 40; Vedāntin's idea of Nāyaka, Bhatta, on Bh's rasa- 136 sūtra 19, 21, 117, 181 ; his muditā (philosophical) 262 lost Hrdaya-darpaņa 180- mudrā, a śabdālaņkāra, 204* 82 ; review of his opinion mudrā 240 in later writings 118*, 123* mugdhā (nāyikā) 270 218; on earlier theories Mukula, Bhatta 37*, 103*, of rasa 122-23, 182 ; his 141, 147, 179 doctrine of enjoyment muktaka (kāvya) 76 (bhoga) of rasa (bhukti- muraja-bandha 7* vāda) 123f, 182f, 289 ; naming of rasa (sva-śabda- postulates three functions vācyatā) 130, 153, 161 of words and threefold Nami-sādhu 36*, 115* potency of poetry 123f, tNandikeśvara 17 183f ; influence of Sāmkhya

SUBJECT-INDEX 321

on 125, 183 ; his influence nyāya (logical reason), a basis 127 ; criticised by Abh of classifying figs 73 ; split 130f ; criticism of dhvani- up into three varieties 73 theory his main object 181- nyāya-virodhi, a doșa 10 .82 ; but his failure to sup- nyāyādapeta (defective logic), plant it 175; relation to a doșa 9 Mbh's work 181 ; accepts ojas, a guna, in Bh 14, 97 ; in rasa-dhvani but not vastu- Bhã 46, 79 ; in D 81, 82, dhvani 183; regards rasa as 97,98; in Vm 94, 95, 96; sva-samvedya 184 ; on the in Dhk 168, 171; defined mental activity involved in by post-dhvani writers the bhoga of rasa 220, 261 172, 219*, 220 ; compre- māyikā, the heroine, classifica- hends ślesa, samādhi and tion of 270-71 ; her quali- udāratā of older writers in ties 272 ; comes topically Mmt 172, 219 ; appro- under ālambana-vibhāva priate to certain rasas 273, 274 220 ; produced by certain 'New School the, after Mam- letters 220 ; in Bhoja and mața 213. See Post-dhvani Vin 238 ; in Jg 256 writers osara (kāvya) 77 meyatva, opposite of D's pada-doșa 88, 220 arthavyakti 80, 87 pada-parārdha-(or pratya-) va- neyārtha, a doșa 10 kratā 186 .nibaddha (kāvya) 76 pada-pūrvārdha-vakratā 186 nidarśana, a kind of kathā 76 padābhyāsa 27 nindopamā, 6 padānuprāsa 80* nirāgama, a doșa 88* padārtha-doșa 88 nirukta, a lakşaņa 249 padya, verse, distinguished nirveda, the sthayi-bhava of from gadya (prose) 45, śānta rasa 273, 275-78 76* nişpatti. See rasa-nispatti pakşa-sattva (logical) 198 nyāya, logic, Naiyāyika i. 9f, parakīya nāyikā 270; two 125; influence on Sań- kinds of 270 ; the highest kuka 125 type of heroine in Vaisnava nyaya, logical correctness of lyrics 270* poetry 10, 47 para-mukha (rasa) 263* 44

322 SANSKRIT POETICS

parikathā, a kind of kathā 76* or āhlāda, q. v.) in poetry parivrtti-sahatva, implied in 40; involved in rasa (q.v.); the theory of anvaya-vya- produced by vakrokti 187; tireka (q. v.) 72 Jg on 253-54. See camat- parușā, a vrtti, in Ud 55, 64, kāra and lokottara 221 ; in Rdt 64 poet the, his qualifications, paryāyokta, a fig, in Bhā 8 his education etc. See kavi- fn. 20, 53 ; in Rdt 60 ; ex- śikşā cluded by Vm 101 poetic figure. See alamkāra pāka, a guņa in Agp 204 Poetics, Sanskrit, unknown pāka, theory of, in Vm 240 beginnings of 1, 213 (also (śabda- and artha-) ; equi- i. 1-17) ; earliest outlines of valent to sauśabdya in 1 ; its development from Mangala 240, 241; Vid Bh to Bhã 25-31 ; original and Vin accept artha-pāka standpoint of 33 ; first de- only 241 ; corresponds to finite scheme of 32; Bhoja's praudhi-guņa 241 ; schools of 32-33 ; as an earlier opinions on 241; objective and empirical etymological sense of 242 ; discipline 33f ; its classification of 240-42 posteriori character 91, pāñcālī (rīti) in Vm 90, 202; 159, 167, 272 ; its learned in Rdt 60 ; in later writers and dialectic tendency 104, 202f, 210, 227* 43-44 ; Vāmana's care- Pārāśara 38 fully outlined scheme pāțhya (kāvya) 77, fn 7 of 89 ; the position of the phala, a lakșaņa 249 Dhvanyāloka in 175 ; in- Pictorial poetry 158. See fluence of grammar on citra-kāvya 142; progress of the pihita, a fig, 63* principal systems of 213f; pīthamarda, the comrade of disappearance of its nor- the hero 270 mative character 214 ; the pīțha-mardikā 270* art of suggestion in 214; plagiarism (haraņa) 287, 296- not coextensive with rhe- 97 toric 189,* 285 *; standar- pleasure, disinterested or dis- disation of a complete sociated (alaukika ānanda scheme of 175, 215; a

SUBJECT-INDEX 323

scnolastic stage of 215; prakrama-bheda, a doșa, decline of the systems of 199*, 282 215f, 252; psychological Prakāśavarșa 211* analysis displayed in 272; pramāņas of Jaimini, ad- what it involves 285 ; 'mitted as figs 210 Rājaśekhara's idea of its prasāda, a guņa, in Bh 12, divine origin 292 ; rasa as 97; in Bhã 46, 79* ; in D its aesthetic foundation 80,89,82; in Vm 94,96; 165. See rasa accepted by Dhk 168* and Poetry, Classical Sanskrit 34* defined by his followers Poetry, purpose of. See 171*, 219 ; in Mmt 219 ; kāvya-prayojana in Agp 2C4; appropriate Poetry, theories of. See under to certain rasas 220 ; in kāvya Jg 256 Post-dhvani writers, their atti- praśamsopamā 6 tude towards dhvani- praśastatā (?) a guņa 204 theory 178 80; Mmt and praśasti, a fig 205 Rk as the principal writers pratibha. as the source of among 213, 217, 218, 235 ; poetry 41-42; may be classification and charac- sahajā or utpādyā 42; as teristics of 215-16 ; reac- the source of alamkāra or tionary tendencies in 217 ; vakrokti, see kavi-pratibhā their scholastic bent and pratijñā (logical) 10 want of originality pratināyaka, rival of the hero 215, 217; commentaries 270 composed by 215; pratipādya, communicable their extraordinary activity meaning, may be vācya or 247 vyańgya 156 prabandha-vakratā 186, 194 pratiședha, a lakșaņa 249 Prabodha-candrodaya 278 prativastūpamā, a fig 86 prabhu-sammita 40, 109 Pratīhārendurāja, his inter- Pracetāyana 38 pretation of Ud 58 ; influ- pragalbhā nāyikā 270 ence of Vm on 58; on prahelikā a fig, 84, 204 rasa 113-14 ; his criticism prajñā 42 of dhvani-theory 179 prakaraņa-vakratā 186 pratīpa, a fig 70, 101

324 SANSKRIT POETICS

pratīti (cognition) of rasa 131f not of poetry 45; three 136, 261 kinds 76* pratīyamāna artha. See vyan- proșitapatikā nāyikā 271 gyārtha Pulastya, 38 pratyaksa (perception), not a punarukta, a doșa 10; cases means to cognise rasa 136 where it is not a doşa but pratyaya-vakratā. See pada- a guņa 64, 172 ; in Mbh parārdha° 199*, 282 pratyāyana a name for sugges- punaruktavadābhāsa, a fig 64 ; tion 157 in Mmt 211 pratyayopamā, a fig, 65 pūrva, a fig 63* praudhā, a vrtti 64 ramaņīyatva or rāmaņīyaka, praudhī, a guņa 204, 241 used synonymously for praudhokti. See kavi-praudo- vicchiti (q. v.) 42 ; in poe- kti try, defined 253-54 ; as an pravahlikā (kathā) 76* element of alamkāra or prayojana, special motive in vakrokti (q. v.) 257 lakşaņā (q. v.), never ex- rasa, Bh on 17f ; pre-Bharata pressed but understood doctrine of 18; influence 148, 155 of Bh's teachings on 18f, preksya (kāvya) 77 21f, 108, 117f, 165; drama- preyas, fig, Rdt on 60 ; turgic rasa-systems in rela- excluded by Vm 101; tion to poetic rasa-theories idea of rasa included in 16-17, 18 and fn, 47, 108, 111 117, 127-28, 165, 176, 266 ; preyas, a guna in Bhoja and in Rājaśekhara's kāvya- Vin 238 purușa 35 ; place given to preyas, a rasa, in Rdt 114, it in Alamkāra and Rīti 275, 277* ; in Bhoja 209, 276 ; in Vaisnava theology systems 52, 108f, 116-117 ; Bhã on 32, 108-9 ; D on (called also sakhya) 267 110f ; Vm on 95, 102, 112- prīti, as an object of poetry 13; Ud on 56, 113f; Rdt 39, 40* on 114-16 ; controversy on prīti, also called dāsya, a rasa Bh's sūtra on 19, 21, 108, in Vaisnava theology 267 117f; Lollata's explanation prose, opposite of verse but of 117f; Śankuka on 119f;

SUBJECT-INDEX 325

Bh N on 122f, 182.183; (q. v.) 137 ; theory of, as philosophical colouring finally determined 137, given to theories on 124f, 259 ; guņa (q. v.) and 136 37, 183-84; Dhk on alaņkāra (q. v.) in relation 127f; defined as an inward to 169f, 219* ; rīti (q. v.) experience of the spectator as accessory to 105f ; Knt 120, 125, 127, 130, 133-34, on 189f ; only anumeya 135, 138, 260, 261, 264 ; as in Mbh's opinion 196; an aesthetic foundation of prominence given to, by poetry, as of the drama Agp, Bhoja and erotic 127f, 165 ; worked up into rasa-writers (q. v.) 202, the dhvani-theory as an 206, 247, 265f; post- element of the unexpressed dhvani theorists on 259f ; 129, 138, 161-62, 163f, 177 ; classification or divi- collateral suggestion of sion of rasa how possible 162-63, 224-25; sometimes 273* ; the original number an alamkāra 192-93; Abh of, in Bh 23, 202, 262, 275 ; on the pratīti or abhivya- the addition of sānta 114, kti of 131-32, 177f ; vibhā- 194, 275-78; whether vas (q. v.) how realised rasas are eight or nine 218, 133f, 261; vibhāvas 275 ; ten rasas admitted not identical nor are by some 114 and fn, 275; laukika causes of 132-33, māyā as 276* ; also preyas 161-62, 261 ; significance (q. v.), śraddhā, bhakti of 'taste' involved in 135, 277* and vatsala as 276 ; 137 ; nature of its enjoy- bhāva (q. v.) as the basis ment 124, 126f, 132*, 133, of 278 f; as laukika and 134f, 184, 254-55, 259-60 ; alaukika 263 ; conflict of 263-64 ; mental activity in- 280f (see rasa-virodha) volved in the appreciation rasa-bhanga, impropriety in of 220, 261 ; the term the depiction of rasa, a untranslatable 135; not case of aucitya (q. v.) 282, compassable by ordinary 284 processes of knowledge rasa-bhāvanā. See bhāvakatva 136f, but vouched by the rasa-dhvani (see suggestion of experience of the sahrdaya rasa, under rasa), in rela-

326 SANSKRIT POETICS

tion to riti 103; not re- poetics 137-8 ; erotic rasa- cognised by Dandin 111 ; writers 247, 265f Rdt on 115; Dhk on 117, rasa-sparśa 224 127f, 160f, 164f, 173 ; not rasavat, a fig, its function in compassed by abhidhā Alamkāra system 52, 162, 153; Anv on 164f; Abh 192 ; Knt on 194 ; ignored on 166, 177f, 224, 259 ; by Vm and Rdt 61, 97, accepted by Bh N 181, 101, 115f ; idea of rasa 183; Knt on 192; as an included in, by Bhā, D important element of and Ud 108, 109, 111f, poetry 214; Abh's posi- 113; involves apprehen- tion regarding, accepted sion of a suggested rasa with modification by 162; comprehended in post-dhvani writers 259; Vis's extreme theory of guņībhūta-vyangya 162,

224f 192 ; its sphere distinguish- ed from asamlakşya-krama rasa-doșa, includes cases of dhvani 192f ; later differ- rasa-virodha or anaucitya ent views on 193* in the depiction of rasa Rasārņavālaņkāra of Prakāśa- 88*, 281, 282, 284 varşa 211* rasana, taste or appreciation Rasārņava-sudhākara 266 of rasa 21*, 131, 137 and rasa-virodha (opposition of fn, 260 (see rasa) rasa), how it occurs and rasa-nispatti, in Bh's sūtra on how avoided 280f ; classed rasa 20-21, 23 ; explained as rasa-doșa (q. v.) as °utpatti, 'anumiti, Rasa-gangādhara 252. See °bhukti and °abhivyakti Jagannātha respectively by Lollata, rasābhāsa 225, 256 ; defined Śankuka (and Mahima- 279 bhatta), Bh N and Abh rasādi, the term as used by 118, 120, 123f, 130 Anv 165 Rasa-school (see rasa), in rasāvaha meaning of, in D 110* connexion with drama 17- rasika 43, 135, 260. See 19, 108; influence on saħṛdaya other schools 108f ; merges rati, love, as the basis of into the Dhvani school of śrňgāra 112

SUBJECT-INDEX 327

rāga, a guņa in Agp 204 the essence of poetry in Rājaśekhara, cites Nandikeś- Vm 35*, 90f, 116; three vara as an authority on rītis in Vm 90-91 ; rasa 17 ; on kāvya-purușa whether vaidarbhi was a 35*, 292 ; the tradition foll- reaction to the older owed by 292, 295 ; scope gaudī 91 ; not equiva- and content of his work lent to the English word 292f ; on the origin of Poetics 292; his division 'style' 92; the doctrine of 102; discredited in of literature 293; on the later times and modified classification of poets 293- 103f, 202, 226; how far 94 ; on literary examina- accepted and justified by tion 43, 294; on pāka dhvani-theorists 103f, 167f; 241-42, 294; on poetic in non-dhvani systems 104, theory 294f ; on classifi- 202f, 210, 245; oppo- cation of speech 295; sition to its definition and on sources and topics of unprofitable classification poetry 296 ; on rīti 105, 106-7, 167f in relation to 295 ; on plagiarism 296; rasa 105-6; distinguished value of his work 297-98 from vrtti 104 ; Vis on reactionary or anti-dhvani 226; used synonymously writers 179f with the term mārga (q. v.) Rhetoric, not coextensive with Rīti-system (see rīti), why so Alamkāra-śāstra 74, 189 . called 75. 214 ; in relation See Poetics to Alamkāra-system 67, 68, Thyme, in Sanskrit poetry 7 102f ; recognises the im- Rīti, ignored by Bh 16, 79, Bhã 45-46, 60, 79* and Ud portance of alamkāra 68 ; as in D and Vm, old in 58; Rdt on 60, 104, 115, tradition 74, 93-94, 97; un- 203 ; D on 66-67, 78f ; its known exponents of 74; technical sense not very of Vm, an advance on D's old 75 ; D's two types 79, 112f ; its recognition of 202 ; how named 79*, 91 ; rasa in guņa 102 ; its in- in relation to guņa 79f, fluence 103f ; its contri- 90f, 116; in relation to butions to general theory alamkāra 82f, 99; as 105f; its fundamental

328 SANSKRIT POETICS

doctrine of riti as the 257, but not a follower of essence of poetry, not Knt or of the Alamkāra accepted in later times school, 232, 234; scope 106 ; its formal conception and nature of his work of poetry 107 228-30 ; adherence to Rudrabhatta, applies rasa to dhvani theory 229, 234; poetry 128*, 265-66; scope his survey of older views and content of his work on dhvani 228-29; on Knt's. 265-66 ; rasa the keynote idea of dhvani 192*, 229 ; of his system 116; accepts on vrtti 105* ; his relation nine rasas 275, but takes to Ud 232; on ślesa the erotic as principal 232-34 ; his reactionary 268 ; classification of tendency 234 ; criticised śṛngāra and nāyaka 269 by Jg 253 Rudrața, on yamaka 7*, 64; rūdhi, usage, in lakșaņā 146. on the purpose of poetry rūdhi-vaicitrya-vakratā, com- 39 ; on śakti or pratibhā prehends dhvani 191 42 and fn ; belongs to the Rūpa Gosvāmin, scope and Alamkāra school 59f, 66, object of his work 267-68 ; 116, but follows a tradi- on ujjvala or madhura rasa tion different from that of 267; bhakti regarded as Bhã and Ud 61, 65; his rasa by 267 attitude towards rasa and rīti and dhvani 59, 60, rūpaka, a figure (metaphor), in Bh 5-6; in Bhã 27, 53 ; 114-115, 229 ; scope and in Rdt 61 ; in Agp 204; content of his work 60f, in Bhoja 210 66, 114-15 ; on alaņkāras rūpaka (drama) 17. See dra- of śabda and artha 61-62; maturgy on general topics of poetics sabhańga ślesa 64 78 ; accepts śānta and sadrśī upama, the simile preyas as rasas 114, 275, based on similarity 6 277 ; indebtedness of Mmt sahajā. See pratibhā to 63 sahokti, a fig 204 Ruyyaka, accepts and applies sahrdaya, the man of taste, Knt's theory of alamkāra the critic or connoisseur 171, 189, 217, 228, 230f, 43, 135, 215 *; his taste

SUBJECT-INDEX 329

vouching for rīti 93, for or guņa 59, 103*, 107* vakrokti 187, for pāka 168*, 226 240, for dhvani 174, for samghāta (kāvya) 76* rasa 135, 137; likened samhita, proximity of letters to a yogin 137; as the 239* name or title of the Dhk samlaksya-krama vyangya 161 141* ; śānta-rasa and samkara, distinguished from. bhakti not the province of samsrsti, 56 262, 267 samketa, convention implied: sakalakathā (kathā) 77* in abhidhā 146f, 152-53 sakhya, also called preyas, a samkīrņa, a doșa 88* rasa in Vaișnava theology saņkhyāna, a fig 29 267 samkşepa, a guņa 239 samatā, a guņa, in Bh 13 ; in samkșepa, a fig 205 D 80, 82; in Vm 13*, sammitatva, a guņa 239 94; rejected by Mmt samnidhi, proximity of words. 172, 219* ; in Bhoja and 149 Vin 238 sampradāya, usage, as a samavāya, inherence 100 means of establishing vic -- samādhi, a guņa in Bh 16-17; chitti 258 in D 13*, 81, 82, 96, 97; samsrsti, in earlier writers. in Vm 13*, 94, 96 ; com- 56; distinguished from prehended by laksaņā 160, samkara 56 ; as a basis of 205 ; included in ojas classifying fig 73 172 ; in Bhoja and Vin, an saņśayopamā, a fig 86 echo of D's 239 samvidhāna, a guņa 204 samāhita, a fig, 86 samvit (philosophical) 126, samāpatti 137 137* samāsokti, a fig, 28, 52, 53, samyoga, conjunction 100 101,162 the term as used in Bh's samāsopamā, a fig 65 sūtra 21, 23 sapaksa-sattva (logical) 198 sañcāri-bhāva, (also called saņbhoga-śrngāra, love-in- vyabhicāri°) the accessory union, 115, 269 feelings, 204 ; mentioned. samdhi (dramatic) 4 by Ud 114 ; classified 23- samghațanā, in relation to rīti 24 ; defined 274

330 SANSKRIT POETICS

sañcāri-rasa 281 sādharmya, used synonymous- Sarasvatī-kaņthābharaņa 201f, ly with sādrśya or aupamya 206f, 211*, 276. See Bhoja 73, 234 sarga-bandha (mahākāvya) 44; sādhāraņya or sādhāraņī- 76, 77,111 karaņa, generalisation, sarvatobhadra 85 134 ; corresponds to bhāva- sasamdeha, a fig 86 katva (q. v.) 264 sasamśıya, a doșa 10 sāhitya, as a description of satkavitva 41 literature 37*, 293 sattva or sāttvika guņa, 24* Sāhitya-darpaņa 224f. See 125,126 Viśvanātha satyā, a guņa 204 Sāhitya-sāra 247 saubhāgya, a guņa 204 sāmayikatā, a guņa 204 saugatas, on abhidhā 146 sāmajika, the audience 260f saundarya, used synonymous- sāmānya (guņa) 203 ly with kāvya-śobhā 102, sāmānyā nāyikā (or veśyā with alamkara in the q. v.) 270 wide sense of embellish- sāmānya-śabda-bheda, an ment or beauty 99, with upamā-doșa 65 vicchitti (q. v.) 106, 230, Sāmkhya-doctrine, influence 232, 257 of i.9f; 126,183 saukumārya, a guņa, in Bha- sāmya, a fig 63* rata 14 ; in D 14*, 80, 82 ; sāttvika bhāva, eight kinds in Vm 14*, 95 ; as a mere of 24 ; meaning of 24*, as doşābhāva (opposite of a special class of anubhā- śruti-kasta) 96, 172, 219* ; vas 260, 274* in Agp, Bhoja and Vin sāttvika guņa. See sattva 204, 238 Schools of poetics i. 328-29 ·sauksmya, a guņa 239 siddhi, a lakşaņa 249 Sauśabdya 37 ; equivalent to skandhaka (kāvya) 77 śabda-pāka 240 ; a guņa in smaraņa or smrti, a kind of Bhoja and Vin 238. See knowledge 125, 136 grammatical correctness speech-figure, distinguished sādṛśya, a fig 204 from poetic figure 74, 231 sādrśya, similitude. See aup- sphota-theory, in relation to amya dhvani-theory 142-44, 201

SUBJECT-INDEX 331

sthairya, an excellence of the Knt 28-9, 49, 49*, 187; hero 271 distinguished from vakro- sthayi-bhava, in relation to kti 49, 49*, 85 rasa 21-22, 260, 261, 272 ; svarāroha and svarāvaroha eight kinds of 23f, 273, 17*, 239 274f ; addition of śānta or svarūpa or svabhava, a fig, nirveda to the number 273, 204 275, 276f ; recognised by Ud 114 ; Lollața on its rela- sva-śabda-vācyatā (of rasa), a

tion to rasa 118; its cog- doşa 8*

nition as rasa by inference svāpnika (rasa) 263*

119f; its generalisation and svīyā or svakīyā nāyikā 270 śabda, along with artha, enjoyment (bhoga) 124; forming the 'body' of poe- exists as vāsanā (q. v.) 133 ; how developed and try 37f, 44, 76, 78, 116*

generalised 133f, meaning 173 ; alone, connoted by the term 'poetry' 222, 253; of the term 133* study and experience neces- guņas in relation to (see śabda-guņas) 93, 106, 169, sary for poetry 41 Subandhu i. 14-15* ; ii. 29* 173, 220*, 223 ; alaņkāra in relation to, 100. 106, subject of poem, in relation 115, 169, 173 (see śabdā- to the poet's mind 50* laņkāra) ; not a vyañjaka suggestion or suggested sense but an anumāpaka 197 in poetry. See vyańgyārtha śabda-brahma, theory of 142* or dhvani śabda-camatkrti 255 :sūkșma, a fig, rejected by Bhā śabda-citra. See citra-kāvya but illustrated by Bhatti śabda-guņa 15, 16, 82 ; dis- 28*, 50; in D 86 ; excluded tinguished from artha-guņa by Vm 101 ; mentioned by by Vm 93; this distinc- Rdt 62 tion explained by dhvani- svabhāva. See svarūpa and post-dhvani theorists svabhāvokti also called jāti, 169f, 220* ; Vm's table of a fig, in D 28, 49*, 73, 94 ; in Bhoja 209 ; in Agp 85, comprehended in Vm's 204. See śabda and guņa artha-vyakti· 15*, 95, 91 ; śabda-doşa 88* disfavoured by Bhã and śabda-hīna, a doșa 9, 10

332 SANSKRIT POETICS

śabda-pāka 240. See pāka śakti, a term used for abhidhā śabda-ślesa 38, 56, 64, 233 146 śabda-vyāpāra or °vrtti, func- śama or nirveda, the sthāyi- tion of word, enumerated bhāva of śānta 273, 276f as three by Bh N 123; Śańkuka, explains Bh's rasa- as explained by dhvani- sūtra 19, 21, 117, 119 ; his theorists 145f; special opinion reviewed in later works on 149 literature 118*, 119f, 218 ; Śabda-vyāpāra-paricaya 149 his theory of rasa 119f ; śabdābhyāsa, or padābhyāsa criticised by Govinda and applied to yamaka 7, others 120f; a naiyäyika 125. 27 śațha nāyaka 269 śabdālamkāra, its treatment Sauddhodani, on rasa 17*, in older writers 6-7; not 247, 291 explicitly distinguished from arthālamkāra by Bh, śayyā, repose of words 240 śānta (rasa) an angin in the- 7, or by Udbhata 61 ; the epics 194 ; not mentioned distinction implied in by Bharata 262, 275 ; Bhã 37-8, D 27, 37* accepted by some writers- 84, but first clearly as additional rasa 114 and adopted by Rdt and Vm fn, 202, 209, 275, 276f ; 62 ; raison d'être of the nirveda or śama as its. distinction 72f ; Rdt on 62, sthāyin 275, 277 ; in rela- 64 ; number always limi- tion to bībhatsa rasa 275 *; ted, largest in Bhoja 72, rejected by others 262, 210 ; discredited by later 276 ; involves philosophi- theory 72; treatment of, cal mental states 262; by D 84, Vm 100, Agp according to Vaisnava 204, Bhoja 210, Mmt 221 ideas 267 śabdārtha-guņa 82 ; first ad- Šāradātanaya, his Bhāva-pra- mitted and classified by kāśa 266 Agp 204 Śeșa 38 śabdārthālaņkāra. See ubhayā- Śińga-bhūpāla(Siņha-bhūpāla) laņkāra his Rasārņava 234; ac- śakti, used synonymously with cepts eight rasas 277 ; on pratibhā 42 rasābhāsa 280

SUBJECT-INDEX 333

śithila, opposite of D's śleșa mental and treated in 80,87 detail by erotic rasa- ślesa, a guņa in Bh 12; writers 265f, 268 ; appears, whether a guna in Subandhu after Bh's description, 29* ; in D 80, 82,95; in as madhura or ujjvala Vm 94f ; comprehended by rasa in Vaișņava trea- ojas of Mmt, 172, 219 ; tises 267 ; classification of in Bhoja and Vin 238 269 :śleșa, a fig, in combination Śṛngāra-prakāśa 202, 209, with other fig 56, 85, 163, 211, 266, 276. See Bhoja 233-34 ; controversy on its Śṛngāra-tilaka 265. See classification 56, 64-5, Rudrabhațța 232f ; as underlying vak- śruta, See vyutpatti rokti fig in Rdt 63 ; both sruti-kasta, unmelodiousness, a verbal and ideal fig, as a doșa 11, 11*, 173, 183 ; well as a basis of classi- saukumārya as its opposite fication of ideal figs in Rdt 219* 62, 64-65; Ruyyaka's śruti-dusta, indecent, a doșa 11 development of Ud's idea śrutyanuprāsa, involved in of śleșa 232-34 D's mādhurya 80, 110 ·śobhā, a lakșaņa 249; an tadvat, a doșa 88* excellence of the hero tadvid-āhlāda, pleasure or 271 appreciation of the sahr- śraddhā as rasa 277* daya (q. v.) 187 śravya (kāvya) 77 tadutpatti (syllogistic) as śńkhalā, a basis of classify- applied to the inference of ing figs 73 dhvani 197 śrngāra (rasa), implied in Bh's tarka-nyaya 73 udāra-guņa and Vm's Taruņavācaspati, on D's view kānti 15 and fn; in D of alamkāra 83 111-12; Rdt on 114-15; tațastha-laksaņa 153 should avoid the dosa Tauta, Bhatta 42, 122* śruti-kasta (unmelodious- tādātmya (syllogistic) applied ness) 172-73 ; extolled by to the inference of dhvani Agp and Bhoja 202, 209, 197 268, 276 ; taken as funda- tādātmya (or tādrūpya), iden-

334 SANSKRIT POETICS

tity of dhvani and bhakti ment of figs 27*, 55f; om 153, 197 utpreksā 29 ; scope and tätparya (import), whether it content of his two works can comprehend the sug- 54 ; admits vrttis in anu- gested sense and rasa i.183; prāsa 55; omits rīti but 136, 153 ; its meaning 148; his vrttis correspond to not universally accepted Vm's rītis and Anv's. as a function 149, 150 guņas 58, 169, 221* ; how tāțasthya 122 far recognises rasa 56, tejas, an excellence of the 113f, 114 ; accepts śānta. hero 271 rasa (q. v.) ; relation to Tilaka, Rājānaka 232 Bhã 54-55, 57; influence transference (of sense). See of 57 ; how far recognises lakşaņā suggested sense 52f, 229; trivarga, as the aim of poetry on samghatanā and guņa 39 59, 168 tulyayogitā 56 uddīpana- vibhāva 274 ubhaya-śleşa 233 ujjvala (or madhura) rasa, a ubhayālaņkāra (miśrālamkara name for śrngāra in the or śabdārthālaņkāra) first limited sense of bhakti- admitted by Agp 72, 205 ; rasa 267 Bhoja on 210; Mmt on Ujjvala-nīla-maņi 267. See. 221 ; in Vin 238 Rupa Gosvāmin udāra, a guņa in Bh 15, 113; ukti, a guņa 239 relation to Vm's kānti 15*, ukti, śabdālamkāra 204* 97, 112-13; D on 15*, 81, ukti-vaicitrya 97, 189, 230. 82, 97; in Agp 15*, See vaicitrya 204; in Vm 95, 96; unexpressed (or suggested ele- comprehended by ojas of Mmt 172; in Bhoja and ment) in poetry 129f. See suggestion Vin 238 udātta, a guņa 238 upacāra 191. See lakșaņā

udātta, a fig 63 *; excluded upacāra-vakratā comprehends.

by Vm 101 some aspects of dhvani 191 Udbhata, on upamā 6, 55; omits yamaka 7* ; treat- upanāgarikā, a vrtti, 55. 64 221*

SUBJECT-INDEX 335

upamā, a fig, in Bh 6, 27 ; in 29, 30, 86; pratīya- Medhāvin 30 ; in Yāska 6* māna utprekșā 61 ; in (i. 3-6) ; its grammatical Agp and Bhoja 204, 210 subdivisions in Ud 55, in utprekşāvayava 86 Rdt 65, rejected by uttara, a fig 257 Appayya Dīkşita 69; faults ūrjasvin, a fig, idea of rasa. in its use 30 and fn, 65, included in, by D 111 ;. 88 (see upamā-doșa) ; its excluded by Vm 101 subdivisions in D 85; vacchomī, a rīti 104 taken by Vm as involved vaicitrya or ukti-vaicitrya. See in arthālamkāra 101, 163 ; vicchitti and vakrokti compared with rasavat vaidagdhya-bhangī-bhaņiti, a 193* ; in Agp and Bhoja name for vakrokti (q. v.). 204, 210; in combination 186 with ślesa 233-34 vaidarbha (kãvya), Bhā on 46, upamā-doșa, older writers on 75,79* 30, 30*, 65 ; included in vaidarbha mārga, in D 67, 79f. general doşa 210. See See gauda mārga upamā vaidarbhī rīti, in Vm 90f; upamāna, a fig, distinguished mentioned by Rdt 60; from upamā i. 110 ; ii. 210 whether it was a reaction upamā-prapañca, a collective from the gaudi 91 *; in name for all ideal figs in Agp 202f; in Bhoja 210; Vm/101 in Vis 227 . See gaudī rīti: upama-rūpaka, a fig 86 and rīti uparūpaka, its introduction as- vaișamya, an upamā-doșa, in' cribed to Kohala i. 21 Rdt 65; opposite of samatā. Upavarșa quoted i. 11 in D 80,87 upekșā (philosophical) 262 vaiśeşika (guņa) 203, 209 Utathya 38 vaișnava ideas, influence of utkalikāprāya, a kind of prose 276; five degrees of bhakti- 76 rasa, according to 267 utpatti-vāda in Bhoja 208. vakratā (or vaicitrya or vicchi- See Lollata tti or bhaniti), turn of ex- utpādya, rasa as 118, 132 'utprekşā, a fig, in older writers pression peculiar to poetry 48, 51, 185f ; see vakrokti

336 SANSKRIT POETICS

vakrokti, a fig of sense, in Vm varņa (colour) attibuted to 48 ; a verbal fig (equivoca- each rasa 273* tion), in Rdt 48 and in all varņa-vinyāsa-vakratā, includ- writers after Mmt 64 es vrtti, anuprāsa and vakrokti, etymological mean- yamaka 191 ing of 48 ; collective desig- varņābhyāsa 27 nation of all alamkāras varņānuprāsa, distinguished in D and Bhā 50, 85; from śrutyanuprāsa 80* relation to D's samādhi varņāvrtti, opposite of śruty- 82 ; as an embellishment anuprāsa 80 of poetic speech 50f, 185f; vartman, used for märga (q.v.) nature of vakratā, vaici- 78 trya or vicchitti involved vastu-dhvani 129, 163, 164, 48, 50-52; produced by 173, 177 ; not conveyed by kavi-pratibhā (q. v.) 48, abhidhā 152; included in 186f ; synonymous with samlakşya-krama vyangya vaidagdhya-bhangī-bhaņiti 161 ; in samāsokti 162 ; its 186 ; relation to atiśayokti place in a complete scheme and the nature of atiśaya of dhvani 163-64; Abh's (q. v.) or lokātikrānta- view on 166, 177-78 ; pro- gocaratā or lokottaratva bably not accepted by Bh involved 50, 51, 187 ; the N 183 ; can be vācya and so-called figs mere aspects anumeya 195f ; in Vis 224 .of, 185-86 ; as the essence vastu-rasa (=agrāmyatva) in and only embellishment of D's mādhurya 80, 110-11 poetry in Knt's theory vastu-vakratā 194 185f; rasa and dhvani vatsala as rasa 267, 276 included in 52f, 191 92; vācaka 150* this view of, developed in Knt but disappears in vācika abhinaya 17

later literature 54 ; coex- vācya or vācyārtha, the ex-

tensive with alamkāra 49*, pressed element in poetry 129, 150* ; the province of 185-86 Vakrokti-jīvita or vakrokti- abhidhā (q. v.) 145-46;

jīvita-kāra 39, 184f. See an aspect of pratipādya sense 156; accepted by Kuntaka Mbh 195

SUBJECT-INDEX 337

vācyāvacana, a doșa 8*, 282 Vāmana, on cūrņa- padas 13* ; vācyopaskāra 229 on iva of utprekşā 29; Vāgbhața, the older, mentions on pratibhā and sources two rītis 194; does not of poetry 41f; object and admit dhvani 244 ; on defi- content of his work 89* ; nition of poetry 245; advance on D 89, 93-4; accepts ten guņas 245 and his theory of rīti as the . nine rasas 275 ; content of soul of poetry 35*, 90f; of his work 246 ; unortho- on the essentiality of guna dox standpoint of, but in- 93f ; his scheme of śabda- fluenced by Mmt and his and artha-gunas 15f, 93, school 206, 252 94f, objected to by Mmt Vāgbhața, the younger, and others 97, 219*, 221 ; accepts three rītis 104; on on figs 99f, on the term dhvani 244-45; adopts alamkāra 99; on the some of Rdt's figs 63, relation of guņa and alam- 244 ; limits gunas to three kāra 99; on drama and 254 ; general standpoint rasa 2, 102, 112 ; his rītis 206, 245-46, 252 ; content equalised with three gunas of his work 246; accepts of Dhk 104, 168*, 171, nine rasas 275 ; includes 219*, 220; on dhvani practical treatment of 229 ; on pāka 240 Poetics 291 Vāmanīyas 33*, 98 vāg-rasa (=śrutyanuprāsa) in vārttā, a fig 28, 86 D's mādhurya 80, 110-111 vāsaka-sajjā (nāyikā) 271

vāg-vikalpa, figs so named vāsanā, latent impression, when devoid of suggestion natural or acquired, 126, 159, 171, 188 ; Knt's justi- 133, 134, 263f fication of 171, 188 vāstava, reality as a basis of vākovākya, a fig 204* classifying figs 62, 73 vākya-doșa 88, 220 vātsalya. See vatsala vākya-nyāya 73 Vedānta, influence of 136 Vākyapadīya 142, 201 veśyā 270 vākya-vakratā 192 verse, in relation to prose in vākyārtha-doșa 88 Sanskrit poetry 45, 76*

45

338 SANSKRIT POETICS

vibhāva, defined by Bh 20; work 235f ; accepts śānta known to Ud but not to Bhã 109, 114 ; relation to rasa 276 ; agrees with Dh's view on rasa 260, 262 rasa, explained by Lollața 118f, by Śankuka 119, by Vidyānātha, includes cra- BhN 122-23, by Abh 130; maturgy 2, 237; accepts bow generalised and be- Bhoja's classification of

comes alaukika 122, 132- guņas 211*, 236, 238 and

33, -161, 261, 264; as Rk's treatment of figs

vyañjaka of the unexpres- 236 ; on guņa and sam- sed rasa and helper of ghațanā 224 *; character

rasābhivyakti 130, 136, and content of his work 161-62; two kinds of 273- 23(-40; on pāka and 74 ; detailed account of, śayya (q. v.) 240; on in the case of each rasa, dhvani 237 in Bh and others 273* vikāsa, mental activity invol- vicāryamāna-ramaņīya (artha) ed in śrgāra rasa 262; 59*, 288 also called vyāpti, asso-

vicchitti (or vaicitrya, or vak- ciated with prasāda 172,

ratā, q. v.) poetic charm 220*, 256, 262

involved in vakrokti (q. v.) vikșepa, involved in raudra 262

or alamkāra 48, 52, 185, vilāsa, an excellence of the 186*, 230f, 257; kavi- hero 271

pratibhā as its source 48, vipakşa-vyāvartaka (logical) 186, 230, 231, 257-58 ; its 198

etymology 186* viparīta-kalpanā, a doșa 88*

vidagdha, as opposed vipralambha-śrńgara 115, 269 to vidvat 186* vipralabdhā nāyikā 271

Vidagdha-mukha-mandana virasa. a doșa 88*

85*, 291 vīra (rasa), in D 111; its vidheyāvimarsa, a doșa 199*, vibhāvas etc 273*

282 virodha (of rasa). See rasa- vidūşaka 270 virodha

Vidyādhara, the virodha, a basis of classifying content, figs 73 scope and model of his virodha, a fig 204

SUBJECT-INDEX 339

virodhi ( deśa-kāla-loka-), a content of his work 283f ; a doşa 10 on nātyālaņkāras and visamdhi, a doșa 9, 10, 88 ; lakșaņas 4, 249 ; on bhāva explained 239 279 ; value of his work vistara, a guņa 239 22 .- 28

vistāra mental activity asso- Visveśvara, his work Alam- ciated with ojas 172, 220*, kāra-kaustubha 251 262 ; involved in vīra rasa vita 270 262 vivakşā. See intention vișama, a doșa 9 viśeşaņa-vaicitrya 73 vivakşitānyapara-vācya (dhva-

viśeşokti, a fig 101 ni), based on abhidhā 160;

viśrānti 126, 137* includes rasa in its scope 160f Viśvanātha, includes drama- turgy 2 ; on kāvya-sarīra vrtta-gandhi, a kind of prose

36 ; on rīti 106-7, 226; 76

on Lollata's view of rasa vrtti, sound adjustment con-

119; on Dhk's definition nected with anuprāsa 55,

of poetry 157* ; influence 64, 80*, 105 *; two kinds in

of Abh's theory of rasa Bhã, three in Ud and five

on 166, 167, 178, 217, 224, in Rdt 55, 58, 64 ; not con-

259, 262; on rasa as the nected with dramatic vrtti

essence of poetry 224f, 55 *; related to rasa by later

262f ; on rīti and vrtti writers 55, 105*, 227* ; its

227 ; Jg's criticism of 166, relation to rīti 104, 105,

178, 224 ; admits śānta 169*, 220-21

and vatsala 276f, as rasas; vrtti, style of dramatrc com- on rasābhāsa 225, 279; position (to be distinguish- on vāsanā 263; on fig ed from anuprāsa-vrtti) rasavat 193 ; on Mbh's 55,105*

theory of anumana 198f; Vrttivārttika 149

cites Agp 202; on Mmt's vrttyaňga 4 detinition of poetry 222; alliteration indebtedness to Mmt 222, vrttyanuprāsa,

225, 227 ; scope and based on vrtti (q.v.) 55, 64, 221*, 227*

340 SANSKRIT POETICS

vyabhicāri-bhāva. See sañcāri- vyāpāra 140; theory of, bhāva how related to sphota- vyakti-vāda. See Abhinava- theory 142-44 ; parallels

gupta to its idea of manifestation

vyangya artha, the suggested 144; comprehends tātparya

sense, older writers on 150 ; defined and explained

52-4, 60, 116, 139f ; the 150; denied by some

idea of, not traceable in writers 151; grounds for

early speculative literature postulating 151f; traced

139, 145 ; not compassable back to other recognised

by abhidha 15lf ; nor by vrttis 180; not identical

tātparya 153; nor by with abhidha 153, or with

lakșaņā 153-54; its relation lakşaņā 153f, or with both

to the expressed sense 154, * combined 154, or with

162, 174 and to the indica- tātparya 153; not com-

ted sense 154f ; not accep- passed by anumāna 155f;

ted when the prayojana of as the special doctrine of

lakşaņā is not charming the dhvani-theorists 157;

155; whether inferrible how explained by reac-

(anumeya) 155-56, 195 ; as tionary writers 179, 184,

an aspect of pratipādya 191-92

artha 156; when called vyartha, a doșa 9 dhvani par excellence 150, 157 ; as the 'soul' of poetry vyāja-stuti, a fig 52, 53, 86

157f ; classific ition of 157- vyāpti-kāraņa (ojas) 172. See

59, 160f, 255f; three vikāsa

aspects of 163f, 195f vyutpatti, culture, as the

vyañjaka 130, 150* source of poetry, called

vyañjanā-vrtti, the function of śruta by D 42

word and sense which vyutpanna, a doșa, opposite

reveals the unexpressed or of D's prasāda 80, 87 rasa 129, 136, 180, 225 ; its yamaka, a fig in Bh 5; treat- non-recognition in early ment of, in early writers speculation 139, 144, 152 ; 7*, 62, 64, 85, 100 ; the explained as anyathāsiddha term śabdābhyāsa used for

SUBJECT-INDEX 341

7 ; distinguished from anu- yati-bhrașța, a doșa 10 prāsa 27 ; omitted by Ud yaugikī, a guņa 204 64 yāvadarthatā, a fig 205 yathāsamkhya, a guņa 204 Yāyāvarīyas 292, 293 yathäsamkhya, a fig (also call- yogyata, compatibility of ed saņkhyāna or krama) words 149 29 ; rejected by Knt and yukti, a lakşaņa 249 Rk 230 yukti, a śabdālamkāra 204

ADDITIONS AND CORRECTIONS

Vol. i. p. 99. Add the following as fn. 4 to 1. 12 : For a review of the Alamkara-section in the Agni-purana see Suresh Mohan Bhattacharya in Bulletin of DCRI, xx, pt. i, pp. 42-49. p. 202, fn 1, 1. 2. Read śabdā- (for śadhā-). p. 292, No. 69. Read BHAŞYA. p. 294, fn. 1. Read Catalogue after Tanjore. Vol. ii. p. 141, l. 28. Read kuśāgrīyayā. p. 173, 1. 2 and p. 183, l. 28. Read śruti-kasta in both places. p. 204, l. 13. Read praśastatā.