1. Sanskrit Poetics Edwin Gerow ( See A Glossary of Indian Figures of Speech)

A HISTORY

OF INDIAN LITERATURE

EDWIN GEROW

INDIAN POETICS

OTTO HARRASSOWITZ . WIESBADEN

A HISTORY OF INDIAN LITERATURE

EDITED BY JAN GONDA

VOLUME V Fasc. 3

1977

OTTO HARRASSOWITZ · WIESBADEN

EDWIN GEROW

INDIAN POETICS

1977

OTTO HARRASSOWITZ . WIESBADEN

A HISTORY OF INDIAN LITERATURE

Contents of Vol. V

Vol. V : Scientific and Technical Literature, Part II

Fasc. 1: J.D.M. Derrett Dharmaśāstra and Juridical Literature Fasc. 2: H. Scharfe Grammatical Literature Fasc. 3: E. Gerow Indian Poetics M. Hahn Metrik C. Vogel Lexicography

CIP-Kurztitelaufnahme der Deutschen Bibliothek A history of Indian literature / ed. by Jan Gonda NE: Gonda, Jan (Hrsg.] Vol. 5. Scientific and technical literature : P. 2. Fasc. 3-> Gerow , Edwin : Indian poetics Gerow , Edwin Indian poetics. (A history of Indian literature; Vol. 5, Scientific and technical literature; Fasc. 3) ISBN 3-447-01722-8

C Otto Harrassowitz, Wiesbaden 1977. Alle Rechte vorbehalten. Photographische und photomechanische Wiedergabe nur mit ausdrücklicher Genehmigung des Verlages. Gesamtherstellung: Friedrich Pustet, Regens- burg. Printed in Germany. Sigel: HIL.

CONTENTS

1. Introduction 217 2. Pre-history of Indian poetics 220 3. The historicity of poetics 223 4. Bhamaha, Daņdin and Bharata: Alamkarasāstra 226 5. Vāmana and Udbhața 233 6. Rudrata 238 7. Nāțyaśāstra: Rasa in dramatic criticism 245 8. Rasa and bhakti: the Dhvanyāloka .. 250 9. After the Dhvanyāloka 258 10. Rājaśekhara .. 260

2. Kuntaka and Dhanamjaya 262

3. Abhinavagupta 264 13. Mahimā, Bhoja, Rudrabhatța 268

4. Mammața 271

5. The post-Mammata period 274

6. The late theorists 283

Bibliography 289

Index .. 297

Edwin Gerow

INDIAN POETICS

1. Introduction

Speculations on literature are a persistent theme of Indian literary history from its earliest beginnings. The Brahmanical literature may be seen as the exegesis of the Vedic samhita, particularly the Rk, once the latter had been canonically fixed in the context of the sacrificial performance. The earliest upanisads present themselves as 'secret' expositions of the purport of certain well-established ritual acts, as codified and rationalized in the brāhmana texts. We know how remarkably seminal the upanisads have been in provoking the most diverse interpretations by later Indian philosophers, who sought once again to "reveal" the sense of those texts in the context of novel religious and social phenomena. The early history of India, from this 'formal' point of view, often seems little else than a series of attempts to readjust a text that had become authoritative to a situation that no longer called for such fixed inter- pretation. Out of that impasse emerged a new authority, and so the process was repeated, renewed. The history of early India is, as often remarked, its literary history.1 And this is true also in the material sense that we are vouchsafed little original data beyond the texts themselves; but texts that are typically not concerned with and do not convey the factual data on which we build our histories. The texts are very revealing on the macrocosmic issues of form and purpose, but nearly silent on the microcosmic issues of occasion and authorship. So while we are dealing with a historical problem that disposes of few but literary resources, the literature is certainly one of the least adaptable to historical concerns. The tension is evident in any "History of Classical India"

1 iha śistānuśistānām šiștānām api sarvatra / vācām eva prasādena lokayātrā pra- vartate // Daņdin, Kāvyādarśa (KĀ.) 1,3. "Thanks to words alone the affairs of men progress, the words (of the) first learned, or (the) learned thereafter, or even (of) their appendices (students)." Even here the pun is inevitable.

218 E. Gerow . Indian Poetics

but becomes almost a paradigmatic study in circularity when literature itself is confronted as an historical problem.2 Literature used as a or as the source in the study of extra-literary dimensions of Indian culture is at least subject to certain probabilistic checks that are grounded on more 'objective' observation or theory; in the history of literature, however, both the goal: an ordered chronological explanation of individual works and theories, and the presuppositions: distinctive chronological identifi- cation of individual works are interdependent and inseparable. The theory of literature that developed on the Indian soil might otherwise provide some independent check on the literature, but is itself exclusively concerned with purposes and forms of literature, and not at all with its occasion: it is, in other words, literary philosophy or aesthetics, rather than criticism. Also the theory of literature little helps the chronological interpretation of literature itself, both because the literature presents itself as largely anonymous and authorless, and because the literature as well as the theory takes for granted that litera- ture, being Sanskrit, is a paradigmatic creation, whose reality is therefore in the relation to the model, or norm, rather than in the accidents of its origin. It should be very clear then that in attempting to write a 'history' of Indian speculations on literature, we are doing something to the material that is quite

2 The two standard histories are P.V. KANE, History of Sanskrit Poetics, Delhi, 3rd revised ed. 1961; and S.K. DE, History of Sanskrit Poetics (two volumes), Calcutta, 2nd revised ed. 1960; both of which were originally published in 1923 (the former as an "Introduction" to Kane's Sahityadarpana), just a year after the appearance of Winternitz' third volume. No serious work on the subject can be undertaken without their aid, and this essay is no exception. Differing in focus, the two Histories are partly complementary; Kane deals with the intricate and often inconclusive issues of text-chronology; De is more concerned with the history of ideas that may be said to be demonstrated by the sequence of texts (especially his second volume), a theme further developed in his Sanskrit Poetics as a Study of Aesthetic, Berkeley, 1963. Their agreement is our siddhanta. Our work is not quite however a résumé of their views; the reader will find little made of the distinction between text and idea: I am trying here not only to summarize the history of the texts of poetics, but to develop through them a sense of the meaning of the terms (alamkara etc.) they convey. De and Kane, if they share a fault, both appear to think that text and idea are two different problems, not two necessary aspects of the same problem. From Kane we often seem to get a history of the texts that means nothing, and from De, a sense that the texts are interesting only insofar as they fulfill some predefined potential of ideal aesthetic "progress". If this history differs from theirs it will be in the attempt to discover what the history is about through the writers' arguments themselves; we will not, for example, be able to give au préalable, a definition of figuration, and then trace its history, for our history, if it is successful, will end in a notion of figuration, etc. Another ambitious effort, the History of Alamkara Literature, by S.N. DASGUPTA, pp. 513-610 of DASGUPTA and DE, History of Sanskrit Literature (Classical period), 2nd ed. Calcutta 1962, is flawed both by errors of fact and judgement. Cf. the review by M. EMENEAU, in JAOS 71, p. 86-7. See KALIPADA GIRI, Concept of poetry. An Indian approach, Cal- cutta 1975.

Introduction 219

alien and perhaps even incompatible with its essential form. History, in the sense that we will be using the term here, is indeed an imposition of recent Western scholarship on the literature of India, ably seconded now by several generations of Indian scholars who are at least much closer to the spirit and the matter of their subject. It can run the risk of replacing that dimension of the original material most central to its formation: tradition itself, samprādaya. Indian poetics is a tradition, therefore a self-validating activity. In attempting to fix a point at which we may affirm with some confidence that an Indian theory of literature exists, it behooves us to identify the kind of literature the theory purports to comprehend. In a trivial sense all written documents are 'literature', and therefore, in a sense not quite so trivial, all Indian literature, insofar as it is traditional, is a theory of literature also (at least of the preceding kind of literature, as pointed out above: the history of India is the history of Indian literature). The history of Indian poetics (which term we will substitute freely for "theory of literature" henceforth) is rightly thought to begin in some kind of quite circumscribed relation to the literature we call belletristic, i. e. literature as a fine art, kāvya.3 The distinction of "literature" in this specific sense, from "literature" in its all-encompassing meaning is both a socio-historical problem (for the Indians themselves identify this category only in terms of a certain epoch from which more authoritative 'Vedic' literatures are excluded) and a formal one (for any effort to define belles-lettres seems to presume the discovery that literature, the written (or spoken) word, is a thing in itself, and can be judged in terms differentiable from the criteria applicable to the things it talks about, its content; that discovery is usually reinforced by the insight that literature for-its-own-sake can provoke and is uniquely adapted to a peculiar pleasure not associated with other-oriented uses of language). Later Indian critics never stray far from this central problem: defining the difference (visesana) of literature vis-à-vis other conventions of expression, science (sastra) and narrative (itihasa or katha) chiefly. In this, they do not differ materially from their modern counterparts.4 But the Indian materials present us with no theory of art for art's sake, which would simplify our task: the first works indubitably impressed with

3 Which includes for the purposes of this very general discussion, nātya (drama)- for this is also in part a verbal art; in what follows, however, the distinction will preoccupy us. 4 The literature on literary theory in the West is vast: cf. R. WELLEK and A. WARREN, Theory of Literature, New York 1956, and its bibliography pp. 307-47; for some attempt to apply this method to Indian literature see E. GEROW, A Glos- sary of Indian Figures of Speech, The Hague 1971, Introduction, p. 9-89. On sastra and narrative, see inter alia S.K. DE, The Ākhyāyikā and the Kathā in Classical Sanskrit Literature, in BSOAS, 3, p. 507; L. RENOU, I.C., p. 1358-69, 1517-18.

220 E. Gerow . Indian Poetics

the stamp of the kavya style, the Ramayana and the epics of Aśvaghosa5, are nonetheless also first rate pieces of narrative or apologistic (so-called sectarian) literature. It is simply the case that the first kāvyas are also first do- cuments of renascent and self-conscious socio-religious movements: classical Hinduism and Sanskritized Buddhism. Earlier Indian literature is often marked with characteristics we associate with self-conscious literature; Renou would even put the Vedic hymns in the context of poet-sabhas, where compe- tition and reward were not functions of piety but of image, metaphor, rhythm and double-entendre6 (all persistent motifs of Indian poetic creation: witness the later Bhojaprabandha? and even contemporary panditic relaxations, samasyāpūrti,s etc.). Poetic theorists remain acutely aware of the extrinsic purposes and rewards (kirti, etc.) of the poetic act; and except for a few off-beat remarks and ironies,9 both poets and poeticians seem both to acknowledge and to serve the vital context in which the poet lives and works. Kāvya, though it is a kind of language different in itself, is not seen apart from and indifferent to the social and intellectual dimensions of language, or to society itself. If anything, its difference is seen in its preeminence, in its being truly what other language is only in potentia: fully exploited expression-whose principle is not subordinate to an external standard-indeed, reflecting the authoritative status of the Sanskrit language. The poetic statement involves us in the model or norm of language at its most self-reflective level; beyond the poem is only silence.10

2. Pre-history of Indian poetics

The beginnings of kavya have more links with the Veda than they suffer points of difference; indeed it may well be the Veda, understood as Veda, as self-existent utterance (mantra) (i.e. reinterpreted in Brāhmana and early

5 Ca. first century A.D .: L. RENOU, L'Inde Classique (I.C.) 820, 444; cf. NABEETA SEN, Comparative Studies in Oral Epic Poetry and the Vālmīki Rāmāyaņa in JAOS, 86, p. 397; SUKUMAR SEN, The Language of Asvaghoșa's Saundarananda Kāvya, in JASB 1930, p. 181. 6 L. RENOU, Études Védiques et Paninéennes, passim; esp. I, p. 18; also KANE, H.S. P., p. 326. 7 Attributed to Ballala; tr. LOUIs GRAY, in American Oriental Series, vol. 34. New Haven, 1950. 8 Where scholars compete, more or less ad libitum, to complete a verse part of which has been fixed as "problem" (samasya); the Bhojaprabandha has many examples, including several fabulous ones about Kalidasa; cf. KEITH, H. S. L. p. 344. 9 Like the famous one of Bhavabhūti: ye nāma kecit ... , Mālatīmādhava, 1, 8 (Nirnaya Sagar Press ed., Bombay 1936; commented on by S.K. DE, Sanskrit Poetics as a Study of Aesthetic (S.P.S.A.), p. 62ff. 10 See L. RENOU, 'Sur la Structure du Kāvya', Journal Asiatique, 1959, pp. 1-114.

Pre-history of Indian poetics 221

Mīmāmsa exegetics), that created kavya by serving as its model. The rsis are also kavi.11 The first kavyas show evidences of self-conscious organization of language whose delineation and rationalization constitute the matter of the later poetic traditions: too well crafted to be the products of a naive lyricism-just as Renou argues the Vedic poems cannot be folk-romanticism. By definition, then, the first poetics is implicit in the first poetry. At the same time it would be futile to begin the study of Indian poetics (śastra) with the first kavyas, for that is to confuse our understanding of the literature with the historical problem of India's understanding of its literature; the latter problem surfaces when the self-consciousness in the poetry itself is made the object of study.12 And it would seem, despite our well-grounded con- viction that such secondary literature must have existed,13 that the production of the first several centuries (0 -- ca. 500 A. D.) has been irrecoverably lost. The initial conditions of this sastra are thus not different from those of any other Indian śāstra.14 Despite this, certain terminology, while not testifying to poetic concerns per se, does demonstrate some of the links that poetics has with other and more general Indian investigations of literature. We find, for instance in the Nirukta of Yaska, a notion of simile (upama) understood technically in a way that prefigures the usage of those who wrote on the figures of speech nearly a millenium later.15 Panini also seems familiar with this analysis16 of the basic figure into four "elements": the subject of comparison (upameya or upamita) ; the thing with which it is compared (extra-contextual: upamāna); the property or standard of similitude (sāmānya, or samānadharma); and the adverbial or other grammatical indicator of comparison (sāmānyavacana or dyotaka).17 These early concerns for figuration do not establish a poetic in our sense; rather they testify to the important exegetical and even so-called rhetorical status that the figures enjoy concurrently with their poetic implementation-in all probability the link between a more general notion of sāstra (esp. mīmāmsā) and the sastra of poetics. Yaska is concerned with the contextual employment

11 Renou, E.V.P. I, p. 26-7; IX, p. 15-16; J. GONDA, Vedic Literature, in this History, I, p. 71; 74; for more fanciful views see P.S. SASTRI, 'Rgvedic Theory and Treatment of Rasa and Dhvani', Poona Orientalist, 9, p. 111. 12 Contra KANE, H.S.P., p. 338, who argues that the existence of poetics is established by the style of poetry. 13 The earliest texts refer to "predecessors", Bhāmaha to a Medhāvin, for example (2,40; 88). 14 M. WINTERNITZ, Geschichte der Indischen Literatur, III, p. 4-5. 15 Citing even a definition of simile: upamā yad atat tatsādrsam "not that but similar to that" attributed to a Gargya, who therefore must qualify as the first ālamkārika: Nirukta 3, 13. 16 Aştādhyāyī, 2, 1, 55-6; 2,3, 72; 3, 1, 10 etc. Cf Bhartrhari, Vākyapadīya 590 and K.D. SHASTRI, Bhartrhari on the relation between Upamāna and Upameya, Viśveśvaranand Indological Journal, 2, p. 87. 17 DE, H.S.P., I, pp. 3-8; GEROW, G.I.F.S., s. v. upamā.

222 E. Gerow . Indian Poetics

of upamā, and particularly whether the status of the upameya is greater or less than the upamāna; Panini needs the grammatical analysis of simile to explain certain grammatical operations that express similes: compounds, deverbative and comparative suffixes, etc. These references are very important in showing us where to look for the intellectual inspiration that later emerged full-blown in a proper poetic: grammar and Vedic interpretation (mīmāmsā).18 Poetics is (in India) a development of an interest in certain kinds of ex- pressive devices that are grounded in language. Some devices are specifically linked to the Sanskrit language (metrics, prosody, alliteration, etc.)-what the poeticians will term sabdālamkāra; others-the arthālamkāra, as the simile of Yaska and Panini-are not specific to the language as sound, but concern rather the expressive content of that language (while not being any the less determined grammatically, syntactically: the comparative compounds of Pāņini). Inscriptional evidence is often adduced19 from the second century, where "alliteration and other tricks of words" (Kane) are as consciously employed as they are in the 'degenerate' kavyas of the late classical period (Bhāravi and Māgha). But again, this proves no more the existence of a "theory" than would the same observations applied to the RgVeda.20 A third area of speculation concerning the 'pre-history' of poetics is occa- sioned by the stray occurrence of certain terms that later are central to the technical poetics, in poetry of an earlier period. Aśvaghosa mentions rasa;21 Bana several of the 'figures.'22 The strongest argument that can be made in this case is that an explicit poetics did indeed exist from which these words were borrowed-an argument that is likely enough; but the argument cannot produce any systematic treatment of these stray terms, and thus is everywhere

18 GEROW, Mimamsa and the evolution of Sastra, article forthcoming, in JAOS. 19 KANE, H.S.P., p. 336; A.B. KEITH, A History of Sanskrit Literature, pp. 48-51, etc. 20 Unless the conceits of an artist are everywhere to be referred to a pre-existent inventory-a proposition that denies the possibility of naive or inspired art, or the naive universality of certain kinds of forms. This issue much vexes the Indian tradition as well as modern critics: poetry in its "inspirational" mode is often discussed with reference to the term pratibha; poetics as an instructional mode, via the term kaviśikşā: cf. DE, S.P.S.A., p. 36, 76, and a voluminous literature in the periodicals: T.N. SRIKANTIAH, Imagination in Indian Poetics, in IHQ 13, p. 58-84; G. KAVIRAJ, Doctrine of Pratibha in Indian Philosophy, in ABORI 5, p. 1-18, 113-32; K.A.S. AIYAR, Pratibhã as the meaning of a sentence, PAIOC, 10, p. 326; Pt K. MURTHY, Observations of Sanskrit Literary Critics on Poetic Imagination, Poona Orientalist, 9, p. 123; V. RAGHAVAN, History of Bhavika in Sanskrit Poetics, IHQ, 14, p. 787; F.W. THOMAS, The Making of a Sanskrit Poet, in Bhandarkar Commemoration Volume, p. 375; K.C. SHASTRI, Requisites of a Sanskrit Poet, Jour. of the Dept. of Letters, Calcutta Univ. (JDLC), 26, no. 2; KUPPUSWAMI SHASTRI, Alamkara Sastra and its bearings on the creative aspect of poetry, PAIOC, 4, 2, p. 57. Pratibha and sastra are first opposed by Bhamaha, 1,5. 21 In the compound rasantaram, Buddhacarita 3,51. 22 KANE, H.S.P., p. 340.

The history of poetics 223

unable to prove the impossibility of the counter-argument, namely that a conventional but not technical sense is implied by these terms. The arguments about the origins of poetics tell us in sum very little about poetics, save that it is unlikely to have emerged full-blown in the earliest extant works of Bharata and Bhamaha. They do identify, as we have said, certain persistent contextual relationships of the sāstra with other language- oriented inquiry and with notable types of language use (prasasti, etc.). These relationships do not materially change in the subsequent 1,500 years of ex- plicit history. With the first texts, however, emerge genuine historical issues that influence the theory of intellectual development we argue in the sastra itself. These issues can conveniently be subsumed under three heads: the subject matter of poetics; its relation to Buddhism and the Bhattikāvya; and the question of independent composition vs. compilation. Poetics, when it emerges into (text-)history, involves not one subject matter and method, but reflects several different and radically distinct concerns, whose clarification constitute the historical problem of the first period of poetic theorizing. Only near the close of the first millenium A.D. does a unified theory of the mimetic art emerge (the Dhvani theory).

3. The historicity of poetics.

1. In its broadest sense, poetics at its text-inception was concerned with two different subject matters-nātya, the 'Sanskrit' drama, and kāvya, Sanskrit court (esp. epic) poetry (So strongly distinctive are the subject matters, by style, by means of composition, by aesthetic effect that it is only in a nominal sense that we speak of "a" poetics at this period; in fact, there are two poetics, each with its "proper" subject matter and methods (and this continues to be the case even later, though not in the radical sense we dimly perceive ab origine). Nātya-śastra, faced with a genre whose means of expression were largely, and perhaps even primarily, non-verbal (abhinaya, characterization, setting, etc.) seems to have identified its first problem in the question, "What indeed did (or 'could') the drama express?". And to have sketched a solution (vide NS 6,7 GOS ed.) in terms of a different kind of emotional response (termed rasa) than the emotion (bhava) that is commonplace in every ex- perience.23 Alamkara-sastra,24 on the other hand, employs only one means of expression-the verbal-whose expressive capacity is not subject to any naive doubt. Other possible means of expression-even abhinaya-have to be conveyed in words. The problem of kavya-sastra was then seen in differentiating

23 Nāțyaśāstra (NS.), chs. VI and VII. 24 Viz. Kāvyasāstra, understanding here the specific meaning rather than the generic one that eventually becomes synonymous with "poetics".

224 E. Gerow . Indian Poetics

that particular expression we call "poetic" from other verbal means, sastra and narrative. And it sought a solution to this problem not in a theory of response but in theory of language that could reconcile grammaticality (for it is language) with a comprehension or an understanding that did not appear to be derived from it. It is taken for granted that in the uses of language we consider normal, the understanding derived from a sentence is founded directly upon the powers of the words or their combination (the respective views of the two Mīmāmsa schools most cogently at issue here), a relation that can be conveniently termed denotative, on which is based the propositional truth value of the utterance. The distinctive feature of poetry, in this basic sense, lies in its sentences or propositions that are by that standard simply not true, yet which in terms of the knowing purpose of the poet are not only true, but usually are even more striking than the truth.25 Bhamaha's mention vakrokti (2.66) is an early witness to this mode of thought-but is not in itself an 'aesthetic'.26 2) Relation of poetic texts to other known works. The issue does not concern the intellectual relationships between poetics and its allied śāstras, but rather the explicit relations between early poetic texts and other works: the Buddhist logics of Dinnāga and Dharmakirti27 (for Bhamaha is widely thought to be a Buddhist on the basis of his father's name28); the academic poem of Bhatti (Bhattikāvya) that systematically illustrates the same set of alamkāras given in Bhamaha and Dandin; and finally whether Dandin is also the author of the

25 "Striking," rather than "convincing," and in that sense we are everywhere dealing with poetic, not rhetoric. Unless they can convince us that persuasion was a major concern of the Indian poets, historians of Indian poetics should cease using the term "rhetoric" as a pejorative synonym of "poetic." Modern writers who do this are guilty of bad rhetoric, for this is part of an effort to persuade us that no theory of expression concerned chiefly with words can truly be called a "poetic". In contrast to bad poetic (viz. rhetoric) stands good poetic: aesthetic. S.K. DE (Sanskrit Poetics as a Study of Aesthetic) is often taken as representative of this view, which is in fact shared, perhaps unthinkingly, by most others-with the notable exception of HERMANN JACOBI: ZDMG 56, p. 392; also Ueber die Vakrokti and über das Alter Dandin's, ZDMG, 64, p. 130. See GEROW, G.I.F.S., p. 13ff. 26 Alamkāra, a "making adequate", a literal interpretation clarified by J.GONDA, The Meaning of the Word Alamkara, in A Volume of Eastern and Indian Studies in Honour of F. W. Thomas, Bombay 1939, p. 97ff. Bhamaha testifies to a discussion vivid among his predecessors as to the scope of alamkāra (1, 13-14); he takes a middle ground, defining "ornament" neither wholly in grammatical or in intentional terms. In fact, he may well have been the first to understand the sense of alamkāra that dominates the later "school", formulating a theory to link alamkaras of "sense" (artha) with those of "sound" (sabda). See also KANE, H.S.P. on The Alamkara School (pp. 372-8) esp. the references, p. 375; H.R. DIWEKAR, Fleurs de Rhétorique, Paris 1930; H. JACOBI, Über Begriff und Wesen der Poetischen Fi- guren ... , Göttingen 1908. 27 See S.P. BHATTACHARYA, Neo-Buddhist Nucleus in Alamkāraśāstra, JASB, 1956, pp. 49-66; and KANE, H.S.P., pp. 64-6. 28 Viz. Rakrilagomin; see KANE, H.S.P., p. 83; nevertheless Bhāmaha is criti- cized by Sāntarakșita!

The historicity of poetics 225

Daśakumāracarita and Avantisundarīkathā.29 Involved in these questions is the toughest chronological problem of the alamkārasāstra: the relative priority of Bhāmaha and Daņdin.30 3) Two kinds of texts are found in both the nātya and the alamkāra traditions: one, by definition older, is the puranic and authorless compilation: the Na- tyasastra itself, but also the sections of the Agni-and the Visnudharmottara puranas that deal with poetics;31 the other, works that by tradition and scholarly judgment are the compositions of individual authors.32 It is clearly more difficult to argue chronology among texts that are ex hypothesi not the product of a fixed time; when such texts are interspersed with datable texts, relations become quite complex if not indeterminate. The oldest 'dramatic' text (NS) is a compilation (though attributed to Bharata)33 the oldest two texts of the alamkāra tradition (B, D) are clearly not. One of the purānas (Agni) bears evident relations to B and D; the other seems rather to look to Bharata (or Bharata to it). Given their strict separation of theoretical outlook, the problems of chronology of the dramatic and strophic poetics can to some extent be mitigated by refusing to see them in a single development.34

29 An excellent summary in KANE, H.S.P., pp. 94-99; see also J. NOBEL, Die Avantisundarīkathā, ZII, vol 5, p. 136. 30 Infra, pp. 227 ff. 31 KANE, H.S.P., pp. 3-10, 66-72, and Antiquity of Agni Puraņa, IA, 1917, p. 173ff .; S.K. DE, Date of the Alamkara Section of the Agni Purana, Poona Orientalist, 2, fasc. 1, p. 15; V. RAGHAVAN, Rīti and Guņa in the Agni Purāņa, IHQ, 10, pp. 767-79 (reply to P. C. LAHIRI, Concepts of Riti and Guna in Sanskrit Poetics, Dacca, 1937); S.M. BHATTACHARYA, Alamkāra Section of the Agni Purāņa, Bulletin of the Deccan College Research Institute, 20, p. 42. 32 A distinction made by the tradition as well: works of the first type are ge- nerally attributed to rsis or munis, indicating a relation to the Veda; the latter to individuals, historical and laukika. 33 KANE, H.S.P., pp. 10ff ;. DE, H.S.P. II, pp. 1ff .; M. GHOSH, The Date of Bharata Nāțya Śāstra, JDLC, 25, fasc. 4, pp. 1-54. 34 As is often done, when Bharata is judged oldest, because he knows four figures, the Vișnu is put next because it knows 17, Bhamaha next, who defines over 30, etc. (KANE, p. 71); whatever the significance of this 'progression', it is of historical value only given the assumption that augmentation or complication is the only 'historical' process that need be considered; but given the fact that the NS. is both a compilation (by 3) and chiefly oriented to the drama (by 1), we adduce a prima facie countervailing reason why that text might treat the figures only dinmatram, so to speak. Note the v. l. to NS. 16-43 cited in the NSP ed. (p. 260) alamkārās tu vijñeyāś catvāro nātakāśrayāh. It is not possible in the compass of this essay to reproduce the intricate arguments that have been advanced in support of various chronologies of early Indian poetic texts. The work of P.V. Kane, by its compre- hensive breadth and fine judgment, we take here as the authority, complemented by S.K. De's work of the same name, and of course by various works of V. Ra- ghavan. We will attempt to indicate what we feel to be crucial arguments that divide authorities; we think it important to emphasize the problematical and imponderable character of the ultimate material, which emerges from the lack of agreement of even the most judicious authorities. This is illustrated perhaps too well in the case of the Nātyaśāstra. Kane, fully cognizant of the purānic and com-

226 E. Gerow . Indian Poetics

4. Bhūmaha, Dandin and Bharata: Alamkāraśāstra

It is conventional to begin the text history of Indian poetics with the Nāțyaśāstra, a purāņa of materials pertinent to the staging of the classical (Sanskrit) drama. In its various recensions (none of its editions can yet be said to be critical) three of its approximately thirty-six chapters deal with the key issues of the two poetic traditions: rasa-bhāva (6,7) and alamkāra-guņa (16). Other chapters touch on related topics: plot (19), genre (18,20), metre (15), the latter reserved to a separate sastra. By and large the text pertains to dramaturgy in its practical aspect, and thus does not concern us here.35 Immediately later writers, Bhāmaha, etc., mention Bharata by name36

pilatory quality of the text, concludes that "a work" at least similar to the present NS. existed before 300 A.D. (p. 47). Now the only solid evidence for putting it that early is the highly speculative alamkāra argument: "all ancient writers on alamkāra, Bhațți (between 590-650 A.D.), Daņdin, Bhāmaha, Udbhața, define more than thirty figures of speech. Bharata defines only four, which are the sim- plest ... " (p. 47). Since other portions of the NS compilation cannot be put con- veniently before 300 A.D., Kane settles for that date, thinking 300 years enough to devise 26 alamkāras. To comment on this procedure is to demolish it. Positive indications in the poetic sections themselves suggest that the rasa sūtra (kernel of chapter VI) had been commented on by 750-800 A.D., by Lollata, whose com- mentary is given in résumé by Abhinavagupta (ca. 1000 A.D.) and Mammața (ca. 1100). But these two arguments taken singly would separate the beginnings of dramatic and poetic speculation by four centuries! And to bridge the gap an unacknowledged assumption is made: that the diverse portions of the text enjoyed a coherent and coeval development. I cannot see that this is in any way justified. 35 It is noteworthy that those chapters of Bharata, especially XIX on itivrtta, that bear closest resemblance to our own classical poetics (esp. Aristotle, based also on the dramatic form) are not inventoried by modern writers seeking the origins of Indian poetic theory! Modern historians of Indian poetics do not consider plot an important problem of Indian aesthetics. We wonder whether this is not a function of the things historians are looking for in the Indian poetics. 'Plot' does continue to be a concern of writers on dramaturgy (Dhanañjaya). It is clear that medieval commentators on the drama were quite familiar with its technical ana- lysis. 'Plot' in a way becomes the subject of a sub-section of the later dramatic literature, epitomized by the Nāțakalakşaņaratnakośa of Sāgaranandin (DE, I, p. 310). That it did not find its way into "poetics" until Viśvanātha (15th century) indicates the limited character and aims of this enterprise: an understanding of the poetry where plot is not an issue -- the strophic poetry of the classical period. Even in its 'epic' guise (Raghuvamsa, etc.) the unit of composition remains the individual verse. The story is pretext, a substitute for anthology (also an important mode of "composition"). In time, a true dramatic poetry is written in India, under the impulsion of the bhakti cults of the middle ages; the 'poetic' changes in response to this new challenge, as we shall see. 36 Śāradātanaya distinguishes at least two Bharatas (Bhāvaprakāsa, GOS 45, p. 36), one called "Bharatavrddha", whose views are evident in our NS .; see S.K. DE, The Problem of Bharata and Adibharata, Journal of the Sanskrit College, Calcutta, 1, part 2; H.R. DIWEKAR, Quelques Sutras de Bharata, JA, 1935, p. 314 (possibility of a sūtra version of NS.). The older alamkārikas are conspicuous for their lack of deference to Bharata: Dandin even says teşām (nātakādīnām) anyatra vistarah (1,30); also 2,367 and Bhāmaha 1,24; see KANE, H.S.P., p. 93.

Bhāmaha, Daņdin and Bharata: Alaņkāraśāstra 227

but appear to associate him with the "other school"; his chapter on alam- kāraśāstra (16) is by most arguments of internal chronology considerably older than the chapters on rasa, natyasastra per se (6,7: see footnote 34). We have for this reason, and because the otherwise oldest datable texts (Bhamaha, Dandin) are of the alamkara tradition, chosen to begin this history with an account of the kavya-oriented poetic. On this score, Bharata's contribution to the tradition is rather mediocre. His list of only four alamkāras (upama "comparison", rūpaka "metaphorical identification", dīpaka "enlightener", several parallel phrases being each com- pleted by a single unrepeated word, and yamaka "word play by means of repetition") treats indifferently figures of imagery, syntax and 'rhyme'. His analysis of upama is immediately abandoned by later writers; dipaka and yamaka are subjected to a mechanical analysis that appears to anticipate the worst in Indian kāvya, not the best. The list of 36 laksanas, or "charac- teristics" of dramatic poetry is largely ignored. Only Bharata's treatment of guna "good quality, merit" and dosa "defect" appear to have struck a responsive chord, and are developed by Dandin and Vamana especially. It is disconcerting that the beginnings of alamkāraśāstra have to be sought in a text on nātyaśāstra. The Nātyaśāstra, as fixed probably in the eighth century, indeed became the creative base of a tradition, beginning with Lollata, of aesthetic speculation on the rasasutra; on the other hand it provoked discussion of dramatic characterization and plot construction-issues that dominate the Daśarūpa of Dhanamjaya.37 But its sixteenth chapter seems to have fallen on barren ground: it was either a late effort to suggest the import for drama of speculations current in the other, kāvya-oriented poetics (alam- kāra-šāstra) (much as some alamkāra texts, Rudrata notably, make passing reference to the chief topics of natyasastra) or, if early, represents a 'poetics' that had apparently to be abandoned (the emphasis on laksana particularly, absent in any creative way in later works)38 in order for alamkārasāstra to originate. We will deal with Bharata's views on rasa in the next section. The alamkāra-tradition may then be said to originate in the works of Bhā- maha and Dandin, whose chronology and mutual relations we now discuss. The Kāvyālaņkāra and the Kāvyādarśa are remarkably similar in point of view, content and purpose; modern scholarship, attempting to sort out the chronological relation of the two texts, has emphasized the differences rather than the fundamental agreement.39 Both works have a single basic purpose: to

37 Infra, p. 263f. 38 laksana: an effort at ad hoc characterization, randomly focusing on content, goal, and method; seemingly an empirical list of the characteristics a play may have. See V. RAGHAVAN, Laksana, in Some Concepts of the Alankāra Sāstra, Adyar, 1942, pp. 1-47. 39 See A.B. KEITH, Dandin and Bhamaha, in C. R. Lanman Commemoration Volume, Cambridge Mass., pp. 167-85, and in general KANE, H.S.P., pp. 78-133, with a detailed résumé of the chronological arguments pro and con, pp. 102-33; also H. JACOBI, Bhamaha und Dandin, SPAW 1922, p. 210.

228 E. Gerow . Indian Poetics

define the mahakavya, or sargabandha40 as it was then termed; their method focusses on those qualities of language and thought that impart on the one hand a distinction to poetic utterance in general, but also define the integrity of the utterance that is mahakavya. In point of view, both writers see their subject matter largely in terms of the distinction gadya: drsya (verbal and dramatic art), relegating drama to the purview of the "other theory" (as above). Again, in format, the works are remarkably similar, often quote one another or appeal to a common source tradition. The definition of mahākāvya is an instance in point. But also they appear pointedly to disagree if not with each other, at least with doctrines espoused by the other: Bhamaha rejects Dandin's views on guna (1.31-2) and Dandin's figures hetu, etc; Dandin, on the other hand appears to reject Bhamaha's views on the difference between katha and ākhyāyikā (1.23-5)41 and to argue against Bhamaha's view that poetry must be vakrokti (2.244) allowing varta or svabhavokti an important place.42 Bhāmaha may be dated with some certainty before the Buddhist writer Santaraksita (who condemns him) and after Jinendrabuddhi (ca. 700 A.D.), author of the Nyāsa to Kāśikāvrtti (whom he quotes unmistakably), thus between 700-750 A.D. Dandin's date is less certain, but it seems obvious that he must be a close contemporary of Bhämaha, whether one reads the close arguments as estab- lishing a slight priority for Dandin43 or Bhamaha.44 No certainty exists regard- ing the locale of their activity, though a tradition associates Dandin with Vidarbha (Berar). It is as though a dialogue had developed between the two authors, a difficult assumption perhaps, but one that would account both for their fundamental agreement and the acerbity of their disagreements. Both works begin with the usual attempt to situate kāvya in the universe of discourse: patronage and educational prerequisites, languages and other genres.45 The major thrust of both works follows:46 a discussion of the distinctive qualities (guna), forms (alamkāra) and debilitating detractions (dosa) of poetic assertion. Poetry is language, and it is language caught in rather small compass: the individual verse forms of the varnavrtta metrics; the stanza is its unit of composition, the whole in which its perfection is to be sought,47 Thus is clarified both the close attention paid to the capacities of assertive language, and the lack of

40 Bhāmaha, 1, 18ff .; Dandin, 1, 14ff. 41 See S.K. DE, The Akhyayika and the Katha in Classical Sanskrit Literature, BSOAS, 3, p. 507. 42 KANE, H.S.P., p. 112, 107-8. 43 So Kane. 44 So S.K. DE, H.S.P., I, p. 62. 45 B. 1,1-30; D. 1,1-39. 46 B. 1,31-4-end; D. 1,40-end. 47 RENOU, Sur la Structure du Kavya, JA 1959, pp. 1-114; D.H.H. INGALLS, An Anthology of Sanskrit Court Poetry, HOS 44, Introduction; GEROW, "Indian Poetics" in Indian Literature, An Introduction, Chicago, 1974.

Bhāmaha, Daņdin and Bharata: Alaņkāraśāstra 229

interest in poetic issues that have more general reference: in content, context, or poet. Issues that emerge in larger frames presume 'works' that transcend in form and purpose their constituent language. Given a poetic culture (based doubtless on the verbal culture that sought pithiness as the very power of speech: mantra, sūtra) that knew no works, it is superficial to demand a poetic suited to works-yet many modern writers tax the classical theorists for just this lapse. There is no disagreement on the central place occupied by figures of speech (alamkāra), although Dandin gives far more space to those figures that are defined as regularities of phonetic or phonemic features (sabdālamkāra) e.g. 'rhyme' (yamaka), than does Bhāmaha; both deal extensively with arthālam- kāra,48 figures that consist in striking modes of assertion, expression of meaning. The distinction is basic in all subsequent alamkara literature. Their differences on this point do not lie chiefly in the kind or quality of figures admitted (despite the overblown controversy on the figure hetu) but seem more to be a function of the organization and presentation of the materials. Bhamaha divides his alamkāra in four groups that are represented as layers of a tradi- tional development. The first five are the four mentioned in Bharata (ch. 16) plus alliteration (anuprasa). It is widely assumed that the Medhavin mentioned by Bhāmaha (2.48) was a writer on alamkāra;48a just as natural is the pre- sumption that the four groups of figures (anyair udāhrtah) represent anterior attempts at compilation, older layers of sastra; Bhamaha either attests a genuine historical elaboration of figures of speech,49 or he sought to bolster his own compilation, which may have been one of the first comprehensive efforts to define poetry, by adducing the partial or disjointed character of the subject as formerly treated. Evident is the relation of Bhamaha's figures to those illustrated in the tenth book of Bhattikavya (one must be a calque on the other, but no certain evidence is available either way).50 Given the Bhatti-like quality of certain passages in Aśvaghoșa-five centuries earlier, and long be- fore Kālidāsa, where aorists e.g. are introduced class by class and morpholo- gical 'puns' are freely made,51 it is again only surprising that we have to wait until the 7th or 8th century for a work wherein this knowledge is codified and coherently presented.

48 Esp. the paradigmatic arthālamkāra, simile (upamā): 2, 14-65. 48a The judgment is ultimately founded on Namisadhu's statement ad Rudrata 1,2. 49 So DE, H.S.P., II, pp. 27-31. 50 KANE, H.S.P., pp. 73-4; DE, H.S.P., I, p. 52-6; perhaps the safest conclu- sion is that of De, who posits another source for both. Even the apparent mockery in Bhämaha 2, 20 and Bhatti 22, 34 can be read either way. See also H.R. DIWEKAR, Bhāmaha, Bhatți and Dharmakīrti, JRAS, 1929, p. 825. (Bhamaha precedes

20, p. 351. Bhatti); C. HOOYKAAS, "On Some Arthālamkāras in the Bhattikavya X", BSOAS,

51 Cf. Saundaranandakāvya, Canto 2.

230 E. Gerow . Indian Poetics

Dandin, on the other hand, while accepting approximately the same number of figures (and the same figures) as does Bhamaha, seems rather to explore the variations provided by context and intention that differentiate each figure internally. He notices thirty-two types of simile; lesser refinements of other important figures, viśeşokti, rūpaka, ete.52 This effort to look at figuration in terms of context, in terms of refining the figure to the shadings of context, while interesting, and perhaps closer to 'criticism' than any other early theore- tical work, was rejected by subsequent authors in favour of classifications based on the forms of the figures themselves, for it is clear that figures, seen contextually; i.e., in exemplification and not in their abstract capacity to assert, are endless.53 Dandin, like Bhamaha, is aware of the relations among figures in this formal sense: rūpaka is a simile (upamā) in which the similitude is suppres- sed:54 in other words where the similitude, implying difference between the things compared, is posited rather as identity; neither author, however, attempts to follow out the implications of such parentage. Both authors discuss the category that had come to represent the inverse of "embellishment": defect, or dosa.55 Just as certain modes of discourse serve to distinguish speech as poetic, so certain deficiencies, both positive and negative, destroy that distinctiveness. The discussion of the dosas is very much tied, of course, to grammatical issues, for lapses of this sort, in a highly artificial language may fairly be said to compromise intelligibility itself. But it is not syntactical fault-the lapse-that retains the poetician's attention; rather it is the faults of exaggeration, of style-that seem functions of trying too hard, and with more energy than skill, for the distinction that is truly poetic: language too ornate to be understood in its plain intention; language so savant as to require sastraic decipherment; language clumsily used so that the powerful engine of double-entendre is not kept under control, etc.56 The dosas are not the contraries of the alamkaras, for the references of the notions are different; the alamkara may be seen as a way of achieving a meaning that is suitable and yet not direct; the dosa accounts for counter- vailing obstacles57 to the achievement of that meaning. The third category,

52 Such as the common tetrad, jāti, kriyā, guņa, dravya; for these figures and other explicit references to figures, see GEROW, GIFS, sub voce, or GERO JENNER, Die poetischen Figuren der Inder von Bhamaha bis Mammata, Hamburg 1968. 53 anantā hi vāgvikalpāh: Ānandavardhana: emphasized by DE, SPSA, p. 15. 54 Dandin, 2, 66. 55 See V. RAGHAVAN, Bhoja's Srngaraprakasa, Madras 1963, ch. XV; K.K. MOORTY, the doctrine of dosas in Sanskrit Poetics, IHQ, 20, p. 217. 56 B. 1,37-59; ch. 4; D. 3, 127-end (fourth pariccheda in some editions). The diffuse views on dosa are, as one might expect, admirably collated and summed up by Mammata, chapter 7. 57 A notion that takes a new lease on life as the rasavighna: obtacle to the awareness that is rasa: Abhinavagupta on the rasasutra: GOS 36, p. 280 (see infra, p. 264-8).

Bhāmaha, Daņdin and Bharata: Alaņkāraśāstra 231

guna,58 "quality" illustrates the divergence of the two authors. Dandin devotes an important section of his first pariccheda "chapter" (preceding his discussion of alamkāra in 2 and 3) to Bharata's ten gunas or qualities of style, which serve not so much to discriminate poetry from non-poetry (as do the dosas and alamkāras) but kinds of poetry from each other.59 Dandin, in this sense, refers to two mārgas or styles, to which he gives the geographical designations Vaidarbhī (from Vidarbha, modern Berar) and Gaudi (from Gauda, modern Bengal).60 The problem of distinguishing one kind of poetry from another, or put more bluntly, of poetry from itself, is, as may be surmised, not a congenial issue in Indian poetics, concerned as it is with the form of poetic utterance in its generality. Dandin's essay in this area is almost a dead issue, being taken up by only one early author, Vamana, who extends the concept of 'style' and also dilutes it.61 The importance of the gunas lies in their service as charac- teristics, as 'plus-features,' of poetry whose alternative is not necessarily non- poetry. In other words, the contrary of a guna may be and usually is another feature whose presence marks another kind of poetry (rather than non-poetry). Thus ojas "vigour" (sp. use of long compounds) marks the Gaudi style, its absence marks the Vaidarbhi. This model is not universally applicable, but serves often enough to indicate the peculiar place of guna in the theory.62 If the 'styles' truly exist, it is (perhaps) as regional habits that have become stereo- typed and thus imitable, much as are the scenic Prākrits, as languages. It is not that a character speaking Sauraseni comes from Gujerat, but that Gujerati has a recognized literary status in expression of certain social or contextual values, all over India. Bhamaha in contrast rejects the guna approach as unfruitful, is an unabashed ālamkārika: his concern is with the form of poetry, not its variations.63 He is also a precursor of the widely held view that the gunas are three (not ten) (a case of historical reduction?) and are nothing but

58 See RAGHAVAN, op. cit. ch. XVI; P.C. LAHIRI, Concepts of Riti and Guna in Sanskrit Poetics, Dacca 1937 (reprint 1974), and Lahiri's search for guna in various early texts of alamkāraśāstra: IHQ 6, p. 345; 7, p. 57; 9, p. 448 and 835. 59 Bharata did not appear aware of this internal differentiation of poetry or drama in adducing the gunas (16,97ff., NSP ed., Kāvyamālā 42). Mārga is not mentioned; instead the gunas appear to be desirable in all composition, and do complement the dosas rather than the alamkaras. But Bharata's use of these terms is often at variance with later tradition. 60 See S. BHATTACHARYA, Gaudi riti in theory and in practice, IHQ, 3, p. 376; S.K. DE, A note on Gaudi rīti, New Indian Antiquary, 1, p. 74 (latter republished in Some Problems of Sanskrit Poetics, Calcutta 1959). 61 The question is raised again by Bhoja, several centuries later, and by several of his followers, in the context of dramatic poetry; see RAGHAVAN, op. cit., ch. XIV. 62 Note that the figures are complementary, not contrastive in this sense. 63 His remarks seem so pertinent to Dandin's peculiar treatment that it is difficult, with Kane, not to accept that he was commenting on it: 1,31-2, 2,1-3 (KANE, H.S.P., p. 112-13).

232 E. Gerow . Indian Poetics

varieties of alliteration.64 But while rejecting the usefulness or the importance of a marga analysis of poetry, he does, reimport, in a final chapter-to which no correspondent is found in Dandin-a notion of sausabdya-grammatical appropriateness in poetry-which in general corresponds to the issues a theory of 'style' might help to define. But he does it with specific reference to gram- matical authority, rather than to general or stereotyped variations of usage that are ultimately impressionistic or statistical. This final chapter in Bhamaha seems not to have provoked a response by Dandin (if Dandin is posterior). And Dandin's heavy emphasis on ornaments of sound (sabda) is not present in Bhamaha at all. Dandin discusses alliteration under madhurya guna,65 and the bulk of the third pariccheda is devoted to an exhaustive treatment of citrakavya-or adhama-(rhyming [yamaka], visual poetry [mantra and citra] and conundrum [prahelika]).66 Alliteration and rhyming are not ignored by Bhamaha (they are his first two figures, in fact); it is rather the full development in Dandin that is striking, together with the emphasis on poetic forms that seem academic or even decadent. Perhaps Bhamaha also sought to respond to this kind of interest in grammatical and phonological strata of language, (which he appears to have deplored, along with S. K. De) by refocussing our attention on the true problem of grammar in relation to poetry: namely its helpful character as an adjunct (and its dangerous character if under- or over-developed!).67 In this context a question is raised regarding Bhamaha's relation to Buddhism (and to Buddhist logic). Chapter five is an inquiry into poetic defects that spring from logical fallacies. It is in general an interesting and quite original contribution to alamkāraśāstra, especially in view of the well-developed (later) view that one type of figuration is defined by the manner in which logical relations (inference, identity, cause- effect, etc.) are misused.68 There is a limit to the poet's power to set aside universal laws of reasoned discourse, and the test established in later works is the implicit clarity of the resulting understanding, which like Aristotle's probability is a unique creation of the poem that only thereby exists in this world. The poet does not wish to speak nonsense; his ultimate declaration is as rational and reasonable as that of any other man. Poetry does not therefore

Mammata's ch. 8. 64 A śabdālamkāra and Bhāmaha's first alamkāra; the reduction is clearest in

65 1,52-60-prefiguring Mammata's reduction? 66 Some of which seems to be reflected in the style of the Daśakumāracarita: especially the niyama of Mantragupta's episode (ch. 7), no labial stops at all (see JACOBI, ZDMG 40, p. 49). Dandin's 'poetic' seems better to correspond to the poetry of Mägha and Bhäravi. His identification with the poet Dandin rests unfor- tunately on wholly conjectural evidence: the independently determined congruence of their dates, and the elegance of their styles. Since there is nothing against it, we may as well accept it. See KANE, H.S.P., pp. 94ff. for discussion. 67 A concern that reappears in the 'recent' strata of Indian poetic speculation, especially in the Citramīmāmsā of Appayadīkșita. 68 Ruyyaka, especially, but also Rudrata (qqv.).

Vāmana and Udbhața 233

lie in the poet's intention as such, but in the unusual means he adopts to declare his meaning.69 This homily puts poetry properly on both sides of the distinction between logic and illogic, a matter again which Bhāmaha shows familiarity with. But historians have found this chapter curious for another reason: Bhamaha's logic appears to have reference to the Buddhist logics then current, and therefore tends to suggest that he was a Buddhist, supporting the odd dedication of his work to Sarva,70 the Buddha. The issue also joins that of his date, for certain scholars71 identify traces of Dharmakirti (fl. 643-95) in the arguments, which would tend to confirm Bhamaha's dating by reference to Nyasa (making it in other words likely that it is the Nyasa of Jinendrabuddhi cited in 6.36). But others72 disagree, and assert that the logic involved can be understood by reference to Dinnaga. Which would make it possible to put Bhamaha somewhat earlier. The weight that Kane73 attributes to Santaraksita's criticism of Bhamaha's logic in arguing that Bhamaha was not a Buddhist, appears to rest on the rather shaky ground that one Buddhist never criticizes another. That Bhamaha uses Buddhist logic (or is familiar with it) is not in itself remarkable: logic as a discipline originated in Buddhist schools of the period.)

5. Vāmana and Udbhața

The early history of the sastra falls naturally into such pairs. Vāmana and Udbhata in precise and pregnant ways develop the differences that separate Dandin and Bhamaha (whose basic agreement and apparent debate we have alluded to), do so themselves as contemporaries (ca. 800) and very likely as rivals at the court of Jayapīda of Kashmir (799-813);74 yet in their works there is no trace of mutual reference or apparent controversy. We will treat them together, for the intellectual development of the Indian poetic tradition depends on the antitheses their works illustrate.

69 Intention, tatparya, "the that about which" is the focus of a continuing controversy in poetics, and seems to be one of the key points on which the poeti- cians differ from the grammarians and the mimamsakas. In these latter schools a sentential meaning (vākyārtha) is deemed relatable to word meanings, but poeticians are unanimous (pace Mahimabhatta) in locating their problem-and the possibility of poetry-in the non-relation (visesana) of those two levels. See Mammața 2, 6 and Abhinava's Locana ad Dhvanyāloka 1, 4 (edition: KUPPUSWAMI SASTRI, Madras 1944, p. 110). 70 Confirmed by R. GNOLI, SOR 27, p. xl. 71 Esp. JACOBI, SPAW, 1922, p. 212. 72 G. TUccI, Bhamaha and Dinnaga, IA 1930, p. 142, which is in fact a reply to K.B. PATHAK, Dharmakirti and Bhamaha, ABORI, 12, p. 372. 73 KANE, H.S.P., p. 84. 74 KANE, H.S.P., p. 146-7; DE, H.S.P., I, p. 80.

234 E. Gerow . Indian Poetics

In different ways Udbhata and Vamana represent the first efforts that have survived to encompass or organize the theory of poetic diction under a principle. Both authors, however, continue the major thrust of the alamkāra or kavya-oriented tradition of speculation. The notion of rasa, as in Bhamaha and Dandin, is referred to, but never as a principle.75 No trace of the dhvani synthesis is found, through which the rasa will be elevated to an all-encompassing principle. Rasa remains a subsidiary in the composition of verses, each of which may indeed "have a rasa" (rasavat) but whose identity as poetry in no way depends on that adjunct, and which indeed must pre-exist76 the rasa that is perceived. Udbhata is indebted to Bhāmaha, of whom he is the only known commentator (except for D.T. Tatāchārya's "Udyanavrtti"77)-recovered in fragmentary form and published by Gnoli -; 78 his list of alamkaras defined follows that of Bhamaha. It may be presumed that Udbhata in fact supplanted Bhamaha, accounting for the near eclipse of the latter (of which a Ms did not come to light until 1906).79 It may then be asked, in what particular does Udbhata's treatment excel that of his mentor? It might appear that the reason had something to do with the form of the work, whose illustrative verses (of the figures as defined) themselves constitute a portion of a kāvya written also by Udbhata (another Kumārasambhava). But in this Udbhata does not innovate, probably taking his model from the relation between Bhāmaha and Bhattikāvya. No, the answer lies in what Udbhata does not do; the tradition, which pre- ferred him to Bhamaha, apparently appreciated that Udbhata concerned himself only with alamkara; of the other topics giving Bhamaha's text such a balanced appearance, none is represented in Udbhata's: no discussion of guna, or dosa, or grammatical purity. It is possible that we are dealing with only a frag- mentary text (though it seems complete in its own terms); if the rest has been allowed to lapse by the commentatorial tradition, it only strengthens our point, that Udbhata achieves distinction as an uncompromising representative of the view that alamkara is the central issue of the kavya-poetic. The sastra stands or falls on this issue, none of the others have any decisive import. We are not to conclude that Udbhata had a 'theory' about alamkāra. He does not even mention Bhamaha's term vakrokti that modern historians have

75 A verse is rasavat "accompanied by a rasa", and this accompaniment is admitted as an alamkara, among many others. In the NSP Udbhata, some remarks of his commentator, Pratīhārenjurāja, are attributed to Udbhata, including a view that "rasa is the soul of poetry". See DE, H.S.P., II p. 58-9, and infra, note 132. 76 As the vakrokti of Bhāmaha: 2,85 (Tatacharya ed.) sā (viz. atiśayoktir) eşā sarvaiva vakroktir; anayā artho vibhāvyate // ... ko 'lamkāro 'nayā vinā? and 1,36: na nitāntādimātrena (reference is to the alliteration in 2,5?) jāyate cārutā girām / vakrābhidheyasabdoktir iștā vācām alamkrtih. 77 Tiruvādi 1936; see KANE, H.S.P., p. 81 (we use this edition). 78 Serie Orientale Roma, 27, 1962. 79 KANE, H,S.P., pp. 135, 80.

Vāmana and Udbhata 235

erected into a theory.80 De's hope that the Bhamaha Vivarana might include a disquisition on this term is not borne out by Gnoli's fragments.81 Udbhata, on the basis of his extant work, which is nothing but definitions and illustra- tions of Bhāmaha's alamkāras, appears to have been the most uncompromising ālamkārika and at the same time the least theoretical: for him there is no subject of study other than the figures in their concrete differentiability.82 Indeed, if the older alamkāra poetic is, as we assert, a study of the capacities of poetic diction,-little purpose would be served in expounding a universal theory, (such as "ambiguity,") for it is only in variety that figuration con- stitutes a "universe" of discourse. Bhāmaha's mention of vakrokti is a summa- tion, not a 'theory.' It is as tautological as the implied contrast (1.30) with svabhavokti would suggest. Put in another way, telling a poet that he must speak "deviously" does not help us understand what in fact he says. The interest, both for poet and for audience, is in how he does it in practice, not in his intention. Vamana and Dandin also are often grouped together as the two ancient exponents of the riti or guna school of poetics.83 Like Udbhata, Vāmana should be seen as a follower who brought not only an analytic interest to the study of poetry but attempted for the first time to offer a rationalization of the subject; unlike Udbhata, rather than singling out a single principle for inquiry, he attempted to find a way of relating in a single organized whole the various principles that had been discussed by his predecessors. He seems most closely indebted to Dandin (rather than Bhamaha) in the sense that he assigns great importance to the notion of guna, or "stylistic element";84 but he differs markedly also from Dandin, not only in trying to organize his subject but in appearing to find in the notion of guna, and style itself (riti) that very principle that permits the integration of the other principles of analysis (including most particularly alamkāra) in a holistic view of poetry.

80 DE, H.S.P., II, p. 250, S.P.S.A., pp. 23ff .; The Vakroktijīvita, Calcutta 1961, Introduction. 81 Op. cit., p. 36: verse 2, 84 is rather fully commented by Udbhata; the remaining fragments do not appear pertinent to the issue-and the next page of Gnoli's text has already proceeded to 3,14-suggesting that the Ms discovered is not only defective but incomplete. Cf. DE, H.S.P., II, p. 54. 82 The copious notes of the editor, N.D. BANHATTI, of the BSPS edition, 1925 (vol. 79) of Udbhata are useful both for the text, and for situating Udbhata's doctrine. 83 DE, H.S.P., II, ch. III; RAGHAVAN, Some Concepts ... , Riti, pp. 131 ff .; KANE, H.S.P., p. 379-80 (links explored but Dandin not considered a member of the 'school'; P.C. LAHIRI, op. cit., p. 84. On the use and abuse of the terms 'school' in reference to the variations of opinion expressed in different Indian poetic texts, see the strictures of S.N. DASGUPTA, H.S.L., pp. 574-7. Much less decisively DE, H.S.P. II, p. 32, ft 1. 84 See V.V. SOVANI, History of the Gunas in Alamkara, Poona Orientalist, 3, fasc. 2, p. 88; SURESH KUMAR, A Note on Sanskrit Stylistics, BDCRI, 24, p. 65; and RAGHAVAN, Śrngāraprakāśa, ch. XVI.

236 E. Gerow . Indian Poetics

He does this by seeming to develop a chance reference of Dandin (2,367) wherein the word alamkāra is used in the non-technical sense of "embellish- ment," into the first effort to define kāvya in terms of its special features: kāvyam grāhyam alamkārāt (1,1,1); alamkāra is here not Udbhața's "figure," but saundaryam "beauty" (1,1,2). By this generalization, he is able to consider any "excellence" that can be linked to the charm proper to poetry, as an alamkāra, which notion he explicitly develops in defining the relation between guņa and alamkāra (in the narrow sense) (3,1,1; 2). Out of this synthetic effort emerges the first attempt to characterize the sine qua non of poetic diction, in any rigorous way: ritir ātmā kāvyasya; rīti represents for Vāmana that collocation of gunas and judiciously subjoined alamkāras (and implicitly the negative factor of no-fault, no dosa) that produces saundaryam (or sobha) the peculiar consequence of poetry, kavya. In this reappraisal are several striking novelties caught in the guise of commonplaces: first is the definition of kävya itself, here attempted for the first time in a form that links the specific verbal dimension of poetry with an effect that is also specific ("beauty"). Bhāmaha speaks laconically śabdārthau sahitau kāvyam (1,16) and the only notion he offers as a generalized quality of poetry (vakrokti) is clearly in the diction, not in the hearer. Thus Vāmana, in effect, is the first ālamkārika writer to focus attention on the "aesthetic" effect as something other than a mere understanding of words and to regard this as constitutive of the genre kāvya. The effort to synthesize principles in the notion riti is itself unprece- dented, and this synthesis is possible only in terms of the notion of aesthetic effect spoken of (kavyasobhāyāh kartāro dharmāh gunāh 3,1,1). And finally the effort to isolate, in this view of poetry as an organic whole, the "animating prin- ciple," the ätman-a rather striking departure from the usual manner of seeking to understand the poetic visesana, the differentia of poetry. Because poetry in a full account requires more than a theory of language, and must in some sense come to grips with the purposes or intentions of the poet himself, the poetician broadens his inquiry to include the proper effect of poetry, and refocusses the definition in terms of that effect, seen as an organic process of all the elements.85 However felicitous the general outlook of Vamana's work, it must be observed that the detailed working out of his ideas received little or no accep- tance in subsequent centuries; his theory is one of the significant dead-ends in the history of Indian poetics.86

85 See note 69, supra; the scope of the inquiry is still the individual verse: this discovery of the poet's intention suggests more a skill at finishing, at combining, than a full-fledged 'imagination', suited to the explanation of larger poetic wholes. The term riti is pregnant with these overtones: rinanti gacchanti asyām guņā iti; rīyate kşaraty asyām vānmadhudhāreti vā rītih (Gopendra Bhūpāla ad Vāmana 1,2,6, KSS vol. 209, p. 15): riti is the "going" or the "flowing together" of the elements of a poem. 86 Unless one looks to the impulse, rather than to the result: cf. B. BHATTACHAR- JEE, The Riti School and Anandavardhana's Dhvani Theory, JASB, 1951, p. 5.

Vāmana and Udbhața 237

The reason for this is probably to be sought in his crucial but imprecise notion of the relation between guna and alamkara: "quality of style" and "figure of speech." This relationship is, of course, central in his attempt to work out an organic theory of the poetic whole (the whole of poetic discourse). Vamana appears to go astray on two points, where his very notions of guna and alamkara are influenced by the theory of relation he is trying to work out; it may be said that Vamana, in being the first who tried to comprehend the elements of poetry in one whole, distorted the parts unrecognizably in the interests of congruence. In the first place, guna and alamkāra are related to each other through their external relationships to riti, the level on which the poetic whole, the proper poetic 'beauty' may be said to be palpable. Both guna and alamkāra are judged constitutive elements of that riti, but in different ways. Relying on time-tested analogies, Vāmana speaks of guņa relating to rīti as moral vir- tues (like courage) relate to the man; alamkāras, as ornaments or embellish- ments. The former are constitutive87, the latter are adventitious, may aug- ment poetic beauty, but do not determine its absence by their absence. In this, the nature of both guna and alamkāra, as understood by previous tradi- tion is radically transformed. Dandin, it has been noticed, thought of guna as a factor enabling discrimination of various styles; there was nothing in- herently "good" about a "quality" as such; the absence of a quality might be as crucial in defining a particular "style" as its presence (ojas, for example). And poetry is not the same thing as style, in the rather obvious sense that there are at least two possible styles (though Dandin prefers one) neither of which exhausts poetry. By Vamana's redefinition, a guna is good in itself (kavyaśobhākāra ... ), and thus its presence or absence determines not a style within the general framework of poetry, but poetry itself.88 Thus Vamana's notion of style turns out to be empty, as he indeed admits 1,2,11 samagragunā vaidarbhi: There is truly only one style, only one riti, characterized by the presence of all the gunas. Though in a superficial way he admits more styles (three) than did Dandin (two), the others turn out to be extreme and defective poetries, characterized by lack of balance (12, 13) and opposed to one another. But the notion of a guna that discriminates only poetry, only the essence of poetry is very difficult to comprehend, particularly in that the entire preceding discussion of the early alamkāra-sāstra is cast into limbo. All his predecessors deemed figures of speech forms of utterance wherein the poetic differentia was to be sought. What then becomes of the figures in Vamana's reevaluation? He redefines them-that had been considered essential-as adventitious, as earrings that one may doff or don without destroying the basic integrity of the

87 In the sense of the samavaya relationship: without virtue there is no man (-liness). 88 This appears to have been the sense in which guna was used in the Nātyasāstra as well, where it is opposed to dosa: cf. DE, H.S.P., II, pp. 12-16.

238 E. Gerow . Indian Poetics

face. But the analogy is terribly distracting. If I take away the 'simile' of the Meghaduta, that the cloud is like a messenger, the poem becomes vārttā-a statement of fact, which furthermore is wholly preposterous (that yaksa is standing on a hill talking to some cloud!). In the figure is the very credibility of the poem, not to speak of its poetic impact, its 'beauty'.89 We cannot of course be so bold as to say that Vämana is in error in attempt- ing these reevaluations in the interest of comprehensibility; but it will be clear that his view, in its detailed working out, sits very ill with the prevailing direction of Indian theory. He further attempted to systematize (and also here was the first to try it)90 the domain of figures of speech under a constitutive principle (as opposed to an expressive property that they had in common, like Bhamaha's vakrokti). Accepting the division of figures into those of sound and those of sense, he sought to comprehend all the latter group under the cate- gories suitable to simile 4,2,1-(upamā). The distortions involved in defining hyperbole, or a poetic version of the cause and effect relationship, as simile, can be imagined. As in the case of guna, the effort to find a constitutive principle of definition, was welcome, and anticipated later developments (Rudrata, Ruyyaka and others); but the particular solution offered was inadequate, and quite evidently so. Whether Vāmana was induced to develop his notion of figure as simile because of his general redefinition of figure in relation to guna, is not clear. It is another point of obscurity that again makes Vāmana a most enigmatic writer.91 Vamana's treatise closes with a chapter on grammatical usages in relation to style, in which he elaborates on the last chapter of Bhamaha's work.92

6. Rudrata

In the Kavyalamkara of Rudrata, we come to the culmination, in one sense, of the early alamkāra tradition that focussed on the classical kāvya. Rudrata carries the penchant for systematic analysis to a more satisfactory conclusion,

89 V. RAGHAVAN defends a similar point of view in Use and Abuse of Alamkāra in Sanskrit Literature, Some Concepts ... , pp. 48ff .; though his interest is 'later' focussed on the notion of aucitya, "propriety" that the dhvani theorists employ to regulate the subordination of the figures (Dhvanyāloka 3,33). Aucitya in this sense is often the entree to the topic of the "education of the poet" (kavisikşa: supra, note 20) as in the Aucityavicāracarcā of Kșemendra (DE, H.S.P., II, pp. 283 ff.). 90 Contra, KANE, H.S.P., p. 153. 91 An intriguing question is whether these peculiarities may be related in any way to the "other" Vamana, the well-known grammarian and co-author of the Kāśika. Like the grammarian, the poetician seems interested in constructing wholes (upamā) out of elements (upameya etc.) by rules of conditioned applicability. Despite the congruence of dates and views, the identity is rejected by most authori- ties. Cf. KANE, H.S.P., p. 147. 92 KANE, H.S.P., p. 83.

Rudrața 239

and for the first time, includes the topics of natyasastra that are suitable to kāvya (hero, heroine, etc.). Udbhata obviously and Vāmana no less certainly restrict themselves exclusively to kavya-sastra, and by that fact belong to the very oldest layer of speculation on kāvya. Rudrata evidences the growing syn- cretism of poetics, based no doubt on the increasing practical difficulty in keeping the genres distinct.93 In presenting a theory of kāvya, based on alam- kāra, he feels obliged to provide an appendix on dramaturgy. No synthesis is attempted; the major thrust is in development of a traditional view of alamkāraśāstra. Rudrata's work is remarkably anonymous. He mentions no other authors and cites none (his examples, like Udbhata's, are of his own composition). We are reduced to thinking him a Kāśmīra largely because of the suffix (-ta) on his name. Oddly, his work has been cited in an astronomical text (Utpala on Varähamihira) permitting rather certain dating in the quarter century before 880 A.D.93a Rudrata is the first successful systematist.94 After the conventional intro- ductory chapter on the purposes and prerequisites of poetry, the next nine chapters offer an account of the topics of the alamkāra tradition: guna, dosa, alamkāra-organized around a theory of figuration. Rudrata, in effect, does what Vāmana tried to do, by taking alamkāra, rather than guna, as his constitu- tive principle. In Rudrata's account, the 'styles', or the modes of existence (vrttis)95, as they come to be called in extenso, are explained as varieties of śabdālamkāra-not now reduced to one of the known sabdālamkāras (as Mammata will do: alliteration)-but considered in effect a sixth grammatical figure.96 Riti, though it does not fit so well into this neat progression from 'supra-segmental' to the 'syntagmeme' as basis for unexpected repetition, is defined as the effect gained by employing compounds variously (2,3-6). It does not have to do with repetition of given morphemes (pun), but can be thought of as repetition of stem-classes that results from systematically deleting certain kinds of morpheme formants (the inflections of Sanskrit). Rudrata, by empha- sizing the guna ojas, is thus able to integrate the riti theory in the wider alamkāra theory; in this he will be followed by most writers, notably Mammata, who goes even further in denying a distinctive "place" to riti in his system.

93 In this context the effort of some later writers on rasa-Dhanika, Sāradā- tanaya-to resurrect the distinction between kävya and nätya on the ground of the "unplayability" of the new (ninth) rasa, sānta, is most interesting (Bhāvaprakāśa, p. 26, GOS vol. 45). 93a KANE, H.S.P., p. 154-5. 94 See P.K. GODE, The Problem of the Classification of the Alamkāras, ABORI, 2, p. 69, and GEROW, G.I.F.S., pp. 35ff. 95 A fourth, latī, is added. 96 The other five being neatly predicated on the 'levels' of grammatical analysis: (a) on supra-segmental intonation (vakrokti: NB., see GIFS), (b) on recurrences of partially or completely identical phonemes (anuprāsa, alliteration), (c) on repeti- tion of whole syllables (yamaka), (d) on congruence of morphemes (śleşa, pun), and (e) on repetition of syntactic (verse) units (citra, the "visual" figures).

240 E. Gerow . Indian Poetics

Rudrata's treatment of the sabdālamkāras is intellectually elegant because of the clarity of the underlying principles of relation. Equally elegantly, he accounts for the arthalamkaras (7-10), by grouping together figures of assertion according to the natural or logical categories that underlie that assertion:97 vastava "factual": aupamya "similitudinary", atisaya "hyperbol(ic)" and the familiar slesa, "pun" or more precisely here: "sense-congruence". Rudrata's classification disputes Vamana's apparent conclusion that all figures are based on simile; it is one thing to say (with Vamana) that all figures may have or suggest a touch of simile, and another to say (with Rudrata) that only some figures involve an assertion that is a simile. Rudrata's first category appears counterproductive98 but the key to its understanding appears to be Rudrata's notion "simile", always considered the typical or the most excellent figure (Dandin, and others)99. In a fourfold classification of this sort, what is the peculiarity of simile? Rudrata's definition as such does not differ materially from that of the tradition: a simile is a relation (similitude) between two things that are different, yet share some aspect of sufficient note to per- mit us to overlook that difference. Note the logical relations implicit in the very definition: simile is a relation, of two subjects; therefore implying a difference in re: yet the relation is specific: partial identity, or identity in property; but all things "have something in common" (if it's only existence or "being"). What is it that identifies the poetic relation of similitude as worthy of note? Here a complex of considerations intervenes: in many similitudes, the relation of identity is indeed the point at issue, the literal content of the obser- vation: potassium is like sodium (in entering into compounds, etc.). The difference between sodium and potassium, though not denied, is not relevant to the aims of the utterance. Such similes are "non-poetic", indeed their purpose is in- struction and may be considered scientific. Those of the law-courts, similarly, are persuasive-as to culpability or innocence in this crime. They also bring us to consider what makes a simile "poetic". "My luve is like a red, red rose";100 the relation of partial identity becomes "striking" in the poetic sense when the difference in subject is not 'irrelevant' or ignored, but transcended-funda- mentally relevant; and of course it is the relevance, relation to context, that the poetic simile asserts strongly. Since the difference of subject is not ignored in poetic simile, the distinctive character of both is recognized: one 'subject' is definitely in context, the upameya (Burns' "love", to whom he addresses his thoughts) the other some- where else (the upamana, the "rose"), brought into context only through the

97 Expressed here are concerns that parallel those of the 18th century "uni- versal grammarians", James Harris et al. 98 Bhamaha in fact considered "figuration" and "factuality" contraries (1,30); see DE, H.S.P., II, p. 49. Svabhāvokti is read as non-vakrokti. 99 See V. SOVANI, History and Significance of Upama, ABORI, 1, p. 87. 100 ROBERT BURNS, 'A Red, Red Rose', stanza 1.

Rudrața 241

simile, and only because the contrast serves the single purpose of enlightening the proper contextual subject. In Burns' simile the relationship is not stated and must be guessed: bright as a rose, fragrant as a rose, joy-giving as a rose, and so on. Because the subjects of poetic simile are both same and different, this counter-factuality lends itself to much play and to much intentional varia- tion. Dandin's list of thirty odd varieties of simile are basically the changes rung on the propositional model by varying the temper of the speaker: let us speak as though identical, or if different, reversed, or if not identical, indistinguishable; let us propound the simile to the detriment of one or the other term, etc.101 These 'variations' of simile are when sufficiently conventionalized, con- sidered figures in their own right: rūpaka is the metaphorical identification of the two subjects (and the consequent deletion of the adverb "like" or its representative: " ... noon, the implacable bassoon" ... 102 It is clear that the four constituent elements of the simile (the two subjects, the relation of similitude, and the adverbial expressor) need not be explicit in every simile: but they can always be supplied, indeed are supplied in the full comprehension of the simile's import. Such observations link the so-called grammatical analysis of simile in the early writers (as Udbhata, followed by Mammata) with the discovery of the dhvani-theorists that many similes re- quire a "suggestion".103 Another important conventional variety of simile is utprekșā, wherein the deleted term is the upamāna, the non-contextual sub- ject; here the common property or mode of behaviour is so formulated as to recall that deleted term to mind when predicated of the true subject: so many idiomatic phrases are of this sort that it seems more a matter of colloquial language than of poetry: night falls (like a heavy object), sudden, decisive. Rudrata of course did not discover the relationships between these figures; the vocabulary of analysis is used by every prior writer. But he did indicate clearly that such considerations were basic in our understanding the nature of figuration. He has given a precise meaning to the ambiguous term vakrokti, showing the types of contrafactuality that indeed function in the major kinds of figure. And so it is important to inquire what is implicit in his other three categories of arthālamkāra. · The notion atisaya, exaggeration or hyperbole, appears to involve not the relation between two subjects, but the relation between one subject (perforce relevant) and its predicates; descriptive attributes or verbally designated

101 Cf. résumé of RENOU, IC, 1569, presenting the similitudinary content of various figures-but it is a good example of what Dandin has in mind for upama itself. 102 E.E. cummings. 103 See K. KRISHNAMOORTHY, Germs of the Theory of Dhvani, ABORI, 28, p. 190. The issue is crucial in the Dhvanyaloka's proof of the dhvani and recurs passim throughout the first three uddyotas.

242 E. Gerow . Indian Poetics

actions. The underlying 'form' both grammatical and logical is predication: A is B, A Bs (where B is an adjective or verb);104 the adjunction of simile is of a different sort, involving two subjects (and thus at least implicitly) two predications: it is apposition: inherently more complex than hyperbole, and it may also involve hyperbole. What makes this form poetic is in this case (and the name testifies to the 'deviation' here) quite clear: predication is the basic form of assertion itself: in many views of language,105 the independent variable of linguistic analysis. Its poetic version is signalled by its falsity. A "skyscraper so tall they had to put hinges on the two top stories so to let the moon go by".106 But of course it is not a falsity that is capable of being detected in the truth tables; for the falsity, the 'exaggeration' is only apparent, only for effect. For the lie in hyperbole is so framed as to conceal a greater truth, and a truth urgently required by the context: as these skyscrapers are the tallest things that man ever built ... The ability to formulate these distortions in such a way that they are self-defeating, indeed redound to the greater credit of the argument, reveals of course the poet's basic attitude to, command of, language, and his essential irony. Under hyperbole would fall many relationships between a subject and itself (or 'its property') that say something pertinent about the subject. The property is larger or smaller than in nature (hyperbole); the property may be separated from its subject, or be many when its subject is one; it may conceal its subject, or it may not be perceptible though its subject is. If the property be thought of as an effect (of the subject as cause) another range of distortions is made pos- sible that impinge on another category we have yet to discuss107. By treating śleşa as an arthālamkāra (as well as the śabdālamkāra) Rudrața takes a lead from Udbhata108 as well as Dandin, who placed his discussion of ślesa among the arthālamkāra, rather than in the sabdālamkāra of chapters I and III. But by generalizing it into a kind of alamkāra he means something more. It is accepted that if a pun involves no more than a span that can be read as two separate utterances (whether this be by redivision of the words109, or rereading the same words in different senses110) it is is a sabdālamkāra since its distinctive characteristic is based on a grammatical rather than a rational or propositional feature. Rudrata is apparently the first to observe explicitly

104 But this hyperbole can be confused with utpreka, for the exaggerated verb- predicate can easily suggest another subject. So Rudrata 8, 32ff., 9, 11 ff. 105 Again, James Harris, John Searle ("speech acts"). Mīmāmsā disagrees, of course: vidhi-the injunctive-is basic. Chomsky seems to agree, but is "NP-VP" an assertion? 106 CARL SANDBURG. 107 vastava, infra; see GEROW, G.I.F.S., p. 58-9 and Glossary for examples. 108 DE, H.S.P., II, p. 56. 109 Sun's rays meet / sons raise meat. 110 Don't labour under a misconception; support abortion reform. Dandin may have intended only the former type as a sabdālamkāra.

Rudrața 243

that there is a further dimension to pun that situates it among the true ar- thālamkāras, and indeed (as James Joyce has led us to rediscover) is close to the very form of literature itself: the division of slesa on morphological grounds assumes that the two meanings resultant are unrelated-that the "charm" of pun is entirely a function of the clever manipulation of grammar, and the unexpected bonus of a second meaning at the very time a first is perceived. But no author of substance uses puns in so lighthearted a manner: in 'serious' punning, the two meanings are closely related, in context, and in effect pro- pound a simile, except that a single language span is used. But the pun, par contre, is not simply a simile, though its words express both upameya and upamāna: the simultaneity of apprehension adds an unmistakable Glanz und Pracht to the figure which itself may be thought of (and no doubt was by Rudrata) as the exaggeration essential to hyperbole. The apprehension of the similitude is accompanied by a distortion so grave as to become hyperbolic. The distortion applies not to the similitude but to its apprehension. Rudrata has 'explained' the significance of the pun in classical kāvya: pun is both the ideal, or perfect figure (from the side of the śastra) for it achieves the maximum density of contextual and non-contextual rapprochment, and in it is realized the poetic difference (when viewed from the point of view of the expressible)-for here the "speaking slyly" (the vakrokti) is not a something superadded to a meaning otherwise comprehensible-it is inseparable from that comprehension: there would be no simile without the pun.111 The density of expression that is characteristic of the classical kāvya-the essential in- weaving of overtone and allusion-reaches a formal limit in the artha-slesa that makes of the latter the laksana of the style itself. The slesa is also an ideal by which the skill of the poet and the sensitivity of his audience are put to the test; having understood ślesa and its place in the structure of poetic utterance, it is not surprising that Rudrata-the quintessential ālamkārika -- turned away from the guņa rīti dialogue over 'style', as superficial to the study of poetry.112 Rudrata's purpose in establishing his fourth (in fact his first, chapter VII) category of figuration, vastava "factual" can be guessed from what has preceded. On the surface he has provided a rationale in which the long argument over

111 Cf. N. CHOMSKY, Syntactic Structures, p. 86, where "constructional ho- monymity" is advanced as a (or the) test of 'linguistic level', which is termed the "central notion in linguistic theory" (p. 11). Few indologists have accorded 'pun" the place of interest it deserves: RENOU, Art et Religion dans la Poésie Sanscrite: le jeu des mots et ses implications, Journal de Psychologie, 1951, pp. 280-5, KEITH, H.S.L., p. 351 (brief!); GEROW, On the Pun as Poetry, Journal of the G. Jha Kendriya Sanskrit Vidyapeetha, 27, pts 3-4, p. 79. Cf. also V. RAGHAVAN, Śleșa in Bāņa, NIA, 1, p. 214, for a more textual approach. 112 Classical Indian theorists, concerned with the inward force of the language itself could not reasonably be expected to dwell on accidental conventions of an already conventional language.

244 E. Gerow . Indian Poetics

svabhāvokti113 can be resolved. It is a figure (agreeing with Dandin) but of an essentially different sort than those captioned by vakrokti (agreeing with Bhamaha). The force of the fourfold classification leads us to define as vastava, figures that involve neither simile nor hyperbole (thus occupying the polar position to ślesa, which involves both). But how can one speak of a figure that appears to involve neither of the constituent elements of figuration? Have we expended our cleverness only to define a null category ? The answer is in Rudrata's propositional calculus that in fact underlies his treatment of simile and hyperbole, as mentioned. The propositional form underlying simile is association or apposition; that of hyperbole is predication. Both involve sentences of the type: A is X, and differ basically as to whether X is a noun or a qualification.114 But the relations of similitude (aupamyam) and gunagunibhava (relation between quality and object) do not exhaust the repertoire of propositionally utterable relationships: and whatever there may be among the latter that are capable of the peculiar distortion that makes for ; poetry, will have to be considered also poetic form, an alamkara. Among these "other relations" is cause and effect, and Rudrata concerns himself here with poetic versions of causation (hetu): More interesting are the figures of accu- mulation (as opposed to adjunction) (samuccaya "accumulation", parikara lit. "entourage", a figure in which the adjectival qualifications are multiplied; - et alia). Repeated references to the contextual associations of an event for example, 'suggest' the event. It is not clear whether Rudrata in these matters is directly responding to a dhvani-based theory, or whether out of such discus- sions may have sprung the problematic of the dhvani itself. It is also quite possible that Rudrata is considering such 'accumulations' as evidence of relations other than cause and effect, relations such as "inference", often based on a notion of cause and effect, and championed by Mahimabhatta as an alternative to accepting the dhvani as an independent function (vrtti). Among the vastava figures is Dandin's jati: where the accumulation is that of minute pictographic detail and the "suggestion" is that of a complex na- tural moment frozen in words)(as the hawk swooping on the piece of meat in the candala courtyard)115. Discussion whether such svabhavokti "charac- terization of a natural or typical individual" is poetic or not misses the point: the picture itself, though natural (but the understanding in any figure is 'natural') is a peculiar word-product: its frozen completeness a function of the epithetical density of the language, here far from (deviating from) imprecise descriptive prose. It is as though accuracy itself were 'unnatural'. It is Rudrata's accomplishment, not only to have provided a comprehensive and systematic account of the figures (De ii, p. 62), but to have shown the

113 See V. RAGHAVAN, History of Svabhavokti in Sanskrit Poetics, in Some Concepts ... , pp. 92 ff. 114 Including the verb, of course. 115 INGALLS, An Anthology ... , 1150, and discussion pp. 326-8.

Nāțyasāstra: Rasa in dramatic criticism 245

stringent conditioning imposed by the functions of language on poetic dis- course. It is this topic that is taken up, and developed separately by the dhvani theorists-and is a preoccupation that persists to the very close of the creative period of the sastra in the 18th century, a topic that in one way or another overshadows the discussion of the figures that gave it birth, throughout the subsequent history of poetics. At this point it will be necessary to sketch the development of the other poetics-the natyasastra-highly speculative though it is, up to the emergence of the rasa-dhvani theory, for, in terms of theory, the dhvani for the first time seeks to formulate a truly integrated view of both alamkārakāvya and rasa- nātya, via which, of course, the dhvani itself is presented (also for the first time) as a comprehensive successfully poetic principle.

7. Nātyaśāstra: Rasa in dramatic criticism

The term rasa is well attested in early Indian literature in the allied senses of "sap/essence" and "taste"-both notions are deeply imbued with over- tones of 'liquidity', inasmuch as both the sap, the 'essence' of plants, etc., is liquid, and the capacity to taste takes that peculiar form of 'liquidy' as its object (as, for example, in sugar cane), so that rasa is even seen in the physical theories as the very differentia of the liquid element ('having taste' marks the liquid as, e.g., 'having smell' marks earth and 'sound' marks ether (ākāsa)). Citations of the objective reference of rasa go back all the way to the Rgveda, where, not unexpectedly, it often designates the "essence" (scil. "power", in German "Kraft") of the soma plant; the subjective reference (to "taste") goes back at least to the Satapathabrāhmaņa (sarvesām rasānām jihvaikāyanam: e.g. 14,5, 4,11 in the Brhadāraņyaka). The question of when the specifically aesthetic flavour of rasa first appears in the literature is disputed. For rasa to have the dignity of a technical term-a precursor of its subsequent destiny-implies in some sense the development of a technical aesthetics in which the terms might function. The oldest work to concern itself (in some way) with the rasa as a definable aesthetic principle is the Nātyasāstra of Bharatamuni-a compendium on the theatre and the dance which is usually dated not later than the 6th cen- tury, but may contain elements as old as the 2nd B.C.116 It is thus roughly contemporaneous with the great flowering of dramatic and other literature under the patronage of the Gupta kings (4th-6th centuries), and it reflects the cultural and aesthetic realities of that flowering. Taken as a whole, the sketch of rasa in the Nātyasastra suggests strongly that the rasa developed its first 'aesthetic' overtones in the context of the

: 116 Supra, note 33.

246 E. Gerow . Indian Poetics

Sanskrit dramas of the classical period. It emerged first as a principle in debates about the nature and function of drama as a discrete genre. It was not at first propounded as a universal principle-at least the texts give us no leave to this speculative conclusion-but rather in some sense as a specific differentia of one well-established and highly valued genre. It is, in terms borrowed from Indian logic, introduced as the laksana117 of drama: an invar- iably concomitant attribute of the drama which thus serves (in the first place) to designate or mark drama apart from all else (as the dewlap does the genus cow). What else the rasa is, or may be conceived to be, probably originates in speculation about the reasons for that invariable relationship, and is thus inextricably linked to the nature of the drama as an art form of distinctive purposes and properties. The rasa does not begin its career either as a psycho- logical principle or even as an aesthetic principle-if by this we mean a uni- versal principle-but as a critical principle.118 As the effect of a drama is. as much a function of its complexity, as its unity, there is not one rasa, but eight: those emotional responses that are sufficiently universal to serve to organize an entire drama.119

117 Supra, note 38: though the term is used in different senses, its progression of meaning is clear from these examples. The 'logical' use is much more rigorous. 118 But while the drama declines as an art form, its principle in effect becomes an aesthetic of all art: 'art' is dramatized while drama loses its culturally distinctive character. 119 The eight rasas enumerated by Bharata are: śrngāra (amorous), hasya (comic), karuņa (pitiable), raudra (violent), vira (heroic), bhayanaka (terrifying), bibhatsa (disgusting), and adbhuta (wondrous), to which a ninth, sānta (peaceful) and others are added later. The literature on rasa is quite large; this notion is rightly thought to be both central and specific to the Indian theory of art. Most accounts of rasa however concentrate on its psychological or metaphysical import, rather than its strictly 'aesthetic' significance; a good portion of the history of Indian poetics is however devoted to explaining just how and in what contexts a notion as apparently innocent as "taste" can have such a functional overload. One such effort, that tries to emphasize the aesthetic issues underlying the multiplicity is my article Rasa as a Principle of Literary Criticism, to be published in the Proceedings of the Honolulu

J. BRANDON. Conference on Sanskrit Drama in Performance, eds. R. VAN METER BAUMER and

In my judgment the best single account of the rasa(s) in psychological and literary terms is still M. HIRIYANNA's essay Art Experience, published in a col- lection of essays of the same name, Mysore 1954. The standard texts on the Sanskrit drama all devote space to rasa, particularly in light of the metaphysical and theolo- gical overtones of Abhinavagupta's classic account of the rasa's origin and nature (infra, p. 264-8 and notes). In much of the literature it would seem that rasa is a single thing, an entity (following Abhinava's account of it as a principle, something of this is inevitable); the question of the multiplicity of rasas seems not to have excited much interest, though it is on this level that the aesthetic implications are clearest: each drama realizes one of the eight rasas-not "the" rasa. One such study is that of H.D. SHARMA, Hasya as a rasa in Sanskrit Rhetoric and Literature, ABORI, 22, p. 103. The question of multiplicity is raised chiefly around the new or ninth rasa, santa, reflecting a controversy patent in the texts themselves (in what

Nāțyaśāstra: Rasa in dramatic criticism 247

The hypothesis that we formulate is not original-though some of its applications may seem so. S.K. De asserts120 " ... Bharata's treatment would indicate that some system of Rasa, however undeveloped, or even a Rasa school, particularly in connexion with the drama, must have been in existence in his time." He spells out the implications of restricting the rasa's origin, by noting the existence of other critical theories or schools where the rasa was not cultivated: "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." The rasa should be understood then, in its earliest form, not only as an integrative principle, but as a distinctive feature, of the dramatic genre. That it occurred first in the context of the drama is a crucial, rather than an inciden- tal, factor in its definition. The rasa appears an external factor to the drama, seen as a work of art;, a medium of experience, emotional awareness, "taste" that is first and foremost in or of the audience; no similar need to go beyond the work is felt in the early poetics devoted to strophic poetry (alamkarasastra); (though with devotional lyricism, such a need did become an obvious enough addendum)121 instead the work is analyzed structurally in terms directly relevant to the dimensions of the language that underlies it: word as sound; word as meaning and intention; symbolic and metaphoric statements (images) that achieve truth through essentially non-literal means. The drama appears to require more than this, in the first place, because it is structurally so disparate, in effect, a composition of structures, or a complex structure. Each of the structures may well be- indeed is-analyzed in the course of the Nātyasastra's treatment: the strophic poetry which figures so prominently (but only as an element) in the drama is discussed in terms strikingly similar to its non-dramatic analysis in chapter XVI; the language of gesture (abhinaya), whereby the dance element be- comes immediately expressive, is minutely analyzed; costuming and stage arrangements (spectacle) are treated; seven chapters discuss music and song (28-34, of the 38 chapter text);122 other essential aspects of dramatic language

sense can "inaction" be "dramatic"?). Santa is needed precisely because the rasa. theory in time englobes other literary forms: the epics, esp. the Mahabharata, once they are determined to need a rasa, can have only this rasa. See V. RAGHAVAN, The Number of Rasas, Adyar 1940; J. MASSON and M. PATWARDHAN, Santarasa and Abhinavagupta's Philosophy of Aesthetics, Poona 1969; E. GEROW and A. AKLUJKAR, On Santarasa in.Sanskrit Poetics, JAOS, 92, p. 80. The dhvani the- ory, which we discuss in the next section, so intimately involves and reinterprets the rasa, that much of the literature will be more appropriately noted there. 120 DE, H.S.P., II, p. 18. 121 See GEROW, Indian Poetics, in Indian Literature, an Introduction. 122 RENOU, I.C., 1580.

248 E. Gerow . Indian Poetics

are treated: metre, dialectical variation, intonation (the figures are included in this part of the sastra (15-17)); more pertinently perhaps-from an Aristo- telian point of view-the questions of characterization and plot or "action" (itivrtta: 19). These aspects of the drama are exhaustively analyzed, sometimes with scholastic subtlety-as when the plot, seen as the attainment of a goal, is divided into five states (ārambhādi), which serve as basis for distinguishing their proper subjective representations (bijadi), which by various combina- tions give as many as 64 subjunctures (anga)-a veritable formal ossature beside which Aristotle's 'formalism' pales.123 The structure of Bharata's text suggests, as Renou124 and others have noted, a primary organization of the subject into four elemental topics āngika (or, representation by use of the body), sattvika (emotions and sentiments), vācika (use of language) and āhārya (the "external" aspects: costume, lighting, rouge, stage, etc.). Of these, the 'sentimental' involves chiefly the extraformal, emotional impact of the play, and of course, in terms of the audience's con- sciousness at the time of or as a result of, the play. Judging from the list of topics figuring in the 'complex' play, the rasa may be one among many-a fragment of a whole as are the rest (as the grouping into four suggests). And, indeed, the list of topics with which the sixth, or rasa-chapter begins, reads like a dictionary of terms, among which rasa has no special status, a mere component. But here, as often, the Indian passion for cataloguing reveals a deeper purpose. Bharata seems interested in the rasa-as component-pre- cisely because it offers a rationale for stating the unity of this complex form which is the play: an organizing mode which turns the elements, when properly perceived, into parts; and further, in virtue of that organic reevaluation, can be said to constitute the end, or purpose of the play. The end is not only a result, but a principle of organization which accepts, indeed requires, material to organize, but which more importantly, modifies that material in precise ways to conform to the expectation of the end as pur- pose. Thus the rasa is a mood, an emotional consciousness, wherein all the disparate elements of the play, language, gesture, imitations, scenery, coincide, and are understood after all not to be disparate;125 but also the rasa is the principle which accounts most successfully for the kind of reality that makes the parts dramatic. The elements, though closely related to those in the 'real' world (of concrete experience)-as the moon in the Vikramorvasiya to the

123 See S. LÉVI, Le Théâtre Indien, Paris 21963, pp. 30-62; M.C. BYRSKI, Sanskrit Drama as an Aggregate of Model Situations, Proceedings of the Honolulu Conference, I. 124 Ibid. 125 Hence the importance of the cooking analogy: NS I, pp. 287ff. (GOS. vol. 36). In the cooking process discrete, unrelated 'elements' are composed into a whole, conveying a single taste, which, as soon as it is tasted, is revealed to be the purpose of the process.

Nāțyaśāstra: Rasa in dramatic criticism 249

'real' moon are not simply fictitious, but are 'real' in a different sense-have a reality that only drama can sustain, that in turn defines its mode of distinctive being. Thus Bharata can say na hi rasad rte kas cid arthah pravartate126 meaning by artha the subject, or vastu of the drama-its congeries of elements-for the rasa is their end, in composite; but it also defines the mode of being of that composite tatra vibhāvānubhāvavyabhicārisamyogād rasanispattiḥ. 127 "The rasa exists or is produced from (we would prefer "as") a combination of (its elements, which are already conceived in their dramatically determined mode of existence, denoted by the technical terms) vibhavas "causes of emotion, e.g. the persons and circumstances represented", anubhavas "effects, conse- quences or external signs of emotions" and vyabhicāribhāvas "transitory states (of mind etc.)"".128 But also the rasa is not only a result, and the components are not only causes (as reflected in the popular derivation of bhava-'concrete emotion' "evam bhāvā bhāvayanti rasān ... "129 Bhava is also the first level apprehension of the concrete elements of the play: Rama as manifesting such and such a particular emotion in such and such a context. There is an element of reciprocal 'causation' in which the rasa itself can be said to be "cause" of the play: as the final cause explaining why it was put together in such and such a way130 the rasa is the organic 'root' of the total variety of the play; and its disciplined form, just as a single "essence" underlies the transformation of seed to tree to flower to fruit (to seed).131 Because the rasa is conceived as a mode of apprehension that is both imme- diate (in the theatre) and more general than verbal apprehension (for verbal apprehensions are one element only of the complex that is the play), its im- plicitude in every element of the play makes it quite inappropriately stated as the function of word imagery alone. That would collapse the evident generic distinction between poetry and drama at its most crucial point: the means of expression. The best that can be said of rasa in the context of verbal (poetic) expression is that it is an aspect or element thereof. And it is in this way that the other 'original' school of criticism has treated it: an as alamkara of speech:

126 Op. cit., p. 272 (prose between 6,31 and 32). 127 Ibid. This remark provides the occasion for Abhinava's extensive comment in the Bhāratī (supra, note 119, infra 264-8), translated by Gnoli (SOR, 11). For one of the rare views from the other side see B. SRĪNIVĀSABHATȚA SĀHITYASIROMAŅI, Rasopāsana (in Sanskrit), PAIOC 4, fasc. 2, p. 65. 128 Terms worthy of discussion in their own right; we note here only that what is intended is the congeries of discrete representations-both conditions (vibhāva) and consequences (anubhava) that the audience accepts as 'dramatic' and related in rasa. See my "Indian Poetics", op. cit .; DE, H.S.P., II, p. 20; P. REGNAUD, Rhé- torique Sanscrite, Paris 1884, pp. 266-356. 129 Op. cit. p. 293. 130 This argument seems to be understood by Abhinava as a pūrvapakșa; atra codyavādī svāšayam unmīlayati (ibid.). 131 Op. cit. 6,38; p. 294.

250 E. Gerow . Indian Poetics

rasavat.132 In the dramatic context, the ultimate rationale for the use of such and such a poetic verse is in the rasa of the play; but from the point of view of the verse, abstracted from the play (and many verses were not written in the context of a play at all) the critical concerns of image-analysis predominate (how the verse conveys its poetic meaning) and the rasa-element, if accounted at all, is necessarily reduced to a subordinate position, for the rasa is not the end of (verbal) utterance (as such). It may well, and often does, accompany utterance.133 Likewise, taking our reference in Aristotle's Poetics, because the rasa is conceived as an emotional apprehension implicit in every element of the play, and not a change, or modification of emotion (katharsis),134 it functions as an end and is a result in quite a different way: it does not 'happen' in the course of the play, but is there constant and immovable, subject only to apprehension. The plot structure which is judged suitable to realize that awareness is thus also quite different: instead of a concern with peripeties or turning points (dénouement) we find a concern for modes of sustenance, revivification through contrast, ultimate reintegration (phalapräpti). But these are matters not ger- mane, except perhaps as illustration, to our present concerns. As a critical principle, this rasabija we see sketched in the laconic text of Bharata functions in three important ways: it distinguishes nātya from the poetry of which natya is largely composed; it serves as an organic principle in terms of which the integrity of the drama can be understood; and is also the result of the drama, its "end", understood as a state of awareness peculiar to the drama-distinct from normal worldly consciousness.

8. Rasa and bhakti: the Dhvanyaloka

The second stage of theory concerning the term rasa is marked by the epochal work of Anandavardhana (9th century).135 We see a definite change of emphasis and a progression of concern from capacities of genre, to their ulti- mate reconciliation in a single theory of expression. It is noted on the one hand that though many aspects of the drama involve non-verbal expression

132 Supra, note 75. See M. HIRIYANNA, The problem of rasavad alamkāra PAIOC, 15, p. 267. 133 On the curious notion of the "appearance" of rasa, see S. BHATTACHARYA, Rasābhāsa in Sanskrit Literature, PAIOC, 7, p. 47. 134 See C. KUNHAN RAJA, Aristotle's Katharsis and Bharata's Sthāyībhāvas, Adyar Library Bulletin, 12, pp. 1-16. 135 DE, H.S.P., I, pp. 101-2; There were writers on rasa between Bharata and the Dhvanyaloka; none of their works have survived. But Abhinava gives an admirable résumé of their views in the course of developing his own doctrine: see infra, p. 264-8. T. S. NANDI, The origin and development of the theory of rasa and dhvani, Ahmedabad 1973.

Rasa and bhakti: the Dhvanyāloka 251

(as gestures, scenery, etc.), so too, in an important sense, does verbal expres- sion itself: that is, without trying to sound overtly paradoxical, the expressive goal (or "sense") of verbal art is often not the same thing as the expressive content of the words and sentences employed. In effect, the meticulous ana- lyses of poetic utterance and poetic imagination that were the business of the alamkāra school, seemed to come to precisely this conclusion: insofar as utterance can be considered 'poetic' (the genre distinct), it must involve, as an essential element, some such detour in the usual or normal apprehension process-a vakrokti: an understanding that the refined mind grasps as a result of the word heard or read, but not directly by means of it. We seem to be moving towards a view that generalizes art through a theory of expression (or perhaps more precisely, of "non-expression"). It remained for .Anandavardhana to draw these matters together. 136 But the theory does not develop in a vacuum. The interest in elaborating a theory whereby one can unify the arts (or at least drama and strophic poetry) should presuppose some strategic event occurring in the arts themselves which at least suggests the opportunity for such a rapprochement. It is not difficult to link what we know to have happened to Sanskrit letters in the post-Gupta and pre-Islamic period with the theoretical reevaluation that we find full-blown in Anandavardhana. With the collapse of the stable, wealthy, and sophisticated Gupta monarchy, and the time of troubles that ensued, Sanskrit drama lost both its audience (never more than the society of sophisticates attracted to the imperial courts and cities) and its means of support (what we presume to have been societies of wealthy literati-and of course the king himself).137 Local kingdoms (Harsa) rose to prominence in various regions, but the cultural stability on which this highly artificial art form depended was destroyed. The later reputation of Bhoja, King of Dhara, as liberal patron of science and the arts reads almost like a desperate and forlorn fantasy on how times ought to be-and once were (in the classical age)-but the 'reality' is now dependent on the whim of one man, in one minor court.138 Dramas continued to be written, but we presume, were rarely performed-which is to say in plain language that drama was deprived of its chief distinctive characteristics, and reduced to the status of a written art. The writing of drama, deprived of the medium in which the drama comes alive, came increasingly under the dominance of the poetic styles, where ornate metaphor, difficult language, and unplayable (non- dramatic) stories dominate. But oddly enough, as drama ceased to be a living

136 The vexed question as to whether the author of the karika portion of the Dhvanyaloka was also Anandavardhana is of almost no doctrinal significance, nor even of much historical significance-inasmuch as no trace of separate influence of the kārikā text is preserved. Abhinavagupta accepted different authorship; so does Kane, after a meticulous review of the arguments: H.S.P., pp. 162-199. 137 See B. UPĀDHYĀYA, Royal Patronage and Sanskrit Poetics, Poona Orien- talist, 1, fasc. 2, p. 13; S.N. DASGUPTA, H.S.L. Introduction, pp. c-cxvi. 138 Supra, note 7.

252 E. Gerow . Indian Poetics

art form, and was brought more and more into conformity with the rather academic style of the non-sophisticate pundit, its distinctive principle was being used as the basis for explaining the growing rapprochement of genres. 'Popular' forms of the drama survive shorn of their literary dimension: the various classical dance-forms that exploit the abhinaya of the drama. One ap- parent survival of the drama is the Kudiyattam of Kerala.139 On the side of strophic poetry, changes too were taking place, reflecting the new contexts in which poetry was used. We see growing devotionalism, and the employment of song, and also music, even marked shifts away from Sanskrit to various vernaculars. Poetry was becoming de-sophisticated (though poetry on classical lines continues to be written in an increasingly academic and un-'courtly' vein, paralleling the drama); its socially alive forms did move with the times, and came under the sway of social, inevitably popular religious, movements that were seeking new ways to express their needs. New metres, showing the impact of regular musical beats, rhyming on 'modern' lines emphasizing simple couplet phrasing so typical of song, crept into Sanskrit from the popular levels (Ardhamagadhi dohas of the 8th-9th century Jain canon) achieving a complete, and for later generations, standard formulation in the one poem of Jayadeva, the Gitagovinda (12th century). 140) The question of the genres is posed anew with all its vigour, and it is to Anandavardhana's credit that he not only reevaluated poetics, but redefined its subject matter. The principle of non-literal intention (common to both dramatic devices and to verbal means) is found to be the third, and most characteristic function of language-dhvani, or "suggestion"; and that subject matter which dhvani is suited to express par excellence, is rasa. So poetry, seen as expressor, has won the status-not so much of one means of expression among many,-but is that means wherein the capacity of language to express is most fully, multi-valently realized; and the integrative quality of the drama, since it can never be expressed, literally, but only suggested (expressed non- denotatively), becomes the proper or typical content of that poetic capacity. The genres are downgraded to "external" classifications, and at the same time, essentially non-verbal arts, such as music and painting, can equally well be considered art, since their distinctiveness is also a matter of externals (sound, colour). The importance of Anandavardhana's Dhvanyaloka can be understood in three ways:141 (a) in relation to alamkārasāstra, it offers an explanation of the

139 See CLIFFORD R. JONES, Literary and Epigraphical References to the Ca- kyar-s, Custodians of Sanskrit Drama in the Temples of Kerala, Proceedings of the Honolulu Conference, I. 140 See A.D. MUKHERJI, Lyric Metres in Jayadeva's Gitagovinda, JASB, 1967, p. 232. 141 K. Krishnamoorthy, to whom we owe the only English translation of the Dhvanyaloka (Poona Oriental Series 92, 1955), has also devoted much effort to the theory of dhvani. See his Essays in Sanskrit Criticism, Dharwar 1964, and A critical estimate of Anandavardhana's Dhvanyāloka, NIA, 8, p. 194.

Rasa and bhakti: the Dhvanyaloka 253

vague notion of vakrokti (Bhamaha) founded on a general theory of signification (rather than on a survey of the types and contexts of figuration)-it thus ties poetics into the concerns of Indian linguistic:speculation; (b) in relation to the natyasastra, the Dhvanyaloka offers an explicit, expressionistic, account of the rasa in poetry, for the first time going beyond the impressionistic analogies); founded upon the largely non-verbal context and technique of the drama;142 and (c) it links the notions of rasa and alamkāra in a single coherent, if syn- thetic, theory of literary aesthetics. It is the latter contribution that is most apparent to the historian: after the Dhvanyaloka, the tradition of Indian poetics appears by and large to be a single tradition, with authors either confronting, or revising the basic pro- positions of the Dhvanyaloka. The Dhvanyaloka proper, a work of the late 9th century, is probably only the commentary on the karika verses, which are usually assigned (even by the Indian tradition, including Abhinavagupta) to a nameless Dhvanikāra, who may then be placed a century or two earlier. Controversy still rages on this point and there can be no settled conclusion.143 Despite the weight of evidence favouring two authors, it is odd that the karika text, if so much earlier than the vrtti, received no notice whatever in any author before Anandavardhana. Like many of these arguments, the issue is purely chronological, and does not in any clear way influence our interpretation of the whole text, which we henceforth treat as one. The Dhvanyaloka is something of a tour de force. It solves a range of problems that had arisen in certain lines of speculation by adducing principles borrowed or adapted from others: however syncretistic, its achievement must be traced not only to this happy confluence of principle, but to its recognition of the changes in the poetic context that appeared to demand a new type of speculation. Put briefly, the linking of vakrokti and theories of signification (Mīmāmsa chiefly) takes place via the notion of dhvani,144 lit. "sound" or "echo" (often a synonym of sabda), here interpreted as an expressive function inheront in

H.S.P., II, ch. 5. 142 See A. SANKARAN, The Theories of Rasa and Dhvani, Madras, 1929; also DE,

143 Supra, note 136, and K.A. SANKARAN, The Authorship of the Dhvanikārikās, PAIOC, 3, p. 85 (Anandavardhana); S. BHATTACHARYA, Dhvanyāloka and the Text of the Dhvanikārikās PAIOC, 6, p. 613; K.G. VARMA, Different Authorship of the Kārikagrantha and the Vrttigrantha of Dhvanyāloka, NIA, 5, p. 265; H. JA- COBI, Introduction to his excellent German translation of Dhvanyāloka, ZDMG 1902, p. 405ff .; K.C. PANDEY, Abhinavagupta, 2Varanasi 1963, pp. 202ff. (accepts identity of authorship); K. KRISHNAMOORTHY, Authorship of Dhvanyāloka and Anandavardhana's date and works, IHQ, 24, p. 180, 300. 144 L. RENOU, Le dhvani dans la poétique Sanskrite, Adyar Library Bulletin, 18, p. 6; K. KRISHNAMOORTHY, Germs of the Theory of Dhvani, ABORI, 28, p. 190; K. KUNJUNNI RAJA, Theory of Suggestion in Indian Semantics, ALB, 19, p. 20.

254 E. Gerow . Indian Poetics

language and different from other recognized functions (vrttis) (chapter I);145 the rasa is linked to poetic diction via a meticulously worked out set of exam- ples that show the rasa expressively functional in every discrete stratum of linguistic analysis, from phoneme (varna) to argument (oratio, prabandha) (chapter III). The important, if understated, points in this exposé, are two: the rasa is in all language universally, and not merely in terms of the speaker's intention or the hearer's wit; and also, that the rasa, unlike any other postu- lated expressive content (artha), runs through the entire gamut of linguistic discreteness: unlike "intention" (tātparya) it may be determined in a single phoneme (though the examples really involve morpheme ambiguity, as the -au is the locative singular ending of both i- and u- stem nouns); unlike "deno- tation" (abhidhā) it may be determined in strings of propositions (mahāvākya, or prabandha)-finessing even the Prabhākara view that a single vākya is prior in understanding to its component "words". The core of the Dhvanyaloka is its second uddyota (chapter II) in which the sub-types of this novel expressive function, dhvani, are discriminated on the basis of its nature (vis-à-vis the other expressive functions or vrttis (abhidhā, lakşaņā, and according to some gauņi and tātparya)146) rather than its linguistic tokens. By "nature" of course is centrally involved an appeal to meaning (artha), but since in one sense the "meaning" of any function is the same, that sense being content per se (as my "love" for Juliet can be expressed literally, or by indirection, or by suggestion without modifying the fact), the distinctive _quality of "meaning" as competent to characterize dhvani and only dhvani will have to do with its mode of apprehension. It is this issue (captioned (sabdabodha in the later treatises) that situates the poetic problem crucially-in an area distinguished on the one hand from the linguistic token of expression (sabda), viewed as means in the instrumental sense, and on the other from content, or artha in the external sense. The Dhvanyāloka, in defining a new area for poetic speculation, has really done little more than reinterpret a long-stand- ing concern of the Mimamsa; it is in this area, where the Vedic injunctions, the optative or quasi-optative tokens are synthesized according to principles of co- herence and intention, into mahavakyas competent to refer to the gross content of the Veda, namely the obligatory sacrifices and other ritual acts of the Vedic tradition, that the peculiar and indeed irreplaceable function of the Mimāmsa is to be located.147 How the mode of apprehending a dhvani-based meaning differs from that of another vrtti is largely explained in terms of the other vrttis, which are taken for granted. Only two primary expressive powers are universally re-

145 L. RENOU, On primary and secondary meanings in Sanskrit (in Hindi), Hindianusilan, special number, Prayāg, pp. 298-304; K.K. RAJA, Indian Theories of Meaning, Madras 1963. 146 K.K. RAJA, op. cit. 147 See GEROW, G.I.F.S., pp. 48-50. The first problem that Mimāmsa has is justifying its own reasonableness: adhikaraņa 1: svādhyāyo 'dhyetavyah.

Rasa and bhakti: the Dhvanyaloka 255

cognized in the word; one of course is denotation (abhidhā) ex hypothesi, the other covers that range of usage where there is comprehension (ergo a power operating) in the presence of an explicit malfunction of denotation: secondary or metaphorical signification: lakșanā, as in the standard example "the grand- stands are cheering", which joins an inanimate subject with an animate-only action, or the older Mimamsaka problem "the rocks are swimming; the trees attend on the sacrifice". These examples are interpreted in various ways; it will be obvious that laksaņā presumes abhidhā, in principle (for it is its failure that occasions the problem), and that also in fact abhidha is used, for "grand- stands", "trees", "rocks" etc. are all understood in rather the way Aristotle defined metaphor, as tatsambandhi-, something satisfying logical and denota- tive requirements having a relationship with the thing named: people in the grandstands; the water is moving so fast that the rocks appear to be moving upstream; so enchanting is the sacrifice that even trees (though inanimate) appear to be following it. Many kinds of relation are possible: connection, opposition, or (as the third example shows) attribution of contrary or similar qualities. An a fortiori argument (arthapatti) appears to be involved. The dhvani, in function and terms, is the other meaning that may be attested along with one or the other preceding types; but since it is other, it cannot be explained as having come to be through the same process as the base type, but must in fact involve the base type in its mode of apprehension. The dhvani then is a kind of 'tertiary' apprehension compatible both with the presence (vivakşitavācya) and the absence (avivakșitavācya) of denotation (this alone suggests that dhvani, if the genus is admitted, cannot be identified with lakșaņā). The former is 'based on' abhidhā; the latter, on lakşaņā (1,13 Comm.). How this other meaning is possible is the issue raised in chapter I of the Dhvanyaloka; its resolution is both traditional148 and philosophical149; that possibility, in chapter II, is grounded on a theory of the expressive function that links the domains of kāvya-viz. alamkāras, with nātya-and its differentiae, the rasas. A dhvani or suggested meaning based on but different from laksana (laksyartha) is either derived from the laksitartha (the case of sarcasm, for example) or is compatible only with its opposite (irony). The type of dhvani that is not incompatible with and hence founded on the primary expressive function (abhidha) is the key both to the understanding of the Dhvanyaloka and to an appreciation of its unique syncretism. Its two varieties are discussed in the bulk of chapter II: 2,2-19; 20-30 The inescapable problem in positing a verbal function called "suggestion" consists in demon- strating that the further meanings we apprehend (and indubitably do) in poetic (and certain other) utterances are uniquely dependent on verbal oper- ations (viz. grammatical) rather than on some other-levelled mental function, such as logic or dialectic.

148 As present in any kāvya: 1, 13ff. 149 As implied by former theories of poetry: 1, 1 ff.

256 E. Gerow . Indian Poetics

Psychologically the problem is quite the reverse, fo it appears that this 'poetic' domain is far more natural to language than the scientific and univocal uses that in high culture forms assume more importance. Both Historically and in terms of everyday usage (where practical concerns are not paramount-as in the banter of working-men) poetry (Homer, Veda) seems prior, and it is deno- tation (abhidha) that later is 'discovered' The needs of explicitude are perhaps first felt as traditional conflict resolving systems are raised to the level of self-conscious legal systems (Solon, Manu). 150 The contention that one of the natural functions of language is associated with poetic and non-declaratory utterance in effect resolves the historical- scientific paradox, but it remains a mere contention unless it can be shown that there is an essentially different mode of apprehension associated with 'poetic' as opposed to 'scientific' meaning. Instead of an abstruse discursive argument the Dhvanyaloka offers here one of the boldest reductions in the history of Indian aesthetics, one that in its concrete simplicity fuses the heretofore separate 'schools' of kavya and natya poetics. The best way to show that there was a difference in manner inherent in poetic comprehension would be to show that there is a distinctive kind of meaning otherwise inapprehensible. Instead of "manner" based on the presumption that all concrete "meaning" (or content) is indifferent to the modes of its apprehension (a common sense view that holds the univocal proposition "I love her" not different in content from its poetic counterpart, the poem), we argue content; and it is asserted that the one content that can never be apprehended through a declarative or denotative utterance is precisely that that had heretofore been reserved to the study of natya as its differentia the emotional response to the work of art, the rasa. The rasa is an affect in the psychological sense, is therefore a content only in a secondary sense, as requiring a substantive or primary content (actions, character, words) as the basis of response; it cannot therefore be a part of that basis in any real sense. While I can say "love", I evoke no sense of "loving" by so doing; that takes the play: situation, excitant, consequence: a sense of the ineluctible human presence. The content, so to speak, of drama, is integrated into poetic as establishing a fortiori the mode of apprehension most proper to poetry. The affective response is thus linked with the non-denotative utterance: not accidentally, as an attribute of a content itself poetic151, but essentially, and simultaneously with the content that is apprehended through other modes and thus is not (as such) poetic. The co-existence of poetic (dhvani) modes of apprehension with others (abhidha etc.) thus becomes not an obstacle to the theory, but & token of the complexity of language and of the paradigmatic essential character of poetry itself.

150 An idea suggested by my colleague, James Redfield. 151 rasavad alamkāra.

Rasa and bhakti: the Dhvanyaloka 257

What happens to drama in this reduction? Is it meant that drama becomes the "best" poetry ? We noted that the Sanskrit drama, shorn of its sophisticated patronage, its stable milieu ravaged by frequent social upheavals, became increasingly a literary form. Such a genre-reduction is not intended by Ānan- davardhana; for him at least, it is a form of kavya that deserves the accolade "best"-that form wherein the dhvani-function is itself the chief object of interest, where it-by a not unambiguous application of the Mimamsa notion of "primacy" prādhānyam)-may be said to be the "main" or chiefly intended mode of apprehension involved. One must speak of "primacy", for given the analysis of speech, the dhvani cannot come unalloyed by other functions. Where the dhvani as function is primary, its most characteristic content, then, the rasa, is by definition what the best poetry is about; since the content, in this case, is proper to the function, its apprehension is simultaneous, instan- taneous (asamlaksyakrama [2,3]); but because rasa is an affect it is always accompanied by a content that can properly be said in some sense to be indicated (that is to say a content in the usual sense, a "meaning"). While not asserting that this true content overwhelms the dhvani (the case was covered under the first division of suggestion, where the meanings of the two functions are separable and dhvani is subservient), the theory nevertheless allows that this prior content (usually either a state of affairs, vastu, or a figure of speech, alamkāra) be apprehended also, which is to say, not at the same moment (samlakşyakrama [2,12]) as the primary dhvani, and hence inevitably, as a function of the dhvani. Thanks to this proviso the character of the utterance as dhvani-primary is not affected, and the richness of the resultant poetry is enhanced by reconciling the dhvani with the multifarious content of poetry. The theory of tripartite verbal function (abhidhā, laksaņā, dhvani) would if pressed appear to terminate in the paradox that the best poetry (where dhvani is the function and rasa the "content") cannot have any prior content at all, or at best will have a merelv adventitious content (in the sense that an affect needs logically some basis). While this would neatly turn the original view of rasa in poetry on its head (where the rasa was the adventitious element: rasavat alamkāra), it does not satisfy the justified common-sense notion that the prior content (vibhāvas and anubhāvas in nātya; vastu and alamkāra in kavya) is somehow crucially related to the poetic message and effect. The theory of Änandavardhana, by espousing a grammatical doctrine of significa- tion, and by defining the many-levelled structure of meaning in language, is able in this most crucial "type" of poetry-rasa-dhvani-to relate affect and content directly, as optimally functioning language. The denotative level persists and is not cancelled; a further "content" is suggested via that deno- tation, which suggestion turns out to promote primarily the stable rasa as well. The Dhvanyaloka appears to bring kavya under the aegis of the dramatic rasa theory; but in fact it does the reverse: rasa is reduced to a peculiar kind of content in an overarching theory of poetry, a poetry that ipso facto is

258 E. Gerow . Indian Poetics

dramatic poetry. Culturally, this development coincides not only with the collapse of natya as a performance-medium, but with the emergence of new styles of poetry, sung and danced, that are still associated with the devotional cults of medieval theistic Hinduism. By offering an intellectually serious defence of dramatic poetry, the Dhvanyaloka bridges the gap between classical and medieval; terminates the former and becomes the basis for later specu- lation. The key to this revision, reflecting the change in content and style of the poetry, is the new concern for capacities of language (sakti, vrtti). By adapta- tion of the old grammatical doctrine of abhidhā and laksanā,152 Anandhavar- dhana was able to relocate the poetic problem in the widest context of lan- guage use-to see poetry as language itself-rather than in the circumscribed area of forms of assertion (arthalamkara). Though the basic hypothesis of the two approaches is not all that different, involving a preference for non-deno- tative utterance, the former leads us in the direction of emotionally dominant poetry (rasa and dramatic poetry) while the latter seems unable to transcend the probably too sophisticated notion that slesa is the summum of poetic ex- pression .! In any case, speculation on the vrttis of language becomes a central concern of later poeticians-beginning with two writers who denied the func- tion dhvani itself: Kuntaka and Mahimabhatta.

• 9 After the Dhvanyāloka

Indian poetics, after the Dhvanyāloka, changes its focus substantially without radically modifying the structure of its problematic. Instead of 'schools' of poetics, wherein we find addressed the problems of aesthetic comprehension of distinct genres of expression (kāvya, nātya), we are con- fronted with a set of discrete problems that every author, in some way, at- tempts to resolve. A certain amount of syncretism is evident; yet in the main, the basic intellectual antitheses of the pre-dhvani poetics (if we include the purāņas) continue remarkably vigorous. (a) The interest in alamkāra, or briefly, in the formal expressive repertoire of the poet, as practitioner of a different and uniquely intelligible kind of language, continues through a group of writers who not only dwell on figura- tion to some extent (Appaya Dīksita, Jagannatha) but are greatly concerned

152 One of the earliest works on this subject, the Abhidhāvrttimātrkā, of Mu- kula (teacher of Pratīhārendurāja, q. v.), was in all likelihood known to Ānanda- vardhana (DE, H.S.P., I, p. 74). It influenced later writers, including Mammata, who wrote a similar work (op. cit., p. 148). By this time it already seems that poetics is a subject reserved to Kashmir. See S.P. BHATTACHARYA, Impress of Kasmiri aivism on Alamkarasastra, Journal of Or. Inst., Baroda, 1, p. 245.

After the Dhvanyāloka 259

at identifying in grammatical and logical terms that dimension of language that makes figuration possible (Mahima, Kuntaka). Though the dhvani as a function is not universally accepted by this group, all the writers confront the issue of grammatical or syntactic function (vrtti) and attempt to define "poetry" in non-emotional (rasa) terms. (b) The neo-rasa psychologists. The rasa likewise survives-in a form not restricted to genre analysis (the Dhvanyāloka seems effectively to have ter- minated this type of reasoning in India)-but rather is generalized as a prin- ciple of appreciation by a brilliant series of writers-Abhinavagupta (commen- tator on both the Dhvanyāloka and the Nātyaśāstra (Locana, Abhinava- bhāratī)), Sāradātanaya, Rūpagosvāmin and others who recognize poetry as a kind of meaning carrier with a unique effect on the receptive mind (sahrdaya). The concern is not so much with expressive device as with the whole in which the poem makes sense: these writers explore links not with grammar and mīmāmsa exegetics, but with reality theories and the prescriptive morality of the Indian socio-philosophies, especially Vedānta. (c) The pre-dhvani 'puranic' tradition (if it may be so dignified)-recognized in the compilations of the Visnudharmottara and Agni Puranas-also con- tinues in a much more sophisticated vein in the works of Rajasekhara and Bhojarāja-works so bizarre that they are difficult to include in the history of any subject, especially poetics. They seem to be inclusive encyclopaedias of poetic lore without any overall attempt at coherence, or even organization. They deal with many topics not considered by more conservative writers, and Bhoja especially testifies to a kind of reckless aggrandizement of the subject that is nevertheless of extraordinary value for its novel approach and com- mentary on other (often non-poetic) subjects. Though huge, these works can also be breathtakingly reductivist-witness Bhoja's effort to reduce all rasas to one: śrngāra, love. (d) A large number of works remain that do not easily fall into these groups: works that are encyclopaedic, yet propound at least verbally, a theory of the unity of artistic expression-often rasa. These works by subject are most clearly related to the pre-dhvani poetics-treating of the entire range of previously defined issues: dosa, guņa, alamkāra (always a major topic), dhvani and rasa-even riti. The vrttis also figure prominently. It is of course not their subjects that make these works novel, it is their point of view. The deservedly famous Kāvyaprakāśa of Mammata both originates this style of poetic, and serves as a model for the later practitioners (though with considerable varia- tion): Ruyyaka, Hemacandra, the Vāgbhatas, Viśvanātha. It would(not be correct to designate these strands of post-dhvani poetic speculation as 'schools', for the emphases presume an interwoven content that forms a common background for most authors. Concern for alamkāra is implied both by an interest in vrtti and in the systematic writers; rasa is analyzed both by the psychologists and pauranikas-and included within the scope of the systematists.)

260 E. Gerow . Indian Poetics

The lines of inquiry that we have classified in a-c above emerged from a period of intense speculation of nearly two centuries that followed the writing of the Dhvanyāloka. The works were often of men who not only were con- temporaries but who knew each other and reacted to each other. Though each problematic finds its adepts in later centuries as well (Jagannātha, Rūpa, etc.), the issues were at least defined in this early seminal period of response: Rājaśekhara, Abhinavagupta, Kuntaka, Bhoja, Dhanamjaya, Mahimabhatta and others confronted the import and growing fame of the Dhvanyāloka. It is possible to see our group '(d); as the (inevitable?) systematists' response to. this period of fermentation just described. Other works remain that seem less "poetic" than pedagogical: inventories of figuresand other technical terminology (Candraloka of Jayadeva) or frankly practical advice to the academic poett the kavisiksa works typified by the Aucityavicāracarcā of Kșemendra, The former have often achieved great po- pularity, especially when commented upon Kuvalayānanda of Appayadīkșita); the latter owe much to the atypical Kāvyamīmāmsā of Rājaśekhara. 153

10. Rājaśekhara

The Kavyamīmamsa154 is the first datable work to follow the Dhvanyāloka. No more striking contrast is found in the history of Indian poetics. Many writers express doubt that the Kāvyamīmāmsa is properly considered a 'poetic' work at all155; the fame of the author, and what is rarer, the definite historical personality he has acquired, make it impossible to exclude him. The work is in the puranic tradition of poetic speculation, and marks the emergence of Agni-Purana-like anecdotalism into respectable academic format; though Bhoja is not uniquely indebted to Rājasekhara, he may be said to have carried the style-format to its 'logical' (scil. anti-logical) conclusion. Rājaśekhara, like other authors, is copied extensively by Hemacandra156. Rajasekhara, who quotes Anandavardhana, is himself cited in the Yaasti- laka, dated about 960 A.D .; his literary activity is thus quite tightly circum- scribed to the first half of the tenth century and probably to be associated with Yuvaraja I of the Kalacuri dynasty in Cedi. He is known chiefly as a poet, author of four extant dramas, including the Prākrta Karpūramañjarī. His command of Präkrit is unexcelled; he seems also to have developed consider-

153 DE, H.S.P., II, p. 291. 154 See V. RAGHAVAN, Rājasekhara and his Kāvyamīmāmsā, Journal of Oriental Research, Madras, 37, p. 1ff .; also the excellent French translation (with Intro- duction) by L. RENOU and N. STCHOUPAK (Cahiers de la Société Asiatique, 8, 1946); résumé KANE, H.S.P., pp. 209-10. 155 E.g. DE, H.S.P., II, p. 297. 156 DE, H.S.P., I, p. 118.

Rājaśekhara 261

able interest in the circumstantial variety of the India of his time-a trait he shares with Kalidasa: his knowledge of geography is remarkable. Furthermore, he seems to have been quite interested in himself, preserving more biographical information in his works than is normal; and finally, we might say, his crest- jewel was his wife, the famous Avantisundari, whose poetry he cites with respect)157 The Kavyamimamsa, though an extensive work of 18 chapters158, does not deal directly with those topics that had come to be considered central to the poetic problem (gunas, alamkaras or rasas)-the problem of kavya-rather it is chiefly devoted to the kavi, or poet himself, which leads De to consider it a work simply of kavisiksa159, a work on the education or training of the poet. While the contrast is inevitable, the Kāvyamīmāmsa is far more than just 'advice' to the aspiring poet160. The Kāvyamīmāmsā is that rarissimum opus in the Indian tradition that concerns itself with the circumstances of an activity rather than with the product of an activity; given the normative thing-to-be- created (it seems to say) what kinds of arrangements are conducive to that end? From daily schedules to the appropriateness of plagiarism (also dealt with in Dhvanyaloka, 4th uddyota161), it measures the author against the presumed standard and result of his work. This work is concerned then with the prerequisites of poetry, rather than with poetry per se, and provides us a valuable insight into the Indian version of a distinction that much vexes us: the psychology of individual poetic activity. But far from seeing the work emerge from a 'free' play of independent subjectivity, the Indian 'psycholo- gist', to whom the work is given in a set of normative expectations, proceeds to tie down the 'subjective' component in an almost Pavlovian manner. From this point of view, Rajasekhara is an absolutely essential contributor to the formulation of the Indian poetic tradition, more valuable for his rarity and iconoclasm. In effect, he develops into an entire work that usual and conven- tional prolegomenon of sastra wherein the purposes (money, fame, etc.) and prerequisites (pratibhā, vyutpatti, abhyāsa) of poetry are statedt62 Pratīhārendurāja, ca. 925-950: Though known only for his commentary on. Udbhata, this writer seems to have held independent views on the role of rasa, and to have been an important teacher in the Kaśmira tradition.163

157 KM. (GOS 1), p. 16; KANE, H.S.P., p. 211-6. 158 As are most purānas! 159 II, pp. 291 ff. 160 DE, S.P.S.A., p. 76. 161 A key problem in a tradition where inspiration is less of a desideratum (though necessary) than acceptance of a norm (unavoidable). See K. KRISHNA- MOORTHY, Anandavardhana's treatment of Pratibha in relation to Dhvani, ABORI, 31, p. 143. 162 Kāvyaprakāśa 1 inter alia. 163 Though usually distinguished from the roughly contemporary Bhattendurāja, Abhinava's guru: KANE, H.S.P., p. 205; DE, H.S.P.,I., p. 74; contra, K.C. PAN- DEY, Abhinavagupta, pp. 74, 142.

262 E. Gerow . Indian Poetics

Himself a student of Mukula, the first writer on the vrttis164, he had for a long time the distinction of having his view on the preeminence of rasa incorporated into the text of Udbhata's Kāvyālaņkārasārasamgraha165. He was apparently one of the first writers to attempt to reformulate or reapply the conclusion of the dhvani theory (which he fully accepts) to the older poetics, specifically the view of rasaprādhanyam. Several other ideas expressed in the Commentary identify him as a novel, and important writer, but quite unreliable as an interpreter of Udbhata. He seems to have migrated from his native Konkana to Kaśmīra to learn the alamkāra śāstra, illustrating the central place of Kashmir in the speculation of this period. 166

11. Kuntaka and Dhanamjaya

Kuntaka. Kuntaka makes no direct reference to the doctrine of the Dhvanyā- loka. His work is one of the most independently conceived in the history of Indian poetics. It may be considered not so much a reply to the dhvani theory (as Mahimabhatta's) as an independent effort to deal with the same problem: the grammatical basis of non-literal utterance; Kuntaka's solution, though unique, is considerably less radical than that of the Dhvanyaloka, indeed seems to summarize admirably the presumptions of the early alamkara school: taking the term vakrokti "evasive speech" from Bhamaha, Kuntaka elaborates the notion of non-literal utterance in two parallel ways that in effect generalize the claims of the alamkarika theorists without limiting the argument to alamkaras only. In effect, the alamkara method is here applied to the entire range of poetic concepts, including rasa.167: In the first place, Kuntaka links the vakrokti quite carefully to its effect (in this way seeming to respond to a rasa-concern) on the audience, using such terms as vicchitti, vaicitryam, making clear that a "striking or charming apprehension" is a principle that must qualify the bare figure as non-literal speech: he thus 'solves' the problem that much taxes the older ālamkārikas: for figures have many non-poetic exemplifications and even poetic ones can become stylized and quasi-literal ("he hit the nail on the head").168 In the other direction, Kuntaka finds the notion of 'figure' too narrow a base for a gram- matical explanation of vakrokti. Figures, in the older school were, if artha, propositions of a certain sort, and thus fell under the grammatical category of

164 Supra, note 152. 165 Ed. M.R. TELANG, Kāvyamālā, 1905, p. 77; cf. DE, H.S.P., II, p. 59; I. p. 77-8. 166 DE, H.S.P., II, p. 58. 167 De's "Introduction" to his edition of the Vakroktijivita is essential, esp. pp. xv-lxi (third edition, Calcutta 1961). 168 The idea is adumbrated at Dhvanyaloka 1, 16.

Kuntaka and Dhanamjaya 263

the sentence (vākya); if śābda, they were treated primarily as unnatural regu- larities in phoneme (varna) recurrence or feature recurrence. Only a few figures, for example, could be said to involve the morpheme (lātānuprāsa, a variety of alliteration). Using the six-fold Pāņinean cum Mīmāmsaka analysis of utterance into levels or constituents: varņa (phoneme), padapūrvārdha (or prakti, stem), pratyaya (affix), vākya (assertion, sentence), prakaraņa (topic) and prabandha (work or composition), he demonstrates the vakrokti in every conceivable dimension of utterance.169 His quarrel with the ālamkārikas is thus not one of principle, but one of generality. He continues to deal with vakrokti in contextual terms (audience, vidagdha ("clever, knowing"); and grammatical scope), and does not confront the basic question of expressive capacity itself: vrtti. In the Dhvanyaloka this second issue is faced squarely, and despite its elegance and sensitivity and exemplification, Kuntaka's work was judged uninteresting by the later tradition170. Dhanamjaya (924-996). Just as Kuntaka is in one way the end of the older alamkāra tradition and at the beginning of the neo-alamkāra speculation, this writer seems to occupy an analogous place in the older natya-centered rasa tradition. His Daśarūpa(ka) is an admirable summary of just those elements of the informe Natyasāstra that concern the form and aesthetic purpose of the drama. It has for all later writers replaced the sastra, It deals, in four chapters, with the topics (broadly speaking) body or plot (itivrtta, anga), character (nāyaka etc.), genre (the ten types of rūpaka, nātaka etc.) and aesthetic response (rasa, bhava etc.). One is aware of both what the Daśarupa does, and what it does not. Missing is the entire range of problems having to do with production: abhinaya, elocution (alamkāra), staging-establishing rather positively the sense in which the dramatic art had become kavya by the 10th century. Likewise, Dhanamjaya seems uninfluenced by the specifics of the dhvani-theory (postulating dhvani as a function to unite kāvya and nātya), while being quite aware of the present improbability of separating the genres. He rather accepts Bhattanayaka's bhāvakatva as the function whereby rasa is conveyed in kāvya (according to his commentator Dhanika, a close contem- porary171being a kind of surrogate to the prima facie functionalism of the Nātyaśāstra per se, whereby vibhāvas etc. (the visible actions and scenes of the play) provoke the sense of rasa in the audience.172 In other words, the

169 Cf. Rudrața's treatment of the sabdālamkāra, supra. 170 Recovered only recently by De (supra, note 167); Kuntaka was probably a contemporary of Abhinavagupta, whose Locana on Dhvanyāloka was instrumen- tal in promoting the dhvani theory, to the eclipse of other points of view (KANE, H.S.P., pp. 235-6; DE, H.S.P., I, p. 128). Both were likely Kāśmīra (the title Rajanaka), but few scholars see evidence that they knew each other's works. 171 DE, H.S.P., I, p. 123-5; probably Dhanamjaya's younger brother, and not as Jacobi thought (Göttingische Gelehrte Anzeigen, 1913, p. 303) a mere pseudo- nym. His commentary ad 4,37 is at issue here. 172 Ibid. ad 4,36.

264 E. Gerow . Indian Poetics

problem that the Dhvanikara presumed to deal with is not yet in focus: the need to formulate a comprehensive principle of expression for nātya and kāvya was not admitted)173 Dhanamjaya has some distinction in not being from Kasmira, but from central India174 at the court of Paramara Muñja.

12. Abhinavagupta (ca. 1000 A.D.).

The investigation of the relation between concrete structure ("form") and the rasa-awareness is the aesthetic chef d'oeuvre of the Kasmira Saivite Abhinavagupta,175 and is chiefly expressed in his long commentary on the rasanispatti-sutra of Bharata176. His predecessors had put forth theories that would establish the relation in some determinate way. Abhinava sums up these theories (most of which are otherwise not independently preserved) as an introduction to his quite radical view that no determinate relation is express- ible, for the rasa is more real and more persistent than any of its so-called causes-(has been there all along, in other words. The first theory mentioned is that of Bhattalollata,177 who espoused the commonsensical view that the conjunction (samyoga) of structural elements, which he took to imply "with the sthayibhava" was the cause (karana) of the rasa (as effect: karya). Two stages in this cause and effect theory are recognized: first the preconditions (vibhava) of an emotional experience (including stage set, characters, poetic phrases, etc.) combine to generate the notion of the concrete emotion involved (love, anger, pity, etc.); then this last, by combina- tion with the external consequences of the concrete emotion (anubhava-faint- ing, stuttering, tears, shouting), through the medium of temporary emotional states (vyabhicāribhāva) (sadness, hope, joy, etc.)-becomes the rasa: sthayy eva vibhāvānubhāvādibhir upacito rasaḥ.178

173 Dhanika explicitly rejects the view that santa rasa is possible in drama. He is thus still faithful to Bharata, and not convinced of the propriety of the generic collapse implied by the Dhvanyāloka. See K.C. PANDEY. Dhanañjaya and Abhi- navagupta on Santarasa, PAIOC, 12, p. 326 and supra, note 119. The Dasarūpa t has been translated with an Introduction by G.O. HAAS (Indo-Iranian Series, New York 1912) but his views are "unreliable" (DE, H.S.P., I, p. 124, 126). 174 Malava: DE, I, p. 121. 175 See K.C. PANDEY's monograph Abhinavagupta (2nd ed.) Varanasi 1963, for a detailed account of his life and thought. Pandey's views often suffer exception however by our authorities, Kane and De. 176 Supra, note 127; GOS edition, vol. 36, pp. 272-87: translated by R. GNOLI, The Aesthetic Experience according to Abhinavagupta (SOR vol. 11). Cf. also S.K. DE, The Theory of Rasa, in Some Problems ... , pp. 177ff. and M. HIRIYANNA, "Art Experience", Mysore, 1954 (supra, note 119), also K.C. PANDEY, Abhinava- gupta's Theory of Meaning, NIA, 5, p. 241. 177 See J. PRABHAKARA SASTRI, Lollata's Theory of Rasa, JOIB 15, p. 157; Abhinavabharatī, p. 272; De on this section of the Bharati: H.S. P., II, pp. 117-138. 178 Ibid.

Abhinavagupta 265

The second theory (of Srisankuka)179 recognizes the inadequacy of the first (in Abhinava's eyes) and goes beyond it to a better formulation. The notion of cause-effect is really too powerful to explain the emergence of rasa-awareness. It presumes a realistic determination (thus ignoring that the play is in an important sense a 'fiction' whereas its 'effect' is real) and more importantly is subject to constraints of logical necessity which simply do not hold in the case of fiction, where we often have causes without effect and effects without cause. Śrīśankuka replaces kāryakaranabhava (relation of cause and effect) with the familiar notion of imitation (anukarana) according to which one may infer the rasa from the wholly fictive portrayal on the stage. Rasa is, in fact, this inference ("a state of knowing") based on imitation. This notion, though clear enough to have dominated Western aesthetics for 2,300 years, is unsound, according to Abhinava's teacher Bhattatauta,180 because it too is overtly realistic, and also psychologically untenable. In no true sense can it be said that any of those involved in the drama-the audience, the players, or even the critics-are imitating anything. "Imitation" involves an awareness that one is not something else (which one imitates), and this kind of awareness, ac- cording to Tauta (and implicitly approved by Abhinava) is wholly incom- patible with the kind of awareness that proceeds from the drama-an aware- ness characterized by a thorough immersion in the events of the play, so thorough that the audience, the players-even the 'critics'-lose all sense of their separate psychological identities (as being a member of the audience, or a person so-and-so, or being here and now limited by individual and concrete needs, hungers, memories ... ). So long as the audience remains on this 'false' level of awareness, the events of the play will seem disjoint from his life, "out there"-a representation of reality: the awareness of the play remains discrete and other. So 'imitation' really explains what is unsuccessful in the play rather than what is successful. This paradox can be overcome only by recognizing the elements of the play as already determined in this overwhelming sense of absorption which we all recognize in ourselves as transcending concrete experience and selfconsciousness. In effect, the events of the play are not even understood as such-and-such (Rama, the historical king, etc.) but are already reformulated in this specifically dramatic awareness as "What Rama has in common with ourselves: lover, husband, aspirer after beauty ... " Bhattanayaka, though earlier in time than Tauta181 and not in the guru- paramparā of Abhinavagupta, already seems to have expressed the positive implications of Tauta's criticism of rasa as imitation182: that the structural elements enter into consciousness already as generalized (in the sense spoken of

179 Ibid. p. 272-3; his view seems to have been the "common-sense" view in India, too: cf. Daśarūpaka 1,7. 180 Ibid. p. 274-6: asmadupādhyāyāh. 181 DE, H.S.P., I, p. 41; pp. 111-12. 182 As recounted by Abhinava after discussing the views of his teacher: p. 276-7.

266 E. Gerow . Indian Poetics

above: (sadharanikrta), and that the percipient state of the audience, players, and the like, is in effect their common experience (expressed via the Indian notion of vāsanã, or(incipient capacity to experience some thing as such-and- such1834) their capacity to love or hate-shared with all men in virtue of their being human-and not their loving or hating this or that person (deter- mined love or hate, or what is the same thing, circumstantial love and hate, where time and place are crucial factors in the realization). Which is to say, in effect, that the rasa is a form of general emotional consciousness, similar to the ätman itself (and like the atman rarely experienced as such but rather as determined personally and temporally). The play becomes a unique medium for the statement ofpure" emotional consciousness184 still the atman is not experienced in its wholly pure state, but is still coloured by shadings of general emotional oppositions: love: hate, etc. But Nāyaka's attempt to formulate a generative relationship between structure and effect (rasa) in terms of his insightful psychology is unacceptable to Abhinava, for Nāyaka attributes a novel function to the play (bhavakatva), in order to "explain" why the rasa is what it is in the audience. Here he has to resort to second 'function'- whose only purpose, apparently, is to express the rasa as a product of the first, in that sensitized audience. He calls it bhogikarana -making susceptible to enjoyment. Further these functions appear to apply to the play only for Nāyaka still clings to the ālamkārika view that rasa in kāvya is a subordinate element only: kāvye 'mšatvam na rūpatā185. The circularity and imperspicuous- ness of these novel "functions"hardly need comment. The "new" function, bhogikarana, merely restates the problem; by verbalizing the difficulty as its own solution: that the rasa is enjoyment no one doubts, but how does enjoy- ment come about? come out of what is often quite hateful, and by definition philosophically neutral and generalized ?180 Abhinavagupta accepts the negative elements of Tauta's criticism and Nāyaka's positive résumé) while trying to give a more convincing rationale for the effective relation (in the dramatic as well as the poetic context) of these quite remarkable abstractions. Instead of positing a new 'function,' the transition from concrete awareness to rasa (which, following Tauta and Nāyaka, we accept as generalized emotio- nal consciousness) is clarified by examining the role of a temporal process in this

183 In the Vedanta, explaining the vehicle of transmigration: Nayaka however, having adopted a special function (bhāvanā) operating in just this sense, does not need the vāsanā: Abhinava ad Dhvanyāloka 1, 1 p. 78-9 (edited by S. KUPPUSWAMI SASTRI, Madras 1944). 184 This creative capacity of the play is called bhavana, a "function" peculiar to the work as art, and also accepted by Dhanamjaya.) 185 Abhinava ad Dhvanyaloka, ibid., p. 78. Thus bhogikarana may implicitly be another effort to account for the 'peculiarity' of drama (cf. the santa rasa contro- versy: supra, note 119). See also DE, S.P.S.A., pp. 59-61. 186 Nāyaka was influenced by the Sāmkhya: DE, H.S.P., II, p. 124, 183; but this 'explanation' resembles more a mīmāmsaka arguing an apūrva!

Abhinavagupta 267

'realization.' All of the previously examined theories, realistic or idealistic, have taken for granted the consequence of the rasa experience. And indeed there is a sense in which the play causes the rasa-as prior cause and posterior effect. But Abhinava asserts that the truth of the relationship is obscured rather than clarified by considering the matter on this level primarily-for as long as rasa is considered an effect, that is to say, determined in a temporal process, its being will be wholly inexplicable; rather, we must turn the play upside down to understand it-the rasa is more real than the play which 'caused' it.187 We may have to resort again to the convenient Aristotelian distinctions: the play is not a 'cause' in the sense of an efficient cause, but rather is a precondition, whose relationship to the 'end'-the rasa-is merely hypothetical the play is a cause then only in the sense of the material cause, and it displays the necessity of any potential in relation to the actual that is its goal and realization. All that seems to happen-in the theatre-is that the play permits the spectator to clarify the implicitude of his emotional pro- pensities-propensities which he brought to the theatre with him and which he will take away again; for these emotional dispositions are the very ground of his sentimental or worldly life they are what make it possible to feel such- and-such in the context of this or that determinate situation 188 Of course these propensities are present in everyday life as well-in this sense the play does not seem to differ radically from the world we live in-and the peculiar cha- racter of art resides only in the manner of determination; from this derives the uniqueness and the clarifying strength of the play. The effect on criticism of this effort to bring all art forms (specifically the two canonical written forms) under a universal principle that is a non-denota- tive state of awareness, was profound. Instead of a disciplined search for structural principles, practical aesthetics took on more the character of establishing the rasa in terms of this or that set of circumstances. Structural considerations were relegated to the status of adventitious concomitants, which could be expounded by list as well as any other way. The integrity of the work tay in its"having a rasa, that is,-in the successful combination of the circumstantial structures to awaken or otherwise bring to consciousness the rasa; But the rasa, by definition not a product of, or otherwise strictly depen- dent on, its structural preconditions (for rasa is found even in the phoneme: varna!) is seen, from the point of view of its truth, not of the work at all, but of the soul where it truly is. Art then, structurally or elementally defined, is no more than an excitant: the differences of one art form or another (since they

187 It is, as we say, an affect. K.L. SAHAL, Objective Correlative and the Theory of Rasa', Calcutta Review, N.S. 2, no 2, p. 237, looks at the curious parallelism with T.S. Eliot's notion. 188 Hence the importance of the discussion of the "obstacles" to aesthetic awareness: Abhinavabharati, ibid., p. 280-81: in general links with the world impede or cancel this transcendent (alaukika) state. The core of Abhinava's reply to his predecessors in establishing his own doctrine is on pp. 279-286.

268 E. Gerow . Indian Poetics

are all brought under the notion of rasa) are as such less important.than the fact that it is the genus art. For it would seem a consequence of the theory that rasa is manipulated in a certain way: it "awaits" the work of art as it were. Contrast the way the emotional effect katharsis is according to Aristotle specifically determined by the plot structure: arising from, and therefore presuming the very essence of the story :- development; turning point; denouement.189 The rasa, though it does arise from the 'story' does so only in the tautological sense that without the story we would have no drama, etc. The rasa arises from all the elements, or more accurately, from their combina- tion. Efforts to assess this combination (samyoga) were directed more toward the end (the rasa), and inevitably fell into discussions of this or that element as contributing to the end; of course, not every "combination" will work; the proprieties and contrasts that are successful continue to be explained via the old notions of anga, etc. borrowed from Bharata.) But whether we call them poetry, or drama, or music, or dance-even painting-the determinants as such work indifferently: the awakening structure, however essential in con- crete terms (for without it we would not have the rasa: Aristotle's relation of material to final cause) is defined not only as adventitious, but in truth, as not existing-in the sense that persistent awareness of the structure as such constitutes an obstacle (vighna) to the plenitude which is rasa. A plenitude that suggests to Abhinava, a theologian of note, the analogy between rasasvāda and brahmasvada, which differ only in terms of the 'impermanence' of the former. The door is thus opened to theological interpretations of aesthetics, or more importantly, to aesthetic interpretations of theology.190(Ru 6a).

13. Mahima, Bhoja, Rudrabhatta

Mahima(bhatta)./Perhaps a generation later, 1020-50, than Abhinavagupta and Kuntaka, and also by his title (Rājānaka) a Kāśmīrą, Mahimā (Mahiman) is the first writer known who explicitly takes to task the dhvani theory of poetic interpretation. The dhvani (ignored totally by Mahima's contemporary Bhoja) did not, it seems, gain immediate acceptance; it was only at the close of the period we have labelled "controversial" that the dhvani assumed the decisive place it has had since191 But it is also true that the importance of dhvani as function (vyañjanā) is superseded by its content of preferedce: the rasa. Later writers speak not so much of dhvani as the soul of poetry192 as

189 Supra, note 134. 190 See M. HIRIYANNA, Indian Aesthetics, PAIOC, 1, p. lv, where these notions are related to the jivanmukta ideal; on Rūpa Gosvāmin, see infra, p. 284-5. 191 Due less to Abhinava than to the 'popular' Kāvyaprakāśa. 192 Dhvanyāloka 1, 1.

Mahimā, Bhoja, Rudrabhatța 269

they do of rasa as the ground of aesthetic pleasure193. The argument whereby rasa was generalized as content of all written art is overshadowed by its conclusion: and this Mahima in no way disputes. In fact, he may be said to have been the first to assert it.194. Mahimabhatta accepts the primacy of rasa while disputing Anandavardhana's account of how it comes to be. ! Mahimabhatta focuses attention on the function (vrtti) dhvani; his accep- tance of the substantive conclusions of the Dhvanyaloka195 is motivated by an eristic determination to replace the dhvani with another function (anumiti). ! The work has very much this negative tone. Its three chapters are devoted to a refutation of dhvani (and several other 'supra-literal' concepts, among them, sphota196), the argument being overall an indirect one: of all the posited ways to comprehend suggested meanings in poetry, none can withstand cri- ticism, leaving only anumiti, "inference" (presumably recognized by all Hin- dus at least) as a possible explanation. The second chapter deals again negatively with poetic flaw (dosa)-only as an elaborate introduction to a further attack on the Dhvanikara's definition of dhvani, which will be shown to violate all the canons of good composition! The final chapter goes through all the examples alleged in the Dhvanyaloka to illustrate dhvani, and reexplains them through anumana inference. Mahimabhatta's purpose in fixing on inference as an alternative to dhvani is not clear, or clarified by his own statements. He seems to be returning to the position attributed to Srīśankuka, and effectively (disposed of by Nāyaka, Tauta and Abhinavagupta.197 But these refutations are not dealt with; indeed Visvanatha more or less repeats them in his turn when demolishing the anumiti ) of the Vyaktiviveka198. They reduce to the problem of designating a vyāpti in "poetic inference", without which necessary concomitance the syllogism of inference will not have validity. Bhoja. The equivocal position of Bhojadeva in the history of poetics has been admirably catalogued by Dr. V. Raghavan.199 Like Rājaśekhara, with whom the period of post-dhvani uncertainty begins, Bhoja seems uncommitted to any traditionally defined view. of the subject.(In many details of organiza- tion, his works seem to owe more to the atypical Agni-Purana than to any

193 yo 'rthah (scil. rasah) tasyānvayāt kāvyam kamanīyatvam aśnute: Bhoja, Sarasvatīkaņțhābharaņa, 5, 1 (RĀGHAVAN, Šrngāraprakāśa p. 425), cf. vākyam rasātmakam kāvyam: Sāhityadarpaņa, 1,3. 194 Cf. Dhanamjaya, Bhattanāyaka, supra. 195 Cf. Vyaktiviveka (KSS 121, 1964, p. 47): artho 'pi dvividho vācyo 'numeyaś ca. 196 Op. cit., p. 61. 197 Supra, pp. 264-8. 198 SD (NSP ed. Bombay 1936), pp. 262-71 (5,4-5); cf. DE, H.S.P., II, pp. 197-9. 199 Bhoja's Śrngāraprakāsa, Madras 1963. The work is also essential for situating the general context of poetic theory in which Bhoja finds a peculiar place. See also the edition of G. JOSYER, Mysore 1955-74; four volumes.

270 E. Gerow . Indian Poetics

recognized authority; yet Bhoja cites and is intimately familiar with the definite views of more early poeticians than any other writer.200 He is the author (a dubious distinction!) of the "largest known work on Sanskrit Poe- tics"201, the Srngaraprakasa (now edited by G. Josyer but chiefly known through Raghavan's "precis"-perhaps the world's longest precis!), yet he ignores the revolutionary notion of his day, the dhvani doctrine of Ananda! His date, which is certain, is in part fixed by accepting the traditional identi- fication with Bhoja, King of Dhara, known to have reigned between 1010 and 1055202 and himself subject of much poet-lore (especially as concerns his liberality to the arts.203) An hypothesis that partly orders this jumble of conflicting evidence might be that the dhvani-doctrine had not got much beyond Kashmir even by the time of Mammata, and it was likely the popularity of the Kāvyaprakāsa that indeed carried it to the rest of India. Bhoja might thus fairly be said not to have known of the dhvani, rather than ignoring it.204 Bhoja's works, the Sarasvatīkanthābharaņa and Srngāraprakāśa (on poetics) we have called pauranika, not only because of the links to the Agni-Purāna, but in view of their nature, as being uncritical compilations of the most diverse views. (De; has put it well "The learning which this work parades, though extensive, is ill-assorted and uncritical, its ideas lacking in system and its expression in preciseness.}'205 The doctrine of Bhoja that most retains the attention of later writers is not his tripartite organization of the alamkāras (borrowed from the Agni- Purāņa) or his elaborate classification of the gunas (pace Vidyānātha206); rather he is the first writer (with Mahima perhaps) to state rasa as an all-encompassing principle of composition. He is much concerned with bringing natya and kavya together under the same rubric207, but seems to attempt this through the wholly unprecedented reductio of considering all eight (or nine) rasas one: śrngāra, 'love'. Two aspects retain our attention: in the first place, the argu- ment (if so intended) is a non sequitur, for the multiplicity of rasas (as such) was never intended as a differentia of the dramatic genre; on the other hand, though the rasa of kavya is often śrngāra (Kumārasambhava), the reduction

200 RAGHAVAN, op. cit., ch. 28; KANE, H.S.P., p. 258. 201 DE, H. S.P., I, p. 136. 202 Ibid. 203 See note 7, supra. 204 DE, H.S.P., II, p. 206; but Bhoja may already have been known in Kashmir! (I, p. 146-7): if mentioned by Mammata: KP 10, verse 505 (ASS 66, p. 531). 205 Vol II, p. 211. 206 Ibid., footnote. 207-Much of the SP. is based on the NS .: chapters XII-XIV and XXIIff., which also form the basis for Rūpa Gosvamin's later treatment of bhaktirasa. See DE, Bhakti Rasa Sāstra of Bengal Vaișnavism, IHQ, 8, p. 643-88 and S. BHATTA- CHARYA, Bhoja's Rasa: its influence on the Bengali Rasa-śāstra, JOIB, 13, p. 106.

Mammața 271

appears to confuse statistics with principlees At the same time he elaborates the number of 'subsidiary' rasas (not eight or nine, but twelve in total). Is the emphasis on one rasa intended chiefly to organize his own vast work? (The shorter Sarasvatīkanthabharana is not as clearly committed to the singleness theory). In any case, as De remarks, the views of Bhoja, the last pauranika and perhaps the precursor of the real encyclopaedism of Mammata, were ignored almost totally, while his fame increased to fabulous proportions. (Rudrabhatta. According to modern authorities, this author is not to be confused with Rudrata.209 His Srngaratilaka is certainly an unlikely work to be attributed to the archetypal ālamkārika! Devoted exclusively to the dra- matic rasas (including santa!), it is an important production in its own right, being in effect the first work that singles out this one topic for exhaustive treatment (anticipating the rasa-writers of the 14th century (Vidyanatha and Viśvanātha)); it is conventional-quite unlike the Srngāraprakāsa,210novel chiefly in attesting to the growing influence of the rasa. Rudrabhatta's date is uncertain, "before 1100 AD".211

14. Mammata

Mammata's Kāvyaprakāsa-probably the (most influential single work of Indian poetics212 marks the end of the period of uncertainty. that followed for nearly two centuries the composition of the Dhvanyaloka. The renown, in many ways merited, of this work is probably due to two causes-it appears to have offered the decisive and convincing interpretation of the dhvani-theory (based on Abhinavagupta), henceforth considered in some sense valid; secondly, it originated (and served as model for) the genre we have termed "systematic". It is true that the works of Bhoja and Rājasekhara to a certain extent merit this designation also, but they really do little more than bring into relief the genuine accomplishment of Mammata: his "system" is not based on atypical or idiosyncratic principles (Rajasekhara's attempt to give a socio-history of pcetry (kāvyapurusa) and Bhoja's odd celebration of śrngāra and his equally bizarre classification of alamkāras and guņas), but on a careful and faithful reading of his predecessors' works. In a way the KP seems less systematic than the uncompromising but erratic monolith of Bhoja,

208 bhāvānām uttamam yat tu tac chrngam śrestham ucyate /

(GOS ed.). iyanti (iyranti?) śrngam yasmāt tu tasmāc chrngāra ucyate: Bhāvaprakāśana, p. 48

209 See KANE, H.S.P., pp. 156-59 for an account of the controversy, and the siddhānta. 210 Based much more clearly on Nātyasastra. 211 KANE, ibid. 212 DE, H.S.P., I, p. 156; see DE, "Mammata's Kāvya Prakāśa" in Some Prob- lems ... pp. 108-130.

272 E. Gerow . Indian Poetics

for it precisely attempts to comprehend and integrate the various extant doctrines of Indian poetics in the light of one of them, the dhvani theory-a task that is bound to leave a number of ragged edges if pursued honestly. (So a better term than "systematic" might be "synthetic'. Given the fame of his work213 the Date of Mammata presents more difficulty than is expected at this relatively late time. Its lower terminus is somewhere between 1135 and -1150 A.D., established by citations in Ruyyaka's Alamkārasarvasva and general certainty regarding Ruyyaka's date. But its upper terminus (with the same degree of certainty) is more than a century earlier, between 1005 and 1015 A.D., established by Mammata's quoting the Navasāhasānkacarita and his frequent mentions of Abhinava. Anything more precise rests on(a single reference to Bhojanrpati214, the fame of whose liberality may have spread to Kashmir during his lifetime.215 The verse is in the style of a standard prasasti, and of course may not refer at all to King Bhoja of Dhara 16) But if this flimsy evidence is accepted, Mammata's work may at least be dated, within a roughly 75-100 year period, between 1050 and 1135. By his title and name (Rājānaka Mammata), he is likely a Kāśmīra, and is a worthy propagator of the distinguished local poetic tradition. As we have implied, it is probably Mammata that made Kāśmīra poetics Indian poetics. The form of the work indicates its novel; character. Rather than being an argument in favour of or against a particular view-point, it attempts to justify many different points of view-it attempts to integrate in a positive way at least the conclusions of several antecedent poetic theories, under the general rubric of the dhvani theory (chapters IV-VI). But even the dhvani, faithfully rendered in terms borrowed from the Dhvanyāloka, is 'integrated' into a larger frame of reference, provided by the theory of verbal functions, vrttis, now accepted as siddhanta (chapters II-III, "signifiant et signifié"). Mam- mata's eclecticism is demonstrated in his 'definition' of kāvya (often criticized as being a hodge-podge characterization; rather than a definition) "tad (kā-

213 The unparalleled interest of the tradition (Bibliography: DE, H.S.P., I, pp. 154-177!) is not shared by modern critics, who generally regard Mammata as the beginning of the 'decline' of Indian poetic thought (DE, S.P.S.A., p. 46). The few attempts to confront Mammata's system are referenced ad loc. in the following notes. Mammata is also the author of a work on the sabdavrttis: the Sabdavyāpāra- paricaya (DE, H.S.P., I, p. 156). 214 See 10, 505; other references: DE, H.S.P., I, p. 146. 215 Regnans 1010-55, DE, H.S.P., I, p. 136. 216("Bhoja" is the name of a region as early as the Mahabharata; the term is a common epithet of kings (as early as Aitareya-Brahmana: MW, p. 767. The verse occurs in the text very close to the traditional point-end of parikara alamkāra (ASS ed., p. 541)-assigned to the end of Mammata's contribution to the text, the remainder completed by Alata (DE, H.S.P., I, p. 149; or Alaka, KANE, H.S.P., pp. 271 ff.). But the division point is far from certain)and we may not be dealing with Mammata at all. See H.R. DIWEKAR, The dual authorship of the Kāvyaprakasa, JRAS, 1927, pp. 505ff .; R.J. JAIN, Authorship of the portion from the parikara alamkāra ... PAIOC, 12, p. 331.

Mammața 273

vyam) adoşau śabdārthau saguņānalamkrtī punah kvāpi" (1, 4ab: ("poetry is (a mix of) word and sense free from fault, sometimes provided with qualities and sometimes lacking ornament"). The definition, it is clear, fits Mammata's overall purpose very well, and anticipates the structure of his book. Dosos fault, being any generalized detraction from the comprehension that is poetic suggestion is dealt with in chapter VIIx gunas are treated in VIII, and alam- kāras (of sabda and artha) are defined in IX and X. The latter topics are re- duced to the status of ancillaries, which do not define poetry (the task of chapters IV-VI) but rather detract from (in the case of dosa) or augment its constitutive charm (guņa, alamkāra).217 Poetry per se is defined in 1,4ab; its Internal divisions (differentiated strictly in grammatical terms, according to the vrttis, yet assuming the status of evaluative concepts) are stated in 1,4cd-5: where vyangya (the suggested sense) predominates, we have the highest type, termed dhvani (a type of poetry) (treated in extenso in chapter IV); where vyangya is important but not primary (is an ancillary to the vācya, or fiteral sense) we have poetry of subordinated suggestion (gunībhūtavyangya- treated in chapter V; all these terms are taken verbatim from the Dhvanyā- loka), and finally, poetry lacking dhvani (sabdacitra: chapter VI, the shortest in the Kāvyaprakāśa)-uninteresting to Mammata, though he allows it as poetry (probably contra Anandavardhana) inasmuch as the use of figures of speech per se (without any trace of ulterior meanings, as in the case of the verbal figures yamaka, etc.) is generally allowed to have a certain kind of charm that marks the utterance as different from ordinary speech. And Mammata is an inclusivist: several principles operate simultaneously to englobe as much of the preceding tradition as possible. Mammata thus seems to be the dharmasāstrī or mimamsaka of the poetic tradition: it is he who in fact has formulated poetics as a śāstra, internally coherent, and externally comprehensive, out of the various strands of local traditions and theories that had come each in their own way to be viewed as inadequate. And as elsewhere in India, the 'correct' solution is that one (during the synthetic or sastraic period) which saves most, is permissive and yet faithful to the now recognized broader needs of universalization 218) Mam- mata's particular contribution to doctrine was very limited, but it can be seen that this is not a serious criticism of his accomplishment, which was to define a form of study and a curriculum in poetics that has persisted to this day. The one area of poetic interest-despite the implication of the Dhvanya- loka219-that is not integrated into the sastra, is that of the old dramaturgy (nātyasastra). On this side his solution was clearly incomplete, and several

217 Note the clever adaptation of a view that Vamana so lucklessly upheld-as well as(the notion of dosa modelled on vighna of Abhinava) 218 As in the case of the older Vedic tradition, when it became clear that an all- India standard of ritual practice was required: hence, mīmāmsā, dharmašāstra. }) 219 And even against the example of Rudrata, whom Mammata follows whenever possible in books 9 and 10 (see V. SUKTHANKAR, in ZDMG, 1912, p. 478).

274 E. Gerow . Indian Pootics

later writers (Viśvanātha, most notably) sought to bring that matter too under the rubric provided by the Kāvyaprakasa. In effect(this is one of the major themes of the later history of Indian poetics among the syncretistic writers. The first three karikas of chapter I respond to, in very abbreviated terms, a favourite concern of the early tradition: the external context of poetry viz. its matter (svabhāva), its utility (prayojana) and its prerequisites (hetu) 1,1-3. The rest of the book is concerned with its "form" (svarupa) and thus may be said to constitute a tolerably more accurate account of the kāvya- purusa than Rājasekhara ever imagined.

15. The post-Mammata period

The post-Mammata poetics is not as barren as De would have it220 but it does presume as its standard the Kavyaprakasa version of the dhvani theory. That more and more appears to be copied from earlier texts is interesting, but of little significance, except to intellectual traditions like our own that place a higher value on originality than on truth. The major works of the six or eight centuries following Mammata each have a peculiar contribution to make to the general history of the sastra, even though they do not often try to set it on its head. Ruyyaka, author of the Alamkarasarvasva, and reputedly of several other works on poetics, including a commentary on the Vyaktiviveka221 was again a (Kaśmīra, whose period of major activity can be rather precisely dated as 1135-50 A.D.222. He wrote a commentary on Mammata, showing among other things the immediate fame of the Kāvyaprakāsa-at least in Kashmir. Ruyyaka's canonical work (Alamkārasarvasva) may have been only part three of a three-part work devoted to the examination of dhvanikāvya on lines resembling chapters IV-VI of the Kāvyaprakāśa.223 If so, it constitutes a revision in Mammata's scheme, and as De thinks224, was probably inspired by Kuntaka's treatment of alamkara,rather than Mammata's) Even without this hypothesis, Ruyyaka's interest in alamkāra seems determined by different principles than Mammata's, and is not at all what might appear on the sur- face, a throwback to the pre-dhvani alamkāraśastra.

220 DE H.S.P., II, p. 215. 221 DE, H.S.P., I, pp. 182-4. 222 KANE, H.S.P., p. 285. 223 DE, H.S.P., II, p. 229; the lost "Alamkaramañjari", corresponding to chapters IV and V (dhvaniprādhānyam; guņībhūtavyangyam) is not thought to be Ruyyaka's by KANE, H.S.P., p. 401. The Alamkārasarvasva is one of the few post-Mammata works to benefit from a good Western translation: H. JACOBI, in ZDMG, 1908, three parts. See his Introduction, pp. 289-94. 224 DE, H.S.P., II, p. 230.

The post-Mammata period 275

(Ruyyaka is as committed to the dhvani and the theory of vrtti as was Mam- mata) he uniquely among the older writers has a sense of the historical place of his own work, gives an encapsulated version of the "schools" of thought to which his work responds225. In Mammata's view alamkara is an adjunct; in line with the old analogy, figures may be put on or off-like a woman's je- welry-enhancing with their presence but never cancelling by their absence, the basic beauty of the poem or the woman. Such a view is not easy to recon- : cile with the older alamkarikas passionate concern for figuration as formally defining poetry, and Ruyyaka sets out to bring Mammata and the older tra- dition into accord. on this important point. His concern in dealing with a vast range of figures (over 80) is to show each one as involving essentially some aspect or element of the distinctive charm that differentiates poetic utterance generally. He uses the term vicchitti, recalling the doctrine of Kuntaka, While this distinctive charm is not itself a form of suggestion (dhvani) nor based upon it (assuming indeed the context of Mammata's book VI, citrakāvya), a link is here being forged between the ideal of figuration in Mammata, the dhvani of the Dhvanyaloka and the vakrokti of Vakroktijīvita. Although certain figures were long recognized as involving essentially an element of the dhvani226, it is Ruyyaka who attempts for the first time, following a principle advanced by Kuntaka, to generalize that insight to all figures. In some figures, the words used actually suggest a second meaning, and it is that relation that constitutes the "figure"; clearly the figure is a "form" of dhvani227. In other cases a figure may be "suggested" by another explicit figure (as upama by an explicit dipaka). But in every figure, if properly so called, some remarkable comprehension must result, a comprehension that differs from the plain sense of the words, and marks the intellection as different from inference. In many cases it is just the juxtaposition of ideas so improbably associated that the mind is forced out of its usual paths to seek a new or revised understanding of the relationship, what they can have in common.("The professor winked at me so hard that his face was like a concertina with a hole in it"228). Vicchitti as used by Ruyyaka seems to designate this movement of the mind that may not involve any new apprehension at all (and is thus not dhvani strictly speaking), but simply a revised and 'novel' insight into what is.229 It would be easy to argue that it was in fact a dhvani) this may have been Ruyyaka's strong conclusion, though (in ironic confidence) he leaves it to our insight, observing only tasmād vyangya eva vākyārthībhūtaḥ kāvyajīvitam ity eșa pakso .. .230

225 As an Introduction to his work, pp. 3-19 of the Kavyamala (35) edition. 226 E.g. Dhvanyāloka, 1, 13. Mentioned are samāsokti, ākşepa, višeșokti, paryā- yokta. 227 Also, AS., p. 19. 228 Joyce Cary, The Horse's Mouth. 229 Cf. AS., p. 61-2, on differentiating figures; also Dhvanyāloka, ibid. 230 AS., p. 14.

276 E. Gerow . Indian Poetics

Ruyyaka's other major contribution to the sastra now evolving responds to the problem of its extension (as the former attempt to relate alamkāra to dhvani might be said to involve a principle of unification). He has, following Rudrata, given an account of the standards of classifying figures that has not been improved upon since, and makes clear the limited internal structure of the figurative universe. Accepting the time-honoured distinction between figures of sound and those of sense, which he does not change essentially, Ruyyaka attempts to analyze the figures of sense (arthālamkara) according to similarly structural principles. The fourfold division of Rudrata is only partially accepted; other principles are expounded. Ruyyaka however mentions only briefly the criteria used in his classifications; we are dependent on his commentators and on our own intuition to discover the rationale of many of the distinctions. According to Jacobi's summary, which is literal,231 the arthālamkāras are divided into nine classes, of which one ("similitude") is further subdivided according to the intended character of the difference (i. e. whether the difference between the two terms "compared" is intended to be cancelled (as in rūpaka) or reinforced (vyatireka) or maintained in identity (as in upama itself)X The first type is itself subdivided (continuing the good Advaita terminology) according as the identification takes place by superimposition (āropa, as in rūpaka) or by deletion (adhyavasāya, as in utprekşā>)the "object" of comparison is deleted, leaving only its manner of acting or being), or in a way, by both, where the deletion of term serves to suggest a further adjunction or comparison (as in the sometime sabdālamkāra dīpaka). Ruyyaka is not beholden to the tradition in many of these details, nor in his overall conception. Not only are some śabdālamkāras reevaluated (as dīpaka, taking a signal from Dhvanyāloka), but some of the criteria of earlier systems are fused in new configurations Atisayokti, for example (one of the four basic terms in Rudrata's system) appears here both as a simile (with utpreksa in the adhyavasayagarbha section) and as a type of virodha (contradiction)>"exaggeration" in other words is now being examined in terms of its intent, rather than in its propositional form. And its importance to the structure declines markedly.232 (The Vedanta base of the analysis in clear; here too "intent" plays the crucial role; poetry in a way recapitulates the "errors" of worldly perception, which of course are based on wrong-will (the "this is mine" paradigm). The remaining varieties of arthalamkara figuration appear to be more traditional. Ruyyaka appears indebted to the dhvani analysis (this too would in part account for the 'intentional' character of his system}; instead of the statement of a relationship, viz. similitude, between two subjects, the poet may intend in some cases the discovery of one of the subjects (through the other): certain of

231 ZDMG, 1908, pp. 626-7. 232 Again, the reader is referred to GEROW, G.I.F.S. for definitions and examples of these figures.

The post-Mammata period 277

the dhvani-based figures, samasokti and others are found here. The suggestive function (vyañjana) alone can be appealed to; in the types of similitude instanced above, both subjects are referred to in some way or other (through a property etc.). In other cases, the poet's intention may dwell on relationships other than similitude. Chief among these is the contrary of similitude, contradiction or incompatibility (virodha). Many figures can be so understood (many involving incompatibilities of logic, as in Rudrata's analysis). But the logical or sequen- tial aspect per se is not given much prominence here. The next four types of figuration in Ruyyaka appear to involve the notion of groups of things (as opposed presumably to the foregoing three classes, which dealt with two things (upameya, upamāna; prākaraņika, aprākaraņika) only). And the notion of group does imply at least the possibility of sequence, and hence argument, logic. Various kinds of arrangement are considered: of words (śrnkhalābandha: the māla figures of Mammata), of terms or conclu- sions (the two figures kavyalinga and anumana, where a logical operation appears to be central), of meanings (padārtha as opposed to pada: a group of figures based on conventional relationships at the sentence level, syntax (vākyanyāya)); finally, of things or referents (lokanyaya, where the issue is not so much on the internal or grammatical relationships among several words or their meanings, as on the external arrangements of things themselves-as when two things are said to be at war with one another, or to have fused (pratyānīka, mīlita)); the emphasis is on the things themselves rather than on their relationship, hence we appear nearly to have returned to the domain of simile. The final two classes of figure appear to reflect interest in the vrttis or functions of signification as elements in defining or capturing the dhvani in the figure. The Dhvanyaloka has argued that the third function233 may relate to figuration in various ways, as subordinate or as principal; as having a figure or merely a state of affairs (vastu) as its vyangya, etc., but no effort was made to investigate these matters in specific figures, except in an ad hoc and illustrative way. Indeed Ruyyaka's second class of figures appears to correspond closely to the Dhvanyaloka's distinction of primary: secondary; in all of them, suggestion seems involved as a subsidiary element crucial to the form of the figure (which is the main thingi or principal234). Subsequent discussion of the relation of dhvani (as subordinate) and figura- tion seems to have made clear that other principles might also be involved. Not only is a meaning (artha) capable of evoking another meaning (all the figures of Ruyyaka's second class fit here), but elements of discourse other than

233 vyañjanā, or dhvani (the latter word used loosely both in the sense of function and the result of the function (a kind of poetry)). 234 In samasokti, it is the meaning implicit in the 'irrelevance' of the figure that I focus on: "suggestion" functions to realize the figure and not vice versa)

278 E. Gerow . Indian Poetics

meaning may cause the apprehension of another meaning235. Lastly even the rasa, the dhvani par excellence, which cannot be evoked directly (viz. by abhidha) in any way at all, may nevertheless sometimes not be the principal, even where it is cognized. Its cognition is proof of the operation of the third function, dhvani; but that cognition is sometimes subordinated to the ap- prehension of a concrete meaning (as in rasavat, where the relations expressed in the figure are primary, and the mood it evokes only secondary.) Cases where a rasa or a rasa-type element emerges as a secondary consequence of figuration are treated in Ruyyaka's ninth class. So taken with the second class, Ruyyaka's eighth and ninth classes (appendices perhaps resulting from discussions of the vrttis that took place among poeticians as well as from the effort to treat rasa more consistently within the figurative universe) explore the range of possibili- ties that the third function may subtend to the sabdartha of discourse: it may be grounded either in meanings or in other components of the linguistic spectrum (even sub-linguistic, if gesture be admitted); it may itself become a content (an artha) (as a rasa), but perhaps still subordinated to other functions (or itself as function?). Ruyyaka's treatise stops at just the point where the exploitation of the third function for its own sake (rasa poetry par excellence) might be said to have found opportunity. Ruyyaka has thus suggested an opening for Viśvanātha, to complete finally the major gap in Mammata's encyclopaedism: incorporation of the entire rasa theory (not just its kāvya portion) into the poetic śāstra. Vāgbhața I and II. Two writers on poetics are identified by this name. Both are Jainas from Western India. It is convenient to group them together with Hemacandra, though the date of the second Vagbhata is very imprecise: ... "later than 1150 A.D. He probably flourished in the 14th century".236 The Vagbhatālamkāra is a short work, dating from ca. 1125-43237 whose chief originality is its final chapter, devoted entirely to rasa. The topic of characterization (nāyaka/nāyikā) figures prominently. (VāgbhațaImay have been the first author to look for a way of dealing with the more dramatic aspects of the rasa theory within the structure of the developing sastra.238 The rest of his work is brief, and except for that, conventional. He discusses the purposes and prerequisites of poetry (I), emphasizing pratibhã; the language of kāvya and its defects (II), the ten gunas (III), and the alamkāras (IV: four of sabda and only 35 of artha), in addition to the final chapter on rasa.

235 Gestures, phonemes, contexts: elaborately dissected by Kuntaka, whom Ruyyaka follows here, in defining his eighth class (gūdharthapratiti) which includes both vakrokti (narrow meaning) and svabhāvokti! Beautiful! 236 KANE, H.S.P., p. 296. 237 Ibid., p. 287 238 Apart from Rudrata, whose work preceded the flowering of the dhvani theory, and does not reflect the preeminent place that rasa has had in poetic speculation since. See also Rudrabhatta.

The post-Mammata period 279

Vagbhata II, author of the Kavyanusasana (not to be confused with He- macandra's work), follows nearly the same pattern as his predecessor, save that his second chapter includes the gunas (reduced to Mammata's three) and that his third and fourth are reserved to śabda and artha alamkāras respectively (six of the former and 63 of the latter.239) Hemacandra,(1140)A. D.240, the well-known polygraph, sums up the tendency of this Jaina interlude. His work is perhaps the most eclectic of this period, whose spirit is often dominated by eclecticism.241 Without committing himself to any doctrine (even the rasa/dhvani) Hemacandra succeeds in giving a circumstantial account of the various topics of alamkārasāstra, and also, of the nātyasāstra His work approaches, in other words, the standards of modern critical scholarship, The two Vägbhatas, too, while not adherents of the dhvani viewpoint, reflect its influence. Hemacandra gives one of the most comprehensive accounts of the dramatic rasa theory in the post-dhvani period, and thus sets the stage for Visvanatha's more orderly reintegration of the śāstra. Historians make much of Hemacandra's "lack of originality"242 despite his Jaina emphasis on pratibha.243He borrows not only ideas, but the words they were couched in, and thus truly lays himself open to the charge of compilation. In balance though, his net was very broad, especially in the area of illustrative verses (some 1500)from various authors), and he seems to have possessed also a comprehensive skepticism rare among Indian śāstrīs. In the eight adhyāyas of Kāvyānuśāsana, Hemacandra deals with a slightly augmented version of the Vagbhata framework: purposes and preconditions (pratibhā) (I), rasa (II), dosa (III), guna (IV), śabdālamkāra (V), arthālamkāra (VI), characterization (nāyaka etc., VII), and thetwholes and parts of kāvya (VIII). The episodic character of this list is evident. Despite the importance of rasa, it is not asso- ciated explicitly with the definition of kavya244. Instead it is thought of in somewhat the same way as were the figures in the dhvani theory: as an addi- tional result of the facultative employment of gunas and alamkaras! The treatment of the rasas in terms of vibhavas resembles Bharata, yet other topics related to the natya aspect of the sastra are dealt with in quite another chapter (VII). Overall, the Kāvyānuśāsana resembles the works of the Vāgbhatas, chiefly distinguished by its two appendices on characterization (VII) and on genre (VIII)-which incidentally provide the formal umbrella in terms of which a unified theory of the art might be worked out-but is not.

239 KANE, H.S.P., p. 295. 240 KANE, ibid., p. 290. 241 See V.M. KULKARNI, Sources of Hemacandra's Kavyānusāsana, JOIB, 14, p. 148. 242 KANE, ibid .; DE, H.S.P., II, p. 243. 243 KA, (Kāvyamālā 71), p. 5. 244 Ibid., p. 19.

280 E. Gerow . Indian Poetics

Vidyādhara, Vidyanātha and Visvanātha. In the early fourteenth century, three writers of similar name, all from southeastern India, bring the encyclo- paedic tendency to that state of completeness of which it is capable, by in- corporating, once and for all, the theory of dramatic plot and character into the kavya tradition. Of these, the Ekavali of Vidyadhara differs least from the model of Mammata, save that the examples are all from the hand of the : author himself, and constitute a prasasti of his patron, King Narasimha of Utkala (Orissa). Heraccepts the dhvani theory, attempts to counter the argu- ments of those who deny it, and seems to differ(from Mammata on few points of doctrine, save a certain reliance on Ruyyaka in his account of the figures, The chief excellence of the work lies in its increasingly systematic charac- ter245 and its noteworthiness is in part a function of its being commented upon by the lucid Mallinatha246 ! The Pratāparudrayaśobhūșaņa of Vidyānātha, presents itself also as a prasasti, of Prataparudradeva, King of Ekasila (Warangal) in Telengana.247+ The reorientation to natya is more evident here,(the prasasti of the third chap- ter is itself a small nataka) rather than the usual flattering kavya. The work'st organization also reflects the shift: its nine chapters:concern, in order, the hero İ.(nāyaka)2 kāvya, the drama par excellence {nātaka), +rasa, doșa, guņa, and the final three alamkāra (sabda, artha and mixed). This is a reworked scheme, and Vidyanatha's apparent purpose is to relate kavya and natya through the notion of character mayaka, or hero)248 frather than through rasa, which is still a "content" of the third function, dhvani or vyanjana (though it is assigned a separate chapter). "Character" as a principle of analysis is rare enough in the history of Indian poetics to deserve some emphasis.249 The context of the work of course cannot be ignored: emphasis on nāyaka is also subtle flattery of his patron, a demonstration of the essentially model character of the poli- tical leader. Like Vidyadhara, treatment of the figures often follows Ruyyaka; but an even more important influence on Vidyanatha seems to have been Bhoja. his chapter on the gunas (twenty-four in number, reversing Mammata's tendency to reduce them) is based thereon, as well as certain of his views relating to the preeminence of dhvani250. Likewise, the third chapter, on and illustrating nātaka-as the best type of drama-is based on the Daśarūpaka. In all these areas Vidyanatha is a true encyclopaedist, drawing on varied sources (as Bhoja), yet submitting them all to a single framework of inter-

245 DE, H.S.P., II, p. 235. 246 The Tarala; KANE, H.S.P., p. 293. See P.K. GODE, A note on the historico- literary importance of Mallinatha's commentaries, PAIOC, 3, p. 63. 247 Reigned to 1323 A.D .: KANE, op. cit., p. 294. 248 See S. LÉVI, Le Théâtre Indien, pp. 62-86; S. LIENHARD, Typen der Nāyikā im Indischen Kāvya, WZKM, 52, p. 386. 249((De is so interested in the theme of the decline of aesthetics that he takes even innovations of this period as signs of confusion: H.S.P., II, p. 235-6. 250 gāmbhīrya of poetry is its dhvanimattā: DE, ibid., p. 237.

The post-Mammata period 281

pretation (like Mammata). With Vidyadhara, Vidyanatha's most individual view (nevertheless) appears to be in his account of kāvya,-perhaps a reflection of his attempt to see character as an organizing principle, a principle other than the aesthetic response of the audience. The twin notions of sayya and pāka ("repose" of words; "maturity" of sense) are developed251. Functionally this view presents itself as opposed to the traditional ālamkarika notion that levels of signification-often simultaneous, as in pun (ślesa)-are the richness and "quality" of poetrys instead there appears here, sketchily at least, the odd-for-India proposition that there is but one truly suitable way of saying what it is you mean: this unique fit of words and sense is the sayya, or (repose" of words in their full accomplishments looked at from the point of view of the meaning thus expressed, ipso facto, more completely expressed we have pāka: maturity of expression 252 This curious excursion into what looks like a Western notion of "style", associating quality with uniqueness and inimitability of ex- pression, should be seen, I think, as a rather late attempt to objectify language, to define in the language, rather than in the hearer, those qualities that determine poetry.(The dhvani, though in its origin no less objective (as a vrtti), was in the process of being replaced (see Viśvanatha and later writers) by its "content", the rasa, as a functional notion in defining poetry, and poetics, as understood by Abhinavagupta, was in fact an aesthetic psychology. Vidyānātha, regarding the latter development as inevitable, if not desirable, felt nevertheless cons- trained to offer some criterion whereby it would still be possible objectively to distinguish kāvya and nātya: to define in kāvya some principle analogous to the 'hero' of drama that gives to each poem its objective uniqueness. It is dis- concerting to us that one of the main themes of our criticism is here sketchily suggested, almost as an appendix. Visvanatha's Sahityadarpana, probably the second best known Indian poetic textfris taken by subsequent tradition as the culmination of these efforts to widen the scope of poetics, and systematize it anew.253 Viśvanātha unabashedly accepts the rasa as the constitutive element of all good poetry and drama, hence operating their effective union. His work covers much of the same ground (As Vidyanatha's-including both "character" and "plot" within sahitya (chapters III and VI)-but on the whole, the organization is more traditional showing ten divisions, which are partially those of Mammata (esp. VII and X: dosa and alamkāra). The work is distinguished only by its

251 Ibid., p. 240. 252 See DE, S.P.S.A., p. 21. 253 The English translation by J.R. BALLANTYNE and P. MITRA (Bibliotheca Indica, 1875) is still standard. It is the Sahityadarpana that provided P.V. Kane the occasion for his History of Sanskrit Poetics (chapters I, II and X, with Intro- duction and Notes, Bombay 1910; 2nd ed. 1923 (with the History ... ); 3rd ed. 1951 (reprinted 1961)). See also SATYA VRAT, Viśvanātha Kavirāja: His lost works, JOR, 4, Madras, p. 198. V.'s date is the first quarter of the 14th century: KANE, H.S.P. p. 298.

282 E. Gerow . Indian Poetics

plan; much that is novel and interesting in Vidyanatha is excised in the interest of a greater synthesis .? 54 The notion of sahitya255 itself (here used for the first time in the sense of "literature" rather than "poetry")256 probably reflects the considerations that led Vidyānātha and Vidyādhara to speculate on śayyā and pāka: suitability of word and sense to each other (not just their "conjunction"-as in many earlier definitions, as Mammata's śabdārthau and Bhāmaha's abdārthau sahitau kavyam) marks not only "poetry" but, in this enlarged sense, "brings together' drama as well, under the aegis of a single notion of convention. The term sāhitya, probably from Kuntaka (1,17), explaining Bhamaha's phrase (anayoh, scil. śabdārthayoh) appears also in a context that foreshadows this idea of "suitability" (though his main point is still to establish the differentia-the vakrokti-of poetic speech): it is glossed as anyūnānatiriktatvam ('lacking both deficiency and excess')257. Visvanatha, sauf erreur, does not use the term sahitya except in his title and the colophon verse (10,100); elsewhere the usual term kāvya is employed (vākyam rasātmakam kāvyam (1,3))-even in contexts where clearly the nātya is intended to be encompassed. Nātya, we take it, was nothing more than a written variant form of kavya-its formal theory, the structural theory underlying the dramatic rasa: vibhavas etc., and the samdhis, are thus ripe for systematic inclusion in the overall poetic. In this context, the invention of a term for "literature" is no great accomplishment: genre distinc- tions had in fact been abolished-and Viśvanātha's 'definition' of kāvya: bears graphic witness to this: for in it are confused the two great genres of classical Sanskrit literature; one becoming the subject of definition kavyam and the other sufficing to define the former (vakyam rasatmakam!). Viśvanātha, as has become conventional, discusses earlier definitions of poetry in his preface.258 The defence of his own definition is quite formal: no other suggested predicate (sagunlamkāram, e.g.) can be found invariably associated with kavyatvam, hence the vyapti (invariable concomitance) that ought to underlie every definition is violated. Viśvanātha himself accepts

254 We do not suggest that the two writers, although contemporaries-but separated by several hundred miles (Visvanātha hailing from Orissa), were familiar with each other's works; more likely both represent the 'state of the art' in the early fourteenth century; it is Visvanatha's version, far more traditional in form, that has come to be taken as 'authoritative'. 255 See V. RAGHAVAN, Śrngāraprakāśa, ch. 8; K. KRISHNAMOORTHY, What is Sāhitya, Mysore Orientalist, 1970, p. 57. 256 Unless the Sāhityamīmāmsa is properly attributed to Ruyyaka-which seems dubious on the ground of its content: see KANE, H.S.P., pp. 280-84; the term sāhitya may have been coined by Mukula (RAGHAVAN, op. cit., p. 85). 257 See DE, S.P.S.A., p. 18. 258 See S. BHATTACHARYA, Viśvanātha Kavirāja and his References to Forgotten Alamkāra Writers, JOIB, 13, fasc. 4(Ruyyaka and Abhinavagupta appear to have been the first to think it useful to consult the history of their subject, thus in a way shifting focus of attention from poetry to the definition of poetry (sastra).

The late theorists 283

consequences of his definition-denyiog the possibility of the third and lowest type of kāvya as accepted by the Dhvanikāra citrakavya) because it is defined precisely as that kavya lacking rasa.259 Of course the argument itself is self- serving: if citrakāvya is admitted as kāvya, Viśvanātha's own definition will be avyāpta.260 Kavya is located, explicitly, in the area of vakya, showing to what extent the drama had ceased to be thought of as a performance; this permits, sug- gests perhaps, to Viśvanātha a rapprochement à la śāstra, with the vākyasāstra par excellence: mimāmsā.(2,1), and with the mīmāmsa-nyaya notion of the syntactic unit, or mahāvākya, construed with the aid of the principles ākānksā, yogyatā, and āsatti (vrtti ad 2,1). Thus even the śāstraic basis of poetics, its grammatical doctrine of the three powers (sakti) of words, is determined ex- plicitly in the śāstraic universe of discourse. With Viśvanatha, the period of poetics beginning with Mammata, marked by a concern to regularize and codify alamkāravidyā as a śāstra, may be said to have reached its conclusion. Henceforth the subject-academically-has the dimensions of Visvanatha's treatise: purpose and definition of kāvya (1); vākya and the powers of words (2); rasa and the technique of characterization (3); kāvya and the types of dhvani (4); dhvani established vis-à-vis the other saktis and in reference to other views that limit its primacy (sp. Mahimabhatta; 5); nātya, its plot structure (6); doșa (7); guņa (8); rīti (9) and alamkāra(10). Nothing really important is left out; the work defines as marginal what is excluded.

16. The late theorists

Oddly, or perhaps perversely, the history of Indian poetics in the period following this synthesis of kavya and natya seems again to falkinto two distinct traditions-one concentrating almost exclusively on rasa (often in a context broader than that of natya, to be sure), and one almost as it were fixing on Viśvanātha's summary rejection of citrakāvya, seeking to demonstrate again the verbality of poetic diction, its inseparability from the notion of alamkāra (Appayyadīksita, Jagannātha). The late rasa theorists. These writers and works seem to relate directly to the original Nāțyaśāstra (via the systematic résumé of Dhanamjaya) and do not show the effects of the syncretism that has, by the time of Viśvanātha, inte- grated nātya into an overarching poetic. Somewhat earlier than Visvanatha is the most important of these writers, Śāradātanaya, author of the Bhāvaprakāśa(na). Referred to above in connec-

259 SD., 4, 1 and vrtti ad 4, 14. 260 So Jagannātha, Rasagangādhara, p. 12 (Benares 2 nd ed.)

284 E. Gerow . Indian Poetics

tion with its view that there are two Bharatas, the Bhāvaprakāśana is in- trinsically one of the interesting books of Indian poetics. In it are developed with grace and critical acumen some of the really puzzling and obscure aspects of Indian dramatic theory. Though Saradātanaya cannot be said to have been a greatly innovative writer,261 what we do find is a curiously circumstan- tial wisdom that infuses many of the dry technicalities of characterization and plot development with a vividness that permits us to situata them in the con- text of a real performance, and hence understand both their critical significance, and the playability of the dramas.262 It is as though this alone were sufficient to integrate and to rationalize functionally the many details of the older doc- trine, taking the place, almost, of a living tradition of dramatic performance in provoking fruitful speculation. Following Abhinavagupta in the main, he nevertheless rejects the central tenet-of the latter's theory, the functional equivalence of rasa and dhvani. Much less committed to a theoretical position, he accepts both the expressi- bility (vacyatva) of rasa, and the non-immanence of dhvani returning (it seems) to Nayaka's bhavakatvam and to the notion of tatparya as a vrtti.263 The notion of hava here becomes both that which links the elements of characterization (vibhāvas, anubhāvas etc.) with the dramatic effect (a postulate as old as Bharata), as well as the fact that is being developed and brought to a complete state through the intricate engine of the play's action, or plot development. What the samdhis are, are modes in this statement of rasa, through its neces- sary externals (a sequence of actions and their implicit oppositions). The Bhavaprakaśana is the main link we have with the lost traditions of dramatic performance .! ) The same noteworthy dependence on the rasa theory (Visvanatha can be seen as exporting it to kāvya, perhaps) is also seen in several post-Viśvanatha writers; where in addition, the influence of Bhoja is perhaps more discernable; or perhaps it is the increasingly erotic devotionalism of the times, in the prominence given to śrngāra rasa among the rest. The Rasārņavasudhākara of the Recarla King Singabhūpala (ca. 1330 A. D.)264 and the two well-known works of Bhanudatta (of Videha), the Rasamañjari and the Rasataranginī (ca. 1450) are notable more for their completeness than for their incisiveness, and obviously owe most of their inspiration to the Bhāvaprakāsana and to the Śrńgāraprakāśa of Bhojadeva. A completely new turn to the rasa theory (in its dramatic context) was given by the Vaisnava theologians of Bengal, notably Rūpa Gosvāmin, who took the preeminence of srngara among the rasas and boldly identified that rasa

261 "reproducing the substance of ... Bhoja's work": DE, H.S.P., II, p. 266. 262 See V. RAGHAVAN, Sanskrit Drama in Performance in Proceedings of the Honolulu Conference ... , passim. 263 Bhāvaprakāśanam (GOS 45, 1930), Introduction (by K.S.R. SASTRI SIRO- MANI), p. 14. 264 JOIB, 7, pp. 25-33 (Kane, H.S.P., p. 433).

The late theorists 285

with the sentiment of the worshipful Krsna bhakta, thus in effect turning the real world of religious concerns into a drama, wherein everyone enacts the play of Krsna and the gopis.265 This wholly uncompromising use of an aesthetic doctrine to state a theology, and in effect to state a new relation between the worshipper and his ethical world, is both unparalleled and yet proof of the continuing vitality of aesthetic traditions in India, It is beyond the scope of this essay to deal with Rupa's work (chiefly the Ujjvalanīlamani in this con- text) as theology, but it must be noted that the essential seriousness of aesthe- tics is here stated in terms that develop, while reminding one of, Abhinava- gupta's relation of theatre and the universe (rasāsvāda and brahmāsvāda). While Abhinavagupta seems content to explore this relation as an analogy (for rasāsvāda lasts only as long as the play, and is a glimmering only of moksa), Rūpa Gosvāmin appears to state the relation as a true equivalence, thus cancelling the separation between the religious domain and the play.266 These works activate the rasa into a principle of being, as well as one of response, and are very likely a mode through which the rasa is communicated to contem- porary Indians, especially in Bengal. Rasa at the same time it is employed to organize all literature (Viśvanātha) and oven all life (Rūpa) has with much less distortion also won the domains of the other fine arts, music, dance and painting. These also trace their histories back to Nātyaśāstra, and have at no time been wholly free of a rasa-influenced context. But no longer adjuncts to the drama, they come to be seen as integral art forms in their own right, and in each a suitable form of the rasa-aesthetic is developed to explain the or- ganization and communication of emotion. Indeed in these domains, which do not function through the mediate word, the attractiveness of rasa as an ex- planatory principle is increased. Literature has always been the most difficult area in which to demonstrate the rasa, owing perhaps to its essential lack of immediacy; but the imperium of rasa was not to be denied. Today, we could broaden even Viśvanātha's syncretism: there is not a single aesthetic or even devotional)expression in contemporary India that is not touched by this principle, and indeed were it only hyperbole we would have no qualms in asserting that to be an Indian is to know rasa. The more academic writers after Viśvanātha are concerned almost exclusively with the vexed question of the "powers" of speech, the vrttis, and (from the space devoted thereto) a demonstration anew of the power of alamkara. We are invited perhaps to conclude that history moves in circles. Appayya Dīksita, one of the giants of Indian pāndityam (var. Appa, Apya, Appaya), was a Tamil brahmin, a noted advaitin, whose family was likely

265 S.K. DE, Vaisnava Faith and Movement in Bengal, Calcutta 21961; Bhakti Rasa Sāstra of Bengal Vaișnavism, IHQ, 8, p. 643-88; A. GUHA, Rasa Cult in Caitanya Caritamrta, in Ashutosh Mookerjee Silver Jubilee Volume, III, p. 368. 266 See E. GEROW, Rasa as a Category of Literary Criticism, part 3, Proceedings of the Honolulu Conference, I.

286 E. Gerow . Indian Poetics

associated with the Saiva temple at Chidambaram.267 His approximate dates were ca. 1520-93868, argued on the basis of his patronage by several rulers whose inscriptions have been dated, notably Chinna Bomma of Velūr (Vellore in N. Arcot) and Venkata (either Venkata of Vijayanagar, or one of Penna- koņda, thus both in southern Andhra269). His works in poetics are three:(a fourth may have been discovered in frag- ment270): one, the (Vrtti Varttika, is devoted exclusively to the two 'basic' powers: abhidhā (denotation) and laksana (transference); it is 'poetic' in the sense that the book aims at establishing these two in secure foundation to a poetic expression that involves the third: vyanjana. It is widely presumed that the work is unfinished because it lacks the third chapter on dhvani271; on the other hand it is also judged a youthful work272, and if so, it may be complete in terms of its limited aim-not to "prove" again the dhvani (what could be regarded as less well established ?) but to establish clearly the proper relation between words in their "natural" and grammatical guise, and the "poetic" product.In a way, the work reads like the Dhvanyaloka backwards: given the dhvani, what must we account for in natural language that will establish the latter as a proper foundation for poetic expression? Appayya's second work, the Kuvalayananda, is an elaborate commentary on an earlier collection of figures, the Candraloka of Jayadeva273, and is largely responsible for its popularity. It contains the largest collection of figures of speech yet compiled (115,-fifteen more than the Candraloka itself). Used in schools in South India even today, it is the standard handbook on the subject, and it seems proper to judge it as such, elegant in definition and judicious in illustration, rather than an academic work. Appayya's third work Citrami- māmsā is a delight, being in purpose an effort to rescue citrakāvya from the disrepute (adhamakavyam) it has earned at the hands of the dhvani theory-but even more as argumentation: a witty discussion of a number of figures to show their intellectual interrelationships (chiefly their relations to simile and its structure) inthe work provides Appayya an occasion to justify or criticize the formulation of some very good medieval poetry. The work approaches more closely a stance of 'criticism' than any other (save possibly Jagannatha's). known to us. Appayya's aim throughout is to reestablish the claim that imagery itself is delightful (arthacitra)-that the process of its apprehension by the cultured mind is a sufficient ground for assigning it 'poetic' status-whether or not the work is infused with the larger, more portentous (now even world-

267 His grand-nephew was Nilakantha Dīksita, the well-known satirist (Ka- lividambaņam: Kāvyamālā Gucchaka 1). 268 KANE, H.S.P., p. 319. 269 DE, H.S.P., I, p. 222. 270 Ibid., p. 225. 271 Or vyakti: cf. the mangalaśloka, p. 1, NSP, ed. 272 DE, H.S.P., I, p. 223. 273 1250 A.D .: KANE, H.S.P., p. 292.

The late theorists 287

cosmic) issues of dhvani-rasa. Appayya seems at times almost a sceptic, one of those charming conservatives who delight in showing us how much we lose by taking ourselves (and our psyche) too seriously. The spirit of the classical period seems rekindled in his work.274 Apparently struck that a major writer could be so good without identifying the cosmic principle underlying all utterance, Jagannatha,275 the last of the great alamkarikas, a Tailanga brahmin, who graced the Moghul court of Shah Jahan, ca. 1620-65 A. D.276, devotes himself inordinately to ad hominem at- tacks on his predecessor, often seeming to adopt positions only because Ap- payya has concluded the opposite. One of his works (Citramīmamsakhandana) is entirely devoted to nit-picking Appaya's sometimes dogmatic interpretation of figures and examples.277 But the Rasagangadhara, on which his fame securely rests, though often revealing this trait278, is far more. Indians rank it with the Dhvanyāloka and Kāvyaprakāśa279; stylistically and intellectually it is indeed a classic. Returning to the theme that a single principle must determine or inform the content of kāvya, Jagannātha nevertheless seems to avoid aligning himself with any of the established points of view: attacking each and rejecting each (especially Mammata and Viśvanātha) in turn, in favour of his own intriguing (and (circular) definition: "poetry is words that produce pleasure" (p. 2). He even (argues (sub Viśvanātha, p. 12) that rasa is an improper predicate of poetry inasmuch as sabdacitra is admitted to be poetry but can have no rasa (by definition). The nominal character of the definition and argument is patent, but Jagannatha's arsenal of arguments is far from trivial. Combining the grace of a poet280 with the relentlessness of a naiyayika, the result always seems somehow less important than how it was arrived at. The Rasagangadhara, apparently (incomplete, is divided into two unequal sections, the first devoted to the types of kavya and to the rasas; the second to the types of dhvani (hence the discussion of vrttis) and to alamkara. This last is broken off in the midst of the discussion of the figure uttara, and Nageśabhatta's commentary also breaks off at that point.281 Jagannätha's interest in the figures is a function of his

274 See RAMARANJAN MUKHERJEE, Contribution of Appaya Diksita to Indian Poetics, Calcutta 1971 for a comprehensive and sympathetic account. 275 See V.A. RAMASWAMY SASTRI, Jagannatha Pandita, Annamalai Sanskrit Series 8, 1942: "best since Abhinavagupta". 276 KANE, H.S.P., p. 324. 277 See R. MUKHERJI (MUKHERJEE?), A note on the controversy between Appaya and Jagannatha, PAIOC, 24, fasc., 2, p. 261. 278 Especially in the lively discussion of Bhartrhari's verse viseşasmrtihetavah ... ", Benares ed., pp. 173ff., where he contests Appayya's view that the non-contextual (aprākāraņika) meanings of proto-puns (homonyms) are also produced by the vrtti abhidhā. 279 Kane, H.S.P., p. 321. 280 And he was a good one: Bhāminīvilāsa. . 281 KANE, H.S.P., p. 322.

288 E. Gerow . Indian Poetics

view of poetry: he judges them essential in producing the "pleasureableness" associated with the third vrtti: vyanjana; he (thus returns to Ruyyaka and Kuntaka, speaking of this inherent delight-intellectual in form-tvicchitty etc.) as something apart from and complementing the rasa-now a mere content or vyangyärtha in the poem. To this end, Jagannātha also assigns importance to the notion of "inspiration" or pratibha, that faculty whereby the poet constructs his figurative universe (as something different from the world and its everyday, self-interested construction); thus here also one of the favourite themes of the earlier ālamkārikas is resurrected.282 Pratibhā is an independent principle, not reducible to training (vyutpatti) or practice (abh- yāsa), which indeed may function without them (as in the case of genius or prodigy) and is always involved in their highest accomplishments. It is a suitable note on which to end this history of Indian poetics.

282 Supra, note 20.

BIBLIOGRAPHY

Abhinavagupta Sushil Kumar De, Some Problems of Sanskrit Poetics. K.L. Mukhopadhyay, Calcutta, 1959. The text of Kāvyāloka - Locana IV, pp. 236-265. Text (edition princeps) of Abhinavagupta's Locana (of Ānandavardhana's Dhvanyāloka). Uddyota 4 only;

Chapter. The Theory of Rasa in Sanskrit Poetics, pp. 177-235; gives text of with notes.

A. Bh. commentary on Bharata NS, ch. VI, with variants, pp. 219-235. Raniero Gnoli, The Aesthetic Experience According to Abhinavagupta. Serie Orientale Roma, XI, Rome, Istituto Italiano per i Medio ed Estremo Oriente, 1956. XXXII, 122: Text and translation of section of Abhinavabhārati that com- ments on Nātya Sāstra VI, prose after v. 33. Text in Roman transliteration, trans- lation into English; and App. I and II: Texts in Roman transliteration, translation into English of A. Bh.I, p. 37 and I, p. 36-37 (from M.R. Kavi ed. N with A.Bh. Gaekwad Or. Ser. XXXVI, 1926).

Agnipurāņa Manmatha Nath Dutt Shastri, Agni Puranam, a prose English Translation. Two volumes. Chowkhamba Sanskrit Series LIV, Varanasi, 1967. Rajendralal Mitra, Agnipurana. 3 volumes. Bibliotheca Indica, Asiatic Society of Bengal, Calcutta, 1873, 1875, 1878. Ānandavardhana Bishnupada Bhattacharja, Dhvanyāloka (Uddyota I) edited with English exposition. 2nd ed. rev. and enlarged. K. L. Mukhopadhyay, Calcutta, 1965. Pt. Durgaprasad and Kasinath Pandurang Parab, The Dhvanyaloka of Anan- davardhanāchārya, with the commentary of Abhinavaguptāchārya. Kāvya- māla Series, 25, The Nirnaya-Sagara Press, Bombay, 1891. Hermann Jacobi, Ānandavardhana's Dhvanyāloka, in Zs. d. deutschen mor- genl. Ges., Leipzig, 1902-03. Kārikās and Vrtti. Band 56: II (1902) pp. 392-412: introduction (in German); 56: III (1902) pp. 582-615: translation into German, Kap I. 1 - II. 23; 56: IV (1902) pp. 760-789: translation into German, Kap. II. 24-III. 9; 57: I (1903) pp. 18-60 translation into German, Kap III. 10 .- III.33 ; 57: II (1903) pp. 311-343 translation into German, Kap. III. 35-IV. 17. Hermann Jacobi, Ānandavardhana's Dhvanyāloka (Die Prinzipien der Poetik). Separately printed from ZDMG (Band 56, 57). (Separatabdruck aus Bd. LVI und LVII); Leipzig, 1903. Kārikās and Vrtti. 1-19 (introduction), 20-159 (translation into German, with footnotes Kapitel I. 1-Kapitel IV. 17). Badarinatha Jha and Shobhit Yadunath Misra, Dhvanyaloka with Didhiti commentary. Haridas Sanskrit Series 66, Chowkhamba Sanskrit Series, Varanasi,

Kuppuswami Sastri, Dhvanyaloka with Locana of Abhinavagupta and Kau- (3rd. ed.) 1964.

mudi of Udayottunga. 1st uddyota only, Madras, 1944. Appaya Dīkşita · Jagadisa Chandra Misra, Citramīmāmsā of Appaya Dīkșita with the Sudhā

290 E. Gerow . Indian Poetics

Commentary of Dharananda and Bharati (Hindi) Commentary. Kashi Sanskrit Series 207, The Chowkhamba Sanskrit Series, Varanasi, 1971. pp. 41 (introduc- tion Hindi). Richard Schmidt, Appayadikşita's Kuvalayānandakārikās; ein Indisches Kom- pendium der Redefiguren, mit Asadhara's Kommentar. Berlin, 1907. Pandit Sivadatta and Kasinath Pandurang Parab, The Chitramīmāmsa of Appadīkshita and the Chitramīmāmsā-Khandana of Jagannāth Pandit. Kā- vyamālā series 38, Nirņaya-Sāgara Press, Bombay, 1893. -. The Vrttivārtika of Appaya Dīkshita. Kāvyamālā Series 36, Nirņaya- Sāgara Press, Bombay, 1893. Aśvaghoșa E.H. Johnston, The Buddhacarita; or Acts of the Buddha, edited and translated with notes. 2nd edition, 2 volumes; first publication: 1935-36, Munshiram Manoharlal, Delhi, 1972. Pt. 1: Sanskrit Text. Pt. 2: Cantos I to XIV translated into English from Sanskrit, supplemented by Tibetan version. Critical app., introduction, and notes in English.

Ballāla Louis H. Gray, The Narrative of Bhoja (Bhojaprabandha) translated from Sanskrit. American Oriental Series, Vol. 34, American Oriental Society, New Haven, 1950. 1-11 (introduction, etc.), 13-97 (translation into English), 98-102 (Poets quoted in Bhp.), 103-109 (metres, lexicographic additions). Kasinath Pandurang Parab, The Bhojoprabandha of Ballala. 2nd revised edition. Nirnaya-Sagara Press, Bombay, 1904. 1-80 (text), iv (index). Bhāmaha Kāvyālaņkāra: see also under Vidyānātha and under Udbhața: R. Gnoli. P. V. Naganatha Sastry, Kāvyālankāra of Bhāmaha, ed. with English transla- tion and notes. 2nd edition. Motilal Banarsidas, 1970. xx, 134. D.T. Tatacharya Siromani, Bhāmaha's Kāvyālamkāra, with Udyānavrtti. Tiruvadi, 1934. K. P. Trivedi, Pratāpa Rudra, appendix VII: Bhāmahālamkāra. Bombay, 1909. Bhānudatta Rasatarangini: see under Bharata Muni: P. Regnaud. Rama Sastri Tailanga, Rasamañjarī by Bhanudatta with the commentaries Vyangyārthakaumudī of Ananta Paņdit and Prakāśa of Nageśa Bhațța. Benares Sanskrit Series Nos. 83, 84, 87; Braj B. Das and Company, Benares, 1904 - Published in three fascicules. 9 (vișayasūcī), 6 (śloka index), 9 (pratipadyavi- șayasuci), 248 (text and the two commentaries). Jivananda Vidyasagar, Rasa-mañjarī in Kāvya-samgraha. 2nd edition; Cal- cutta, 1886. Bharata Muni Rene Daumal, Bharata, L'Origine du Theatre, la Poesie et la Musique en Inde; traductions de textes sacrés et profanes. Editions Gallimard, Paris, 1970. L'ori- gine du Théâtre de Bharata, pp. 13-36, translates into French the first chapter of Nāțyaśāstra, with notes. Manomohan Ghosh, The Natyasastra, a Treatise on Hindu Dramaturgy and Histrionics ascribed to Bharata-Muni, completely translated for the first time from the original Sanskrit with an introduction and various notes. I, chapters 1 to 27; Bibliotheca Indica, 272. Calcutta. The Royal Asiatic Society of Bengal, 1950; II: chapters 28-36, Calcutta, 1961.

Bibliography 291

-, -. The Natyasastra, ascribed to Bharati-Muni, edited, introduction, v. 1. I, chapters I-XXVII. Manisha Granthalaya, Calcutta, 1967. 1st edition. Manomohan Ghosh: The Natyasastra, ascribed to Bharata-Muni edited, II: chapters XXVIII-XXXVI; Bibliotheca Indica 272 A, Calcutta, 1956. [vol I publ. Calcutta 1967]; xxx (introd. & index), 1-218 (chs. XXVIII-XXXVI). Joanny Grosset, (édition critique). Traité de Bharata sur le Théâtre, texte Sanskrit, i, pt. i, Paris, 1898, (ch. 1-14 only, not completed), in Annales de l'Université de Lyon. M. Ramakrishna Kavi, (ed.) Nātyasāstra of Bharatamuni; with the commentary Abhinavabharatī by Abhinavaguptācharya. I: chapters 1-7 first edition 1926; second edition Rev. K.S. Ramaswami Sastri Siromani 1956; XXXVI Gaek- wad's Oriental Series, Oriental Institute, Baroda II, chapters 8-18. 1st edition 1934; LXVIII ibidem, III, chapters 19-27. 1st edition, 1954; CXXIV ibidem,

CXLV ibidem. IV, chapters 28-37, ed. Kavi, M. Ramakrishna and Pade, J.S. 1st edition 1964;

Paul Regnaud, La Metrique de Bharata; texte Sanscrit de deux chapitres du Nāțyaçāstra, publié pour la première fois, et suivi d'une interprétation française, pp. 65-130, in Annales du Musée Guimet, 2, Lyon, 1881. Preface, final 68 slokas of ch. XV ("intitule Chandovidhana"), and ch. XVI NS text in Romanized transcription, translation into French, notes. -, -. La Rhétorique Sanskrite, exposée dans son développement historique, et ses rapports avec la Rhétorique classique, suivie des textes inédits du Bhāratīya- Nāțya-Çāstra (sixième et septième chapitres) et de la Rasataranginī de Bhānu- datta. Ernest Leroux, Paris, 1884. 1-52 (livre ler: les différents sens des mots), 53-258 (la Poésie), 267-394 (le Rasa), 1-70 (Textes Sanskrits: NS adh. 6 and 7). Rasatarangini: tarangā I-VIII, text, Romanized, notes, variants: 43-70. Batuknath Sarma, and Baladev Upadhyay, Natya Sastra (in 36 chapters) and Abhinavabharatī. Chowkhamba Sanskrit Series, Benares, 1929 Madhusudan Shastri. Natyashastra of Bharatamuni, with the comm. Abhinava- bhāratī of Abhinava Guptāchārya, and Madhusūdanī (Sanskrit) and Balakrīda (Hindi) coms. I, Samskrita Sahitya Anusandhan Series: 4, Banaras Hindu University, Varanasi, 1971; (introduction, alphabetical index ślokas, index titles, Hindi introd., text of NS adhs. 1-7, relev. Abh. and other two coms.). Pt. Sivadatta and K.P. Parab, Nāțyaśāstra and Abhinavabhāratī. Kāvyamālā Series, 42. Nirņaya-Sāgara Press, Bombay, 1894, adhy. 1 to 37. Bhartrhari K. V. Abhyankar and V.P. Limaye, Vakyapadiya of Bhartrhari. Text of three kāndas, v. ls. University of Poona Sanskrit and Prakrit Series II, Poona, 1965 (introduction, text, appendices, glossary, kārikās, quotations, sources etc.). Pt. G. Sastri Manavalli, Vakyapadiya, A Treatise on the Philosophy of Sanskrit Grammar by Bhartrhari, with a commentary by Punyaraja. Benares Sanskrit Series 11, 19 and 24. Braj B. Das and Company, Benares, 1887. Raghunath Sharma, Vakyapadīyam, with the commentary Ambakartrī (Sans- krit). Sarasvati Bhavan Granthamala 91, Varanasi Sanskrit Vishvavidyalaya Varanasi, 1963, 68. Pt. 1: Brahma Kanda, 1963, 1-269 (text and commentary), 270-72 (śloka index). Pt. 2: Vākya Kaņda, 1968, 1-582 (text and commentary), 583-591 (śloka index). K.A. Subramania Iyer, Vakyapadiya of Bhartrhari with the commentary of Helaraja, Kanda III, Part 1. Deccan College Monograph Series 21, Poona, 1963. 1-370 (text of III:I, seven samuddeśas with Helarāja's commentary); 371-407 (indices: verses, quotations, authors, ideas etc.). -, -, -. The Vākyapadīya of Bhartrhari, Chapter III, part i, English transla- tion. Deccan College Silver Jubilee Series 71, Poona, 1971.

292 E. Gerow . Indian Poetics

Bhațți Ramavadh Pandeya, Bhatțikāvyam with commentary (Sanskrit) Jayamangalā of Jayamangala, comm. (Sanskrit) Mugdhabodhini of Bharatamallika, and comm. (Hindi) Kaśika. Part I: sargas I to IV, Motilal Banarsidas, Delhi, 1968. Part II: sargas V to VIII, with Hindi explanation by R. Shukla, Motilal Banarsi- das, Delhi, 1971. G.G. Leonardi, Bhatțikāvyam, translation and notes. Leiden, E.J. Brill, 1972. Bhavabhūti Sir R.G. Bhandarkar, and V.V. Mirashi, Mālatīmādhavam nāma prakaraņam; Mahākavi Bhavabhūti praņītam; Jagaddharakrtațīkayā samāvetam: edited with notes etc. Bombay Sanskrit and Prakrit Series, XV, Bhandarkar Oriental Institute, 3d edition, Poona, 1970. Ludwig Frike, Malati und Madhava; ein Indisches Drama von Bhavabhūti. Translation into German. Reclam's Universal Bibliothek, 1884, Leipzig (? 1883). R. K. Tailanga (ed.), Mālatī Mādhavam, Nirnayasagar Press, Bombay, 1936. Bhoja, King of Dhārā Anundoram Barooah, Sarasvatī-Kanthabharana, five paricchedas. Sanskrit text, notes. 1st publ. 1880; reprinted by Publication Board, Gauhati, Assam, 1969. T.R. Chintamani, Sarasvatīkaņțhabharana of Bhojadeva, Madras University Sanskrit Series 11, Madras, 1937; with index of sūtras and Bhoja-Panini con- cordance. K. Durgaprasad and V.L. Pansikar, Sarasvatī Kanthabharana; with commen- tary of Ratnesvara (i-iii) and Jagaddhara (iv). Nirnaya-Sagara Press, Bombay, 1925, 1934. G.R. Josyer, Mahārāja Bhojarāja's Śrngāraprakāsa, the Great Tenth Century work on Sanskrit and Prakrit Rhetoric. G.S. Josyer, Mysore. I: Prakāśas I to VIII. 1955; II: Prakāśas IX to XIV. 1963; ii (Prakaraņa verses rendered); III: Prakāśas XV to XXIV, 1969; IV: Prakāśas XXV to XXXVI, n.d. V. Raghavan, Bhoja's Śringāra Prakāśa (a complete publication of the author's exposition of Bhoja's magnum opus). Punarvasu, Madras, 1963. with translation, index etc. Caņdīdāsa Sivaprasada Bhattacarya. Kāvyaprakāsadīpikam Caņdīdāsaviracitam; III. Sa- rasvati Bhavana Granthamala 46, Varanasi Sanskrit Vishvavidyalaya, Varanasi, 1965.

Daņdin S.K. Belvalkar and R.B. Raddi, Dandin's Kavyādarsa edited with new Sanskrit commentary and English notes. II, 1st half: pariccheda 2, up to 368: Bombay Sanskrit and Prakrit Series, LXXV, The Department of Public Instruction, Bombay, 1920; II: 2nd half: Notes to pariccheda 2 pp. 67-220 (Arabic). ibidem, Bombay, 1920. O. Böhtlingk, Dandin's Poetik (Kāvjādarça), Sanskrit und Deutsch, Leipzig 1890. (Skt. text with running German translation). Georg Bühler, The Daakumāracharita. I: Pūrvapīthika ucch. 1-5 and ucch. 1-3, notes. Bombay Sanskrit Series X, Government Central Book Depot, Bom- bay, 1887. -, -. Dandin. Die Zehn Prinzen, ein altindischer Roman. 13 chapters. Jakob Hegner, Köln, 1971. N.B. Godabole and K.P. Parab, The Dasakumāracarita of Dandin, with three commentaries, viz. the Padacandrikā of Kavīndra Sarasvatī, the Bhūșaņa of

Bibliography 293

Śivarama and the Laghudīpika; edited with various readings. Nirnaya Sagara Press, Bombay, 1903. 4th rev. edition. J.J. Meyer, Dandin's Daçakumāracaritam; die Abenteuer der zehn Prinzen; ein altindischer Schelmenroman. Translation into German. Lotus Verlag, Leipzig, 1902. Arthur W. Ryder, Dandin's Dasha Kumara Charita, The Ten Princes, translated from the Sanskrit. The University of Chicago Press, 1927; republished Jaico Publishing House, Bombay, 1956. G. Harihara Sastri, Avantisundarī Kathāsāra. Kuppuswami Sastri Research Insti- tute, Mylapore, Madras, 1957. Pt. Rangacharya Shastri Vidyabhusana, Kāvyādarsa of Dandin, ed. with an original commentary, Prabha. Government Oriental Series, A, 4; Bhandarkar Oriental Research Institute, 1970. Pt. Premachandra Tarkabagisa, The Kavyadarsa of Śrī Dandin, with a comm. Malinyapronchanī (Sanskrit), ed. and comm. Bibliotheca Indica, Asiatic Society of Bengal, 30, 33, 38, 39, and 41, Calcutta, 1863. Anantalal Thakur and Upendra Jha. Kāvyalakșaņa of Dandin (also known as Kāvyādarśa), with comm. Ratnaśrī of Ratnaśrījñana. Mithilavidyāpīțha Gran- thamalā, I: Pracinācaryagranthavali section, IV. Mithila Institute of Post Graduate Studies and Research in Sanskrit Learning, Darbhanga, 1957.

Dhanamjaya T. Venkațāchārya, The Daśarūpaka of Dhanamjaya with the commentary Avaloka by Dhanika, and the sub-commentary Laghuțīkā by Bhatțanrsimha. The Adyar Library Series 97, Madras, 1969. Fitz-Edward Hall (with Avaloka), Bibl. Indica 1861-65. Transl. into English, with transliterated text, introd. and notes, by G.C.O. Haas in Columbia Univ. Indo-Iranian Series, New York, 1912.

Harșa Bak Kun Bae, Sri Harsa's Plays; translated into English with full Sanskrit Text. (Texts of Nn Rv & Pd, with notes and running English translation). Nāgānanda, Ratnavali & Priyadarsika. Indian Council for Cultural Relations and Asia Publishing House, Bombay, London, New York, 1964.

Hemacandra Rasiklal C. Parikh and V.M. Kulkarni, Kāvyānuśāsana [with Alaņkāracūdā- mani and Viveka], with two anonymous Tippanas. Sri mahavira Jaina Vidyalaya, Bombay, 1964, 2ed. (appendices: udāharaņa index, Skt. chāyās of Prakrit verses, name, place index etc.). MM Pandit Sivadatta and Kasinath Pandurang Parab, The Kāvyānuśāsana of Hemachandra with his own gloss. Kavyamāla series 70 (sic; correctly 71), Nir- naya-Sagara Press, Bombay, 1910.

Jagannātha Pandit Durgaprasad and Kasinath Pandurang Parab, The Rasagangadhara of Jagannātha Pandita with the commentary of Nageśa Bhațta. Kāvyamālā series 12, Nirnaya Sagara Press, Bombay, 1888. ed. G. Sastri (with Nagoji's comm), Benares Skt. Series 1885-1903.

Jayadeva Duncan Greenlees, The song of divine love, Gita Govinda. Translated into English poetry in 1945, with a life of Jayadeva, and a running comm. Madras, Kalakshetra Publications, 1962.

294 E. Gerow . Indian Poetics

V.M. Kulkarni, Jayadeva's Gitagovinda with Mananka's commentary. Lalbhai Dalpathbhai Series 7, Bharatiya Sanskriti Vidyamandira, Ahmedabad, 1965 (with Mananka's Tippanika). Jinendrabuddhi Swami Dwarikadas Shastri and Pt. Kalikaprasad Shukla, Nyāsa or Pañcika Commentary of Ācārya Jinendrabuddhipāda and Padamañjarī of Haradatta Misra on the Kāśikāvrtti (Comm. on the Atādhyāyī of Pāņini) of Vāmana - Jayāditya. Part I to VI, 1965-67. Pracya Bharati Series 2-7, Varanasi.

Kșemendra Aucitya-vicāra-carcā, Kāvyamālā Gucchaka series 1, 1886.

Kuntaka Sushil Kumar De, The Vakrokti-Jivita, a treatise on Sanskrit Poetics by Rā- jānaka Kuntaka, with his own commentary; ed. with notes. 3d. rev. ed. K.L. Mukhopadhyay, Calcutta, 1961.

Mahimabhațța T. Ganapati Sastri, The Vyaktiviveka of Rājānaka Mahimabhațța with introd. notes and an anonymous comm. (attributed to Ruyyaka), Trivandrum Sanskrit Series 5, 1909.

Mammața Sivaprasad Bhattacaryya, The Kavyaprakasa of Mammata, with the commen- tary of Śrīdhara (Kāvyaprakāśaviveka), edited with introduction and notes

II. ibidem XV. I. Calcutta Sanskrit College Research Series VII. Calcutta, 1958. Ullasas 1 to 4.

R.C. Dwivedi, The Poetic Light: Kāvyaprakaśa of Mammața, Text, with translation [into English], and Sampradāya prakāśinī Comm. of Šrīvidyācakra- vartin, Text, Comm., transl. of text, notes; I, 1966; II, 1970, Motilal Banarsidas, Delhi. Sir Ganganatha Jha, MM Dr., The Kavyaprakasha of Mammata with English translation (revised). I. Bhāratīya Vidyā Prakāshan, Varanasi, 1967. Chapters I to VII, entire Kārikāvalī (142). (Text and running English translation). II. ibidem, 1967. Chapters VIII to X. (Text and running English translation), 491-504 (index of Sanskrit terms), lix (appendices, Skt. version of Prākrit verses, index of illustrative verses). A.B. Gajendragadkar, The Kāvyaprakasa of Mammața, 1st, 2nd, 3d and 10th ullasas, edited and translated into English. 3d edition, rev. S.N. Gajendragadkar, Popular Prakashan, Bombay, 1970 (with running English translation). Sivaprasad Bhattacharya, The Kavyaprakasa of Mammatacarya with comm. Dipika of Candīdasa ed. Fwd. Notes. I: Princess of Wales Sarasvati Bhavana 46, Govt. Skt. Library, Benares, 1933, ull. 1-4; II Sarasvati Bhavana Grantha- mala Series 46, Varanasi Skt. University, Varanasi, 1965. pp. 137-196. ull. 4; III ibidem 46, Varanasi, 1965, pp. 197-570; ull. 5 to 10. V.R. Jhalakikar and R.D. Karmarkar, Kāvyaprakāsa: with Balabodhi Comm. (comm. by Jhalakikar). Poona, Bhandarkar Oriental Institute, 1965.

Rājaśekhara C.D. V. Dalal and Anantakrishna Sastry, Kāvyamīmāmsā, a work on poetics by Rajasekhara. Gaekwad's Oriental Series I, Baroda 1st publ. 1916; reprinted 1924. Translated into French by Louis Renou and Nadine Stchoupak, Paris, 1946.

Bibliography 295

Rudrabhațța R. Pischel, Rudrața's Çrngāratilaka and Ruyyaka's Sahrdayalīlā. C.F. Haessler, Kiel, 1886.

Rudrața Pandita Durgaprasada and Kasinath Panduranga Paraba, The Kāvyālankāra (a treatise on Rhetoric) of Rudrata with the commentary of Namisādhu. Kā- vyamālā series 2, Nirnaya-Sagara Press, Bombay, 1886. Rupagosvāmin Kedarnath and V.L. Pansikar, Ujjvalanilamani with Locana rocani of Jiva Gosvāmin and Ānanda-candrikā of Viśvanātha Cakravartin, Nirnaya Sagar Press, Bombay, 1913. Ruyyaka MM Pandit Durgaprasad and Kasinath Pandurang Parab, The Alankarasarvasva of Rājanaka Ruyyaka with the commentary of Jayaratha. Kāvyamālā series 35, Nirnaya-Sagara Press, Bombay, 1893. Hermann Jacobi, Ruyyaka's Alamkārasarvasva: ZDMG, 1908, pp. 289-336, 411-58, 597-628; reprinted in Hermann Jacobi, Schriften zur Indischen Poetik und Ästhetik, Darmstadt, 1969. Kumari S.S. Janaki and V. Raghavan, Alamkara-Sarvasva of Ruyyaka with Sanjīvanī, Comm. of Vidyacakravartin, text and study. Mehar Chand Lachhman Das, Delhi, 1965. Sāgaranandin The Nāțakalakşaņaratnakosa of Sagaranandin; ed. Myles Dillon, London, 1937. Śāradātanaya Yadugiri Yatiraja, (Swami of Melkot) and K.S.R. Sastri, Bhāvaprakāśana, ed. with an introduction and indices. Gaekwad's Oriental Series No. 45, Oriental Institute, Baroda 2d ed., 1968. Śińgabhūpāla Kalikumar Datta Shastri, Nāțaka Paribhāșā of Šingabhūpāla. Skt. Sahitya Pa- rishat series 30, Calcutta, 1967. Udbhața Raniero Gnoli, Udbhața's Commentary on the Kāvyālamkāra of Bhāmaha. Serie Orientale Roma XXVII, Istituto per il Medio ed Estremo Oriente, Roma, 1962. Texts of Udbhata's Vivarana (as recovered), fragments of the Raghuvansa recovered with above. Text with Pratiharenduraja's comm. by M.R. Telang, Nirnaya Sagara Press, 1905, 1915 (this edition is useful for its comm., but some verses occurring in the comm. are given mistakenly as Kārikā verses). Ed. N.D. Banhatti, with the comm. of Pratīhārendurāja, Bombay Skt. Series, Poona 1925. Our references are to Telang's edition, unless otherwise indicated. Vāgbhața I Pandit Śivadatta and Kasinath Pandurang Parab, The Vāgbhațālamkāra of Vāgbhața with the commentary of Simhadevagaņi. Kāvyamālā series 48, Nir- naya-Sagara Press, Bombay, 1895. Jīvānanda Vidyāsāgar (Bhattacarya), Vāgbhațālamkāraḥ Śrīvāgbhațaviracitah. Six Paricchedas. 4th ed., rev., enlarged and annoted by Vidyabhusana, Asubodha and Vidyaratna, Nityabodha, and published by them, Calcutta, 1917.

296 E. Gerow . Indian Poetics

Vāgbhața II Pandit Sivadatta and Kasinath Pandurang Parab, The Kāvyānuśāsana of Vāgbhata, with his own gloss. Kāvyamālā series 43, Nirnaya-Sagara Press, Bombay, 1894.

Vāmana Carl Cappeller, Vamana's Lehrbuch der Poetik. Kāvyālamkāravrttiḥ. Hermann Dufft, Jena, 1875 (to 5, 2, 92). Ratna Gopal Bhatta, Kāvyalankāra Sutras with glossary by Paņdit Vāmana and a commentary called Kāvyālankārakāmadhenu by Sri Gopendra Tripura- hara Bhupala; Adhikaraņa 1, adhyāya 1 to Adhikaraņa 5, adhyāya 2. Benares Sanskrit series 134 and 140; Braj B. Das and Co. and Chowkhamba Sanskrit Book Depot, Benares, 1907, 1908. Vidyānātha K.P. Trivedi, The Pratāparudrayasobhūșaņa of Vidyānātha, with the commen- tary Ratnapaņa of Kumārasvāmin (son of Mallinātha), and an appendix con- taining the Kavyālankāra of Bhamaha. Bombay Sanskrit and Prakrit Series LXV, Department of Public Instruction, Bombay, 1909. Vișņudharmottara Purāņa Priyabala Shah, Vișnudharmottara Purana: Third Khanda, critically edited with notes etc. I: Text, critical notes pp. 411 G.O.S., cxxx. 1958; II: Introduc- tion, appendices, indices, etc. pp. 243, 92, GOS cxxxvii, 1961. Oriental Institute, Baroda, India. Rajendralal Mitra: Vișnudharmottara, 3 vols., (Bibliotheca Indica) 1873, 1876, and 1878. Translation into English in two vols, by Manmathanath Datta. Calcutta, 1903-4. Viśvanātha P.V. Kane, The Sāhityadarpaņa: Paricchedas I, II, X, Arthālankāras with exhaustive notes. 1st ed. Motilal Banarsidas, 1965. J.R. Ballantyne and Pramadadasa Mitra, The Mirror of composition, a treatise on Poetical Criticism, being an English translation of the Sāhitya-darpana of Visvanatha Kaviraja. Baptist Mission Press, Calcutta, 1875. In 10 chapters. Durgaprasada Dviveda: The Sāhitya-Darpaņa by Viśvanātha Kavirāja, with the Commentary of Rāmacharaņa Tarkavāgīśa Bhattāchārya; 6th ed: Nirņaya Sagar Press, Bombay, 1936 (also the sections on Poetics in Agni Purāna).

INDICES

Numbers in parentheses refer to footnotes on the pages indicated by the preced- ing number; where no parenthesis is given, the reference is to the body of the text. The main reference (sub works and classical names) is given in hyphenated form (e.g. 246-51) and refers to the locus of principal discussion in the text; other references are ancillary.

Index of Names (Classical)

Abhinavagupta 226 (34); 230 (57); 233 Bhoja 230 (55); 231 (61); 251; 259; 260; (69); 246 (119); 249 (127), (130); 250 268; 269-71; 272; 280; 284 (261); 284 (135); 251 (136); 253; 259; 260; 263 cf. 1) Sarasvatīkaņthābharaņa 2) (170); 264-8;269;271;272;273(217); Śrngāraprakāśa 281; 282 (258); 284; 285 Chinna Bomma 286 cf. 1) Locana; 2) Abhinavabhāratī Ānandavardhana 230 (53); 250-8; 260; Daņdin 217 (1); 224; 225; 226-33; 235; 236; 237; 240; 241; 242; 244 269; 270; 273 cf. 1) Kāvyādarśa; 2) Daśakumāraca- cf. Dhvanyāloka rita 3) Avantisundarīkathā Appayadikşita 232 (67); 258; 260; 283; Dhanamjaya 227; 260; 263-4; 266(184); 285-7; 287 (278) cf. 1) Kuvalayānanda; 2) Citrami- 269 (194); 283 cf. Daśarūpaka māņsā; 3) Vrttivārttika Dhanika 239 (93); 263; 264 (173) Aśvaghoșa 222; 229 Dharmakīrti 224; 229 (50); 233 cf. 1) Buddhacarita; 2) Saundaranan- Dhvanikāra 253; 264; 269; 283 da cf. Dhvanyāloka Ballāla 220 (7) Dińnāga 224; 233 cf. Bhojaprabandha Gārgya 221 (15) Bāna 222 Gopendra Bhūpāla 236 (85) Bhāmaha 221 (13); 222 (20); 223; 224; Harşa 251 224 (26); 225; 226-33; 234 (76); 234; Hemacandra 259; 260; 278-9 235; 236; 238; 240 (98); 244; 262; 282 cf. Kāvyānuśāsana cf. Kāvyālaņkāra Jagannātha 258; 260; 283 (260); 283; Bhānudatta 284 286; 287-8 cf. 1) Rasamañjarī; 2) Rasataranginī cf. 1) Rasagangādhara; 2) Bhāminīvi- Bharata 223; 225; 226-7; 229; 231; 231 lāsa (59); 245-50; 264; 268; 279; 284 Jahan (Shah) 287 cf. Nāțyaśāstra Jayadeva 252 Bhäravi 222; 232 (66) I cf. Gītagovinda 260; 286 Bhartrhari 287 (278) II cf. Candrāloka cf. Vākyapadīya Jayapīda 233 (Bhatta) Lollața see Lollața Jinendrabuddhi 228; 233 Bhațanāyaka 263; 265; 266 (183); 269 cf. Nyāsa (194); 269; 284 Kālidāsa 229 Bhattatauta 265; 269 cf. Kumārasambhava; Meghadūta; Bhațțendurāja 261 (163) Vikramorvaśīya Bhavabhūti 220 (9) Kșemendra 238 (89); 260 cf. Mālatīmādhava cf. Aucityavicāracarcā

298 E. Gerow - Indian Poetics

Kuntaka 258; 259; 260; 262-3; 274; (169); 271; 273 (219); 276; 277; 278 275; 278 (235); 282; 288 (238) cf. Vakroktijīvita cf. Kāvyālaņkāra Lollata 226 (34); 227; 264 Rūpagosvāmin 259; 260; 268 (190); 270 Māgha 222; 232 (66) (207); 284-5 Mahimabhatta 233 (69); 244; 258; 259; cf. Ujjvalanīlamaņi 260; 262; 268-9; 270; 283 Ruyyaka 232 (68); 238; 259;272;274-8; cf. Vyaktiviveka 280; 282 (256), (258); 288 Mallinatha 280 cf. Alamkārasarvasva cf. Tarala Šāntarakșita 224 (28); 228; 233 Mammața 226 (34); 230 (56); 232 (64), Śāradātanaya 226 (36); 239 (93); 259; (65); 233 (69); 239; 241; 258 (152); 283-4 259; 270; 270 (204); 271-4; 275; 277; cf. Bhāvaprakāśa(na) 278; 279; 280; 281; 282; 283; Śingabhūpāla 284 287 cf. Rasārņavasudhākara cf. 1) Kāvyaprakāśa; 2) Šabdavyā- Śriśańkuka 265; 269 pāraparicaya Udbhața 226 (34); 233-8; 239; 241; 242; Medhāvin 221 (13); 229 261; 262 Mukula 258 (152); 262; 282 (256) cf. Kāvyālaņkārasārasamgraha cf. Abhidhāvrttimātrkā Vāgbhața 259; 278-9 (I and II) Nāgesabhațța 287 cf. I) Vāgbhațālaņkāra; II) Kāvyā- Namisādhu 229 (48a) nuśāsana Narasimha 280 Vāmana 227; 231; 233-8; 239; 240; 273 Nīlakaņțhadīkșita 286 (267) Pānini 221; 222; 263 (217)

cf. Așțādhyāyī cf. Kāvyālaņkārasūtravrtti Varāhamihira 239 Paramara Muñja 264 Venkata 286 Prabhākara 254 Pratāparudradeva 280 Vidyādhara 280-1 cf. Ekävali Pratīhārendurāja 234 (75); 258 (152); Vidyānātha 270; 271; 280-1; 282 261-2 cf. Pratāparudrayaśobhūşaņa Rājaśekhara 259; 260-2; 269; 271; 274 Viśvanātha 259; 269; 271; 274; 278; cf. 1) Kāvyamīmāmsā; 2) Karpūra- 279; 281-3; 284; 285; 287 mañjarī Rudrabhatța 271; 278 (238) cf. Sāhityadarpaņa

cf. Śṛńgāratilaka Yaska 221; 222 cf. Nirukta Rudrața 227; 232 (68); 238-45; 263 Yuvarāja I 260

Index of Names (Contemporary or Western) Aiyar K.A.S. 222 (20) Aklujkar A. 247 (119) Chomsky N. 243 (111) Dasgupta S.N. 218 (2); 235 (83); 251 Aristotle 248; 250; 255; 267; 268 Ballantyne J. (and Mitra P.) De S.K. 218 (2); 219 (4); 220 (9); 221 (137) 281 (253) Banhatti N.D. 235 (82) (17); 222 (20); 224 (25); 225(33), (34);

Bhattacharjee B. 236 (86) 226 (36); 228 (41), (44); 229 (49), (50);

Bhattacharya S. 231 (60); 250 (133); 253 230 (53); 231 (60); 232; 233 (74); 234

(143); 270 (207); 282 (258) (75); 235; 235(80),(81),(83); 237(88);

Bhattacharya S.M. 225 (31) 238 (89); 240 (98); 242 (108); 244; 247; 249 (128); 250 (135); 253 (142); Bhattacharya S.P. 258 (152) Byrski M.C. 248 (123) 260 (153), (155), (156); 261; 261 (159),

Cary J. 275 (228) (160), (163); 262 (165), (166), (167); 263 (170), (171); 264 (173), (174),

Indices 299

(175), (176), (177), (178); 265 (181); Krishnamoorthy K. 241 (103); 252 266 (185), (186); 269 (198); 270 (201), (202), (204), (205), (206), (207); 271; (141); 253 (143), (144); 261 (161); 282 (255) 271 (212); 272 (213), (214), (215), Kulkarni V. 279 (241) (216); 274 (220), (221), (223), (224); Kumar S. 235 (84) 279 (242); 280 (245), (249), (250); 281 Kunhan Raja C. 250 (134) (251), (252); 282 (257); 284 (261); 285 (265); 286 (269), (270), (272) Kunjunni Raja K. 253 (144); 254 (145), (146) Diwekar H.R. 224 (26); 226 (36); 229 (50); 272 (216) Kuppuswami Shastri 222 (20); 233 (69); 266 (183) Eliot T.S. 267 (187) Lahiri P.C. 225 (31); 231 (58); 235 (83) Emeneau M. 218 (2) Lévi S. 248 (123); 280 (248) Gerow E. 219 (4); 221 (17); 222 (18); 224 Lienhard S. 280 (248) (25); 228 (47); 230 (52); 239 (94); 242 Masson J. and Patwardhan M. 247 (119) (107); 243 (111); 246 (119); 247 (121); 249 (128); 254 (147); 276 (232); 285 Moorty K.K. 230 (55) Mukherjee R. 287 (274), (277) (266) Mukherji A.D. 252 (140) Ghosh M. 225 (33) Murthy K. 222 (20) Giri K. 218 (2) Nandi T.S. 250 (135) Gnoli R. 233 (70); 234; 235; 249 (127); Nobel J. 225 (29) 264 (176) Pandey K.C. 253 (143); 261 (163); 264 Gode P.K. 239 (94); 280 (246) Gonda J. 221 (11); 224 (26) (173), (175), (176) Pathak K.B. 233 (72) Gray L. 220 (7) Raghavan V. 222 (20); 225 (31), (34); Guha A. 285 (265) 227 (38); 230 (55); 231 (58); 235 (83), Haas G.O. 264 (173) (84); 238 (89); 243 (111); 244 (113); Harris J. 240 (97); 242 (105) Hiriyanna M. 246 (119); 250 (132); 264 247 (119); 260 (154); 269 (193), (199); 270 (200); 270; 282 (255), (256); 284 (176); 268 (190) Ramaswamy Sastri V.A. 287 (275) (262) Hooykaas C. 229 (50) Ingalls D.H.H. 228 (47); 244 (115) Redfield J. 256 (150) Jacobi H. 224 (25), (26); 227 (39); 232 Regnaud P. 249 (128) (66); 233 (71); 253 (143); 263 (171); Renou L. 219 (4); 220; 220 (5), (6), (10); 274 (223); 276 221; 221 (11); 228 (47); 241 (101); 243 Jain R.J. 272 (216) Jenner G. 230 (52) (111); 247 (122); 248; 253 (144); 254 (145); 260 (154) Jones C.R. 252 (139) Sahal K.L. 267 (187) Josyer G. 269 (199); 270 Sankaran K.A. 253 (142), (143) Joyce J. 243 Sastri J.P. 264 (177) Kane P.V. 218 (2); 221 (12); 222 (19), Sastri P.S. 221 (11) (22); 224 (26); 225(29), (31), (33), (34); Sastri Siromani K.S.R. 284 (263) 226 (36); 227 (39); 228 (42), (43); 229 Satya Vrat 281 (253) (50); 231 (63); 232 (66); 233 (73), (74); Searle J. 242 (105) 234 (77), (79); 235 (83); 238 (90), (91), Sen Nabeeta 220 (5) (92); 239 (93a); 251 (136); 260 (154); Sen Sukumar 220 (5) 261 (157), (163); 263 (170); 264 (175); Sharma H.D. 246 (119) 270 (200); 271 (209), (211); 272 (216); Shastri K.C. 222 (20) 274 (222), (223); 278 (236), (237); 279 Shastri K.D. 221 (16) (239), (240), (242); 280 (246), (247); Sovani V.V. 235 (84); 240 (99) 281 (253); 282 (256); 284 (264); 286 Srikantiah T.N. 222 (20) (268), (273); 287 (276), (279), (281) Srinivasabhatta B. 249 (127) Kaviraj G. 222 (20) Sukthankar V. 273 (219) Keith A.B. 220 (8); 222 (19); 227 (39); Tatacharya D.T. 234; 234 (76) 243 (111) Telang M.R. 262 (165)

300 E. Gerow - Indian Poetics

Thomas F.W. 222 (20) Varma K.G. 253 (143) Tucci G. 233 (72) Wellek R. (and Warren A.) 219 (4) Upadhyaya B. 251 (137) Winternitz M. 221 (14)

Index of Works (Classical)

Abhidhāvrttimātrkā (of Mukula) 258 Kāvyānuśāsana (of Vāgbhața II) 279 (152) (of Hemacandra) 279 Agnipurāņa (anon.) 225; 259; 269; 270 Kāvyaprakāśa (of Mammața) 259; 261 Aitareya Brāhmaņa (anon.) 272 (216) (162); 268 (191); 270; 271-4; 287 Alaņkāramañjarī (of Ruyyaka) 274 Kumārasambhava (of Udbhata) 234 (223) (of Kālidāsa) 270 Alamkārasarvasva (of Ruyyaka) 272; Kuvalayānanda (of Appayadīkșita) 260; 274-8 286 Așțādhyāyī (of Pāņini) 221 (16) Locana (of Abhinavagupta) 233 (69); Aucityavicāracarcā (of Kșemendra) 238 259; 263 (170) (89); 260 Mahābhārata (anon.) 272 (216) Avantisundarīkatha (of Dandin) 225 Mālatīmādhava (of Bhavabhūti) 220 (9) Bhāminīvilāsa (of Jagannātha) 287 (280) Meghadūta (of Kālidāsa) 238 Bhatțikāvya (of Bhatti) 223; 224; 226 (34); 229; 234 Mīmāmsā (darśana) (anon.) 221; 222; 224; 253; 254; 257; 263; 273; 273 Bhāvaprakāśa(na) (of Šāradātanaya) 226 (36); 239 (93); 271 (208); 283-4 (218); 283 Nāțyaśāstra (of Bharata) 223; 223 (23); Bhojaprabandha (of Ballala) 220; 220(8) 225; 225 (34); 226-7; 237 (88); 245- Brahmana (literature) (anon.) 220 Brhadāraņyaka (anon.) 245 50; 259; 263; 270 (207); 271 (210); 283; 285 Buddhacarita (of Aśvaghoșa) 222 (21) Navasāhasānkacarita (of Padmagupta) Candrāloka (of Jayadeva) 260; 286 272 Citramīmāmsā (of Appayadīkșita) 232 Nirukta (of Yāska) 221 (15) (67); 286 Khandana (of Jagannatha) 287 Nyāsa (of Jinendrabuddhi) 228; 233

Daśakumāracarita (of Daņdin) 225; 232 Pratāparudrayasobhūșaņa (of Vidyā- nātha) 280 (66) Daśarūpa(ka) (of Dhanamjaya) 227; Rāmāyaņa (of Vālmīki) 220

263-4; 265 (179); 280 Rasagangādhara (of Jagannātha) 287-8

Dhvanyāloka (of Ānandavardhana, Rasārņavasudhākara (of Šingabhūpāla) 284 Dhvanikāra?) 238 (89); 250-8; 259; 261; 262; 262 (168); 263; 266 (183- Rasataranginī (of Bhānudatta) 284

185); 268 (192); 269; 271; 272; 273; Rasamañjarī (of Bhānudatta) 284

275; 275 (226); 276; 277; 286; 287 Rk-Samhita (anon.) 217; 222; 245 Sabdavyāpāraparicaya (of Mammața) Ekāvalī (of Vidyādhara) 280 Gītagovinda (of Jayadeva) 252 272 (213)

Karpūramañjarī (of Rājaśekhara) 260 Sāhityadarpaņa (of Viśvanātha) 269

Kāśikāvrtti (of Vāmana and Jayāditya) (198); 281-3

228; 238 (91) Sāhityamīmāņsā (of Ruyyaka?) 282

Kāvyādarsa (of Dandin) 217 (1); 227-33 (256) Bhoja) Kāvyālamkāra (of Bhāmaha) 227-33 Sarasvatīkanthabharana (of 270-1 (of Rudrața) 238-45 Kāvyālaņkārasārasamgraha (of Udb- Śatapathabrāhmana (anon.) 245

hata 233-5; 262 Saundaranandakāvya (of Aśvaghoșa) 229 (51) Kāvyālaņkārasūtra(vrtti) (of Vāmana) Śrngāraprakāśa (of Bhoja) 269-70; 284 233-8 Śrngāratilaka (of Rudrabhațța) 271 Kāvyamīmāņsā (of Rājaśekhara) 260-2 Tarala (of Mallinātha) 280 (246)

Indices 301

Ujjvalanīlamaņi (of Rūpagosvāmin) Vișņudharmottara purāņa (anon.) 225; 285 Vāgbhațālaņkāra (of Vāgbhața I) 278 259

Vakroktijīvita (of Kuntaka) 262-3; 275 Vivaraņa (of Bhāmaha) 235

Vākyapadīya (of Bhartrhari) 221 (16) Vrttivārttika (of Appayadīkșita) 286

Vedānta (anon.) 276 Vyaktiviveka (of Mahimabhatta) 269; 274 Vikramorvaśīya (of Kālidāsa) 248 Yaśastilaka (of Somadeva Sūri) 260