1. The Foundations Of Indian Poetry Nobel

The Rāmakrishna Mission Institute of Culture Library Presented by Dr. Baridbaran Mukerji

11332

Calcutta Oriental Series, No. 16. E. 9

THE FOUNDATIONS OF INDIAN POETRY

and their Historical Development

(General Outlines)

BY

J. NOBEL, PH. D.

Docent for Indian Philology at the University of Berlin,

Librarian, Prussian State Library

1925

Published by Mr. R. N. Seal, B.A. 107, Mechuabazar Street, Calcutta

R M I C LIBRARY Acc. No. 1332 Class No. Date St. C'r Class. Cat. Bk. Card Checked

Rs. 5/-

London Agents, Messrs. Luzac & Co. 46, Great Russell Street, London, W. C.

Printed by Mr. N: C. Paul at the Calcutta Oriental Press, 107, Mechuabazar Street, Calcutta.

PREFACE

Indian Poetry is of a peculiar kind. It is so closely connected with the theory of poetry, that it can only be understood fully when considered from this standpoint. The Indian poet is not merely a poet, he is also a scholar.

The domain of poetics is a wide one, and in the various periods many learned men have written on and studied this subject: This book does not present the tenor of the alamkārasāstra (poetics) of a limited period, but it is an attempt to describe the historical development of and the connection between the ideas of the older masters of the sāstra. Moreover, only the general questions are treated here, the examination of the special questions as guṇa, doṣa, alamkāra etc. being reserved for a later occasion.

There are four gentlemen to whom I am especially indebted and to whom I wish to express my thanks here : the learned editor of this series, Dr. Narendra Nath Law for his willingness to publish my little work ; Prof. Benoy Kumar Sarkar, who showed a warm interest in my work and acted as intermediary between Dr. Narendra Nath Law and me ; my colleague in the Prussian State Library at Berlin, Dr. Cl. Sherwood who spent many morning hours in correcting and amending the English ; and Dr. S. K. De at Calcutta, who not only read the proofs but gave much important informa-

ii

tion on various questions and on some difficult passages in the Sanskrit text.

Berlin, April 1923. J. Nobel

CONTENTS

Introduction ... ... 1

The Effect of Poetry ... ... 37

The Causes of the Kāvya and the Poet ... 43

The Definition of Kāvya ... ... 78

The Nature of Alamkāra ... ... 85

The Soul of Poetry ... ... 94

The Rīti ... ... 98

The Division of Poetry ... ... 126

INTRODUCTION

The poetic literature of India, with the exception of that which is of a merely narrative kind, is known under the name of kāvya; but as a rule, only the epic and lyric works are generally known by that name, while the dramatic literature takes up a more separate position, being a category of its own. It is, however, a fact that the Indian writers on Poetics positively state that the drama also is to be considered as a sort of kāvya. On the other hand, the ālamkārikas in their śāstras never deal exhaustively with the drama, but refer the reader to those works which particularly treat this subject. A book on poetic works in all branches, indeed, would have to be very extensive if the whole of dramatic literature were included. This department of literature has thus become a separate study1.

I The oldest book on dramatic art is the Nāṭyaśāstra of Bharata-muni. There is to this day no critical edition of this book, which has always remained the standard work of its kind. The edition of the Kāvyamālā (Vol. 42) can only be called a preliminary edition. Some chapters are edited by P. Regnaud and J. Grosset in Annales du Musée Guimet, Tome 2, Paris 1880, and Annales de l' Université de Lyon, Fasc. 40, Paris 1898, respectively. Another important work is the Daśarūpa by Dhananjaya, mainly based on the Nāṭyaśāstra. The author lived about the end of the tenth century A. D. A good edition was published by Fitz-Edward Hall, Calcutta 1865. The translation of the Daśarūpa by G. C. O. Haas (New-York 1912) contains

2 THE FOUNDATIONS OF INDIAN POETRY

In speaking of kāvya we mean that kind of poetry which claims to be in accordance with certain theories, which are given and founded by a proper science called the Alamkāra-śāstra. Consequently, the kāvya cannot be understood and estimated to its full extent but by men of a certain literary accomplishment, who are themselves well-skilled in the conception of the aesthetic beauty of poetry, the learned ones or the men of taste, as they are called in the works on Poetics.

It is not necessary to prove that poetics are possible only under the supposition that Poetry existed already, and that Poetry is older than Poetics; but if we try to go back to the beginning of both Poetry and Poetics, we find that the way ends in obscurity. That, however, is the general fate of all branches of human knowledge. As in the literature of other peoples, only the more precious works remain; so it is also in India. Not only the less valuable works have been lost, but, what we regret a good deal more, also all those which, valuable in themselves, were replaced later by better works.

As regards poetry, our search appears to assume a better aspect at the outset because Indian many mistakes (cf. Jacobi's review in Göttingische Gelehrte Anzeigen 1913, p. 302 ff). A third important work on the same subject as well as on poetics is the Sāhityadarpana composed by Viśvanātha Kavirāja (about the end of the 14th century). Of this there are several editions, e.g. that of E. Roer and R. Ballantyne, Calcutta 1851, and a translation into English by R. Ballantyne and Pramadadāsa Mitra, Calcutta 1875.

tradition itself calls Vālmīki's Rāmāyaṇa the ādi-kāvya1, but it must be confessed that we do not gain very much by this statement. Let it be granted that according to the researches of Prof. J a c o b i, the Rāmāyaṇa probably belongs to the fifth or even the sixth century B. C., and that it is older than the Mahābhārata3, still it remains quite uncertain to what extent the individual parts are to be considered as going back to yet older texts. Besides, there is still the difficulty, which up to this day nobody has been able to solve, namely, that of restoring out of the mass of the Rāmāyaṇa, as it is

1 Not everywhere though. So in Rājaśekhara's Kāvya-mīmāṃsa (Gaekwad's Oriental Series No. 1, p. 7) the Rāmāyaṇa is called itihāsa and the (Mahā)bhārata samhitā (sa [Vālmīkih] tu mahāmuniḥ pravṛttavacano Rāmāyaṇam itihāsaṃ samadrbhat. Dvaipāyanas tu ślokaprathāmādhyāyī tatprabhāvana śatasāhasriṃ samhitāṃ Bhāratam). A few lines before we are told that Vālmīki invented the śloka and in the second adhyāya (p. 3) Rāmāyaṇa and Bhārata are spoken of in a similar way as being a kind of itihāsa belonging to the category of purāṇa. (Sargaḥ pratisaṃ-hāraḥ kalpo manvantarānvi amśāvidhilh | jagato yatra niba-ddhaṃ tad vijñeyam purāṇam iti | "purāṇapravibheda eve-tihāsah" ity eke | sa ca dvidhā parikriyāpurākalpābhyām | yad āhuḥ | parikriyā purākalpa itihāsagatir dvidhā | syād ekanāyakā pūrvā dvaitīyā bahunāyakā // tatra Rāmāyaṇam Bhāratam codāharaṇe /). In the opinion of Rājaśekhara, therefore, the Rāmāyaṇa is older than the Bhārata.

2 Sec H. J a c o b i's work : Das Rāmāyaṇa, Geschichte und Inhalt, Bonn 1893, p. 100 ff. H. O l d e n b e r g in a posthumous work : Das Mahabharata, Seine Entstehung, sein Inhalt, seine Form, Göttingen 1922, p. 53 ff, does not agree entirely with the theories of Jacobi, without giving

known to us, the old form of the ‘ādi-kāvya’. Every manuscript varies from the other to such a degree, and in different places the epic poem has been altered to such an extent, that we are accustomed to speak of various and different versions.

It is, however, a task not quite hopeless, to examine the older literature, and find out whether there is any connection between it and those literary compositions that are kāvya-like, as, e.g., the Rāmāyana is a kāvya. Though I cannot deal fully with the matter here, a few remarks may be allowed.

Even in the Rigveda we find several hymns, in which the poets endeavour to leave the low level of mere invocations and in which they are trying to show a certain skill in poetic matters.

A real kāvya-style, however, can scarcely be found in any of Rigveda hymns. For this, the distance in time between the Vedic poetry and the Sanskrit period appears to be too great.

But going further and investigating the texts of the later and the latest Vedic period, certain passages in the Brāhmaṇas, which are, however, not very numerous, and especially a little epic work, very important in the development of poetry, the Suparnādhyāya1,

however, strict arguments against Jacobi's statements. It is a matter of course that many passages in the great epic go back to older times, but the question is at what time the Mahābhārata or rather Bhārata as a whole, as a Sanhitā, as Rājaśekhara calls it, was completed.

That seems to have been the case later than the composition of the Rāmāyaṇa.

1 Edited by Elimar Grube, Berlin 1875. Published also, but only the text, in Albrecht Weber's Indische Studien, Vol, 14.

(which shows the characteristic feature of the Vedic time and, on the other hand, bears t'le stamp of the kāvya-style) we may find that there is an uninterrupted line leading to that sort of poetry which we meet in the Rāmāyana of Vālmīki. Perhaps, Vālmīki was the first who created a perfect and great poetic work, and many things may have been introduced by him, so that his work with some right can be regarded the type of literary composition, and thus as the ādi-kāvya.

Large passages of the Rāmāyana could very well have a place in the kāvyas of the later period without disagreeing with the poetic theories of the ālam-kārikas. On the other hand, there can be little doubt that many poetic factors, similes, and other details are taken from older works which are lost for ever. On account of the high reputation attached to the Rāmāyana from the beginning, it must be regarded as a matter of course that the less poetic compositions in the period before Vālmīki could easily fall into oblivion. So the Rāmāyana obtained a position of marked importance in the development of the kāvya.

The Mahābhārata which, with regard to the text, presents even more difficulties than the Rāmāyana, is not called a kāvya by Indian tradition; but can anybody deny that also in this great epic there are many passages to be found showing very clearly all the characteristics of the kāvya-style? As a whole, the Mahābhārata is, as has been pointed out by Prof. Jacobi, younger than the work of Vālmīki, yet nothing is said thereby with regard to the various

6 THE FOUNDATIONS OF INDIAN POETRY

parts the epic consists of; and a history of the kāvya has to examine the Mahābhārata as well as the Rāmāyana.

Thus we may call, with some reserve, the Rāma-yana of Vālmīki a mark-stone in the historical de-velopment of the kāvya-literature. From here the road leads to Aśvaghoṣa, Bhāsa, Kālidāsa, Bhāravi, and Māgha, only to mention the most important representatives of poetry. The significance of some kāvyas was so great that the less famous works written before them have been forgotten. The fact, however, that during long periods no kāvya is known to us, does not allow us to conclude that during these periods there was no cultivation of the kāvya at all. Prof. Max Müller's theory of the "renaissance" of Indian Poetry has been refuted by Prof. Georg Bühler, who after having examined the inscriptions of the Gupta kings has pointed out that in the period before that of Kālidāsa the kāvya-style was on the contrary highly cultivated, which is shown also by the compositions of Aśvaghoṣa1 and Bhāsa2, whose dramatic works were discovered lately.

Indian Poetry is written mostly in Sanskrit, and this fact shows that the kāvya was not a poetry for the great mass of the people, but only for the educated classes. Besides that, we can read in every kāvya that the poets wish to satisfy the wise ones, who alone are able to understand and estimate poetry fully.

1 See H. Lüders : Bruchstücke indischer Dramen, Berlin 1911.

2 Edited by Ganapati Sāstrin in the Trivandrum Sanskrit Series.

INTRODUCTION

In India all branches of human knowledge, with the exception of that of history in which the Indian people in the older times never took a warm interest, have always been adapted to scientific systems ; and the same was done with Poetry, and Poetics soon became a separate system. The poet could not but respect the theories and the rules which were presented by the learned ones in their śāstras, if he wished to be acknowledged. Far from making himself independent of these theories1 he endeavoured anxiously to be in accordance with every thing laid down by the laws of Poetics and to be, so to speak, mathematically exact in inventing new and varying old forms of similes, in selecting words, phrases etc., which had to agree fully with the rasa of the matter in question.

This science is named Alamkāra-śāstra. If we try to determine its age, we must unfortunately admit that the attempt is a good deal more difficult than to give an historical account of Poetry itself. It has been said before that the writers on Poetics regard the drama as a branch of the kāvya, but probably dramatic writing and the art of epic and lyric poetry in the older period were independent of each other, dramatic works being the oldest. For it must be conceded

1. It is therefore not correct, when R. Pischel (Kultur der Gegenwart, Berlin u. Leipzig 1906 1, 7 p. 201) says that Kālidāsa in the Kumārasambhava has made himself almost entirely free from the rules which are applicable to the mahākāvya, whilst in the Raghuvaṃśa he has followed them. That this view is wrong requires no proof. Besides, there can be little doubt that the Raghuvaṃśa was written after the Kumārasambhava.

8 THE FOUNDATIONS OF INDIAN POETRY

that the essential features of the drama and particularly of the older drama belong solely to dramatic art and could not be transferred to epic and lyric poetry. Later on, however, when the drama was assuming a more or less epic and lyric character, when the story was interrupted by verses which in a sentimental way described all sorts of situations, then of course dramatic writings and alamkāra-śāstra came together. Thus it became quite natural that in the works dealing with the drama, Poetics were also touched on as far as it was necessary to the dramatic poet. These metrical passages were, however, by nature strange to the drama, because the verses, as it were, tend more to hinder the course of the story than to promote it. The strict kāvya-style could have but a subordinate place in the drama, which was enacted before the eyes of the spectators and thus had to be written in an easier style. Matters however, changed. Already in the age of Kālidāsa and to a much larger extent in that of Bhavabhūti the character of the nāṭakas grew a good deal more lyric and epic, thus approaching that kind of poetry which is to be found in the so-called mahākāvyaS, and so dramatic writing was regarded by the ālamkārikas as a kind of kāvya.

Now the oldest book treating of poetic matters is a work on dramatic art, viz. the Nāṭya-śāstra. It is said to have been composed at a remote period by Bharata-muni. According to more recent views, however, it is not very much older than Bhāsa, and it must be added that the state of the text as it is printed in the Indian edition1 is rather bad. Much

1 Kāvyamālā no. 42. See above note on p. 1.

research and investigation there must be before that important work will take its proper place in the history of Indian Poetry. The great importance of the book cannot be denied, if one bears in mind that it was always regarded as an authority with respect to all matters belonging to the art of dramatic writing and that many of the later books dealing with the same subject are based on it, e.g., the Dasarūpa of Dhananjaya. Besides the doctrines of dramatic writing, some poetic matters are also dealt with, so far as they are necessary for dramatic poets.

Considering the whole mass of books that have been written by Indian scholars on poetics it can be easily seen that the alamkāra-śāstra is not a doctrine the system of which shows uniformity at all. On the contrary, from the beginning to the modern times there has been a steady growth and development. The views of the older works were rejected or modified, one theory was substituted for another, and poetry was regarded from quite different points of view ; in short, there was scarcely one theme that did not assume a new aspect in the course of the historical development.

Compared with other branches of human knowledge this progress and growth was throughout natural and intelligible and in accordance with the progressive methods employed in treating abstract matters.

The refinement of taste for poetic matters became more and more visible. Even poets who are chronologically not far apart show slight differences in dealing with the poetic material. These differences, however, present themselves a good deal more clearly, if literary compositions of different

ii

10 THE FOUNDATIONS OF INDIAN POETRY

periods are compared with each other. For instance, it appears a thing hard to believe that the Rāmā-yana and the Siśupālavadha belong to the very same sort of poetry, the kāvya. It is not possible to enumerate these differences shortly here. The quint-essence of a poetical work seems to have become different. Regarding Vālmīki it is evident that the plot occupies the main interest and that the poet endeavours to narrate a series of actions, which in their totality have an interest of their own, and these events in the history of Rāma and Rāvana are described in a poetical and pleasant manner. The em-ployment of poetic forms is to be considered, as Bhāmaha would say, as a mere exterior (bāhya) orna-ment. The epic element was regarded as the main part. With respect to Māgha, on the other hand, the description of actions is placed into the back-ground, existing as it were only on account of their being indispensable for another purpose. Poetry has ceased to be a poetical description of the deeds of heroes, poetry has become desirable in itself. The plot of a poem could be told in a few lines. The brilliance and elegance of style, the mathematically and logically exact congruence of the separate parts of similes, and all the other things which are described in the śāstras, was the end the poet was aspiring to. Hence it comes that the stanzas following each other are much less coherent in themselves. On the contrary, every verse stands so to speak like a monument polished by the skill of the highly learned poet, who shows his genius (pratibhā) and his wide experience (śakti) in all matters that have to do with poetry.

As a matter of course these differences, which can be observed during the long period of the kāvya, correspond with the theories of the ālamkārīkas. If we wish to get a correct idea of the whole kāvya literature and to be qualified to understand the Indian kāvi without prejudice, it seems to be necessary to know the views of the theoretic scholars. We will therefore look into the doctrines of the alam kāra-śāstra, giving a sketch of the views on poetic matters and pointing out by which ways the older theories are assumed and specified by the younger representatives of the śāstra. This literature being a very large one, we shall but deal with the older writers on poetics, examining the more recent books only in such cases when it is required for the correct interpretation of the opinions of the classical period.

Before going into detail we will cast a look at the poetical theories the ālamkārīkas deal with and at the way in which their systems are presented. Leaving aside the Nātyaśāstra of Bharata-muni, which as a book on dramatic writing must take its place in the history of that art, the oldest work we know of is the Kāvyālaṃkāra by the famous Bhāma-ha. Some older writers we know only by name. The most remarkable among these apparently was Medhāvin or Medhāvirudra, as he may be called more correctly1. He is not only mentioned by Bhāmaha, but also in the later works on alaṃkāra, and in commentaries he is frequently cited as a man of great

1 See e. g. Rājaśekhara, Kāvyamīmāṃsā p. 12, l.1; Namisādhu on Rudraṭa's Kāvyālaṃkāra II, 2 ; XI, 24.

12 THE FOUNDATIONS OF INDIAN POETRY

authority. He is said to have been blind1. Regarding the fact that many works which seemed lost have been discovered in Indian libraries we may hope that Medhāvirudra's book will come to light some day.

The Kāvyālaṃkāra by Bhāmaha, which was edited for the first time some fifteen years ago, did not lose its renown even when poetry was looked upon from a higher point of view. The work is divided into six adhyāyas. After having opened with an introductory verse the author praises the effect and describes the causes of good poetry. Then a highly important question is touched and discussed, viz., that of the position the alaṃkāra is attaining in poetry. Then the definition of the kāvya is given, and thereafter its divisions.

The following verses are devoted to a very grave and interesting matter, the rīi (style or diction as the Sanskrit term is sometimes, but not quite correctly, translated). Bhāmaha opposes the opinion that there are some different rītis. Now, not every way of expressing the sense is to be called a beautiful and correct one. The kavi has to avoid certain doṣas or faults. It is stated, however, that the doṣas must not in all cases be considered to make a kāvya worthless. There are some things that are far from being strictly correct but are in some way or another sanctioned by tradition.

The second book begins with a rejection of the doctrine of the guṇas, the rejection being understood, however-

1 Rājaśekhara l. c.: Medhāvirudra-Kumāradāsādayo jātyandhāḥ kavayaḥ śrūyante.

ever, indirectly. For the term of guṇa itself is not used in this connection by Bhāmaha, who speaks only of mādhurya, prasāda, and ojas. After having settled this question in only three stanzas, while in other works (e.g. in the Nātyaśāstra or in the Kāvyādarśa) it is dealt with much more extensively, the author goes on to define and illustrate the alamkāras, which he considers to be the essence of poetry, as we may guess even from the title Kāvyālamkāra. This theme is exhausted at the end of the third adhyāya. The alamkāras are divided into two groups: śabda and abhidheya-(artha-) alamkāras. When the doctrine of the alamkāras will be treated at length, we shall deal with the question whether something can be guessed from the manner of enumerating the alamkāras, and with certain other things belonging to the same matter. In the fourth book the author gives a full account of the so-called doṣas, which, as we said before, in many cases cease to be faults. In the next adhyāya we meet with the elaboration of a logic of poetry (nyāyanirnaya), while in the last chapter the work winds up with a description of grammatical correctness (śabdaśuddhi). It may be mentioned that later on Vāmana will end his Kāvyālamkārasūtravṛtti with a chapter of a similar content.

The next writer on poetics is Daṇḍin. On account of his being connected very closely with Bhāmaha there cannot be a correct judgment of the Kāvyādarśa without knowing the latter's Kāvyālamkāra. Daṇḍin appears to be an

opponent of Bhāmaha, and his manner of polemics against his predecessor bears a rather personal stamp, whic'1 the reader of the Kāvyādarśa can be hardly fail to see. Thus, in my opinion, we may be right in assuming as a fact that Dandin, though he wrote his work later than Bhāmaha, was a younger contemporary of Bhāmaha. If this statement is correct (and I hope the reader will come to the same conclusion after having examined the following pages wherein the views regarding the foundations of poetry are dealt with fully) it will become necessary to place both Bhāmaha and Dandin (and not only one of them) either before or after the time of Kālidāsa. This question, a highly important one for the history of the Sanskrit and Prakrit literature, has not yet been solved with absolute certainty to this day. The only certainty is the life-time of Kālidāsa. He must have flourished about the middle of the fifth century A. D. when Kumāragupta I (455-480) was reigning1. Did Bhāmaha and

1 It would fill a big volume, if I should attempt to give the literature concerning the date of Kālidāsa and so I will not add a new essay to the older ones. I may mention that in my opinion the title Kumāra-sambhava can only be completely understood, if we assume that the author has chosen it with respect to king Kumāragupta, and that the title Vikramorvaśī in a similar way refers to the surname of Candragupta (Vikramāditya), and that Mallinātha (commenting Meghadūta 14) takes for granted that the well-known philosopher Dinnāga lived before or during the same time as Kālidāsa. By the way it may be mentioned that the commentator Vallabha,

Dandin live before or after this time? I confess that formerly I was inclined to put both alamkārikas before Kālidāsa. I cannot prove it by direct arguments, but after having read the work of Bhāmaha, I had the impression that it must be a very old book. The authors and works Bhāmaha mentions (Asmakavamsa I, 33; Rāmaśarman II, 58; II, 19; Acyutottara II, 19; Śakavardhana II, 47; Rājamitra II, 45; III, 10; Ratnahaarana III, 8) are all quite unknown to us. Indeed, the many arguments which were brought forth to strengthen the contrary view do not settle the matter. It is true, however, that Bhāmaha mentions the clouds (jalabhrt) among the things which should not be represented as messengers. What else could Bhāmaha have in view but Kālidāsa's well-known Meghadūta? But, on the other hand, would Bhāmaha have found fault with it, had he known Meghadūta, which famous and splendid poem must have convinced every man that, on the contrary, a cloud was very well fit for acting the part of a messenger? Some have tried to find out some chronological relation between Bhāmaha and Māgha. Using the term śabdārthau in Siśupālavadha II, 86 Māgha is said to refer to the definition of kāvya

who does not give the interpretation concerning Diñnāga, is not an old author, as E. Hultzsch thinks, (see introduction to his edition of Kālidāsa's Meghadūta), but belongs to the twelfth century, as is shown by K. P. Pathak in the introductory pages to the second edition of the Meghadūta, Poona 1916.

16 THE FOUNDATIONS OF INDIAN POETRY

as given by Bhāmaha¹, but if this is considered as an argument I may be allowed to take as an argument of the same force that Kālidāsa in Raghuvaṃśa I, 1 by the words vāgārthāviva sampr- ktau refers to the very same definition of Bhāmaha, and that by this fact Bhāmaha's priority may be taken for granted. By using the simile of śabdārtha, however, Māgha probably refers to that of Kālidāsa or perhaps to some other view and not to the definition of kāvya in Bhāmaha. This argument, however, is of no great value, because the

1 K. B. Pathak in Journal of the Bombay Branch of the Royal Asiatic Society, Vol. 23, p. 31. Against the state- ments made here, see P. V. Kane in a paper Bhāmaha, the Nyāsa and Māgha ib. Vol. 23, p. 91 ff. Pathak in the essay mentioned above has made an attempt to show that the term nyāsakāra used by Bhāmaha, Kāvyalaṃkāra VI, 36 refers to Jinendrabuddhi, a Buddhist commentator on the Kāśikāvṛtti who must have lived about 700 A. D. and he has defended his arguments against P. V. Kane (referred to above) in a paper Dandin, the Nyasakara, and Bhamaha, Indian Antiquary 1912, 232-37. Compare also Trivedi's Introduction to the Pratāparudrīya (Bombay Sanskrit Series Vol. 65) p. xxxv. I am sorry to say that Pathak's arguments in my opinion fail to settle the matter. His polemics against Kielhorn, too, who in Journal of the Royal Asiatic Society 1900, part I, pp. 499-502 dealt with the state- ments in Siśupālavadha II, 112 and pointed out that by the words vṛtti and nyāsa the author of the Kāśikāvṛtti Jinendrabuddhi, is alluded to (what Mallinātha has already said), is far from convincing. The theory of Dandin's priority to Bhāmaha, assumed by K. B. Pathak and other scholars, needs no further refutation. That the contrary is correct will be shown by the pages below.

so-called definition of kāvya (as consisting of word and sense united) may go back to some older ālamkārikas and because, as will be shown later on, that definition in the same or a similar form is given also by scholars posterior to Bhāmaha. Nevertheless, Bhāmaha (and Dandin) seem to have written after the time of Kālidāsa. Prof. Jacobi13 draws my attention to the fact that Bhāmaha in V, 28, 29 refers clearly to Dharmakīrti's Nyāyabindu III, 138, 13914. In other cases, however, Bhāmaha seems to be dependent on the philosopher Diṅnāga, who is older than or rather contemporary with Kālidāsa and is, according to Mallinātha, alluded to in Meghadūta 14. To return to D a n d i n, the Kavyādarśa 15 con-

13 In a letter dated March 29, 1922. Prof. Jacobi's paper on this subject entitled Bhāmaha and Dandin, ihr Alter und ihre Stellung in der indischen Poetik, is now printed in Sitzungsberichte d. preuss. Akad. d. Wiss. 1922.

14 Bhāmaha V, 28, 29 : dūsanā(ṃ) nyūnatādyuktir nyūnaṃ hetvādinātha ca | tanmūlatvāt kathāyās ca nyūnaṃ nestam pratijñayā // jātayo dūsanābhāsās tāḥ sādharmyasamādhy(adr)ayah | tāsāṃ prapañco bahudhā bhūyastvād tha noditah // Dharmakīrti's Nyāyabindu (Bibliotheca Buddhica Vol. VII, p. 94), III 138, 139 : dūsanāni nyūnatādyuktīh // ye pūrvam nyūnatādayah sāhanadosā uktās teṣām udbhāvanam dūsanam | tena parestārthasiddhi-pratibandhāt //

15 Edited several times, e. g. at Calcutta 1863, with a very good commentary by Premacandra Tarka-vāgīśa. Another Indian edition was published by

iii

18 THE FOUNDATIONS OF INDIAN POETRY

sists of three books. Opening with an introductory verse and after having referred in a general way to the older works on poetics, Daṇdin praises the advantages of good poetry, in a manner similar to that of Bhāmaha. In order to explain what is meant by good poetry, śāstras have been composed, which deal with the body (śarīra) of poetry and its ornaments (alaṃkāra in the wider sense of this word). Daṇdin next gives an account of the body. We meet with the definition and division of kāvya. These explanations are, to a greater part, to be considered as a refutation and a correction of the views of his predecessor.

Then Daṇdin goes on to give a detailed account of the doctrines of both rīti and guṇa, thus indicating a point of view different from Bhāmaha's treatment of the matter, as will be shown below. In the last part of the first book the author points out what is necessary and desirable for becoming a real poet (kavi).

The second adhyāya, which contains not less than 368 verses, is wholly devoted to the definitions and illustrations of the arthālaṃkāras, while in the first 96 stanzas of the last book all the different species of śabdālaṃkāra are dealt with, the illustrative verses being to some extent horribly difficult. Daṇdin had a special liking too for the riddle (prahelikā) and gives its divisions in 28 verses.

Vidyāsāgara ; fifth edition, Calcutta 1911. There is also a German edition of the text, together with a German translation, by O. Boehtlingk, Leipzig 1891.

rest of the adhyāya (51 stanzas) has the doctrine of the so-called ten doṣas for its subject.

The ālaṃkārikas, who followed Daṇḍin chronologically, were Vāmana and Udbhaṭa. Though they were contemporaries (they both lived at the court of king Jayāpiḍa, 779-813), their works have but little in common. Perhaps their respective standpoints would have been clearer to us, if the other works of Udbhaṭa were known to us. Both Udbhaṭa and Vāmana are well acquainted with Bhāmaha.

Vāmana was the first to treat the matter in sūtra-form. These sūtras were of course, hardly comprehensible by themselves ; a commentary was needed, the author of which is also Vāmana. Thus his work goes under the name Kāvyālaṃkāra-sūtra-vṛtti. It is worth noting that this title may have been chosen with reference to Bhāmaha, whose work bears the same name Kāvyālaṃkāra. It is divided into five adhikaraṇas : the first containing three, the second and third two, the fourth three, and the last two adhyāyas.

16 Edited as No. 134 and 140, in the Benares Sanskrit Series, 1907-08 ; as No. 5 in the Sri Vani Vilas Sastra Series, Srirangam,1909 ; as No.15 in the Kāvyaṃālā, Bombay, 1889 ; in Germany by C. Cappeller, Jena, 1875.

The same scholar also gave the text and translation alone of the last book of Vāmana's, under the title : Stilregeln, Strassburg, 1880. A translation of the entire work of Vāmana is given by Pandit Gaṅgānātha Jhā, Indian Thought Series, no. 3, Allahabad, 1911-12. There are, besides, some other editions, which I will not enumerate here.

Vāmana opens with the statement that the kāvya is acceptable on account of its embellishments (alam-kāra in the wider sense). Then is pointed out what is meant by the word alamkāra, i. e. the avoidance of the doṣas and the employment of the guṇas and alamkāras (in the narrower sense). In the last sūtra of the first adhyāya the effect of a good kāvya is spoken of. Then the question is answered as to who should be taught to compose a kāvya. Next, Vāmana deals with the rīti, which according to him is the soul of poetry. In doing so, he presents, on the one hand, a contrast to Bhāmaha and, on the other hand, he accepts to some extent the views of the Kāvyādarśa. In the last adhyāya of the first adhikaraṇa Vāmana describes the so-called aṅgas of the kāvya, this matter forming the subject of twenty sūtras. Here we find everything mentioned which is to be studied by one who intends to become a kavi. The first adhikarana winds up with the divisions of poetry. The author then proceeds to present a full account of the doṣas. There are three kinds of doṣas, viz. pada-, vākya-, and vākyārtha-doṣas. As a pendant of the doṣas Vāmana in the following sūtras deals exhaustively with the guṇas, which are divided into śabda (or bandha) and artha-guṇas. The whole fourth adhi-karaṇa gives the doctrine of both śabda- and artha-alamkāras, the last being regarded as mere sub-divisions of the main figure : upamā. After the example of Bhāmaha the work closes with the

doctrine of poetic conventions (kāvya-samaya) and of the correctness of words (śabda-śodhana).

With respect to U d b h a t a, it is a matter of regret that only his Kāvyālamkāra-sāra-samgraha

has been edited17, which work is closely connected with the poetics of Bhāmaha, as is already suggested

by the title. The six books of the work consist of the treatment of the śabda- and arthālamkāras. Udbhata

wrote some other works, too, the most important and interesting one being his commentary on

Bhāmaha. As a manuscript of it is extant we may hope that it will be published as soon as possible.

A good deal more extensive than the works mentioned before is the Kāvyālamkāra of R u d r a ṭ a18

bearing (which fact is interesting) the same title as the poetics of Bhāmaha. There are many views with

regard to the time of that ālamkārika19. Some say that his work was written in the second half of the

11th century A.D., others are inclined to assume that Rudraṭa did not live after the middle of the 9th

century while others again would find out that he lived about 950 A.D., At present, we may take it

as a matter of fact that he lived about or rather before 850 A.D., because an opinion of his is discussed and

17 The text is given by J a c ó b in Journal of the Royal Asiatic Society, London 1897, a great part of the

verses also by T r i v e d i in his notes to Vidyānātha's Pratāparudrīya (Bombay Sanskrit Series No. 65.).

18 Kāvyamālā vol. 2, Bombay 1886.

19 See R. P i s c h e l, Rudraṭa's Śṛṅgāratilaka and Ruyyaka's Sahṛdayalīḥt, Kiel·1886, Introduction.

22 THE FOUNDATIONS OF INDIAN POETRY

rejected in the Kāvyāmīmāmsā30 by the famous Rājaśekhara, who belongs to the end of the 9th or the beginning of the 10th century. We must further bear in mind that the Kāvyālamkāra of Rudraṭa, though dealing with the whole matter of poetry, does not say one word with respect to the doctrine of dhvani, which is discussed fully in the Dhvanyāloka and which is referred to in all later works on poetics. Since the Dhvanyāloka was written about the middle of the 9th century, the Kāvyālamkāra cannot be placed after that time.

The work of Rudraṭa is one of the more difficult books on alamkāra. It consists of sixteen adhyāyas, composed in the āryā-metre. After a short introduction, the author speaks of the effect of good poetry and goes on to deal with all a kavi should know. The second book begins with the definition of kāvya (as the union of word and sense; śabda and artha). With this we have an indication of nearly the whole Kāvyālamkāra. Rudraṭa intends to discuss first the nature and qualities of the word (śabda) and then those of the sense (artha). With regard to the noun, Rudraṭa, after some general remarks on the word, discriminates between two kinds, compound and non-compound, this division being a highly important one because thereupon are based the differences of the rīti. Then the author defines the vākya which, on the one hand, may be divided into prose (gadya) and metre (chandogata) and, on the other, into further divisions according to the language in which the kāvya is written. The remaining chapters discuss the śabdālamkāras, the

20 Adhyāya 7, p. 31.

śabdadoṣas and the vākya-doṣas. The first part of the definition of kāvya being herewith completed, Rudraṭa takes up the second part, artha. Under this heading there is presented a full description of arthālamkārās, which are to be regarded from four different stand-points, viz. vāstava, aupamya, atiśāya, and śleṣa. This treatment of the matter seems to have some connection with a similar opinion of Vāmana, for we may remember that this ālamkārika regards all the the figures of speech from the point of view of aupamya. Then the artha-doṣas are described and especially that of upamā. The doctrine of the rasas, which really had its place in dramatic poetry—for the persons of the drama show in fact all shades of sentiments and the reflection of most inner feelings—and which, as a matter of consequence, has never been treated in fulness by the older ālamkārikas Bhāmaha, Daṇḍin, Vāmana, and Udbhaṭa, takes up much room in the Kāvyālamkāra of Rudraṭa. Among the rasas the śṛṅgāra-rasa is treated exhaustively, and the doctrine of the nāyaka and the nāyikā and the like is discussed at length. The last book of Rudraṭa's deals with the various kinds of literary compositions (prabandha).

Comparing the alamkāra-śāstra of Rudraṭa with those of his predecessors it must be confessed that he has endeavoured to present new ideas. Though many subjects he treats of are spoken of for the first time by him and though he has given new shapes to old views, Rudraṭa must be regarded as belonging to the older school of the alamkāra-śāstra. As a matter of fact he is entirely untouched by the doc-

trine of the dhvani, which was gaining a great influence upon the further development of poetics.

This aesthetic criticism was brought into a system by the author of the Dhvanyāloka about the middle of the 9th century A.D. 21. It is now settled as a fact that Ānandavardnana is the author of the vritti only, and not also of the kārikās 21a.

That the doctrine of the dhvani, however, had in some way or other been treated already before the time of the Dhvanyāloka, may be gathered from the words of the author himself. There is, however, little doubt that in dealing with the new views and developing them at length, the Dhāvanikāra and Ānandavardhna of Kashmir played the chief part. The idea is this : poetry is of value only when the matter the poet wishes to deal-with is presented to the hearer by mere suggestion ; simple description of events in plain terms has nothing to do with poetry. In one word, the unspoken is the

21 The text is edited in Kāvyamālā No. 25, Bombay 1891. Translated into German by H. Jacobi, Zeitschrift der Deutschen Morgenlaendischen Gesellschaft, Vol. 56 and 57, 1902-03, Ānandavardhana's name (Ānanda) is quoted in the Kāvyamīmāṃsā of Rājaśekhara, p. 16 (prati-bhāvyutpatti-yoh pratibhā śreyasi ity Ānandah | sa hi kaver avyutpatti-krtam doṣam āśeṣam ācchādayati ; tad āha | avyut-pattikrto doṣaḥ śaktyā samvriyate kaveḥ, yas tv aśakti-kṛtas tasyā jhaṭity evāvabhāsate.) See Dhvanyāloka, p. 137 (3rd uddy.).

21a See also the excellent paper by S. K. Dé : The Text of Kāvyāloka-locana IV. Reprinted from the Journal of the Department of Letters vol. ix. Calcutta University pr. 1923, p. 2 f

essence of poetry. With this idea wrote Kāli-dāsa and the other great poets long before the real quintessence of poetry was analysed and reasoned out in books on this subject. The correctness of the new idea being acknowledged by all who had a taste for poetics, the doctrine laid down in the Dhvanyāloka could not fail to influence all the later works on alamkāra.

We cannot pursue the theme further here, where we are going to inquire into the development of the ideas as they were presented by the old masters. Though the doctrine of the dhvani attained a very great significance for poetic criticism, the older theories are by no means put aside, but they hold their place as constituting the foundation of the kāvya.

The big work of Mammaṭa, the Kāvyaprakāśa22, shows clearly the influence that the dhvani-doctrine was gaining. Mammaṭa lived in the 11th century. Though his work is written in verse throughout, the expression is as brief and condensed as possible and thus rather bears the character of the sūtra-style. We are therefore not surprised that Mammaṭa himself composed a commentary on it22a.

22 Edited many times in India, e. g. with a great commentary by Bhaṭṭa Vāmanācārya B. R. Jhalakīkara, Bombay 1901 (2nd ed.). For an English translation we are indebted to Gaṅgānātha Jhā, Benares 1897-99 (From Pandit, Vols. 18-21).

22a There are, however, some doubts about the authorship of the Kāvyaprakāśa. See V. Sukthankar, Miscellaneous Notes on Mammaṭa's Kāvyaprakāśa, Zeitschrift der Deutschen Morgenländischen Gesellschaft, Vol. 66, p. 477 ff and 533 ff.

iv

26 THE FOUNDATION OF INDIAN POETRY

The Kāvyaprakāśa consists of ten ullāsas. The author opens in the same way as his predecessors by pointing out what the effect of good poetry is, and what attributes, besides learning, are to be regarded as necessary for becoming a poet. After that, the definition of the kāvya is given, and Mammaṭa turns to the divisions of poetry. Here we meet with the higher criticism of dhvani. According to the degree in which the suggested meaning (dhvani) prevails there are three kinds of poetry. This question is discussed further, up to the end of the sixth ullāsa, starting from the nature of words and ending with the doctrine of rasa, which is dealt with from the standpoint of dhvani. Then the various kinds of doṣas are enumerated, and in conneotion with this, the views regarding the guṇas are criticised. In the ninth and tenth ullāsas Mammaṭa treats of the alamkāras, which in the meanwhile have become a good deal more numerous.

By mentioning Mammaṭa the chronological line of the writers on the alamkāra-śāstra has been interrupted. The most peculiar feature of the book, however, which I am now presenting to the reader, justifies me in speaking of the Kāvyaprakāśa so soon. We have to deal now with a theoretical work of the well-known dramatist Rājaśekhara, who lived about 900 A. D. He is younger than Rudraṭa and Ānandavardhana, who are both mentioned by him. The work is entitled Kāvya-mīmāṃsā. This highly interesting book, which also gives important information about other matters, is edited as the first volume of the brilliant Gaekwad's Oriental Series, 1919.

The text now published is but a small part of the whole Kāvyamīmāmsā, bearing the title Kavirahasya. This is evident from Rājaśekhara's own statements, especially in the introductory lines, but since a manuscript of the other parts has not yet been found, we are not quite sure whether the author did not get further than the end of the first adhikarana, so that the work must be regarded as incomplete, or whether other parts are unknown, because the manuscripts have not been found or are lost. The copious notes the learned editors of the Kāvyamīmāmsā have added to the text show that the author has very often referred to older works, both kāvyas and śāstras.

The style of Rājaśekhara's book differs a good deal from that of his predecessors, the main part being written in simple prose, but not in sūtra-form which Vāmana or the Dhvanikāra used. The prose, however, is interrupted by more or less large passages in verse, which, on the one hand, support the ideas presented in the prose-parts, and, on the other, in a more independent way give some further details. Though in dealing with the matter, the author shows a certain raciness of expression and tells plenty of interesting facts, he is in many cases not free from pedantry, particularly in finding out new groups and divisions.

The Kavirahasya (as the first adhikarana of the Kāvyamīmāmsā is called) consists of eighteen adhyāyas each of them having titles. In the first (śāstrsamgraha) the various kinds of poetic matters are enumerated. The second adhyāya (śāstranirdeśa)

shows that there are two kinds of literary composition (vāñmaya), viz., śāstra and kāvya, the śāstra being considered as previous to kāvya. With respect to the former, Rājaśekhara gives subdivisions and explanations, speaking of the style of the śāstra and of the various forms and designations of commentaries etc. With the next adhyāya (kāvyapuruşotpatti) the author begins the principal matter, the doctrine of the kāvya. There are we are told that the goddess of Speech, Sarasvatī, longing for a son, underwent severe penitential exercises. She, then, bore a son, the Kāvyapuruṣa. Rājaśekhara gives a story of the encounter of Vālmīki, the inventor of the śloka and author of the itihāsa Rāmāyaṇa, with the lonely wandering Kāvyapuruṣa. In telling this, the author mentions also Dvaipāyana, who was the first to study the ślokā (invented by Vālmīki) and composed the samhitā Bhārata, which is said to consist of 100,000 ślokas. In course of time the bride of the Kāvyapuruṣa became Sāhityavidyā (vadhū). On account of her wandering through many countries, some poetic forms evolved themselves, the most important among them being the three rītis, Gāudīyā, Pāñcālī, and Vaidarbhī. These stories may be regarded as introductory. In the fourth adhyāya (the title of which seems to be incorrect, being the same as that of the sixth chapter, padavākyaviveka) goes on to deal with the nature of the kavi. The students of kāvya are said to have different degrees of understanding. The differentiation appears to have been influenced by a passage in the Arthaśāstra of the famous Kauṭilya, whose theories Rājaśekhara sometimes makes use

of also elsewhere. Then the author continues to point out what things are requisite and desirable for everyone who wishes to become a kavi. Speaking of the pratibhā (the inborn and working fancy of the poet) Rājaśekhara attaches much value to its two-fold nature, 'viz., as kārayitṛī and bhāvayitṛī. These remarks show the influence of Vāmana very clearly. Hereafter, in the fifth chapter (kāvyapāka), the vyutpatti (the literary education) is dealt with. Rājaśekhara describes the relation between vyutpatti and pratibhā, quotes the views of other authorities (among these also the view of his wife A v a n t i s u n d a r ī), states several groups of the kavi, going a good deal into detail, and criticises again and again the theories of his predecessors. The sixth adhyāya (pada-vākyaviveka) has the doctrine of pada (word) and vākya (sentence) as the subject. The definition of poetry (kāvya) is presented here. In the next chapter (pāthapratiṣṭhā) Rājaśekhara discusses the various kinds of expression with respect to the speaker (gods, man, and so on), the language (Sanskrit, Prākrit, Bhūtabhāṣa) being dependant on them. After having mentioned the varieties of ṛti (Gauḍīyā, Pāñcālī, Vaidarbhī), the author deals with another factor which in his opinion has a peculiar interest, namely recitation, kāku or pāṭha. As a matter of fact, this passage is a highly important one, because it shows what kind of pronunciation the various peoples of India used. That an opinion of Rudraṭa regarding the same subject is rejected by Rājaśekhara, is of importance with respect to the chronological relation between these two writers on poetics. The mere

title of the eighth adhyāya (kāvyārthayona­yah) calls to mind a certain passage in Vāmana's Kāvyālam­kārasūtravṛtti. Here we find all matters communicated which are important with regard to the contents of a kāvya (śruti, smrti, itihāsa etc.) It need hardly be mentioned that the author adds some new groups to the old ones, but it must be confessed that in inventing new subdivisions Rājaśekhara here (as well as elsewhere) shows a good deal of overmuch pedantry. The next chapter (arthavyāpti) discusses the contents of kāvya. The opinion of Drauhiṇi that with respect to the contents there are three kinds, viz, divya, mānusa, and divyamānusa, is refuted by declaring that seven kinds must be assumed, since pātā­līya, martyapātā­līya, and divyamartyapātalīya have to be added. Rājaśekhara is careful to illustrate the seven kinds by examples. After having done so, the author gives some explanations of descriptions (of rivers, mountains, towns, separation etc.) enlivened by the rasas. A few remarks about the nature of things as they are in reality and as they are represented by the fancy of the poet are of a certain interest. Then Rājaśekhara speaks of two forms regarding the contents, viz. muktaka (a single stanza, closed in itself) and prabandha (a whole composition); and the chapter ends with another subdivision according to the reality or non­reality of the events the kavi describes. The tenth chapter bears the name kavicaryā rājacaryā ca, and, at the outset, points out what branches of learning are necessary to know before one is able to compose a kāvya : viz; grammar, metrics, dictionaries and

so on. Besides these, the kavi has to know the accessory sciences (upavidyā), as e.g., the arts etc. Rājaśekhara does not forget to lay stress on another thing to which the kavi must turn his attention : that is the care of the body. In studying these passages, the reader will be rather amused and compensated for many tiresome explanations and divisions, though here also the descriptions bear the stamp of pedantry which exhausts all possibilities in a wearisome manner. "As the kavi, so the kāvya". Further, the ideal dwelling house of the kavi is described in full detail. Many facts mentioned here are of great interest for us in more than one respect. The names of poetesses are given, for instance ; but the chapter culminates in a detailed description of the kavi-examination held by a committee of men selected for this purpose, who sit or stand in the examination-hall in order of rank.

The adhyāyas 11 to 18 deal with the borrowing (harana) from older poets. The author points out cases in which borrowing is or is not allowed, and goes on to treat all such possible cases. In Rājaśekhara's opinion, harana must have been of great importance, since the matter is dealt with in more detail than is agreeable to the reader. The fourteenth and fifteenth chapter (jātiraya-kriyāsamayasthāpanā and guṇasamayasthāpanā) investigate all kinds of poetic licences and customs, which, though not correct in the strict sense, are sanctioned by tradition. Among many other things we find here the well known relations between colours and affections or conditions of mind and the like.

In the next chapter (svargyapatāliyakavirahasya-sthāpanā), the author presents some special doctrines for the svargya and pātāliya poet. The seventeenth adhyāya is more interesting and important, because we glean a geography of India from it. The details given here, however, were not unknown even before this discovery of the Kāvyamīmāṃsā, for the Jaina monk and polyhistor Hemacandra and another writer on alamkāra, the younger Vāgbhaṭa, have almost the same deśavibhāga included in their works. That both have borrowed from Rājaśekhara we did not know before the latter's work was published. The Kavirahasya winds up with the eighteenth adhyāya (kālavibhāga), which gives a description of the division of time.

We stayed a little longer with Rājaśekhara than with the ālamkārikas before him. This we did because our author deals with poetics in a rather different way and illustrates matters which are not to be found in the works of his predecessors. Later writers on alamkāra, Hemacandra and Vāgbhaṭa, have borrowed long passages from Rājaśekhara, often almost verbatim.

Of the later writers on poetics only the more important names shall be mentioned here. Rājanaka Ruyyaka or Rucaka, who lived in the beginning of the twelfth century, has treated only the alamkāras in his Alamkārasarvasva. The way Ruyyaka takes up the matter and the exact and correct manner in which he delivers the doctrines of the figures of speech and groups them, show clearly what progress had been made regarding

poetry. The form of expression is somewhat similar to that of Vāmana, or rather to that of the Dhvani-kāra : the sūtra-text and a commentary on the sūtras which, however, is much more exhaustive than the commentary of Vāmana and constitutes really the main part.

The older Vāgbhaṭa, who, like Ruyyaka, also belongs to the beginning of the twelfth century, treats of the whole domain of poetry. His work is entitled Vāgbhaṭālaṃkāra and is written in verse throughout. Vāgbhaṭa begins with the definition of the kāvya and, thereupon, speaks of the purpose of poetry and the causes of it. After giving a division of the kāvya, he presents the doctrines of the doṣas, the guṇas, the dhvani, the śabda- and arthālaṃkāras, the rītis, and the rasas. It must be confessed that the importance of Vāgbhaṭa is not great.

The compendium of poetics, the Kāvyānuśāsana, of the celebrated Jaina monk, Hemacandra (1088-1172), is a good deal more extensive. He also wrote a great commentary on his own work, the Kāvyānuśāsanaviveka. It is a matter of course that Hemacandra could not present original ideas of his own, as he unfolded a somewhat astonishing literary activity. He was a complete master of the knowledge of his time, and had thus in many cases only registered the theories and results his predecessors had found. The sources, however, which Hemacandra used, are partly unknown to us. The discovery of the Kāvyamīmāṃsā of Rāja-śekhara has shown that Hemacandra in his commen-

tary has borrowed long passages from it ; but the name of Rājaśekhara is never mentioned. It is possible that other parts of the Kāvyānusaśanaviveka may go back to other books of the Kāvyamīmāṃsā which have not yet been found.

Hemacandra begins in the usual way with the topic of the effect and cause of poetry. The definition of the kāvya (adoṣau saguṇau sālamkārāu ca śabdār-thau kāvyam) indicates at the same time the mode of his treatment of different topics. It is explained what guṇa, doṣa, and alamkāra are, and it is pointed out in which way these ideas are connected with the rasa. For, though the rasa is a factor of immense importance, it is not mentioned at all in the definition of poetry, because in the opinion of Hemacandra the kāvya as consisting of word and sense (śabdārtha) is sanctioned by tradition.

The author then presents an account of the doctrine of the word, in course of which the dhvani doctrine is touched on, apparently under the influence of the Kāvyaprakāśa. Hereafter the rasa, and then all kinds of doṣas (rasa-, pada-, vākya-, pada-vākya- and artha-doṣas) are characterised. The description of the guṇas, of which only three are mentioned, also shows among many other things the influence of Mammata. The śabda- and arthālamkāras are detailed in the fifth and the sixth book respectively.

This is followed by the doctrine of the nāyaka (hero) and his female counterpart, the nāyikā (heroine). The division of the kāvya forms the last part of Hemacandra's work.

In the 13th century Vāg bhata the

younger—the older Vāgbhaṭa has been mentioned above—composed a work on poetics. In his own commentary on his work, Vāgbhaṭa is influenced also by Rājaśekhara. Vāgbhaṭa's book, though a very large one, is of no great importance. New ideas are scarcely given. In the beginning are discussed the very well known question about the offects and the causes of poetry and the definition of the kāvya, which discussion takes the same form as that of Vāgbhaṭa's predecessors. Before illustrating the various parts of the definition, the author speaks of the divisions of poetry. The second and third adhyāya contain the doctrine of the doṣas (śabda-, vākya-, artha-doṣas), of the guṇas (the number of which is given as ten, after Daṇḍin), of the arthālambanākāras. There are 62 arthālambakāras, while Hemacandra deals with only 29. The absence of doṣas and the presence of guṇas and alamkāras form the body (śarīra) of the kāvya, its soul (prāṇa) being the rasas, which are detailed in the fifth adhyāya.

After this succinct account of more the important representatives of Indian writers on poetics23,

23 A complete account of all ālamkārikas (with the exception of Rājaśekhara, whose work was not discovered then) is given by Hari Chand in his book : Kālidāsa et l'Art poétique de l'Inde, Paris 1917. There the reader will find further details with respect to the less important writers on poetic matters and the time they lived in. A work very often referred to is the Sarasvatī-kaṇṭhābharana of Bhojarāja (about the middle of the 11th century A. D.). It is, however, a mere compilation and, therefore, is of very small interest. Bhojarāja has

36 THE FOUNDATIONS OF INDIAN POETRY

let us now deal with the historical development of the theoretical foundations of Indian poetry and see in what manner old and simple ideas are changed in course of time. We shall consider in the following pages the effects of good poetry and the causes of it, and shall gain an idea of the qualities which the poet should possess. After this, the various attempts to define the kāvya will be illustrated, and then the general position of the alamkāra will be touched on. A further question is : what is to be considered as the soul of the kāvya ? A large space will be taken up by the doctrine of the so-called ritis and the guṇas (though not the whole theory of guṇas) closely connected with these. Finally, we must treat of the division of the kāvya. In this chapter we shall meet with important questions, which in more than one respect are of the greatest interest.

borrowed long passages from Daṇḍin's Kāvyādarśa. Another author of the same time, Kṣemendra, wrote two works on the alamkāraśāstra the Aucityavicāracarcā and the Kavi-kanṭhābharana, which last title reminds one of the work of Bhojarāja. The value of these works is not great. On account of its clearness of description the Candrāloka of Jayadeva is highly valued in India. The author lived in the 13th century and also wrote a drama, the Prasanna-rāghava. The works of Vidyādhara and Vidyānātha (about 1300 A. D.) the Ekāvalī and the Pratāparudrayasobhūṣaṇa, both edited in the Bombay Sanskrit Series, Nos. 63 and 65, are also of some importance. Jagannātha (17th century) whose Rasagangādhara, though incomplete, takes the first place among all the works of the preceding centuries on account of his independent and original researches, belongs to the modern eriod.

THE EFFECT OF POETRY

Before entering on the discussion of the kāvya itself there are some things of a more general character which the writers on alaṁkāra deal with in an introductory way. When anything is to be understood, it is asked what effects and what causes may be connected with it. The oldest author, Bhāmaha, treats the question of the effects of good poetry in connection with the question about the qualities of the poet. Not to disturb the context, the verses of Bhāmaha in question will be given when we speak of the kavi. Hero we can therefore be short.

Bhāmaha says I,2 : dharmārthakāmamokṣeṣu vaicakṣaṇyaṃ kalāsu ca / pritiṃ karoti kīrtiṃ ca sādhukāvyanibandhanaṃ // "The composition of a good kāvya presents cleverness in (the caturvarga, i.e.) right, wealth, love and liberation, and in the arts, and (presents further) pleasure and fame."

Dandin sees the matter from another point of view, declaring in Kāvyādarśa I, 3-6 iha śiṣṭānusiṣṭānāṃ śiṣṭānāṃ api sarvathā / vācāṃ eva prasādena lokayātrā pravartate // īdam andhaṃ tamaḥ kṛtsnaṃ jāyeta bhuvanatrayam / yadi śabdāhavyayaṃ jyotir āsaṃsāraṃ na dīpyate // ādirājayasobimbam ādarśaṃ prāpya vāṅmayam / teṣāṃ saṃnidhāne 'pi na svayaṃ paśya nāśyati // gaur gauḥ kāmadughā samyak prayuktā smaryate ciram / duṣprayuktā punar gotvaṃ prayoktuḥ saiva śaṃsati //

"By the favour of words, as well of those which are taught by the grammarians (Sanskrit and Prakrit) as of the others1 (the various dialects which are not expressly taught by the scholars), business in the world is prospering.

"If the light, called word, did not shine into this existence, then these three worlds would be in complete darkness.

"The idol of fame of the first kings, which manifests itself through the mirror of words, does never vanish even when (the kings) have passed away.

"Well-composed speech is called a cow, which grants every desired object ; but composed in a bad way, speech manifests only the ox-nature of the composer2."

Dandin states the relation between the kāvya and the fame of the king as follows. According to him, the main purpose of a poem is to narrate and praise the life and deeds of the king, the kavi being thus, generally, a court poet.

Vāmana appears to depend on Bhāmaha, saying in Kāvyālankārasūtravṛtti I. 1. 5 :

Kāvyam saddṛṣṭādrṣṭārtham pritkīrtitihetutvāt.

"Because the kāvya is the cause of pleasure and fame, its effect is (twofold) : to be seen and not to be seen."

Pleasure is visible, but fame is not visible, as

1 Premacandra has a second interpretation of the word śiṣṭa : svataḥ śiṣṭānām sañjñā-rūpāṇām. But the first appears to be the better.

2 Premacandra : gotvam vṛṣabhatvam mūr-khatvam-ityarthah.

it manifests itself mostly after the poet has passed away. Fame, however, stands higher than pleasure, as Vāmana points out in the following verses :

"The merit of composing a kāvya is considered as an uninterrupted road (leading) to fame. The ridiculousness of being a bad kavi is said to be a way to shame.3

"The learned ones designate fame as something leading to heaven in the end, shame, on the other hand, as leading to the places of hell. To attain fame and to avoid shame, the "king-poets" ought to study the contents of (this) Kāvyālaṃkāra thoroughly."

The matter in question is dealt with in a much more detailed way by Rudrata, whose words run thus (Kāvyālaṃkāra I. 5-13) :

"When in course of time the temples etc., erected by the kings are fallen to decay, then even the names (of the kings) would not remain, if the kings had no good poets.

"Does not a person render a good service to another by displaying his fame, which will last for ever and will be pure and faultless and desirable for all the world ?

"And the learned ones4, who have studied the highest things, agree with each other in stating that to bestow benefits on a person is of quite a special merit.

3. Compare Raghuvaṃśa I. 3: mandal kaviśaḥprārthī gamiṣyāmy upahāsyatām.

4. Vādinah. Or should it be translated : " men, who in other cases dispute with each other, agree in this matter ... "?

40 THE FOUNDATIONS OF INDIAN POETRY

That is the very idea of which Dandin has spoken :

the kavi par excellence is the court poet of the

king, and the king must bear in mind what the

poet may become to him ! But besides that, the

blessings of good poetry are evident also with respect

to other things :

"The poet who has composed beautiful panegyrics

of the gods attains wealth, liberation from evils,

fortune beyond measure, in short, all he may

desire.

"So some poets by praising Durgā have crossed

(the stream of) misfortune which is hard to do ;

others have turned away sickness ; others again

have got a much-longed-for gift.

"The gods who have fulfilled the poet's wishes

as a reward for the panegyrics are the same even

to this day, though the kings have changed.

"Yet, why shall I say much ? Who in the world

may be enabled to comprehend to which high degree

the kāvya, this ocean with its jewels, brilliant

excellences (the poem is filled with), becomes the

cause of eminent fame ?"

In the next verses Rudraṭa varies the words of

Bhāmaha :

"Therefore the wise ones, who wish to gain a

full success in the (four) manifestations of human

life (dharma, artha, kāma, mokṣa) may compose

a kāvya as faultless as possible after having studied

what is necessary for it.

"For, the fine formation of speech derived from

the acknowledged śāstras (dealing with the correct

formation) of word and sentence, is the fruit of

the learned ones and well-formed speech,on the other hand, matures a splendid kāvya."

M a m m a ṭ a 's statements about this matter may be quoted, as they contain some interesting details. Kāvyaprakāśa 1-2:

Kāvyaṃ yaśase 'rthakṛte vyavahāravide śivetarakṣataye, sadyaḥ paranirvṛtaye kāntāsammitatayopadeśayuje.

The vṛtti thereon runs thus :

"The composition of a kāvya leads to fame, to wealth, to understanding of all worldly business, to removal of all sorts of evils, to immediate or later pleasure joined with an instruction similar to that, which may be given by a beloved woman5)."

"Fame, as in the case of Kālidāsa and others, wealth in the case of Dhāvaka etc. from king Harṣa, understanding of all wordly business for kings etc., removal of all sorts of evils, as in the case of Māyūra etc. from the Sun etc. The crown, however, of all is the complete pleasure, which is caused by the taste of rasa and which takes away all other sentiments."

The last words show that no little progress has been made in judging poetry: the last cause, why a good kāvya is a course of untroubled joy and pleasure, is the rasa, the poetic sentiment, which is the strong quintessence of all poetry. Regarding the matter from this point of view Mammaṭa continues to show in which respect poetry differs from other kinds of literary compositions:

5 This stanza is quoted by Mallinātha in the beginning of his commentaries to the five mahākāvyas.

VI

42 THE FOUNDATIONS OF INDIAN POETRY

"By the fact that in the kāvya word and sense (śabdārtha) recede into the background, the unfolding of the single parts of the rasa is placed into the foreground. Thereby the kāvya differs, on the one hand, from the sciences as Veda etc., where the word forms the main part, like the (command) of a king (prabhu-sammita), and, on the other, differs from the mere narrative literature as the Purāṇas etc., where the main part is the sense, in the manner of (an instruction) from a friend (suhrt-sammita). Thus, the kāvya is the work of a poet who is skilled in presenting a matter in a manner that goes far beyond the common way. As a beloved woman by means of her graceful attraction (rasa) gains (her lover's) interest (and so may get from him what she wishes), in the same way the kāvya presents instruction in a fitting manner both for the poet and the knowing ones (the hearers of the kāvya), an instruction similar to that of Rāma and not to that of Rāvaṇa. Therefore one should use the utmost endeavour with respect to the kāvya."6

Mammaṭa here gives a few more details about the matter in question than his predecessors did, regarding poetry from a higher standpoint. The ideas he speaks of in a certain degree belong already to the question as to the real nature of poetry and to the definition of the kāvya.

6 The Ekāvalī, which generally is based on the Kāvyaprakāśa, has borrowed these similes from Mammaṭa in the verses I, 3-6.

THE CAUSES OF THE KAVYA AND THE POET

Scholars agree, as we have seen, that a good kāvya produces many desirable effects for both the author and the hearer. They agree also that these effects can only be expected from the composition of a good kāvya. Thus, the question of the kāvyaphala (effects of poetry) is connected with that of the kāvyakarana (causes of poetry). In other words, how should the poet (kavi) be? Going into the treatment of this matter it cannot be avoided to take up some things here which strictly speaking belong to the preceding chapter, where the effects of poetry were spoken of, but Bhāmaha does not, as mentioned before, separate kāvyaphala and kāvyakarana strictly.

Right at the beginning of his work he says (I, 3-5) :

"What liberality is to the poor, what skill (in using weapons) is to the coward, what cleverness is to the ignorant, that very same is knowledge of the śāstra to a man who is no kavi (by nature).

What is wealth without good conduct, what is night without the moon, what is ability in speech without being a good kavi.1

Even the untalented ones are able to study the śāstra with the aid of the instruction of the teacher,

1 This stanza (Bhāmaha i, 4) is quoted as an example for the arthālamkāra Vinokti in Ruyyaka's Alamkārasarvasva p. 83.

44 THE FOUNDATIONS OF INDIAN POETRY

but the kāvya can be composed only by a man who possesses the pratibhā (the working genius).

Not everybody has the art of producing literary compositions, and one who is not endowed with it by nature can never learn it by studying the śāstra.

The innate genius is called pratibhā, verbatim the ' beaming of the ideas.'

On this occasion Bhāmaha speaks of the effects of good poetry (I, 6-8) :

"Even when those who have produced a good literary composition go to heaven, their body, which consists of the kāvya, remains pure and pleasant (on the earth).

And as long as the poet's fame, everlasting, fills heaven and earth, so long the poet, who has done good work, attains the celestial abode.

One who, therefore, longs for fame that will last as long as the world will remain, may endeavour to compose a kāvya after having studied all that the kavi should know."

There is no doubt that the genius is the conditio sine qua non for the poet; but there are, on the other hand, some other things not less necessary.

Bhāmaha deals with them in the verses 9-11:

śabdaś chando' bhidhānārthā itihāsāśrayāḥ kathāḥ, loko yuktiḥ kalā ceti mantavyāḥ kāvyayonayaḥ². śabdābhidheye vijñāya kṛtvā tadvidupāsanām, vilokyānianibandhāṃś ca kāryaḥ kāvyakriyādaraḥ.

2 With respect to the text of this verse, the end of which is not given in the edition of Bhāmaha, see the notes to Rājaśekhara's Kāvyamīmāṃsā p. 8.

The first line is found in the Dhvanyālokalocana p, 10.

sarvathā padaṃ apy ekaṃ na nigādyam avadyavat,

vilakṣaṇā hi kāvyena duḥsuteneva nindyate.

" (The doctrine of the) word,3 metre, sense of the words,4 stories which are based on the itihāsas, the (knowledge of) world (ly affairs), suitability,5 and the arts, all these are held to be the foundations of the kāvya.

After having made oneself acquainted with (the doctrine of) word and sense, after having devoted oneself to the teaching of the masters who know that, and finally, after having studied the compositions of other poets one should endeavour to make a kāvya.

One should never utter a word which contains something blamable, for on account of a kāvya that shows bad features one is blamed as on account of a bad son.

Bhāmaha is very hard on the bad poet (I-12):

" The fact that a man is no kavi does not imply cate his dharma-lessness, his illness, or even his

3 The doctrine of the word is the grammar (śabdānusāsana); see Vāmana I, 3, 4.

4 The author apparently means the knowledge of the dictionarics; cf., below, Vāmana I, 3, 5: abhidhānakośāt padārthaniścayah. An interpretation different from this is presented by Udbhata in his (not yet edited) commentary to Bhāmaha (Bhandarkar Commemoration Volume, p. 405, note 2): śabdānām abhidhānām adhivyāpāro mukhyo gunavṛttis ca. Trivedi remarks: Should it not be explained as abhidhīnam abhidhāvyāpārah upalakṣaṇam idam tena lakṣaṇavyavajanayor api grahaṇam'ity arthah? For Bhāmaha is not dhvanyabhāvavādī, one who does not believe in the existence of the suggested meaning."

5 Then things are to be considered in a suitable way.

punishment. But the state of a bad poet is called the embodied death itself by the knowing ones."

The postulate which are laid down by Bhāmaha are taken up by D a n d i n, only a little more pregnantly. We read in kāvyādarśa 1-103 :

naisargikī ca pratibhā śrutam ca bahu nirmalam, amandāś cābhiyogo'syāḥ kāraṇaṃ kāvyasampadaḥ

"Innate talent, great and profound scholarship,0 uninterrupted practice7 are the cause of that success of the kāvya."

In the next verses, however, Dandin proceeds to contradict his predecessor in an important point. In the opinion of Bhāmaha the pratibhā (genius) is a necessary condition without which a kāvya can by no means be composed. Dandin denies this statement altogether by declaring 1, 104, 105 :

"If there is not to be found the wonderful genius, which is connected with impressions from preceding births, then, notwithstanding, the vāc (Goddess of Speech) presents some favour, when worshipped by scholarship and effort.

Therefore those who strive for fame should put aside idleness and continually and intensively serve the Goddess of Speech. Even if there is but small talent for poetry a man who shows energy can enjoy

6 The vṛtti runs: bahu anekaṃ chandovyākaraṇakośa- kalācaturvargagajaturagakhadgadādilakṣaṇātmakam ity arthah. Nirmalam sadupadeśena niḥsandehamadhigatya samyak- pariśīlitam ity arthah. Premacandra has taken the first part from the commentary of the Kāvyaprakāśa.

7 Commentary: kāvyajñopadeśena paunahpunyena pra- vṛttih.

THE CAUSES OF THE KAVYA AND THE POET

himself in the company of the learned ones. These statements illustrate very lucidly the manner of Dandin's polemic against Bhāmaha. Also with respect to things about which there should not be a doubt and which are acknowledged by all the later writers on alamkāra the author of the Kāvyā-darśa presents a view opposite to that of Bhāmaha. This opposition seems to have its last reason in some personal dislike against a rival.

The matter is treated much more exhaustively by Vāmana. His words are not without a certain originality. General reflections about the kavi open the second adhyāya of the first book. Vāmana knows two sorts of poets: arocakinaḥ satṛṇābhyavahārinaś ca kavayaḥ "There are poets who suffer from want of appetite and there are poets who eat even grass."8

As the vrtti explains, these terms are used in a metaphorical sense.9 The first group of poets is very difficult to please in all that appertains to poetics. Composing a kāvya they proceed with the greatest diligence. They discriminate accurately be-

8 In Rājaśekhara's Kāvyamīmāṃsā these words are attributed to an author Maingala by name. We read (p. 14): "te ca dvidhā' rocakinah satṛṇābhyavahārinaś ca" iti Mainga-lah. "kavayo' pi bhavanti" iti VāmanIyāḥ. . . . . "tatra vivekinaḥ pūrve tadviparītās tu tatonantarāḥ" iti Vāmanī-vāl. Nothing is known with regard to this Maingala, nor can I find his name in any of the works of the ālamkārikas. Two verses of his are quoted in the Saduktikarnāmṛta, p. 290, as is shown by the editor of the Kāvyamīmāṃsā.

9 arocakisatṛṇābhyavahāriśabdau gaṇārthau.

tween what is fit and what is not fit. The second group, on the other hand, does not differentiate good from bad. In other words : the one is vivekin, and the other avivekin (discriminating and not discriminating).10 Only the former class of poets are worth being instructed on account of their being endowed with the ability of discriminating (I, 2, 2: pūrve śisyā vivekitvāt), whilst the other should not be instructed, because they cannot discriminate (I. 2, 3 : netare tadviparyayāt).11 The śāstra is not made for all : for the second group of poets it will be of no advantage (I. 2, 4: na śāstram adravyēṣu arthavat), as the nut which is employed for clarifying muddy water is by no means able to clarify a swamp (I, 2, 5: na katakam paṅkaprasādanāya).12

The so-called kāvyakarana is the subject of the third adhyāya of the same book. We meet the same subjects which are with but slight variations dealt with by all ālamkārikas. Vāmana calls them kāvyāṅgas. There are, he says, three kāvyāṅgas: the world

10 ko' sāv arthah? vivekitvam avivekitvam iti.

11 tadviparyayād avivecaśīlatvāt.

12 With the nut of the kataka tree the sides of vessels were rubbed, by which procedure the unclean elements of the water were said to segregate on the sides. Cf. Manu vi, 67: phalain katakavṛkṣasya yady apy ambuprasāda-kam, na nāmagrahaṇād eva tāsyā vāri prasīdati.

Kālidāsa mentions the kataka nut in a simile in his Mālavikāgnimitra II. 8: mando'py amandatām eti saṃsargeṇa vipaścitah, paṅkacchidah phalasyeva nikaṣenāvilam payah.

It is often spoken of, too, in the works on medicine.

THE CAUSES OF THE KĀVYA AND THE POET

(loka), the sciences (vidyās), and matters of various kinds (prakīrṇa).

It need not be proved that the stories told in the poem must be probable or possible, that they must agree with the rules valid for this world (I, 3, 2: lokavrttān lokaḥ).

Then, the poet must possess the knowledge of the sciences, which are enumerated in I, 3, 3: śabdasmrtyabhidhānakośacchandovicitikalākāmāśāstradanītipurvā vidyāḥ.

" These are the sciences : grammar, dictionary, prosody, the doctrines of the arts, the doctrine of love, the doctrine of politics, and others."

Śabdasmrti is identical with vyākaraṇa. From the grammar the poet learns the correctness of speech (4: śabdasmrteḥ śabdaśuddhịḥ).13

The dictionary gives security with regard to the sense (5: abhidhānakośāt padārthaniścayah).

The doctrine of prosody takes away the doubts which arise regarding metre (6: chandoviciter vrttasaṃśayacchedạḥ).14

The kalāśāstras give information about the arts (7: kalāśāstrebhyạḥ kalā tattvasya saṃvit). The vrtti mentions singing (gīta), dancing (nṛtya), painting

13 śabdasṃrteḥ vyākaraṇāt śabdānāṃ śuddhịḥ śādhutvaniścayạḥ kartavyạḥ.

14 In the Kāvyādarśa (I. 12) and the Kāvyamīmāṃsā the term prosody is also rendered by the word chandoviciti, and in other works one will find the same. Therefore the hypothesis, according to which the term chandoviciti in the Kāvyādarśa does not denote prosody generally, but a work thereon, composed by Daṇḍin, is incorrect.

vii

50 THE FOUNDATIONS OF INDIAN POETRY

(citra). Apparantly there existed compendiums with regard to the arts referred to.

From the kāmasāstra one gets the knowledge of the usages respecting love (8: kāmasāstrataḥ kāmo-vacārasya [sampriti]).

The compendium of politics which by the vrtti is named arthasāstra15 shows in which cases the sixfold way in foreign affairs should be adopted or avoided (9: dandanīter nayāpanayayoh [samvit]).16

The knowledge of the arthasāstra is of importance, because the intrigues of the plot are founded upon it.17

In I, 3, 11 we are told which matters are included in the subject prakīrṇa.

"(Under) Miscellaneous matters (must be understood) knowledge of examples, practice, reverence for the masters, careful examination, talent, and close attention."

Knowledge of examples is intimate acquaintance with the already existing kāvyas (12: tatra kāvyaparicayo lakṣyajñatvam).18

15 The Arthaśāstra by Kauṭilya is very likely meant.

16 tatra śādgunyasya yathāvat prayogo nayah tadvipari-to' panayaḥ nahi tāv avijñāya nāyakapratināyakayor vṛttạṁ śakyạṁ kāvye nibandhum.

17 Vṛttiḥ: itihāsādih itivṛttạṁ kāvyāśarīrạṁ tāsya kuṭilat-vaṃ tato dandanīteh ābalīyasādiprayogavyutpattimūlatvāt tasyāḥ.—This, in the first place, is valid for the drama, but it has already been mentioned before that the theories of the drama were to a great extent transferred to the kāvya in the widest sense of the word.

18 anyeṣāṃ kāvyeṣu paricayo lakṣyajñatvam, tato hi kāvyabandhasya vyutpattir bhavati.

Abhiyoga is the practice which is to be gained if one, by way of trial, composes kāvyas, or, at least, fragments of a kāvya (13: kāvyabandhadyamo’ bhiyogah).

Reverence for the masters is the willing obedience to the teachers, when they give instruction with regard to kāvya (14: kāvyopadeśaguruśuśrūṣaṇam vrddhasevā).

Avekṣaṇa is the exact examination, whether a certain word should be used or not (15: padādhāno-ddhāraṇam avekṣaṇam).

In commenting upon this sūtra Vāmana also mentions another term, kāvyapāka which is of more importance than appears at first sight. The author says:

“As long as the mind is uncertain, (the process of) taking or avoiding (of words) exists: but when a word is absolutly fixed, then speech is complete.

When the words have come to that state that they no longer admit of being altered, then this is called Ripeness of words by those who are expert in the employment of words (in compositions).”19

Vāmana mentions kāvyapāka a second time, at the end of the third adhikaraṇa, which contains the definition of the guṇas. He quotes three verses (atra ślokāl); the first two are the following:

“When the guṇas are entirely clear, one speaks of ripeness of the kāvya; and it is compared with the ripeness of the mango.

19 This is quoted in Rājaśekhara's Kāvyamīmāṃsā with some variations. See below.

52 THE FOUNDATIONS OF INDIAN POETRY

A kāvya which is endowed with correctly formed nouns and verbs, but where the guṇas of the sense are bad, is like the ripe vrntāka fruit: people dislike it.

This ripeness of kāvya we will meet again later on, when we deal with what Rajaśekhara has to say on the nature of the poet. Here we only wish to draw attention to the fact that kāvyapāka is never mentioned in one of Vāmana's sūtras, but only in the verses of the commentary. It is a matter of interest that the vṛtti goes a good deal farther than the explanation of the sūtras requires.

Talent is the germ of poetry (16: kavitvalījāṃ pratibhānam). This pratibhāna (which, of course, is identical with the pratibhā of Bhāmaha) is an inborn talent and a mental impression, the last cause of which lies in a former existence. With regard to this definition Vāmana appears to be influenced by Kāvyādarśa II, 104. Without the pratibhāna literary composition cannot be produced, or, if a man nevertheless tries to do so, the effect will only be ridiculous.

By avadhāna the attention is meant which is directed to the one and only end, the faultless perfection of a kāvya (17: cittaikāgryam avadhānam).

20 kavitvasya bījam kavitvabījam. janmāntarāgatasam-skāraviśeṣaḥ kaścit. yasmād vinā kāvyam na niṣpadyate niṣpannam vāhāsāyatanam syāt.

21 cittasyaikāgryam bāhyārthanivṛttih tad avadhānam. avahitam hi cittam arthān paśyati. Cf. Kāvyamīmāṃsā p. 11: manasa ekāgratā samādhiḥ. samāhitaṃ cittam arthān Paśyati.

THE CAUSES OF THE KĀVYA AND THE POET

To these definitions the following sūtr′as give some supplementary explanations, which are of some interest on account of their originality: Vāmana declares that the perfect attention depends on two things: place and time (18: tad [avadhānam] deśakāla-lābhyām). The most suitable way for the composition of a kāvya is to do it in loneliness (19: vivikto deśah), the most suitable time is the fourth (the last) part of the night (20: rātriyāmas turīyah kālah), that is according to our idea, the early morning. The commentary Kāmadhenu refers to a couple of past sages in the kāvyas : Kālidāsa says in Ragluvamśa xvii, 1:

paścimād yāminīyāmāt prasādam iva cetanā,

and Māgha in Śiśupālavadha II. 6:

gahanam apararātraptābuddhiprasādāḥ,

kavaya iva mahīpāś cintayanty arthajātam.

Comparing the words of Vāmana with those of his predecessors it appears that the author depends on Bhāmaha. This one, however, is new, namely, that the points the knowledge of which is necesseary for the poet are considered under three headings: loka, vidyā, and prakīrṇa. This subdivision, however, is by no means good, for the most important point, the pratibhāna, which should be named in the first place, is enumerated under the miscellanea, as if it were something subordinate, though Vāmana himself holds the pratibhā to be the very germ of poetry. We need not be astonished that this division of Vāmana's is not referred to in later works on alamkāra.

The various points are, as mentioned before, for

54 THE FOUNDATIONS OF INDIAN POETRY

the greater part the same as those which Bhā- maha considered necessary to anyone who wishes to become a good kavi. Bhāmaha had also taught that the poet must be familiar with worldly business (loka); that knowledge of the grammar (Bhāmaha : sabda, Vāmana : sabdasmrti), of dictionaries (abhi- dhānārtha, abhidhānakosa), of prosody (chandas, chandoviciti), of the arts (kalā) must be assumed for the kavi. The sciences of love (kāmasāstra) and of politics (dandanīti, or, as it is called in the vrtti, arthasāstra) are added by Vāmana. Among the points mentioned under the title prakīrṇa Vāmana's laksyajñatva corresponds to Bhāmaha's vilokyānya- nibandhān; abhiyoga to kāvyakriyādarah; vrd- dhasevā to krtvā tadvidupāsanām; aveksana to the

idea expressed by Bhāmaha in I, 11. Pratibhāna is is only possible when there is pratibhā, Bhāmaha opens his discussions with this point. Though Vā- manā is of the same opinion with regard to pratibhā, because he calls it the germ of poetry, he mentions this most important factor only by the way, as it were, under the title of prakīrṇa. The avadhāna, which Vāmana then speaks of is not mentioned by Bhāmaha, but it is not very different from aveksana. On the other hand, we do not find Bhāmaha's yukti nor his itihāsāśrayāh kathāh in Vāmana's sūtra, which last subject is of still more importance. In the vrtti to sūtra 10, however, these kathās are considered by Vāmana as kāvyasarīra; so they are not, strictly speaking, a kāraṇa for the poet, but for the founda- tions of the kāvya itself.

THE CAUSES OF THE KĀVYA AND THE POET

The ideas of Vāmana only rarely met with acknowledgment by the later writers on poetics, who more or less dealt with the matter in a way similar to Bhāmaha and Daṇḍin. Udbhata in his book on the alamkāras has no occasion to touch this fundamental question, but Rudrata refers to the Kāvyādarśa, when he declares in the Kāvyālamkāra I, 14 ;

tasyāsāranirāsāt sāragrahaṇāc ca cāruṇaḥ karaṇe, tritayam idam vyāpriyate śaktir vyutpattir abhyāsah.

"To avoid the unbeautiful and to take the beautiful there must be these three things for the compositions of a good (kāvya): talent, scholarship, and practice."

These terms śakti, vyutpatti, and abhyāsa are apparently chosen only for the reason that he does not wish to copy Daṇḍin verbatim.

In I, 15 we are informed as to what must be understood by śakti :

manasi sadā susamādhini visphuraṇam anekadhābhidheyasya, akliṣṭāni padāni ca vibhānti yasyām asau śaktiḥ.

"When in a well-concentrated mind many ways, ideas, and words, which are not worn out, spring forth, it is called śakti."

Thus śakti is but another word for pratibhā, which is used in the Kāvyādarśa. It seems, however, that Rudrata in the term śakti includes a little more than the mere pratibhā, or poetical talent, namely also some thing contained in 'ṛuta', for śakti does not refer only to the idea or the sense (abhidheya) which the poet whishes to express, but also to the bearer of the idea, the word. That in this con-

56 THE FOUNDATIONS OF INDIAN POETRY

nection pada is synonymous with sabda and does not mean the sentence (vākya) is shown by its standing in contrast to adhidheya. In the following stanzas the quality of sakti is illustrated : pratibhety aparair uditā sahajotpādyā ca sā dvidhā bhavati, puṃsā saha jātatvād anayos tu jyāyasī sahajā. utpādyā tu kathamcid vyutpattȳ janyate parayā.

" The sakti, named by others as pratibhā, is twofold; innate and to be acquired. Among these two sorts the innate (sakti) is the better one on account of its being produced together with the man.

For that (innate sakti) only seeks for its further accomplishment another cause; but the second sakti is with some difficulty to be produced only by the highest accomplishment itself.

Thereby the author wishes to say that the innate sakti, though through the saṃskāra existing as such, must yet undergo in some way or another a certain development from outside, if it is to be useful for the special purpose of composing a kāvya. This is in accordance with the above stated fact that in Rudrata's sakti something which strictly belongs to the term 'śruta' is already contained. The second kind of sakti, on the other hand, being not yet in existence, must be produced and has thus to seek a primary cause.

Apparently the mode of Rudrata's treatment of the sakti is an attampt to remove the inaccuracy with which the pratibhā is dealt with in the Kāvyādarśa. The naisargikī pratibhā of Daṇḍin is of course the same as the sahajā sakti of Rudrata, and

THE CAUSES OF THE KĀVYA AND THE POET

cannot thus be substituted by another element as is done in the Kāvyādarśa I. 104, 105. Thus the sahaja śakti is set in contrast by Rudraṭa to the utpādya śakti. Strictly speaking, however, Rudraṭa is no more right than Daṇḍin. The matter in question is treated in a satisfactory way only by Bhāmaha.

Vṛutpatti is thus defined by Rudraṭa, I, 18; 19: "Vṛutpatti, in the more limited sense, is the correct distinction between what is suitable and what is not suitable, on account of the knowledge of prosody, grammar, the arts, worldly business, the word, and the sense of the word.

But in the wider sense, is there anything to be found in the world at all, whether that should be expressed (the word), which may not become an element of the kāvya? Therefore this (vṛutpatti in the wider sense) is the knowledge of everything."

Only an illustration, not a definition, is given with regard to the third question, the abhyāsa;23 I. 20: "A learned and talented man should, after having studied all branches of human understanding, under the guidance of a good poet and an expert man continually, by day and night, practise (the composition of) the kāvya."

As Rudraṭa is wholly influenced by his predecessors, so also Mammaṭa presents no new ideas in saying. "Talent (śakti), experience (nipụṇatā) with respect to the world, the śāstra, the kāvyas, etc., and practice (abhyāsa) based upon the instruction (by a teacher) who is expert in the kāvya, are the (three) causes for a kāvya"

22 Namisādhu: abhyāso lokaprasiddha eva. viii

57

58

THE FOUNDATIONS OF INDIAN POETRY

Śakti, which term is apparently taken from the Kāvyālamkāra of Rudrata, is again identified with Bhāmaha's pratibhā. It may be considered, says the author in the commentary, as the germ of kavi-tva, without which nobody is able to compose a kāvya, if he does not wish to produce something to be laughed at23. This last note Mammata has borrowed from Vāmana24.

Nipunatā comprises the knowledge of all the poet should understand, as the business of the world, prosody, grammar, the dictionary, the arts, the so-called caturvarga (dharma, artha, kāma, mokṣa), the śāstras, which deal with elephants, horses, chariots, weapons etc. Finally one must be well acquainted with the kāvyas already in existence25.

Supposing he is in possession of both śakti and nipunatā, the disciple has, under the direction of a well-versed teacher, to practise composing kāvyas.

To show in which way the later ālamkārikas vary old and sanctioned ideas, the definitions and explanations of the older Vāgbhaṭa (12th century) may be given (Vāgbhaṭālamkāra I, 3ff).

"The (innate) fancy (pratibhā) is the cause of the kāvya, scholarship (vyutpatti) (is) its ornament, and practice (abhyāsa) causes productiveness, so the first kavis have said."

23 Mammata says in his commentary: śaktiḥ kavitvabīja-rūpaḥ samiskāra-viśeṣaḥ yam vinā kāvyaṃ na prasaret pra-sṛtaṃ vā upahasitiyaṃ syāt.

24 Cf. above p. 52.

25 lokasya sthāvarajaṅgamātmakalokavṛttasya śāstrāṇāṃ chandovyākaraṇābhidhānakośakalācaturvargagajaturagakhaḍ-gādilakṣaṇagranthānāṃ kāvyānāṃ ca mahākavinibandhānāṃ.

(Pratibhā)

"The (innate) fancy of the good poet, which lets the use of clear26 words be well understood and gives birth to a new tenor, is called pratibhā which is all pervading."

(Vyutpatti)

"The unique, tradition-based knowledge of the system, of words, of (the trivarga) dharma37, artha, and kāma, of politics28, of love, etc., is named vyutpatti."

(Abhyāsa)

"The assiduous devotion, which by the aid of the teachers is to be manifested with respect to the arrangement of a kāvya, is called abhyāsa. A way whereby one can come to it will now be told.

With a series of words which (by form) sustains the beauty of the structure, but has not yet any (new) tenor, one can make oneself acquainted with all sorts of metre with the view of employing them in the kāvya.

Through combinations (of sounds) one should form a length at the end of the word, one should not suppress the visargas, one should avoid bad sandhis; these are the ways which produce the beauty of the structure."

The making of heavy vowels by conjuncture with the following word perfects, says the commentary, a stability of the structure and the visargas produce the guṇa called ojas.

26 The commentary explains prasanna by akliṣṭa,

27 Commentary: dharmaśāstram āgamaḥ.

28 Commentary: arthaśāstraṃ Cāṇakyapraṇito rājanīti-granthah.

60

THE FOUNDATIONS OF INDIAN POETRY

Vāgbhaṭa gives the following examples :

śite kṛpāṇe vidhṛte tvayā ghore raṇe kṛte,

nradhīsā kṣitipā bhītyā vana eva gatā javāt.

" O king, when in the horrible battle you draw

your sword, then the enemies run from fear quickly

into the forest."

There we have the above mentioned faults: the

words stand there one by one, so that the structure

becomes very loose. The harshness can be avoided,

if the poet would produce heavy syllables by con-

tractions. Further the visarga in kṣitipā is sup-

pressed. If it were not, (by placing after kṣitipā

another word than bhītyā) then the guṇa ojas, in

this very case of a peculiar effect, would be taking

place. Finally a bad (though no false) sandhi lies

in nradhīsā,

Now Vāgbhaṭa speaks of the tenor :

" If the construction of a new tenor will not

succeed because the pupil does not possess enough

experience as yet, he should endeavour also in the

conversations to become able to find out a new

tenor,"29

After having illustrated this statement with an

example, the author touches the question of borrow-

ing from other poets.

"Forming the poetical combination of the tenor

of the compositions of other poets may be (some sort

of) exeroise. It is, however, not very fair, because

the poet thereby becomes a thief.

29 arthasaṅkalanātattvaṃ arthasya abhidheyāsya saṃka-

lanātattvaṃ saṃghaṭanārahasyaṃ padyabandhavidhilakṣaṇaṃ

saṃkathāsv api parasparālāpesv apy abhyasyet.

THE CAUSES OF THE KĀVYA AND THE POET

Only in the samasyā30 the borrowing from other kāvyas becomes a merit rather for the poet, for then he produces a new tenor, which combines with the tenor of the (prior poem).

To some other things, which are of advantage to the poet, the author draws attention in the next stanza :

manahprasattih pratibhā prātahkalo' bhiyogitā, anekāśādarśitvam ity arthālokāhetavah.

" Clearness of mind, fancy, early morning, practice, acquaintance with the numerous śāstras: these are the causes for finding the tenor."

The term prātahkāla shows the influence of Vāmana, but with the exception of that the old trinity pratibhā, vyutpatti, and abhyāsa is preserved, the words vyutpatti and abhyāsa being taken, probably, from the Kāvyālaṃkāra of Rudraṭa.

Rājaśekhara deals very exhaustively with the matter in question in his Kāvyamīmāṃsā. Only some of his ideas can be stated here. According to him there are two sorts of disciples: the buddhimat and the āhāryabuddhi. These terms appear to have been borrowed from Kauṭilya's Arthaśāstra31. The buddhimat is a man whose mind of itself follows the śāstra (yasya nisargaṛaḥ śāstram anudhāvati buddhịḥ saḥ buddhimān), whilst the buddhi of the āhārya-buddhi is educated by the occupation with the śāstra (yasya ca śāstrābhyāsạḥ saṃskurute buddhim asāv

(yasya ca śāstrābhyāsaḥ saṃskurute buddhim asāv

30 A sort of poetry where the poet has to complete a stanza the beginning of which is given.

31 I,17: buddhimān āhāryabuddhir, durbuddhir iti punaraviśeṣaḥ,

62 THE FOUNDATIONS OF INDIAN POETRY

āhāryabuddhih). The buddhi is of three kinds: referring to smrti (remembrance of past things), to mati (understanding of present things), and to prajñā (knowledge of future things). This threefold buddhi is the helper (upakartrī) of the kavi. Thé main peculiarity of both the buddhimat and the āhārya-buddhi is the devotion to a good teacher (sugurūpā-sanā).

When anyone composes a kāvya, then, in the opinion of Śyāmadeva,33 says Rājasékhara, the poet's samādhi i.e, the concentration of the mind of one subject (ekāgratā) begins to work, but according to Mañgala34 the abhyāsa, or uninterrupted practice must be considered as the chief requisite for the poet. Abhyāsa is defined as avicchedena sīlanam.

The author of the Kāvyamīmāmsā replies that samā-dhi is the interior and abhyāsa the exterior effort, and that these two together bring the śakti (the active power) into prominence (udbhāsayatah), and this śakti is the sole source of poetry.

Then the author explains the relation between sakti on the one, and pratibhā and vyutpatti on the other hand. The two functions, or objects of pratibhā and vyutpatti, have the śakti as their agens or subject34). In other words: only if there is śakti,

śiṣyumāṇo dharmārthāv upalabhate cāmutiṣṭhati ca buddhimān. upalabhamāno nānutiṣṭhaty āhāryabuddhiḥ, apāyanityo dharmārthaveṣī ceti durbuddhiḥ.

32 Nothing is known about this writer on alaṅkāra,

33 Another unknown ālaṅkārika.

34 śaktikartṛke hi pratibhāvyutpattikarmaṇI.

THE CAUSES OF THE KĀVYA AND THE POET

63

pratibhā, and vyutpatti then a kāvya can be produced35. From this it becomes clear that in the opinion of Rājaśekhara śakti and pratibhā are two quite different things and not synonyms, as is the case in Rudrata's Kāvyālamkāra, and this idea of Rājaśekhara seems to be a new one36.

The author defines the pratibhā as follows:

"Pratibhā is the (agens) which in a pleasant way makes the richness in words, the affluence of sense, the system of alamkāras, the diction, and other things of this kind appear."

For anybody who does not possess pratibhā the richness in words becomes invisible, but for the man who is endowed with it word and sense stand before his very eyes, as it were, even if he be blind37. The great poets, too, do not only see the things of this world about them, but with their intellectual eye see other countries, other continents, and are enabled to describe the deeds of the heroes of kathās. Then follow some illustrations taken from Kālidāsa's works38).

35 The author says śaktasya pratibhā śaktasya vyutpadyate.

36 That the term śakti was employed also by other scholars beside Rudrata as a synonym of pratibhā is said by Rājaśekhara himself, on p. 16 of his work. After having quoted a stanza from the Dhvanyāloka, where śakti means "genius," the author says: śaktiśaktaś cāyam upacaritah pratibhāne varttate: "the word śakti is here used metaphorically for pratibhāna."

37 As in the case of Medhāvirudra (without doubt the famous writer on poetics, who is referred to already by Bhāmaha) and Kumāradāsa (yato Medhāvirudra-Kumāradā-sādayo jātyandhāḥ kavayaḥ śrūyante).

38 Rājaśekhara illustrates deśāntaravyavahāra, dvīpāntaravyavahāra, kathāpuruṣavyavahāra; he gives another ex-

Pratibhā has a twofold aspect, being kārayitrī and bhāvayitrī. This division though not be found in any of the older ālamkārika's works, must have been discussed in a similar way before Rājaśekhara, as the author in his further explanations mentions the views of other scholars, among whom Syāmadeva is quoted by name.

The former term (kārayitrī) can be translated by "creative," but it is difficult to find an English equivalent as regards the latter. We may render it provisionally by "discriminative"39. This classification implies a difference between kavitva and bhāvikatva40).

Regarding the former kind of pratibhā, Rājaśekhara says (p. 12: kaver upakurvānā kārayitrī) "the creative pratibhā helps the poet." It helps the poet when occupied with the composition of the kāvya, its structure, its embellishment, and every thing which is connected with it. Kārayitrī, then, refers to the outer part of the work. This pratibhā ample to illustrate the ādi of deśadvipāntarakathāpuruşādi. The examples are respectively from Śakuntalā VII, 42; Raghuvaṃśa VI, 57; Kumārasambhava III, 67; Raghuvaṃśa VI. 82.

39 This is the rendering by Dr. De, who has been kind enough to give me very valuable explanations concerning Rājaśekhara's treatment of Kārayitrī and bhāvayitrī, in a letter dated Nov. 23rd 1922.

40 Dr. De draws my attention to the fact, that this distinction (Kavitva and bhāvakatva) has a resemblance to that between Imagination and Fancy made by the early 19th century Romantic critics in England. Consequently the terms Kārayitrī and bhāvayitrī can also be rendered by "imaginative" and "fanciful."

is also threefold: sahajā (innate), āhārhā (to be grasped), and aupadeśikī (to be taught). The first comes from another existence and has need of only small cultivation in the present birth; the second is based, too, on former existence, but requires great cultivation in the present existence and becomes manifest by practice in this birth; the third can be acquired only by instruction from mantra and tantra etc., in this world41. It need not be pointed out that Vāmana's and Dandin's ideas turn up here, though not in the old form, because also the bhāvayitri pratibhā was meant by those scholars.

With reference to this trinity Rājaśekhara declares that there are three sorts of poets (kavi), who are called sārasvata, ābhāsaka, and aupadeśika. This trinity, again, refers to the above described three kinds of kavi, viz. buddhimat, āhāryabuddhimat, and durbuddhi. It is evident that the sārasvatakavi is the best poet.

The latter kind of pratibhā is called bhāvayitri, with regard to which Rājaśekhara says: bhāvakasyopakurvāṇā bhāvayitri. sā hi kaveḥ śramaṃ abhiprāyaṃ ca bhāvayati.

"The discriminative pratibhā helps the discrimination, for it brings into effect the poet's effort and intention."

In other words, this pratibhā helps the poet's

41 Janmāntarasaṃskārāpekṣiṇī sahajā, janmasaṃskārayonirāhāryā, mantratanatrādyupadeśaprabhavā aupadeśikī. aihikena kiyatāpi saṃskāreṇa prathamāṃ tāṃ sahajeti vyapadiśanti, mahatā punarāhāryā. aupadeśikyāḥ punar aihika eva saṃskārakālaḥ.

ix

thinking, imagination, or intention42,

whilst the former helps him in the act of composing

the kāvya. Kārayitrī, then, refers to the outer part

of the work, whilst bhāvayitrī has nothing to do with

outward forms, whatever they may be, but creates

the inner value of the poem from within. "Through

this bhāvayitrī pratibhā," the author continues, "the

poet's tree of work becomes fruitful, otherwise it

would be barren."43.

There are, however, some scholars who do not

acknowledge the distinction between kārayitrī and

bhāvayitrī, because the kavi is a bhāvaka, and the

bhāvaka is a kavi;44 and they say:

pratibhātaramyena pratiṣṭhā bhuvi bhūridhā,

bhāvakas tu kaviḥ prāyo na bhajaty adhamāṁ daśām.

"In the world stability (of fame) is of many kinds

according to the degree of pratibhā. The bhāvaka,

however, being a poet, generally does not occupy

an inferior position."

"No, says Kālidāsa, the state of a bhāvaka is dif-

ferent from the state of a kavi, and the state of a kavi

from the state of a bhāvaka, on account of a difference

in their nature as well as their scope. It is said:

One is able to compose words (a kāvya), the other

only to hear them. Your intelligence, fortunate in

both cases, makes us wonder. For, in one object

there are not met with all excellent qualities toge-

42 It produces the "Sohanen," as the German term

would be.

43 tayā khalu phalitaḥ kaver vyāpāraratur anyathā so

vakeśi syāt.

44 Dr. De would prefer bhāvuka instead of bhāvaka

THE CAUSES OF THE KĀVYA AND THE POET

ther: one stone produces the gold, the other is able to test it.

After having pointed out that the bhāvaka need not necessarily be a kavi at the same time, and that therefore a strict distinction between both kavi and bhāvaka should be made Rājaśekara goes on to deal with the bhāvaka. He says:

"The (bhāvakās) are of two kinds, those who suffer from want of appetite (arocakinah), and those who eat even grass (satṛṇābhyavahāriṇaḥ), says Maṅgala. The kavis, too, says Vāmana. The (bhāvakās) are really of four kinds, says Yāyāvara, because they are matsarīn (envious) and tattvābhiniveśin (turning the mind to the truth) besides."

Then matsarīns are rare, but the tattvābhiniveśins are much rarer still. We shall, however, not treat the matter further here, as these statements of Rājaśekhara are of no great importance, being nothing more than some enlargements of the above described ideas of Vāmana.

After having dealt with the most important postulate for the kavi, the twofold pratibhā, Rājaśekhara continues the subject in the fifth udhyāya by speaking of vyutpatti.

The masters have said, the author begins, that vyutpatti is the state of one who knows many things (bahuśrutatā). According to Rājaśekhara, however, vyutpatti is the exact discrimination between what is suitable and what is not suitable (ucitānucitaviveka).

Now there arises the question: is pratibhā or vyutpatti the better? In the opinion of Ānanda45

Ānandavardhana. It is only the case where he is

68 THE FOUNDATIONS OF INDIAN POETRY

it is pratibhä, because it hides a fault which may spring from the poet's non-vyutpatti, and: avyutpattikrto doṣah śaktyā samvriyate kaveḥ, yas tv aśaktikṛtas tasyā jha ity evāvbhāsate46.

"A fault of the poet arisen from his non-vyutpatti is concealed by (the poet's) śakti47; But a fault which is caused by the poet's non-śakti is obvious at first sight."

Maingala, on the contrary, holds vyutpatti to be better, because vyutpatti wholly conceals a fault which the poet makes by not possessing pratibhā. Rājaśekhara endeavours to reconcile these contrary views by declaring that both pratibhä and vyutpatti must come together, and, when united, are both better (pratibhävyutpattī mithah samavete śreyasyau).

After this brilliant "correction" of the masters' view Rājaśekhara pursues the subject a little further: One who is endowed with pratibhä and vyutpatti is called "kavi", and there are three sorts of kavi: śāstrakavi, kāvyakavi, and ubhayakavi. According to the opinion of Śyāmadeva, the kāvyakavi is better than the śāstrakavi, and ubhayakavi better

referred to by name by Rājaśekhara, and is important as regards chronology.

46 Dhvanyālokalocana, uddyota iii to verse 6 (p. 137 in the edition of Kāvyamālā No. 25). avyutpattikṛto doṣah śaktyā samvriyate kaveḥ, yas tv aśaktikṛtas tasyā jhaṭity avabhāsate.

By the quotation of this stanza under the name of Ānanda (vardhana) every doubt about the authorship of the vṛtti is set aside.

47 The term śakti is used here in the sense of pratibhāna. Rājaśekhara says: śaktiśabdaś cāyam upacaritah pratibhāne varttate.

than both śāstra- and kāvyakavi, which Rājaśekhara denies, saying that in their own domain everyone of the three must be considered as the better, "as the rājahaṃsa is not qualified to drink the beams of the moon nor the cakora to distinguish milk from water." The śāstrakavi, according to his special education in the śāstras, "tears asunder (vicchinatti)" the plenty of rasa in the kāvya, whilst the kāvyakavi "loosens" the sense by the splendour (vaicitrya) of the diction (ukti), even if it is hardly to be understood by logic. The ubhayakavi, as a matter of course, is better than the two others, supposing that he is well versed both in śāstra and kāvya, and, consequently, śāstra- and kāvyakavi help one to the other. The author warns against being partial48.

Regarding the two sorts of kavi: śāstra- and kāvya-kavi, Rājaśekhara goes still farther, making divisions of both. This division is of no great interest and shows more the author's pedantry than any progress in poetical matters. While the śāstrakavi is threefold (the first composes a śāstra, the second arranges a kāvya in the śāstras, and the third puts down the sense or tenor of the śāstra on the kāvya) there are eight groups of the kāvyakavi : 1st. the racanā-kavi; he attends especially to the sounds, 2nd. the śabdakavi; of whom again there are three sorts, viz. the nāmakavi, the ākhyātakavi and the nāmākhyāta-kavi, who chiefly employ nouns, or verbs, or nouns and verbs respectively. 3rd. the arthakavi; he is

48 yac chāstrasaṃskāraḥ kāvyam anugṛhṇāti śāstraikapravaṇatā tu nigrhṇāti, kavyasaṃskāro' pi śastravākyapākam anuruṇaddhi Kāvyaikaprayaṇatā tu viruṇaddhi.

70 THE FOUNDATIONS OF INDIAN POETRY

devoted to the manifoldness of sense. 4th. the alam- kārakavi (two sorts); he likes to show śabda- and arthālamkäras. 5th. the uktikavi; he demonstrates elegance of expression. 6th. the rasakavi; his strength lies in the rasa. 7th. the märgakavi; he respects peculiarly the styles (rīti). 8th. the sästrakavi; he manifests cleverness in scholarly matters. It need not be said that these divisions have scarcely any value in themselves, and therefore we will not give the examples the author quotes to illustrate the different kinds of poets. The sorts of poet named above are again divided into three kinds, the kanīyas, the madhyama, and the mahäkavi according to whether they use only two or three, about five, or all gunas. The author then speaks of the ten con- ditions of the kavi, but we will not pursue the matter further here49.

Finally Rājasékhara mentions abhyāsa, but main- ly to deal with another subject, the päka. He says that on account of practice (abhyāsa) the good poet's speech becomes "ripe (päka)"50. As to the definition of päka there are different views (Maingala calls it parinäma, and parinäma is in his opinon the correctness regarding nouns and verbs51). This, however, is not right. This last is identical with

49 Under these new classes we find a mahäkavi again, and besides him a kavirāja.

50 Satatam abhyāsavaśataḥ sukaveḥ vākyaṃ pākam āyāti.

51 "Kah punar aya!! parināmah" ity ācāryāḥ, "Supā! tiṇāṃ ca śravah saiṣā vyutpattiḥ. The term śrava is a little strange. Is it the hearing of the (correctly formed) nouns and verbs?

THE CAUSES OF THE KĀVYA AND THE POET

71

sauśabdya⁵². According to other scholars pāka is a firmness in the employment of the wordś (padani-veśaniskampatā). Here Rājaśekhara quotes a stanza which we find in Vāmana's Kāyālamkārasūtravṛtti 1, 3, 15⁵³. Then the author mentions a similar opinion of the school of Vāmana (Vāmanīyāh), and quotes the second verse in the vṛtti to Vāmana 1, 3, 15⁵⁴. This, however, is also not correct, because this sort of pāka according to Avantīsundarī (Rāja-śekhara's wife) is nothing else than āśakti.

In the opinion of Rājaśekhara's wife ripeness exists in such cases, where the very same subject is expressed in many ways and illustrated by the mahākavis. In other words, when a poetical composition is endowed with the rasa, then indeed, it is pāka⁵⁵. Pāka therefore is far from being a quality of words, its province is rather the sense or the idea, or, still better, the way the sense communicates itself to the hearer. Rājaśekhara renders the view of his wife in the following lines:

"This is in my opinion ripeness of expression (vākyapāka) whereby the guṇas, the alamkāras, the

52 Bhāmaha I, 14. 53 See above p. 51.

54 "āgrahaparigraād api padasthairyaparayavasāyas tas-māt padānāṃ parivṛttivaimukyaiṃ pākah" iti Vāmanīyāḥ. tadāhuḥ, yatpadānī tyajanty eva parivṛttisahiṣṇutām, taṃ śabdanayāniṣṇātāḥ śabdapākaṃ pracakṣate.

The first part of this passage appears not to be a verbal quotation from Vāmana's work (where it is not to be found), but a mere rendering of the meaning of the quoted stanza.

55 Yad ekasmin vastuni mahākavīnām aneko' pi pāṭhal-paripākavān bhavati tasmād rasocitaśabdārthasūktinibandha-naḥ pākaḥ.

style, the diction, word, and sense together become tasteful to the learned ones."

The author also quotes a stanza from Vāmana's work, which here, however, is presented in another context, viz. the Vaidarbha rīti (to I 2, 11):

"There may be a speaker, there may be (good) sense, there may be (a usage of words which is correct according to) grammar, without this (pāka, as Rājaśe- khara means)56 the honey of speech will not flow."

According to the view of Rājaśekhara himself pāka is primarily conveyed by words, and hence taken as śabdavyutpatti or sausabdya ; is chiefly the province of abhidhā; yet it finds its scope only in artha or the idea, which is established by the appreciation of the men of taste57.

After that there follows an enumeration of nine defferent kinds of pāka, which we will not describe.

As regards the poet there are also many other theories mentioned and founded by Rājaśekhara. It is, however, not possible to describe all these things here ; only a few of them I may be allowed to refer to. They are given in the tenth adhyāya, the name of which is kavicaryā rājacaryā ca.

After having carefully studied the sciences (vidyā, viz. nouns and verbs, lexicography, prosody, and the doctrine of the alamkāras) and their accessories (upavidyā, viz. the sixty-four arts), one should en-

56 In the context of the verse in Vāmana's book we must understand "without the Vaidarbhī rīti." See p. 135.

57 "Kāryānumeyatā yat tac chabdanivedyaḥ paraṃ pāko 'bhidhāviṣayas tat sahṛdayaprasiddhisiddha eva vayavahārāñ- gam asau " iti Yāyāvarīyaḥ.

THE CAUSES OF THE KĀVYA AND THE POET

78

deavour to compose a kāvya. Some other things are designated the mothers of the kāvya (kāvya-mātarah), which are the presence of a good poet, news from (or about) the country (deśavārtā), the speeches of the learned ones, the course of worldly life, the meetings of the wise ones, and the compositions of the old poets. The author quotes a stanza, according to which eight things are considered as the mothers of kavi-ship: Well-being (svāsthya), fancy (pratibhā), practice (abhyāsa), devotion to yurus (bhakti), the tales of the wise ones (vidvat-kathā), wide scholarship (bahutrutyā), good memory (smrtidārdhya), and self-reliance (anirvedā58).

Further, the poet should be pure. There are three kinds of purity: of the speech, (vāksauca), of the mind (manahsauca), and of the body (kāyasaūca). The first two have their origin in the śāstras. Regarding the purity of the body the author presents the following particulars: the poet should pare the nails of his feet; he should chew tāmbūla (a leaf of piperbetel) after meals59; he should anoint the body; his garment should be splendid though not excessively so; in his hair there should be flowers; in other words, he should be a perfect gentleman.

58 Some of these things are dealt with by the author on another occasion being considered there from a different point of view.

59 Rājaśekhara says only satāmbūlam mukham, but the meaning is apparently as rendered above. Compare a stanza in the 7th ullāsa of the Kāvyaprakāśa (verse 180): tāmbūlabhṛtagallo' yaṃ bhallaṃ jalpati mānuṣaḥ, karoti khādanam pānaṃ sadaiva tu yathā tathā.

x

74 THE FOUNDATIONS OF INDIAN POETRY

As to his abode Rājaśekhara gives a full description as it should be: it is well cleansed; has rooms fit for every one of the six seasons; its garden preserved by the trees from heat, has a little pleasure-hill, lakes, and ponds, hamsas, cakoras, a bath-room, a pavilion, a palanquin, etc. etc.

The persons who are in this ideal dwelling place must, of course, be endowed with certain qualities in order to support the working poet: there should be a solitary place, where the poet can stay quite alone and undisturbed by anybody's presence. His attendants are skilled in Apabhramśa, his female servants speak a dialect mixed with Māgadhi, the ladies of his harem speak Sanskrit and Prakrit, and his friends all languages. His writer is skilled in all languages, too, and has some other excellent qualities, he should be himself a poet. The master of the house is, of course, setting the fashion also with regard to particulars concerning pronunciation and the like60.

Writing material must be close at the poet's hand. According to the masters there are the "retinue" of the science of kāvya. No, retorts Rājaśekhara, pratibhā is the "retinue."

60 The author narrates some anecdotes: Śiśunāga, King of the magadhas, had prohibited the use of cerebrals with the exception of ṇ, and of ś, ṣ, s and of kṣ. King Kuvinda of the Sūrasenas did the same with respect to harsh groups of consonants. King Sātavāhana of the Kuntalas and King Sāhasāñka in Ujjayinī gave order to speak only Prakrit or Sanskrit respectively. Regarding Sātavāhana and Sāhasāñka compare Kāmasūtra II, 7, 28 and Sarasvatīkanṭhābharana II, 15.

THE CAUSES OF THE KĀVYA AND THE POET

After some other remarks, which will be touched on elsewhere, the author speaks of the division of time. As we have seen, Vāmana has also dealt with this question, but Rājaśekhara goes much more into detail. The poet should rise early in the morning and, after having performed Sandhyā, he should read the Sārasvatasūkta. Then he may, if he likes, stay in the academy (vidyāvasatha) and study the sciences and their accesories some three hours or so (ā praharāt). The second āyāma should be devoted to the composition of the kāvya. About noon he should bathe and eat what is not forbidden. After dinner he may hold a meeting where questions concerning the kāvya are discussed (kāvyagoṣṭhī). The occupation in the third āyāma are of various kinds. In the fourth ayāma the poet should hold an examination of that part of the kāvya he has written before noon, in the presence of some learned persons, and correct and amend what is considered as being less good. In the evening he may again worship Sandhyā and Sarasvatī. By the beginning of the night (ā pradoṣāt) he should write down the corrected kāvya. After having slept well in the first and second part of the night he should rise very early, i. e. during the fourth part on the night, for early in the morning the mind sees things very clearly. This subject is treated a good deal longer still by Rājaśekhara, but we shall not pursue it further here.

Not only the men but also the women should endeavour to compose kāvya, because the samsāra, which is the working cause, is based not on the sexual difference but rather on the ātman, or the soul, if

76 THE FOUNDATIONS OF INDIAN POETRY

we are allowed to render the word in this way. There have been daughters of kings and ministers, courtezans, etc, who have understood the śāstra and have become poetesses 1.

Omitting some other remarks, which are more amusing than important, we shall deal with a more interesting subject not mentioned by any of the former representatives of the alaṁkārasastra in a few words: the examination of the kavi. The king-poet (rājakavih) should arrange a kavi-meeting. In order to examine a kāvya or a śāstra he must have a hall built, with sixteen pillars, four doors, and eight turrets, and a pleasure house (keligrha), which should be attached to the hall. In the middle of the hall there should be a vedikā one hasta high and the floor should be adorned with jewels. During the examination the king sits on this throne. To the north of him the Sanskrit poets take their seats, behind them are the Veda-learned (vedavidyāvid), the logicians (prāmāṇika), the Purāṇa-scholars (paṇ-rāṇika), the scholars in the domain of smṛti (smārta), the physicians (bhiṣaj), the astrologers (mauhūrtika).

and the like. To the north of him there sit the Prakrit poets, and behind them the dancers (nartaka), actors (naṭa), singers (gāyana), musicians (vādaka), vāgjīvanas, kuśīlavas, tālāvacaras, who appear to have been certain groups of bards. To the west of him there come the Apabhraṁśa poets; behind them the painters (citralekhṛt), jewel-setters and similar

61 Compare Kāmasūtra I, 3, 12: Santy api Khalu śāstraprahatabuddhayo gaṇikā rājaputryo mahāmātraduhi-taraś ca.

THE CAUSES OF THE KAVYA AND THE POET

classes (manikyabandhaka, vaikaṭika, svarṇakāraravardhakilohaikāra); and to the south of him the Bhāṭa-bhāṣa poets, behind them paramours, (bhujamga), courtezans (ganikā), rope-dancers (plavaka), saubhī-kas62, wrestlers (malla), and soldiers (sastrājivin).

Then the king opens the discussion and examines the kāvya. In doing so he should emulate the famous "presidents" of older times, as Vāsudeva, Sātavāhana, Sūdrakn, Sāhasānka. He should honour the poets according to their merits. In the great cities the king should establish similar committees (brahmasabhā) in order to have kāvyas and sāstras examined. Whosoever has undergone this parīkṣā should be driven in a particular carriage (brahmaratha) and crowned with a diadem. Rājaśekhara closes this unique chapter with the following verses, speaking of kāvya-examinations in Ujjayinī and sāstra-examinations in Pāṭaliputra.

Srūyate cojjayinyām kāvyakāraparīkṣā— iha Kālidāsa-Menṭhāv atr-Āmara-Rūpa-Sūra-Bhāravayah, Haricandra-Candraguptau parikṣitāv iha viśālāyām. Srūyate ca Paṭaliputre sāstrakāraparīkṣā— atr-Opavarsa-Varsāv iha Panini-Piṅgalāv iha Vyādih, Vararuci-Patañjali iha parikṣitāḥ khyātim upajagmuḥ. ittham sabhāpatir bhutvā yaḥ kāvyāni partikṣate, yasas tasyā jagadvyāpī sa sukhī tatra tatra ca63.

This is, however, only a pretty story, not fact.

62 Concerning the saubhikas see Prof. H. Lueders' very interesting paper in Sitzungsber. d. Kgl. Preuss. Ak. d. Wiss., philos.-hist. Kl., 1916, pp. 698ff. The passage is borrowed to a great extent from the Arthasāstra, where we read (p. 125): matanartakavāgjīvanakuśīlavaplavakasāubhikacāra-nānām...........sarvatālāvacāśarāṇām ca.

63 With respect to these poets, see the remarks of the editors of the Kāvyamīmāṃsā on p. 10, 11 (Notes).

THE DEFINITION OF KĀVYA

No exact definition of kāvya has been found by the older Indian writers on poetios. Only in more recent times scholars have shown that what is said by the older writers cannot be considered as the essential matter of poetry.

B h ā m a h a defines I, 16 : śabdārthau sahitau kāvyam.

Word and sense combined (is) the kāvya".1

D a n d i n says a little more (Kāvyādarśa I, 10) : taiḥ sarīram ca kāvyānām alaṅkārāś ca darśitāḥ, sarirataḥ tāvad iṣṭārthavyavacchinnā padāvalī.

"By these (the older scholars) the body and the ornaments of the kāvyas have been pointed out. With respect to the body it consists of a series of words, qualified by the sense which (the poet) wishes to express".

1 It has already been mentioned before that Kālidāsa in Raghuvaṁśa I, I appears to refer to such a definition of the kāvya. Nothing, however, can be gathered from this fact, because that or a similar definition was common for a very long time. We get no correct idea of Bhāmaha's opinion on the kāvya, if by the unjustified combination of Bhāmaha I, 16 with I, 30 we construct the following definition sabdārthau sahitau kāvyam yuktam vākrasvabhāvoktyā.

This is done by S o v a n i in a paper on pre-dhvani schools of Alaṅkāra (Bhandarkar Commemoration Volume, p. 395). Prof. J a c o b i (Sitzungsber. d. preuss. Ak. d. Wiss., 1922, p. 224) thinks that the term Sāhitya ('poetry') is to be derived from that definition of kāvya.

THE DEFINITION OF KĀVYA

This definition shows some progress when compared with the simple statement of Bhāmaha that the kāvya is word and sense together, without any further explanation concerning the qualification of both śabda and artha. Dandin distinguishes at first between two things which are quite different from each other, and one of which is the body (śarīra). Then there is given a definition of the so-called kāvya-śarīra, which differs from Bhāmaha only in so far that the sense (artha) gets the attribute iṣṭa. The second element of the kāvya consists of the alamkāras, with which we will deal later on. In this context the word alamkāra has not the signification of what we call figure of speech, but means the ornament of the kāvya in general. The guṇas also are thus to be understood under the name alamkāra. The entire first book of the Kāvyādarśa deals with the body of the kāvya : the doṣas, the guṇas, and the rītis are all parts of it.

Vāmana, speaking of the kāvya in the first sūtra of this work, gives no definition. He says

"The kāvya is to be seized on account of the ornament".

In the vrtti, however, the definition of Bhāmaha is clearly alluded to :

kāvyam khalu grāhyam upādeyam bhavati, alaṅkārāt ; kāvaśabdo'yan guṇālaṅkārasamskṛtayoḥ śabdārthayor vartate ; bhaktyā tu śabdārthamātravacano'tra grhyate.

"The word kāvya has the signification of word and sense, which are endowed with guṇas and

80 THE FOUNDATIONS OF INDIAN POETRY

alamkāras ; but from reverence (to the older authorities) it is employed as expressing only word and sense". The word alamkāra is used by Vāmana, as in the Kāvyādarśa in the sense of ornament in general and of the so-called figures of speech. In the sūtra the word has the former signification ; for it is said in I, 1, 2 : saundaryam alamkārah. "Alamkāra (is) beauty".

The following sūtra shows in which way this alamkāra is to be made : sa dosagunālamkārahānā-dānābhyām". This (ornament is to be produced) by avoiding the dosas and employing the gunas and alamkāras. Thus the word alamkāra is used in the double sense.

There is little doubt that Vāmana had the definition of kāvya as given by Bhāmaha in view, and the treatment of the matter shows further what great strength was attributed to the pramāna or authority. For, though Vāmana thinks that the expression śabdārthau is not sufficient to define poetry as it does not show clearly enough, what the characteristic feature of kāvya is, he does not endeavour to present a new definition, which would have contrasted with the hallowed view of the old master in poetics. In the opinion of Vāmana every necessary explication is contained in the old definition.

2 The vrtti runs : sa khalu alamkāro dosa-hīnāt gunālaṅkārādānāt ca sampādyah kaveḥ. The next sūtra says where the poet should be instructed regarding dosas, gunas, and alamkāras : "śastratas tu", and the vrtti : te dosagunālam-hārahānādāne śāstrād asināt.

3 There are many cases in the Kāvyālaṅkāra-sūtravṛtti

The term kāvyāśarīra, which we met in the Kāvyādarśa, is found also in the Kāvyālaṃkārasūtra-vṛtti, but not in agreement with the view of Daṇḍin, in the vṛtti to 1, 3, 10 : the plot of the kāvya is designated by the word kāvyāśarīra4.

D h v a n i k ā r a gives a more accurate qualification of the śabdārtha in the Dhvanyāloka (p. 7), saying śahrdayahrdayāhlādiśabdārthamaya-tvam eva kāvyalakṣaṇam. "What consists of word and sense in such a manner that it pleases the mind of the learned is called kāvya".

The old definition of kāvya we find again in the Kāvyālaṃkāra of R u d r a t a, 1, 2 : nānū śabdārthau kāvyam. "Word and sense (combined is) kāvya".

M a m m a t a is a little more explicit in the Kāvyaprakāśa 1, 4 : tad adoṣau śabdārthau saguṇāv analaṃkṛti punah kvacit. "The (kāvya is) word and sense (combined), and sometimes without alaṃkāras".

It appears that this definition is a combination of the term of Bhāmaha and the explanation of Vāmana. It is remarkable that Mammata says analaṃkṛti punah kvacit. The vṛtti points out that this is said to prevent the false view that when

where the author respects the opinions of Bhāmaha. Sometimes he is quoted verbatim. We must therefore assume that Bhāmaha was a great authority for Vāmana though the latter is a good deal younger than Bhāmaha. Daṇḍin presents quite a different standpoint.

4 See note above.

xi

82 THE FOUNDATIONS OF INDIAN POETRY

there is no alamkāra as occurs at times, the poem ceases to be a kāvya5.

A little more is said by Vāgbhaṭa, though the old definition of Bhāmaha is easily discernible.

Vāgbhaṭālamkāra 1, 2 :

sādhuśabdārthasandarbhaṃ guṇālaṃkārabhūṣitam,

sphuṭārttirasopetaṃ kāvyaṃ kurvita kīrtaye.

"A kāvya which consists of pleasing word and sense, which is adroned with guṇas and alaṃkāras, and which is endowed with distinct rītis and rasas, the poet should compose for fame".

Vāgbhaṭa, however, gives no real definition here like his predecessors, but by beginning his book thus, the verse becomes a sort of explanation of what in his opinion the essence of kāvya is.

For him as well as for the scholars before him the chief element of poetry was śabdārthau, word and sense combined; it is, also, no new idea that śabdārtha is qualified by sādhū and guṇālaṃkārabhūṣita.

Considering finally that the differences of style (rīti) according to Daṇḍin, Vāmana, and others are based on the differences of the guṇas we cannot find ·anything new in the term rīti-upeta.

There remains the introduction of the rasa in the definition. This had not been done by any of the older alaṃkārikas, though, as a matter of course, they must have considered the development of poetic 'sentiment' as an important feature of poetry, but their definitions do not mention it.

The doctrine of the rasa had its place primarily in the drama. Hence it came into

5 Vrtti : kvāpīty anenaitad āha yat sarvatra sālaṃkārau kvacit tu sphuṭālaṃkāravirahe'pi na kāvyatvānih.

the kāvya, and was dealt with in detail by the

younger representatives of poetics. This theme, how-

ever, requires separate treatment by itself, which

is impossible here. This opinion of Vāgbhaṭa has

obviously influenced Viśvanātha, who defines

in the Sahityadarpana 1, 3 the kāvya as follows :

vākyam rasātmakam kāvyam doṣas tasyāpakarṣakah,

utkarṣahetavaḥ proktā guṇālankāraritayah.

"A literary composition, the nature of which

consists of rasa, is called kāvya. The doṣas (faults)

impair (its beauty), the guṇas (qualities), alamkāras

(figures of speech), and rītis (diction) are the causes

of its excellence".

The term vākya is nothing else than the well-

known śabdārtha of the older ālamkārikas. Before

giving the definition Viśvanātha criticizes the opinion

of Maminaṭa concerning the same matter sharply,

and points out that the definition as given in the

Kāvyaprakāśa is false. He declares that in the

best kāvyas there can also be found plenty of doṣas,

though nobody would think that they were losing

their kāvya-character thereby. It is true that the

scholars have found many 'faults' in the poems of

even Kālidāsa. With respect to the term saguṇa,

too, the Kāvyaprakāśa is wrong in the opinion of

Viśvanātha, because the so-called guṇas are

qualities not of the word, as Maminaṭa opines, but

of the rasa. The doctrines of the guṇas had assumed

another form in the course of time, which did not

correspond with that of the scholars of the older

period.

We close this chapter by presenting the definition

of a more modern writer on poetics, J a g a n n ā t h a, who regards the matter from a more correct point of view. He begins his big work, the Rasagangā-dhara, by a discussion of the question of kāvya and states :

ramanīyārthapratipādakah śabdaḥ kāvyam.

"A composition which produces a pleasing sense is called kāvya".

It must be confessed, however, that this is not an absolutely new definition, for in the Dhvanyāloka we read :

sahrdayahrdayāhlādişabdārthamayatvam eva kavyasahṛdayāhlādişabdārthamayatvam eva kavyalakṣaṇam*

It is remarkable that Jagannātha, in the exhaustive commentary which follows the definition, does not mention these words, whilst he criticizes the definitions of other authorities.

The main point in the definition is ramanīya. In commenting it Jagannātha says : ramanīyatā ca lokottarāhlādajanakajñānāgocarātā. "If knowledge brings forth pleasure that goes beyond the common limits, it is ramanīya". In other words, if a literary composition produces a pleasing surprise, a camatkāra, as it is called in the śāstra, we have to do with a kāvya. In the commentary the author deals fully with that camatkāra, which is indeed a characteristic quality of poetry. The following sūtras contain some more details as to the real nature of poetry, pointing out the relation between what is said by words and what remains unspoken ; The quality of a kāvya is determinated by this relation.

6 See above p. 8I.

THE NATURE OF THE ALAMKĀRA

B h ā m a h a, after having spoken of the quali-

ties of the good poet, does not yet go on to give the

definition of kāvya. He discusses an interesting

question, which is of great importance for the under-

standing of poetry and which appears to have been

the subject of controversy : In which relation do the

so-called alamkāras stand to the kāvya1. The author

says I, 13-15 :

rūpakādir alamkāras tasyānyair bahudhoditaḥ,

na kāntam api nirbhūṣaṃ vibhāti vanitāmukham.

rūpakādim alamkāraṃ bāhyam ācakṣate pare,

supāṃ tīṇāṃ ca vyutpattim vācāṃ vāñchanty alamkrtim.

tad etad āhuḥ sauśabdyam nārthavyutpattir idṛśī,

śabdābhidheyālamkārabhedād iṣṭam dvayam tu naḥ.

"Rūpaka, etc. are called by other (scholars) the

alamkāra (ornament) of the (kāvya). The face of a

girl, though she may be handsome by nature, does

not shine without ornament2.

1 14b and 15a are quoted by Premacandra in his

commentary to Daṇḍin's Kāvyādarśa I, 10.

2 With this stanza III, 57 (at the end of the definitions

of the alamkāras) must also be compared :

giram alamkāravidhiṃ savistaraṃ svayaṃ viniscitya dhiya

mayoditah,

anena vāgarthavidām alamkṛtā vibhāti nārtva vidagdha-

maṇḍalā.

"After having settled the matter myself, I have, to

the best of my knowledge, exhaustively described the

alamkāras of speech. The speech of the sense-knowing (poets)

Other (scholars, however,) designate rūpaka, etc. as (mere) exterior ornament (being not closely connected with the essential nature of poetry) : They would have the correct formation of nouns and verbs as ornament of sentences (the kāvya).

This is called saṃśabdya by them, and the formation of sense (the so-called arthālamkārās) is not of this kind. But in our opinion there are two kinds (of alamkārās) : alamkārās of the words and alamkārās of the sense."

These statements show clearly that already in the time of Bhāmaha there were very different views regarding poetry, and that poetics in that period was by no means a recent science. Up to this day, however, nothing is known of the pre-Bhāmaha masters ; certainly there were two opinions : According to the first, rūpaka, etc., were called alamkārās ; but according to the other, these figures of speech are of an exterior kind, the real alamkāra lying according to them only in the vyutpatti (= alamkāra) of the word (noun and verb, not of the sense). This ornament is designated as saṃśabdya. The above mentioned stanzas of Bhāmaha are quoted in the

being ornamented with these (alamkārās) shines like a girl with lovely ornaments.

One is reminded by this verse of Bhāmaha's reading of the following stanza of Vāgbhaṭa (Vāgbhaṭālamkāra IV,1) :

doṣair muktam் gunair yuktam api yenojjhitam vacah. strīrūpam iva no bhāti tam bruve'lamkriyocāyam.

"Alamkriyā (alamkāra) is that, without which, a speech does - not shine though free from doṣas and endowed with guṇas, as the form of a woman does not shine of itself without ornaments".

THE NATURE OF THE ALAMKĀRA

87

sixth ullāsa of Mammaṭa's Kāvyaprakāśa and ex-

plained exhaustively by the modern commentator

Jhalakikara, who has used an older gloss3. The

school according to which rūpaka, etc. are excluded

as alamkāras and only the so-called śabdālamkāras

acknowledged, argues that nothing but the word is

able to produce the camatkāra : the separate words

are combined with each other in a way that an

anuprāsa is produced or another of the śabdālam-

kāras, or that the guṇas (mādhurya etc.) come into

existence, It is only the word on which all depends.

For it is said : "A kāvya is recited, is heard, is sung".

Thus, nothing else but the word can be the thing in

question. The so-called arthālamkāras, on the other

hand, do not possess this particularity, because they

are founded ou the sense. So these may be considered

as something exterior (bāhya), and the term alamkāra

can be ascribed to them only in a metaphorical

3 Reference may be made to the explanation given by

JhalakIkara in his ed. of the Kāvyaprakāśa, 2nd Ed.,Bombay,

1901, pp. 313f : tasya kāvyasya rūpakṛdīḥ.......guṇa eváti

bhāvalh. Then the author gives a quotation from the

sārabodhinī of the same tenor.

The commentary of Mānikyacandra, published

in the Ānandāśramu Sanskrit Series, No. 89, p. 120,

remarks : Gauḍamatam etat. Tad etad iti. Suptiṇātmā-

kaṃ sauśabdayaṃ śabdavaicitryaṃ. Śabdālamkārasūdhikā

nedṛśī. Na suptiṇātmikety arthah. Athavirthavvyutpattir

arthālamkārarūpā. Rūpakādyalaṃkṛtir īdrśy antaraṅgā na.

Kim tarhi guṇā. Ity api vyākhyā. Athavārtha vyutpattir

api kāvyaprayojyatvena matety āha : dvayaṃ tu na iti.

Bhāmahah śabdasyārthasyā cālaṃkārān iṣṭavān paraṃ guṇa-

prādhānyena.

88 THE FOUNDATIONS OF INDIAN POETRY

sense4. The relation between the guṇas (as far as it is possible to speak of guṇas in connexion with Bhāmaha) and the alamkāras is according to Bhāmaha not the same as in the opinion of the later writers. Mādhurya, prasāda and ōjas, which the author defines without using, however, the designation of guṇa, are not sharply separated from the alamkāras ; in one passage the word guṇa is even evidently used in the sense of alamkāra5.

Bhāmaha is not of the opinion of these scholars (evidently the guṇdas)6, but holds, without giving further arguments against the other view, that there are two sorts of homogeneous alamkāras : śabda and arthalamkāras. And this opinion became predominant : the same division is found again in almost all books on alamkāra. Even Dandin agrees with Bhāmaha on this point, though elsewhere he is always inclined to be at variance with his predecessor. As Dandin belongs to the Vaidarbha school as well as Bhāmaha, this agreement of the two ālamkārikas is nothing to be astonished at.

Bhāmaha did not present a definition of alamkāra. That is done for the first time by Dandin, (Kāvyā-darśa II, 1). kāvyaśobhākarān dharmān alamkāran ācakṣate. “Qualities which produce the beauty of the kāvya are called alamkāras”7.

4 Compare the corresponding text of the last note.

5 See below.

6 See also the first words of the commentary of Māṇikya-candra referred to in note 3.

7 That the word alamkāra does not mean “ornament”

This definition is not quite correct, for not only the alamkāras are the causes of the kāvya. There are some other things which have the same effect. The most important among these are the guṇas. Though the guṇas will be treated exhaustively in another place, it is necessary to deal here, at least, with their general definition, as, reckoning from the time of Vāmana, the doctrine of the guṇas is closely connected with that of the alamkāras. It is worth noting that a definition of the guṇas is not to be found in the works of Bhāmaha and Dandin, though they in a part of them are referred to by these scholars.

Vāmana appears to have been well aware that Dandin's definition of the alamkāra is not satisfactory. Using the same words as the author of the Kāvyādarśa he explains it in the Kāvyālaṃkārasūtravṛtti, III, 1, 1-2 : Kāvyaśobhāyāḥ kartāro dharmā guṇāḥ, tadatiśayahetavas tvalamkārah.

'Qualities which produce the beauty of the kāvya (are called) guṇas. The causes of a very high degree (of the beauty of the kāvya are called) alamkāras'. In the vṛtti it is especially mentioned that the guṇas must be considered as the elements which produce the kāvya's beauty if there were only alamkāras, the kāvya would be far from being beautiful8.

in general here, but "figure of speech" especially appears from the enumeration of the alamkāras in II, 2 ff. 8 ye khalu śabdārthayor dharmāḥ kāvyaśobhāṃ kurvanti, te caujahprasādādādayaḥ na yamakopamādayaḥ, kaivalye teṣāṃ akāvyaśobhākaratvāt, ojaḥprasādādīnāṃ tu kevalānāṃ asti xii

90 THE FOUNDATIONS OF INDIAN POETRY

Thus Vāmana has transformed Dandin's definition of alamkāra into a definition of guṇa, and has from this derived one of alamkāra. The idea, however, apparently was not quite new. Even in the time of Bhāmaha the guṇas were, though not the same as, but homogeneous to the alañkāras : these may, so to speak, be considered as a sort of continuation of the guṇas. That at least we can gather from the way the guṇas and alañkāras are treated by Bhāmaha. Later on, we shall deal more closely with the matter. Vāmana is, as far as we know, the first who precisely formulated the question regarding the relation between guṇa and alamkāra. The close connection which Vāmana assumes between both is corroborated by him, when he goes on to illustrate the separate guṇas : they are divided as well as the alañkāras into bandha-(=śabda) and arthaguṇas.

Vāmana's treatment of the matter in question was, however, not acknowledged by M a m m a t a. He concedes indeed that the guṇas are in some way or another connected with the alañkāras, for he presents the definition of the alañkāra after that of the guṇa, because the second follows from the first, but the definition itself rests on quite a different point of view (VIII, 66) :

ye rasasyāñgino dharmā śauryādaya ivātmanaḥ, utkarṣahetavas te syur acalasthitayo guṇāḥ.

"Those qualities which belong to the rasa like heroism, etc. to the soul and which are the causes

kāyaśobhākaratvam. And vrtti to III, 1, 2 : tasyāḥ kāyaśo-bhāyā atiśayaḥ tadatiśayaḥ, tasya hetavaḥ, tuśabdo vyatireke ; alañkāráś ca yamakopamādayaḥ.

of (the rasas') excellence and have a permanent existence, are called guṇas".

In the vrtti the author explains that the guṇas are by no means qualities of the sound [which is the opinion of Vāmana, who distinguishes between bandha-(=śabda) and arthaguṇas] but rather qualities of the rasa, the poetical sentiment, as heroism is a quality not of the body but of the soul.

The term acalashitayah is not new. Even Vāmana said already (I, 3, 3) : pūrve nitye "the first (viz. the guṇas) are permanent", and that Mammaṭa's utkarṣahetaval has its parallel in Vāmana's atiśayahetaval need not be mentioned.

The quintessence, however, is that Mammaṭa considers the theory of guṇa and alamkāra from a different standpoint, the rasa. This fact touches the question of the soul of poetry, which will be dealt with in another chapter.

After having explained the character of guṇa Mammaṭa goes on to define alamkāra (67) : upakurnvati taṃ santaṃ ye'ṅgadvāreṇa jātucit, hārādivad alamkāras te'nuprāsopamādayah.

"Qualities, which sometimes help an extisting (rasa) by means of a link (viz. word or sense), as necklaces, etc. (which are put round the neck of a person and thus adorn him) are the alamkāras, anuprāsa, upamā, etc".

While the guṇas are integral parts of the rasa, the alamkāras have rather an accidental or unessential character, as they aid or adorn the rasa which, without them, is already complete in itself. This ornament can be referred to the sense or to the

THE FOUNDATIONS OF INDIAN POETRY

word which expresses the seuse, and thus we get two sorts of alamkāras : śabda and artha. The commentary says somewhat more than the sūtra9. It happens, says the vṛtti, that in certain cases there is no rasa ; then the alamkāras are used only with the aim of making the verbal turn of expression (uktivaicitrya) interesting. Besides this a third case is possible, when, though there are rasas, the employed alamkāras do not help these rasas, but have nothing to do with them really. The author illustrates the mentioned possibilities by some examples. After having done so he goes on to criticise Vāmana's opinion of the guṇas and alamkāras which he calls false. For, if we assume that Vāmana is right, then the following question arises : Is poetry constituted by the co-existence of all guṇas or only by a part of them ? If the first is the case the Gauḍī and Pāñcālī must cease to be poetry, as in the opinion of Vāmana these dictions have only a part of the ten rasas. This is absurd. In the second case such sentences which contain a few guṇas, but contain nothing which can be called a poetic idea in their structure, would be poetry notwithstanding. In the following example :

adrāv atra prajvalaty agnir uccaiḥ, prajvalat prajvaṃ ullāsaty eṣā dhūmāḥ.

"On this mountain there shines a fire ; thence

9 I shall not consider the question here whether we must conclude from the divergence of the commentary and the sūtras that the author of the text is not the same as that of the vṛtti. cf. V. S u k t h a n k a r, ZDMG. 66, 477 ff., 533 ff.

THE NATURE OF THE ALAMKĀRA

93

rises great smoke" the guṇa ojas is contained, but nobody would be inclined to see any poetic idea in the two lines.

Not less wrong, continues the Kāvyaprakāśa, is Vāmana's definition of alamkāra. For according to the author an alamkāra can occur only where a guṇa already exists. He presents the following instance :

svargaprāptir anenaiva dehena varavarṇinī, asyā radacchadaraso nyakkarotitarām sudhām.

"This beautiful woman embodies the acquisition of heaven on account of this (beautiful) body ; the sweetness of her lips despises nectar."

This contains in spite of the absence of any guṇa the two arthālamkāras Viśeṣokti and Vyatireka and this is without doubt poetry.

THE SOUL OF POETRY

We shall now touch on another interesting subject and consider what the older scholars take to be the soul of poetry. A big step forward is to be observed in the Kāvyādarśa. In the opinion of D a ṇ d i n an essential part of poetry is its 'body' (śarīra) and he declares (I, 10) that this body is opposed to a second thing, which is represented by the alamkāras. These alamkāras adorn the kāvya as ornaments adorn the body of a woman, and in the beginning of the second adhyāya, alamkāras are called those qualities which produce the Kāvya's beauty. Indeed, the entire arrangement of the poetic matter as treated by Daṇḍin leaves no doubt that he considers the alamkāras as the main part of poetry ; and so does B h ā m a h a too. Thus both Bhāmaha and Daṇḍin were still far from looking upon poetry from a higher point of view. In Kāvyādarśa I, 42, however, we are told : iti Vaidarbha-mārgasya prāṇā daśa guṇāḥ smrtāḥ. "These spirits of the Vaidarbha-rīti are called the ten guṇas," but this refers only to the Vaidarbha-style, and the author by no means wished to say that the guṇas should be considered as the spirit in general1. There is no doubt that in the eyes of Bhāmaha and

1 H. J a c o b i, Ānandavardhana's Dhvanyāloka (ZD MG, vol. 56, p. 393). We find the term kāvyaśarīra also in Bhāmaha's Kāvyalamkāra I, 23, but Bhāmaha means thereby the kāvya as a whole. One might be inclined to consider bhāvikatva (III, 52, 53) as the soul of poetry accord-

Dandin the essential part of poetry is formed by the alamkāras, though they do not speak positively of the "soul".

Vāmana has followed Dandin in so far as the sentence (vākya) is called by him the body of the kāvya. He says in vritti to 1, 26 : ritir nameyam ātmā kāvyasya; śarīrasya veti vākyaśesah, but pursuing this idea he goes much farther than Dandin, so that in his view poetry assumes quite a new aspect.

He is the first scholar that speaks of soul (ātman) of poetry. And what is the soul of poetry ? Sūtra 1, 2-6 answers : ritir ātma kāvyasya "the style is the soul of poetry".

It is true that the Kāvyādarśa also contains the doctrine of style and of the guṇas, which are closely connected with it ; but rīti is nothing independent here, being, on the contrary, a part of the doctrine of Kāvyasāriṇa.

According to Vāmana, both rīti (and guṇas) have nothing to do with the body of poetry : rīti is the soul of poetry.

This idea of Vāmana's was progressive but it was a matter of little satisfaction that the style should be the essential and life-giving part of poetry.

As the rīti regards only the expression of ideas and not the idea itself, it regards merely the outside, so to say, and Dandin was right to consider it as a part of the body.

ing to Bhāmaha. This figure of speech does not refer to a single stanza, as alamkāras generally do, but to the whole composition (prabandhaviṣayam guṇam). Bhāvikatva, on the other hand, is regarded by Bhāmaha as alamkāra or, what in this case seems to be the same, as guṇa.

96 THE FOUNDATIONS OF INDIAN POETRY

U d b h a t a, who appears to have lived at the same time as Vāmana but exercised a greater influence, has a more correct opinion regarding the "soul of poetry" designating by this term the rasa. After having defined the alamkāras, bhāvika, and kāvyaliṅga Udbhaṭa says (VI, 17) :

"As a kāvya, which is endowed with rasa and so on, is taken to be a living form, the rasa is called the soul of the kāvya".

And with respect to the alamkāra bhāvika the author remarks (VI, 14)

rasollāsī kaver ātmā svacche śabdārthadarpaṇc, mādhuryaujoguṇapraụdhe pratibimbyā prakāśate,2

"The rasa-bright soul of the kavi shines reflected in the pure mirror of word and sense, endowed with the guṇas mādhurya and ojas".

Though this opinion of Udbhaṭa's certainly shows progress, a quite clear idea of the quintessence of poetry was not yet found, even by Udbhaṭa, as we may conclude from the word ādi in VI, 17. This is not surprising, if we consider that in that period the doctrine of rasa was still in its development.

The doctrine of rasa, taken apparently from the dramatic poetry, was taught by the older ālamkārikas only in connection with a few figures of speech, the tenor of which indicates a certain state of mind. Not long, however, after the time of Udbhaṭa (perhaps still under his influence) there

2 J a c o b (JRAS. 1897, p. 846) has the variants rāsollāsī, mādhuryaujoyutapraṇdhe, and prativimbya.

arose a new system of poetical aesthetics. This is the doctrine of dhvani (the doctrine of the unexpressed), which was given in detail in the Dhvanyā-loka. According to it the dhvani was the soul of poetry. What he teaches is shortly the following :

The soul of poetry is the tenor (artha), and this tenor is twofold : it can be expressed by words (vācya) and can be suggested (pratīyamāna). Only the latter kind is of value for poetry, and thus nothing else than the unexpressed and merely suggested sense can be called the soul of poetry3.

3 Compare also Jacobi's introductory pages to the translation of the Dhvanyāloka (ZDMG., vol. 56 and pp. 9-16 of the same author's introduction to Ruyyaka's Alampkāra-sarvasva. Another work on the same subject is the Vakrokti-jīvita. As H. Jacobi informs me, this work has been found now and has been edited by Dr. S. K. De in the Calcutta Oriental Series.

THE RĪTI

The doctrine of rīti takes up a great deal of space in the older books on alamkāra. The word rīti, the main designation of which is 'manner or method', designates in poetry a certain method of poetical diction, and has no equivalent in other languages. If one renders it by 'style or diction', one must remember that nothing is explained thereby regarding the essential nature of rīti.

That the doctrine of rīti is very old may be gathered from the manner of its treatment in Bhā-maha's Kāvyālamkāra. Though we do not know in which way older authorities have dealt with the matter in question, it is certain that there were two styles opposed to each other: Vaidarbhī and Gau-ḍīyā, and that the Vaidarbhī was regarded as the better one, of course only by the poets in Vaidarbhī rīti. As a matter of fact there was once a period during which Gauḍīyā rīti had famous poets and a great influence upon the development of poetry, but in course of time Vaidarbhī grew superior and determined the direction poetry was going to take, to the advantage of Indian poetic literature, as we may add.

Before the time of Bhāmaha there was lively discussion about the existence of two different styles: Vaidarbhī and Gauḍīyā. Bhāmaha opposes this view saying (1, 31-35) :

vaidarbham anyad astiti manyante sudhiyo' pare, tadeva kila jyāyaḥ sadartham api nāparam.

gaudīyam idam etat tu vaidarbham iti kim prthak, gatānugatikanyāyān nānākhyeyam amedhasām. nanu cāsmakavaṃśādi vaidarbham iti kathyate, kāmaṃ tathāstu prāyeṇa samjñecchāto vidhīyate. apuṣṭārtham avakrokti prasannam r̥ju komalam, bhinnam geyam ivedaṁ tu kevalam śrūtipesalam. alamkāravad agrāmyam arthyaṃ nyāyyaṃ anākulam, gaudīyam api sādhīyo vaidarbham iti nānyathā.

"Some scholars hold Vaidarbha(rīti) as a different (thing). And (they say that) this is the better, (because) it has a good sense, and not the other (Gau-dīya)1. But where is, we reply, the difference between saying this is Gundīya and saying that is Vaidarbha? The very same thing is to be called by different names by the foolish ones according to the method of one who hangs his judgment on the sieve of another2. But (the other answer) is not Cāsmaka-vamśa3 etc. called a Vaidarbha (composition)? That may be so ! Names are usually given capriciously. In the opinion of those scholars, Gauḍīya is of a not well developed sense, has no poetical ornaments4, straight, and tender ; the other (Vaidarbha)

1 Or should we translate: "This (Vaidarbhī) is the better, and not the other (Gauḍīyā), though this (latter) may be of a good sense?" I preferred the above given translation on account of apuṣṭārtham in stanza 34 and arthyam in stanza 35.

2 Gatānugatika.was a proverbial saying. Compare Pañcatantra I, 342: gatānugatiko loko na lokah pāramārthikah.

3 Nothing is known as regards this composition.

4 Vakrokti, verbatim 'curved manner of speaking' has various meanings in poetics. Here it is apparently the same as alamkāra, as may be gathered from alamkāravad in verse 35.

is, as it were, to be sung; only this is agreeable for hearing, has poetical ornaments, is not vulgar, of a good sense, suitable, and not confused. But (so we reply) also Gaudiya is excellent and does, after all, not differ from Vaidarbha.

A much clearer picture of ritis we gain from Dandin's Kāvyādarsa. As Dandin is an opponent of Bhāmaha we are not surprised that the criticizes Bhāmaha's dealing with riti severely. The author of the Kāvyādarsa opens (I, 40) with a polemic note: asty aneko girām mārgah sūkṣmabhedah parasparam, tatra vaidarbhagauḍyau varnyete prasphuṭāntarau.

"There are many kinds of diction (style), very finely distinguished from each other; but of these (many kinds of style) Vaidarbha and Gaudiya are described, (because) their differences are especially manifest."

This verse can only be fully understood when it is taken to be pointed at Bhāmaha: There are not only two sorts of ritis but many; but only two of them, Vaidarbha and Gaudiya, which in the opinion of Bhāmaha have no distinctive features, are described in detail, because, just on the contrary, their differences are peculiarly clear.

Of what kind are these manifest differences between Vaidarbha and Gaudiya? The author says (I, 41, 42):

"Šlega, prasāda, samatā, mādhurya, sukumāratā, arthavyakti, udāratva, ojas, kānti, and samādhi: these spirits of the Vaidarbha style are called the ten gunas. Mostly one sees the contrary of them in the Gauda style."

Thus the doctrine of rīti is closely connected with that of guṇas, and we have therefore to deal with the nature of the above mentioned ten guṇas as far as it is necessary for the understanding of rīti⁵.

In the opinion of Dandin the ten guṇas are peculiarities especially of Vaidarbha rīti, while, if we may say so, the guṇas of Gauḍīya are opposed to those of Vaidarbhī. Bhāmaha, too, mentions some qualities, wherein Gauḍīya and Vaidarbha according to the 'incorrect' opinion of some scholars are said to differ from each other. But they are of another kind. A few of the guṇas enumerated by Dandin we also find in Bhāmaha's work, but not directly connected with rīti. They are not qualities of a certain style (the differences of which are denied by him), but rather of the good kāvya generally.

We shall see that matters are considered in a similar way by Vāmana. But let us see, what we are told about guṇas by Bhāmaha.

In the beginning of the second pariccheda the author says:

"The wise (poets), who claim mādhurya and prasāda, do not employ too many compounds.

Some (poets, however), who are willing to express ojas (prefer) long compounds, as mandārakusumarupīñjaritālakāḥ (whose curls were coloured yellow by the pollen of mandāra flowers)."

Bhāmaha continues (II, 3):

⁵ A fuller description of guṇa will be given in another place, so as not to disturb the context here. I may, however, mention the fact that Vāmana's treatment of the guṇas is totally different from the one we find in the Kāvyādarśa;

102 THE FOUNDATIONS OF INDIAN POETRY

'śrāvyam nātisamastārtham kāvyam madhuram iṣyate, āvidvadañganābālapratittārtham prasādatvat6. "A kāvya, whioh is agreeable to hear and expresses the sense by words which are not too much compunded, is considered as madhura (attractive). The sense (of the kāvya), which is to be understood by the wise7, by women, and by children has prasāda (clearness)." With these three verses the whole question is settled for Bhāmaha; verse 4 contains the first group of alamkāras. From the fact that the author has nothing more to say regarding ojas we may gather that according to him ojas is not a good quality of the kāvya. It is most remarkable that Bhāmaha, in dealing with mādhurya, prasāda, and ojas, does not mention the term guṇa. As a matter of course, the doctrine of guṇa was fully developed even before his time, because it is mentioned in the Nātyaśāstra8.

6 Quoted by Mallinātha and Jayamaṅgala in their commentaries to the Bhattikāvya XI, 1. Both authors read śrāvya instead of śravya as given by Trivedi. Hemacandra in Kāvyānuśāsanaviveka quotes the first line as follows: tena "śravyam nātisamastārthaśabdam madhuram iṣyata" iti mādhuryalakṣaṇatvena śrūyatuam yad Bhāmahenoktam tan na yuktam ity arthah. As Trivedi mentions in Bhandarkar Commemoration Volume, p. 410, a definition quite similar to this is ascribed by Pradīpakāra to an alaṅkārika Bhāskara by name. Pradīpakāra says: Bhāskaras tu śravyatvaṃ mādhu-ryaśya lakṣaṇam āha sma tad ayuktam.

7 Or should we rather read avidvado and take it as an adjective to aṅganā and bāla? Then the sense would be a little more clear.

8. We shall return to this matter on another occasion.

Bhāmaha, therefore, knew the guṇa theory very well, but he totally rejected it as he rejected the theory of 'different styles'. Treating the rīi (I, 31-35) he mentions some qualities which in the Kāvyādarśa are taught as guṇas, but these are also of no peculiar value in the eyes of Bhāmaha. Only two qualities mādhurya and prasāda are of importance, but by no means with regard to one certain sort of diction, but to the kāvya in general.

The whole second pariccheda of Bhāmaha's is devoted to the illustration of the alamkāras. Now the first three stanzas contain mādhurya and prasāda. Are we not justified to conclude from this remarkable fact that two qualities (as we may call them) were in some way or another homogeneous to the alamkāras themselves, though they are not alamkāras in the strict sense of the word? And does not Vāmana derive the definition of alamkāra from that of guṇa?⁹ But even in Bhāmaha's work there is a passage, where no great distinction is made between guṇa and alamkāra.

In III, 1-4 the enumeration of those alamkāras is given, the definition and illustration of which is the subject of the following verses. Stanza 4 closes: bhāvikatvam ca nijagur alamkramiṣumed-hasah. Dealing with the alamkāra bhāvika the author remarks: bhāvikatvam iti prāhuḥ prabandha-viṣayam guṇam, pratyakṣa iva drśyante yatrārthā bhūtabhāvinah. "That quality (guṇa) is called bhāvi-katva, the sphere of which is the composition as a whole (and not a single stanza as it is the case regarding the other alamkāras), where past or future

9 See above p. 90.

things are standing, as it were, before one's eyes. Thus guṇa is in the opinion of Bhāmaha not essentially different from, though not identical with alamkāra. Udbhata is, as regards bhāvikatva of the same opinion as Bhāmaha. He remarks too that bhāvikatva is accompanied by mādhurya and ojas. The term guṇa is not used in this connection by Udbhata10.

Let us, however, return to Dandin's treatment of rīti. Bhāmaha had but little to say on this subject, wholly denying the existence of different styles and also the theory of the ten guṇas. As Dandin is the opponent of Bhāmaha, he teaches not only the existence of several rītis and of the ten guṇas but also goes into the explanation and illustration of the matter at great length. According to his opinion the ten guṇas are qualities of the Vaidarbhī rīti, but not qualities of the kāvya in general. With regard to the single guṇas Dandin gives the following account :

1 Śleṣa. I, 43: śliṣṭam aspr̥ṣṭaśaithilyam alpaprāṇālakṣarottaram, śithilam mālatīmālā lolālikalilā yathā.

"That (is called) śleṣa which does not possess "looseness." "Loose" is what to the greater part consists of syllables which are pronounced with (only) little breath, as: mālatīmālā lolālikalilā11." Dandin explains (I, 44). anuprāsadhiyā gauḍais tad iṣṭam bandhagauravāt, vaidarbhair mālatīdāma laṅghitam bramañair iti.

10 The passage is found in the Kāvyālamkārasaṃgraha VI, 13-15: The later ālamkārikas appear to be dependent on Bhāmaha; cf. Kāvyaprakāśa X, 28; Alamkārasarvasva p. 183; Jayamaṅgala and Mallinātha, commenting on Bhatti-kāvya XII, 1, refer to Bhāmaha.

11 For iti in verse 53 seem's to refer to the whole preceding stanza, and not only to the last pāda.

THE RITI

103

"On account of their being prepossessed in favour

of anuprāsa the Gauḍas are fond of this. The

Vaidarbhas (on the other hand) like on account of

the stability of composition the following: mālatī-

dāma lañghitam bhramaraih."

A sentence like mālatīmālā lolālikulilā is not-

withstanding its looseness, liked by the western

poets, because it contains the śabdālankāra anup-

rāsa, which is very highly estimated by them. The

Eastern poets, on the other hand, attach more value

to the stability of composition, as the instance

mālatidāma lañghitam bhramaraih shows.

2 Prasāda.

I, 45: prasādavat prasiddhārtham indor indivaradyuti,

lakṣma lakṣmiṇ tanottīti pratītisubhagaṃ vacah.

"That is 'clear' the sense of which is well-known,

o.g. a sentence like the following: 'the moon's spot

shining like a dark lotus beauty' is beautiful on ac-

count of its clear perception."

Everybody knows without any further explana-

tion what is meant by poet's indivara, indu, etc.

As to the Gauḍas the author says I, 46:

vyutpannam iti gauḍīyair nātirūḍham apiṣyate,

yathānatyarjunābjāmasadrkṣāūko balakṣaguh.

"If that is based on the words' etymology, the

Gauḍas also like a mode of expression not altogether

clear, e.g.: 'The white-beamer (i.e. the moon) has

a spot which is similar to the unbright (i.e. dark)

waterborn ones (i.e. the lotuses)."

Here the poet employs rare words: arjuna, better

known as a byname of Kārtavīrya than as 'white';

still stranger is the term an-atyarjuna as meaning

xiv

106 THE FOUNDATIONS OF INDIAN POETRY

'dark', and in the same way balakṣagu and abjanma. After all, the words lakṣma lakṣmīṁ tanoti in I, 45 would not be liked by the Gauḍas, because the connection given in I, 46 would not be liked by the Vaidarbhas as, besides the bombastic mode of expression, it contains too harsh combinations of consonants.

3 Samatā. I, 47: samaṁ bandheṣv aviṣamaṁ te mrḍusphuṭamadh-yamāḥ, bandhā mrḍusphuṭonmiśravarṇavinyāsayonayaḥ.

"Samaṭ (evenness) is the non-unevenness in the composition; the compositions that are founded upon the employment of soft, harsh, or mixed (soft and harsh) sounds are soft, harsh, or medium (respectively)."

He adds further in I, 48-50 : "To me comes the malaya-wind, talkative through the cooing of the kokilas, sprinkled by the drops of the mountain-streams, which are quite clear and the drops of which are going with (the wind). This malaya-wind, the fragrance of which is great on account of its relation to the sandle-wood, rivals the breath from the mouths of beautiful girls, because its steadiness increases.

A Gauḍa kāvya-style of this kind, which (as verse 49 shows) takes no care of the unevenness and respects only the splendour of sense and alamkāra acquired a wide extension."

The main point of this guṇa is a certain uniformity in the sequence of the sounds, the combinations of sound being soft, harsh, or medium. The poet, however, should not be heedless and thus cause a

pell-mell of sounds, the effect of which becomes disagreeable to the ear. So it is in the opinion of the Vaidarbhas, while the Gauḍas, on the contrary, pay no attention to this matter, aiming only at brilliant or rather bombastic expressions. As to the instance given by Daṇḍin, 48a appears to illustrate mṛdu, 48b sphuṭa samatā, and 49 viṣamatā as peculiar to the Gauḍīya rīiti12.

4 Mādhurya.

I, 51: madhuram rasavad vāci vastuny api rasasthitiḥ, yena mādyante dhīmanto madhuneva madhuvrataḥ. "Madhura is what contains rasa; (and in this case) rasa lies in both word and sense. By this (rasa) the knowing ones grow intoxicated as bees by honey."

In the treatment of madhura Daṇḍin differs absolutely from Bhāmaha, who as we have seen, spoke of madhura too. According to Bhāmaha mādhurya is in line with prasāda, as in both 'qualities' the length of compounds is the decisive point. Daṇḍin, criticising his rival, states (wholly different as) something the essential matter of mādhurya, which has

12 Tarkavāgiśa and Vidyāsāgara present a quite different interpretation of this stanza: "Śliṣṭa is what to a large extent consists of syllables which are pronounced with only little breath and what therefore is loose, whereat the looseness, however, is not felt too obviously, e.g. mālatīmālā lolālikalilā (a garland of mālatī-flowers covered with swarming bees)." Prof. Lüders pointed out to me that this cannot be the correct meaning of Daṇḍin's words. Mālatīmālā lolālikalilā is not an example of śliṣṭa, but of sithila, which fault the poet should avoid. That this is correct appears from Kāvyādarśa I, 69, where the author refers to I, 43 (See below p. 109).

108 THE FOUNDATIONS OF INDIAN POETRY

nothing to do with compounds. To illustrate and to defend his diverging opinion against Bhāmaha, the author of the Kāvyādarśa treats the subject more exhaustively than he does that of other guṇas.

The second line of I, 51, is, though closing the ‘definition’, an instance for that kind of madhura where the rasa lies in the repetition of similar sounds. The next stanza explains:

yayā kayācic chrutyā yat samānam anubhūyate, tadrūpā hi padāsattih sāñuprāsā rasāvahā.

“An uninterrupted line of words of such a kind that one hears in one way or another (words of) similar (sounds) embraces rasa, because that (line of words) is endowed with anuprāsa.”

Thus, the guṇa mādhurya has for an inherent factor an anuprāsa, especially the so-called śruty-anuprāsa.

Another instance is given in verse I, 53:

eṣa rājā yadā lakṣmīṃ prāptavān brāhmaṇapriyah, tataḥ prabhrti dharmasya loke ‘sminn utsavo ‘bhavat.

“Since that time this king, the friend of the brahmins, had got the royal power, was a feast of dharma in this world.”

In this stanza homogeneous sounds ṣa and ra, ja and ya, da and la, ma and pa, etc. are set close to each other18.

The Gauḍas do not like, as is mentioned in I, 54, this kind of anuprāsa, where homogeneous sounds

13 The commentary says: atra śakārarakarārayor ekasmin mūrdhani evam jakārayakārayos tālau, dakāralakāroyoś ca dant’ uccāiryamānatvāt sāmyam iti śrutyanuprāsah, sa ca dharmavīraparipuṣṭasya rājaviṣayakaratibhāvasya. vyañjaka iti mādhuryarasabhāvah.

are repeated, being, on the contrary, more fond of the repetition of the very same sounds.

In the following verses Dandin goes on to illustrate the second sort of madhurya, where the rasa lies in the sense. He says (I, 62):

kama'n sarvo 'py alamkāro rasam arthe nisiñcati, tathāpy agrāmyataivaïtaṃ bhāraṃ vahati bhūyasā.

This is explained by quoting some instances, wherein by the use of vulgar and obscene words the rasa of sense is lost.

After the remark that in the foregoing verses madhurya is explained in its various forms (vibhaktiṃ iti mādhuryam) Dandin defines the next guṇa.

5. Sukumāratā

I, 69: anişṭhurākṣaraprāyaṃ sukumāram ihesyate, bandhaśaithilyadoṣas tu darśitaḥ sarvakomale..

Example I, 70, 71: māndārtkṛtya barhāṇi kanṭhair madhuraṅtibhiḥ, kalāpināḥ praṇṛtyanti kāle jīmūtamālinī. ity anūrijita evārtho nālamkāro 'pi tādrśah, sukmāratāyaivaitad ārohati satāṃ manaḥ.

14 In the verse referred to by the author the fault of looseness is illustrated by "mālatīmālā lolālikalitā" which is liked by the Gauḍas, because they are very fond of the amṛtāpisa; see p. 105.

110 THE FOUNDATIONS OF INDIAN POETRY

"Having formed their tail-feathers into a circle, the peacocks dance in autumn, (crying) with throats, the sounds of which are pleasant."

In this stanza the tenor is not very luxurious ; only because it contains sukumāratā (the stanza) enters the heart of the wise ones".

Thus sukumārata is a peculiarity of the sound of the words and does not regard the sense. The Vaidarbha style is soft and tender and is in contrast thus to the Gauḍīya, which is fond also of words the articulation of which is rather difficult and rough :

dīpatam ity aparair bhūm nā krcchrodyam api badhyate. nyakṣena kṣayitah pakṣah kṣatriyāṇām kṣanād iti.

"Because they regard it as brilliant, the others (the Gauḍas) use also words the pronunciation of which is difficult e.g. : nyakṣena.....(By Paraśurāma the party of the kṣatriyas was destroyed in a moment)".

The harshness is caused by the frequency of kṣ. This, however, is justified by the fact that the verse contains the so-called vīrarasa. According to the commentary the Vaidarbhas would employ the guṇa sukumāratā also in such a case15.

6. Arthavyakti.

1, 73, 74. arthavyaktir aneyatvam arthasya harinoddhrta bhūḥ khurakṣuṇṇagāsrglohitād udadher iti. mahi mahāvarāheṇa lohitād uddhrtodadheḥ. itīyatvena nirdiṣṭe neyatvam uragāsrjaḥ.

15 Gauḍā hi yatra vīrasādirūpam ojasvi vaingyam tutra parusavṛpais tadvyatijanasyāvaśyakatayā saukumāryaṁ nādriyante, vaidarbhās tu tatrāpi saukumāryaṁ praveśayanti.

THE RITI

"There is arthavyakti where the sense is not to be conjectured, as : Hari (Viṣṇu) raised the earth out of the ocean, which was reddened by the blood of the snakes crushed by his claws. If (in this case) nothing would have been said but : The great boar raised the earth out of the reddened ocean, one would have to conjecture 'the blood of the snakes".

Arthavyakti is not identical with prasāda. As regards the latter, clearness is established in the sense of a word in so far as it is not too unusual, while as to the former a sentence does not contain all that is necessary to understand the connection of ideas wholly. Daṇḍin (I, 75) concedes that the Gauḍas also aim at arthavyakti.

7. Udāratva

I, 76 : utkarṣavān guṇaḥ kaścid yasminn ukte patīyate, tad udārahvayaṃ tena sanāthā kāvyapaddhatih.

"When in a sentence there is perceived a quality of peculiar excellence, then it is called udāra. The style of the kāvya is permeated by this (guṇa)".

The most important word in this definition is pratīyate. The idea of a special excelling quality is not mentioned directly by words, but is rather suggested by other ideas, which are as such of a more subordinate character. If that is the case, we have the best kind of poetry. In this statement of Daṇḍin's we notice some of the so-called dhvani.

Udāratva is illustrated by verse I, 77, 78 :

arthinaṃ kṛpanā dṛṣṭis tvanmukhe patitā sakṛt, tadavasthā punar deva nānyasya mukham īkṣate. iti tyāgasya vākye'sminn utkarṣaḥ sādhu lakṣyate. anenaiva pathānyatra sa mānanyāyam ūhyatām.

112 THE FOUNDATIONS OF INDIAN POETRY

"When the poor eye of the supplicants has once fallen upon thy face, then, O king, it looks no longer on the face of any other. In this stanza the excelling quality of liberality is well recognised. In a way similar to this the udāratva should be expressed also in other cases."

There is, however, still another opinion regarding the guṇa in question, as mentioned in the next verse : ślāghyair viśeṣanair yuktam udāram kaiścid iṣyate, yathā līmbujākṛtāsarohemaṅgadādayah.

"Some scholars hold as udāra what is endowed with epitheta ornantia, as : a toy-lotus, a toy-pond, a golden bracelet, etc."

8. Ojas.

I, 80 : ojal samāsabhyastvam etad gadyasya jīvitam, padyepy adākṣinātyānām idam ekam parāyanam.

"There is ojas where long compounds are employed. This (ojas) is the life of the prose (-kāvya). (But in the opinion) of the Gaudas, (ojas) is the only and highest aim also for (the kāvya in) verse".

As regards the definition of ojas Daṇḍin agrees with Bhānahha, who however, does not use the term guṇa, and does not expressly confine it to prose alone.

In I, 81-85 the author specifies ojas :

"In so far as heavy or light syllables are in the majority, in the minority, or combined with each other, this (ojas)16 is of a higher or a lower kind. It is to be met with in the ākhyāyikā etc.

The Western Quarter whose sun-ray-cover is

16 Tarkavāgiśa explains tad in the text as meaning samāsabhuuyastvam, but in my opinion ojas is meant.

lying on the top of the setting mountain looks like a woman whose red and beautiful garment lies on the large breasts.

In this way the Eastern (poets) employ speeches which are full of ojas, but the others acknowledge ojas of speeches, only if it is not confusing and is pleasant, as in the following verse" :

payodharatotsaṅgalagnasandhyātapāṅśukā, kasya kāntāturāṅ ceto vāruṇi na karisyati.

"Whose mood is not rendered love-sick seeing the Western Quarter, the garment of which the evening sun hangs on the slopes of the clouds (of the breasts)".

The first instance shows the ojas of the Gaudas, who are fond of long compounds also in verse ; besides that, this sort of style is to be recognised by 'the chosen words. In the second example we have apparently another kind of ojas peculiar to the Vaidarbha-rīti. The plain ślistarūpaka (payodhara) bestows a certain charm upon the whole sentence.

The long but soft sounding compound of the first line is balanced by the short words of the second.

9 Kāṇti.

I, 85 : kāntāṅ sarvajagatkatāṅ laukikārthanatikramāt, tacca vārttābhidhāneṣu varṇanāsv api dṛśyate.

"A sentence is kānta, if it pleases all the world by not exceeding the bounds of the natural. And this (pleasing mode of expression) is to be noticed in dialogues as well as in descriptions."

The first kind is illustrated by I, 86 : gṛhāṇi māma tāny eva taporāśir bhavādṛśāh, sambhāvayati yāny eva pāvanaiḥ pādapāmśubhiḥ.

314 THE FOUNDATIONS OF INDIAN POETRY

"Those only are real houses which an ascetic like you honours with purifying dust of his feet".

The second kind (I, 87) :

anayor anavadyāṅgi stanayor jṛmbhamānayoh,

avakāśo na paryāptas tava bāhulatāntare.

"O you with blameless limbs, on your bosom there is not room enough for your swelling breasts.

Dandin comments on these verses (I, 88) :

iti sambhāvyam evaitad viśeṣākhyānasamskṛtam,

kāntam bhavati sarvasya lokayātrānuvartinah.

"Matters like that become, because they are possible and endowed with the diction of a peculiar excellence, pleasant for every one who has to do with worldly affairs".

Something of hyperbole lies, of course, in the above given instances, but as every poetical description is based on atiśayokti, and a matter-of-fact account of the actual state of things has but little to do with poetry, the ideas of the stanzas are pleasing and natural notwithstanding17.

The style of the Gaudas, on the other hand, contrasts greatly with Vaidarbhī as regards the naturalness of ideas, as it employs exaggerations to such an extent that they go quite beyond the usual limit. Dandin proceeds (I, 89-92) :

"If one intends to express an idea by raising it metaphorically beyond the natural limit, then only the (over) clever18 are satisfied, but not the others;

17 There is also an arthālaṅkāra called atiśayokti.

18 Viḍagdhā must be understood ironically, Tarka-vāgiśa says viḍagdhāḥ viḍagdhān manyamānā Gauḍā ity arthah; solluṇṭhanoktir iyam.

Our house has to be revered from this time on as the abode of a god, because its sins are entirely washed off by the falling dust of your feet.

The creator has made too narrow the world, not considering that your breasts would grow to such an amplitude.

This is called exaggeration; it is favoured very much by the Gauḍas, but the method explained above is the kernel of the other (Vaidarbha) style".

Thus kānti of the Vaidarbha and atyukti of the Gauḍa are in contrast with each other.

10 Samādhi.

1, 93, 94 : anydharmas tato'nyatra lokasımānurodhinā, samyag ādhīyate yatra sa samādhih smr̥to yathā. kumudāni nimīlanti kamalāny unmiṣanti ca, iti netrakriyādhyāsāl labdhā tadvācinī śrutih.

"If anybody, respecting the limits of naturalness, transfers a quality of one thing to another, it is called samādhi ; as :

The day-lotuses shut their eyes and the night-lotuses open them19. In this instance we find the metaphorical transference of the function of the eye (to the shutting and opening of lotuses), a term which designates this (the function of the eye)".

Here also, says Dandin, one should respect naturalness. The commentary gives an instance of bad adhyāropa : vrmhanti maṣakā yatra tatra nidrā sudurlabhā. "Where the mosquitos are trumpeting, there is hardly any sleep to be found". By transferring the function of trumpeting, 'peculiar to elephants, to

19 This is, however, not a quite correct rendering of the Sānskrit terms nimīlanti and unmiṣanti.

the mosquitos the limits of naturalness are too far transgressed.

In the next verses Dandin points out that in a metaphorical way even those words can be used,

the placing of which is prohibited in a literal sense,

as nigthin, udgr, vam, which all mean "to spit".

The metaphorical use of more than one quality

at the same time is also permitted (I, 96b-98) :

"Even a transference of several qualities at the

same time is desirable : These rows of clouds, which

are tired by bearing the heavy embryo and grown

(thunder), lean against the lap of the slopes.

The leaning against the lap of the female friend,

groaning, (feeling of) heaviness, fatigue : all those

manifold qualities of a pregnant woman are (meta-

phorically) pointed out in this example".

The adhyāropa or adhyāsa described above màkes

up a great part of poetry, and here lies a certain

difficulty of rendering it into another language, for

by translating in the first instance nimīl and unmiṣ

by 'to close and to open the eyes' we express too

clearly what the Sanskrit terms give in a more

suggestive way. Pointing out the high importance

of samādhi Dandin says (I, 100) :

tad etat kāvyasarvasvm samādhir nāma yo gunah,

kavisārthah samagro'pi tam enam anugacchati.

"The whole troop of poets should aim at the guna

called samādhi, which is the quintessence of the

kāvya".

Dandin's dealing with rīti winds up with the

stanzas I, 101, 102 :

"In this way both styles (Vaidarbhī and Gaudīyā).

differ from each other on account of their peculiar characteristics. But their further differences, as they are met with in (the compositions of) the various poets, cannot be described (on account of their being too manifold).

Great is the difference between sugar-cane, milk, molasses, etc. ; but even Sarasvatī would not be enabled to point out that (difference)”.

Thus Dandin has at full length refuted the view of Bhāmaha, who was inclined to deny the difference of rīti. In the opinion of Dandin the style of Vaidarbha is better, because it is endowed with the ten guṇas described fully by the author. The differences between Vaidarbhī and Gauḍīyā were in fact certainly not so great as Dandin would have us believe, but he is influenced by a prejudice against his predecessor.

Generally speaking, the theory of rīti as described by the author of the Kāvyādarśa was acknowledged by later scholars, though there were many dissenters with respect to particulars. Even Vāmana, who in many cases has followed Bhāmaha, concedes not only the differences of style but is of opinion that the style is the soul of poetry (I, 2, 6 : rītir ātmā kāvyasya). He attempts to present a definition of rīti (I, 2, 8) :

“A specified arrangement of words (is called) rīti.

According to Vāmana there are not two but rather three different styles, the differences of which are evident, the third being Pāñcālī. This statement is of some interest and of some importance with regard

118 THE FOUNDATIONS OF INDIAN POETRY

to kāvya literature. The author has something to say about the origin of ṛtis (I, 2, 9 and 10 vṛtti). "There rises a question : Do the guṇas (which characterise the different kinds) of kāvyas originate in different countries, that they are called after the names of countries (Vidarbha, Gauḍa, Pāñcāla), as certain material products (come from certain countries) ? (We answer :) No ! For (the sūtra runs) : They are called Vaidarbha etc., because they are met with in different countries as Vidarbha etc., (that is) because they are used in their pure form by poets in the countries of Vidarbha, Gauḍa and Pāñcāla, therefore these ways of diction are called after the name of the countries ; but the countries by themselves have by no means any effect upon the (form of) Kāvyas."

Vāmana agrees with Daṇḍin not only on this point that the differences of ṛiti are founded on the guṇas30, but that among the ṛitis Vaidarbhī is the best. Vaidarbhī is endowed; he says, with all guṇas (I, 2, 11 : samagragunopetā Vaidarbhī). With respect to the superiority of Vaidarbhī the author quotes the following ślokas :

asṛṣṭā doṣamātrābhih samagraguṇagumbhitā, vipañcīsvarasaubhāgyā vaidarbhī ṛtir iṣyate. tām etām kavayah stuvanti : sati vaktari saty arthe sati śabdānusāsane, asti tan na vinā yena parisravati vāṅmadhu.

"That ṛiti is called Vaidarbhī which is untouched

20 Vāmana's treatment of the guṇas which differs widely from that in the Kāvyādarśa, will be discussed in another place.

by even the slightest faults, furnished with all the gunas, and which sounds sweetly as the notes of a lute.

The Kāois praise it :

There may be a speaker, there may be a (good) sense (or tenor), there may be (a usage of words which is correct according to) grammar—without the (Vaidarbhī) the honey of speech will not flow".

This style is illustrated by a stanza from Śakuntalā21.

gāhantāṃ mahiṣa nipānasalilaṃ śṛṅgair muhus tāḍitaṃ chāyābaddhakadambakaṃ mrgakulam romantham abhyasyatu,

visrabdhair kriyatāṃ varāhapatibhir mustālṣatih palvale viśrāntim labhatāṃ idam ca śithilajyābandham asamaddhanuḥ.

"The buffaloes may wallow in the water, striking it again and again with their horns ; the flock of deer may ruminate in the shadow of the wood ; the great boars may uproot without fear the pools' grass ; and this bow of mine may rest with loosened string".

This instance shows that Vaidarbhī by no means avoids every compound absolutely but later on we shall meet a special kind of this style, where there are no compounds at all.

Gaudīyā is endowed with but two gunas : ojas and kānti (I, 2, 12 : ajahkāntimatī Gaudīyā). The commentary says a little more :

"As Gaudīyā is devoid of mādhurya and sauku-mārya, it has long compounds and harsh sounding words. There is the following stanza :

The learned ones, well-versed in style, praise

21 Act II, ed. Cappeller, p. 19.

Gaudīyā as a way of expression which has (long) compounds and high sounding words and which therefore is endowed with the guṇas, ojas and kānti".

The following stanza from Bhavabhūti's Mahāviracarita (I, 54) is quoted as an instance of Gaudīyā :

dordan̄dāñcitacandraśekharadhanurdaṇḍāvabhańgodyatas-

taṅkārādvanir āryabalacaritaprastāvanādiṇdimāḥ

drākparyastakapālasampuṭamitabrahmāṇḍabhāṇḍodara-

bhrāmyatpiṇḍitacaṇḍimā katham aho nādyāpi viśrāmyati.

"Alas ! Even now the twanging sound of the bow-string does not stop, (the sound) produced by the breaking of Śiva's bow, which he bent with his arms, (the sound of) a drum for glorifying Rāmn's youthful deeds, (the sound) the heaped-together-wrath of which rumbles through the world in the vessel of the Brahman-egg, (the world) which lies torn asunder between the rapidly burst shells".

Comparing this stanza with the above quoted words of Kālidāsa's the characteristic feature of Gaudīyā becomes very obvious. There is scarcely a possibility of lengthening the compounds still more than is done by Bhavabhūti. The mode of expression appears to be very stilted and the words are not the usual ones²².

Pāñcālī, the last of the rītis mentioned by Vāmana, possesses the two guṇas mādhurya and saukumārya (I, 2, 13). The vrtti says :

ojahkāntyabhāvād anulbaṇapadā vicchāyā ca, tatra ca ślokāḥ :

āśliṣṭaślathabhāvāṃ tu purāṇacchāyayānvitām,

madhurāṃ sukamārāṃ ca pāñcālīṃ kavayo viduḥ.

²² Reading Bhavabhūti's dramas one will find that the stanzas therein are of a type similar to the above mentioned. They are, however, not throughout of the same kind ; but, on the whole, the rīṭi of this poet's work is Gaudīyā.

THE RITI.

"As the Pāñcālī does not show ojas and kānti, it has no high sounding words and is without brilliance. There is a śloka :

The wise ones call Pāñcālī sweet and tender, which is endued with mildness and resembles the style of Purāṇas".

Example :

grāme' smin pathikāya pāntha vasatir naivādhunā diyate rātrāv atra vihāramanḍapātala pānthah prasupto yuvā, tenotthāya khalena garjati ghen smṛtvā priyāṃ tat kṛtam yenādyāpi karañkadanḍapatanāśaukī janas tiṣṭhati.

"Traveller, in this village no wayfarer is allowed to dwell now. (Once) a young traveller slept at night under the roof of this pavilion. When the cloud was thundering, the wicked one rose, remembered his beloved, and did something by reason of which even to-day people constantly fear the falling-down of the skeleton"23

After having spoken of the particular qualities of the single ritis Vāmana goes on to deal with the relation of Vaidarbhī, Gauḍīyā, and Pāñcālī to each other. The poet should cling to the Vaidarbhī, because only this mode of expression is possessed by all guṇas, but not to the two others, as they have but a few of them (I, 2, 14, 15). The opinion of some scholars, according to whom the poet should practise in the other ritis and may so become a master in Vaidarbhī, is refuted by Vāmana, who says that if one limits oneself to the unreal, one cannot attain the real (tattva) ; a weaver, who weaves only with

23 As this stanza appears to have been quoted from a poem we do not know, the.context is doubtful.

XVI.

132 THE FOUNDATIONS OF INDIAN POETRY

jute, how can he get skill in the weaving of silk (I, 2; 16-18).

Thus, the Vaidarbhī is the best rīti, and there is, says Vāmana, a special kind of it, described in I, 2, 19-21 :

sāpi samāsābhāve suddhavaidarbhī, tasyām arthaguṇa-sampadāsyād, tadupāroḥād arthaguṇaleśo'pi.

"Vaidarbhī is called pure Vaidarbhī, if there are no compounds. In this (pure Vaidarbhī) the richness of arthagunas comes to be tasted. Even the slightest arthaguṇa (is to be tasted), when connected with the (pure Vaidarbhī, to say nothing of the entirety of the arthagunas)".

The commentary gives the following lines :

"In the (Vaidarbhī) there is an incomparable arrangement of words, where even what is nothing becomes something, as it were. When it reaches the wise one's ear, it causes delight ; it enters the heart like a stream of nectar.

Such 'ripeness' (pāka), pleasing the hearts of men of taste, results from the Vaidarbha style that the beauty of words quickens, and that even the unreal acquires a reality".

There are two other passages in Vāmana's book where the term pāka is mentioned. Rājaśekhara dealing with vyutpatti, goes into detail concerning the pāka. On a previous occasion the matter in question has been spoken of24.

Finally, Vāmana says with respect to. this most excellent kind of diction (I, 1, 22) :

24 See above, p. 51.

sāpi vaidarbhī tātsthyāt; and the vṛtti : sāpīyam artha-gunasampad vaidarbhitā uktā; tātsthyād ity upacārato vyavahāram darśayati.

"This richness of arthagunas is even called (metaphorically) Vaidarbhī, because (those guṇas are to be found in the (Vaidarbhī and nowhere else)".

Let us now turn to Rudraṭa. Bhāmaha and Daṇḍin have described two rītis : Vaidarbhī and Gauḍīyā; Vāmana presents a third : Pāñcālī; in Rudraṭa's Kāvyālamkāra we meet a fourth : Lāṭīyā.

But there is a remarkable difference between Rudraṭa's treatment of the rītis and that of his predecessors : Rudraṭa does not judge the mode of expression from the standpoint of guṇa, but from that of the structure of compounds.

Rudraṭa in the second book defines kāvya as a 'union of word and sense. After having stated that there are four categories of words, he continues (II, 3-5) :

"Nouns are twofold, compounded or not compounded. When the nouns take the shape of compounds, then we have three rītis.

They are called Pāñcālī, Lāṭīyā and Gauḍīyā, as they have short, medium, and long compounds, respectively.

We have Pāñcālī when there are compounds of two or three words, Lāṭīyā, when there are compounds of about five to seven words; there is Gauḍīyā where compounds are formed by as many words as possible".

Verbs prefixed by prepositions are not considered as compounds in this sense; these prepositions are used only to modify the meaning of the verb

respectively.

124 THE FOUNDATIONS OF INDIAN HISTORY

(II. 6a : ākhyātāny upasargaih samśrjyānte kadācid arthāya).

The above described three ṛtis are in contrast to the diction of Vaidarbhī ; II, 6b :

ṛtter samāsāvā vaidarbhī ṛtir ekaiya.

"When the words are compounded, then we have but one ṛti, the Vaidarbhī".

That is all Rudraṭa has to say about ṛtis. His treatment of matters apparently shows that the doctrine of ṛti has lost a great deal of its former significance and importance, and that it was no longer founded upon geographical differences and upon the existence on contemporary compositions belonging to a certain style. Any poem might have verses in different ways of diction, according to the sort of compounded words met therein. Vaidarbhī was thought the finest style, and had in course of time put all the others in the background. As Rudraṭa's Vaidarbhī contains no compounds, it resembles the suddha-Vaidarbhī of Vāmana.

Rājasékhara in his Kāvyamīmāṃsā (adhyāya 7, p. 31) mentions only three ṛtis : Vaidarbhī, Gauḍīyā, and Pāñcālī, which he considers as the three forms of speech (ṛtirūpaṃ vākyatritayam), without giving any further explanations. Rājaśekhara appears, as regards these three ṛtis, to be influenced by Vāmana. For the author of the Kāvyamīmāṃsā another factor is of much greater importance, that is the manner of recitation and the change of pronunciation which bears the name kāku. As kāku is reckoned under the sabdālankāras by Rudraṭa (which opinion is not shared by Rājaśekhara), it must

be described on another occasion, when we treat the history of the alamkāras.

In the time of Mammata the doctrine of rīti became of still less significance. Though in the Kāvyaprakāśa the whole domain of poetics is treated exhaustively, rīti is touched on only by the way.

It is mentioned twice : first in the eighth ullāsa, where Mammata refuses to accept Vāmana's opinion regarding the guṇas, and secondly in the ninth ullāsa.

Here, the śabdalamkāras are defined. The anuprāsa, the author says, is twofold being cheka- and vṛtya- nuprāsa, the latter consisting in the repetitions of the same consonants.

According to the various sorts of consonants there are three kinds of this anuprāsa : 1. upanāgarikā : the repeated consonants suggest mādhurya ; 2. paruṣā : the consonants are the cause of ojas ; 3. komalā : when the sounds are of a kind different from those of the two mentioned.

Now the author says that in the opinion of Vāmana etc. the three kinds of rīti Vaidarbhī, Gauḍī, and Pāñcālī are based on those three kinds of vṛttyanuprāsa.

It appears from the statements that the doctrine of rīti, though of great importance in the period of Daṇḍin (the greatest defender of the difference of dictions), had retained only a mere historical interest.

THE DIVISION OF POETRY

The division of poetry (kāvya) had always been a matter of great interest for the Indian teachers of poetics. Before going into details we have to examine from which points of view the division of the kāvya has been made.

Bhāmaha, who deals with the matter immediately after the definition of kāvya is given, presents a division according to the following standpoints : (i) prose and verse (gadya and padya). (ii) the language the composition is written in ; that is Sanskrit, Prakrit, and Apabhramśa. (iii) the subject matter : anything which has really happened, which is invented by the poet's imagination, the domain of arts, and, finally, that of śāstra. (iv) The shape of the work as a whole: sargabandha, abhineyārtha, ākhyāyikā, kathā and anibaddha.

In the work of Dandin the divisions are arranged a little better: the author divides according to only two points of view, which are (1) gadya and padya, and (2) the language. The other kinds mentioned by Bhāmaha are brought under the groups gadya or padya, so that sargabandha, both ākhyāyikā and kathā, and abhineyārtha (nāṭaka), and campū (not mentioned by Bhāmaha) are but subdivisions of padya (verse), of gadya (prose), and of a combination of both padya and gadya, respectively.

Vāmana, too, gives only two main divisions, but not in the same way as Dandin; these are :

THE DIVISION OF POETRY

1. gādya and padya ; 2. anibaddha and nibaddha. Anibaddha is that kind of poetry which consists of verses each distinct in itself (e. g. the śatakas are of this kind); nibaddha, on the other hand, is a poem, where the whole of the stanzas forms the content. Poetry in prose is according to Vāmana, threefold : vrttagandhi, cūrṇa, and utkalikāprāya, which division is not respected by later ālaṃkārikas. Sargabandha etc. are not mentioned, and nothing is said about the different languages.

Rudraṭa, again, has another standpoint. A kāvya is a combination of word (śabda) and sense (artha), and as Rudraṭa describes both subjects separately throughout, the division of kāvya is to be met with in two different passages of the Kāvyālaṃkāra: in the second and in the sixteenth adhyāya. Under the heading śabda we find gādya and padya, and, further, the division with respect to the language; under the heading 'artha' the classification according to the content: utpādya (where the subject is invented by the poet) and anutpādya (where the subject is known). This last idea is, however, not quite new, as it had already been presented by Bhāmaha. A second principle of division is the greater or lesser extension of the composition (prabandha). Mahā-kāvya, ākhyāyikā, and kathā are reckoned under utpādya poetry.

Though the ways of classification greatly differ from each other in some respects (the authors of the alaṃkāra śāstra endeavouring to present new ideas), the general principles are the very same everywhere. Then only, when poetry was considered from a higher

point of view (that of dhvani), the above mentioned classifications were no longer placed in the fore-ground. They were, however, not considered as incorrect or false, but they were of only little interest (vaicitrya, which term had been employed already by Vāmana in a similar connection) for the scholars of more 'modern' times. If anybody wished to inform himself regarding those divisions, he was referred to the older masters, who had treated the matter fully.

Thus M a m m a ṭ a teaches that there are three kinds of poetry : superior (uttama), medium (madhya-ma), or inferior (avara) poetry, according to the superiority or inferiority of the 'unspoken'.

In R ā j a ś e k h a r a's Kāvyamīmāṃsā there are ideas we do not meet with in any of the former works on alamkāra. Though the author's explanations are, for the greater part, of little value, and though, strictly speaking, they ought not to be treated here, because they deal with the division not of kāvya but of vākya, we will give a short account of them. Speech (vākya=vacana), says Rājaśekhara, is, according to what the author is, either brāhmya, or śaiva, or vaiṣṇava. There are five divisions of the first kind: the speaker may be Brahman himself, the īśvaras (the Bhṛgus, etc., who are descended from Brahman's mind), the sons of the īśvaras (the ṛṣis), the offsprings of the ṛṣis, and, finally, their sons. Examples may be taken from the Purāṇas. But according to the view of the kavīs (Rājaśekhara continues) speech (vacas), as it is found in the Vedas, is called pāramestvara and

later on, divya, because the gods (deva) and those beings who are of divine origin (devayoni : vidyādhara, apsaras, yakṣa, rakṣas, gandharva, kinnara, siddha, guhyaka, bhūta, and piśāca) act as intermediaries. With respect to this, four kinds of speech are especially of importance : vaibuddha, vaidhyādhara, gāndharva, and yoginīgata. The form of the compounds is partly decisive here. There are some interesting peculiarities : The Piśācas (attendants to Śiva) should speak in their own sphere Sanskrit, but on earth Bhūtabhāṣā; the language of the Apsaras is Prakrit. That is according to Rājaśekhara important, because there appear gods, etc. in the nātakas. Divya is in contrast to Vaiṣṇava. The language called Vaiṣṇava has come to earth by means of the incarnations of Vāsudeva, so that it is commonly called mānuṣa. With respect to the ṛti speech is threefold: Vaidarbhī, Gauḍīyā, and Pāñcālī, but the way of recitation (kāku) gives rise to varieties of these three kinds.

We will return to Bhāmaha in order to see in which way poetry has been divided by him. He says (I, 16): śabdārthau sahitau kāvyam gadyaṃ padyam ca tad dvidhā, saṃskṛtam prākṛtam cānyad apabhraṃśa iti tridhā. "Word and sense combined are kāvya. It is twofold, prose and verse ; it is further threefold, as it may be composed in Sanskrit, in Prakrit or in another (idiom) different (from those), viz. Apabhraṃśa."

180 THE FOUNDATIONS OF INDIAN POETRY

1. Gady a and Padya.

According to D a n d i n there are not two kinds,

prose and verse (as Bhāmaha thought), but three

(Kāvyādarsa I, 11a):

padyam gadyam ca miśram ca tat tridhāiva vyavasthitam

"The (body of the kāvya) is threefold: verse, prose,

and mixed (verse and prose)."

It is also worth noting that Dandin does not

begin with gady a, as Bhāmaha did, but with padya.

As is said in I, 31 the nātakas and other composi-

tions belonging to the miśrakāvyas.

As to padya the author comments (I, 11b, 12):

padyam catuspadi tac ca vrttam jātir iti dvidhā.

chandovicityām sakalas tatprapañco nidarśitah,

sā vidyā naus titirṣūnām gambhiram kāvyasāgaram.

"Padya is a stanza, consisting of four metrical

feet; and the padya is of two kinds: vrtta (i. e. metres

wherein the syllables are counted), jāti (metres where-

in the moras are counted). In prosody all the varie-

ties of metre have been described: this knowledge is

the ship for every one who is willing to cross the

profound kāvya-ocean."1

Vāmana is of the same opinion as Bhāmaha,

stating (I, 3, 21):

kāvyam gadyam padyam ca.

"The kāvya is in prose and in verse."

That the miśra of Dandin is contained therein

follows from the further description of Vāmana ;

gady a is named first (as was done by Bhāmaha),

1 From this stanza some scholars (J a cobi, Peterson)

would gather that Chandoviciti is the title of a third work of

Dandin's; but Chandoviciti is a common term for 'prosody'.

See note 14, on page 49.

because it presents more difficulties in the opinion of the author9. Thus Vāmana corroborates the opinion of the master.

There are, according to Vāmana (I, 3, 22) three kinds of gadya. When verses occur occasionally in the prose, it is called vrttagandhi. This is the misra of Dandin. Cūrṇa is called that sort of prose the words of which are pleasing and do not form too long compounds( I, 3, 24): anāviddhalalitapadam cūrṇam3. Utkalikāprāya being the reverse of the latter has long compounds and high sounding words.

As we have seen before these qualities are peculiar to different styles ; utkalikāprāya shows a very great similarity to gaudīyā.

Padya is, of course, divided into many kinds, but no further detail is given by Vāmana4.

Rudrata as well as Bhāmaha and Vāmana deals only with the two kinds gadya and padya, which, however, are not divisions of kāvya itself, but rather of what kāvya consists of and what is called kāvyasarīra by Dandin. Considering that according to the Kāvyādarśa sarīra is defined as padāvalī (a series of words), and that in Rudrata's opinion the vocal expression of kāvya is the sentence (vākya), it appears that the theories of both Dandin and Rudrata are almost the same in this point. Rud-rata says (Kāvyalāmkāra II, 11):

2 gadyasya pūrvam nirdeśo durlakṣyaviśeṣatvena durbandhatvāt, tathāhuḥ,—gadyam kavīnām nikaṣam vadanti.

3 Vṛtti: anāviddhāny adīrghasamāsāni lalitāny anud-dhatāni padāni yasmims tad anāviddhalalitapadam cūrṇam iti.

4 Vṛtti: padyam khalv anekena samārdha-sama-viṣa-mādinā bhedena upetam bhavati.

vākyam bhavati dvedhā gadyam chandogatam ca.

"The sentence is twofold, prose and verse.'

Mammata does not mention this division in his

Kāvyaprakāśa at all.

2. Sanskrit Prakrit and Apabhramśa

Bhāmaha's second division (I, 16) is determined

by the idiom the kāvya is composed in5. The stanza

is of some interest on account of the word apabhram-

śa6. Apparently Bhāmaha does not understand

by this term one special language, but rather, a

group of several idioms, which widely differ from

Sanskrit. This is suggested by the name itself.

If this is right, prākṛta does not mean a class of

idioms, but a certain dialect; and this dialect is

probably the māhārāṣṭrī, which without doubt in the

period of Bhāmaha had a very great importance as

a language in which poetical compositions were

written.

As in my opinion Dandin was a younger contem-

porary of Bhāmaha the literary facts spoken of in

5 The stanza (Bhāmaha I, 16b) is quoted by Namisādhu,

commenting on Kāvyālaṃkāra II, 11, in the following form

prākṛtam saṃskṛtam caitad apabhraṃśa iti tredhā.

The

incorrectness of this appears already from the anteposition

of prākṛtam. The difficulty lies in anyad. Another transla-

tion of this important verse than the one given above is

scarcely possible.

6 Of the more recent investigations on Apabhraṃśa, I

may here mention two papers by Prof. Jacobi: Bhavisatta

kahā von Dhanavāla (Abhandlungen der Kgl. Bayer. Akad.

D. Wiss., Philos.-philol. Kl., Vol. 29, 4), and Sanatkumāra-

carita (ib. Vol., 31, 2). The former paper is in many respects

supplemented and corrected by the latter.

THE DIVISION OF POETRY

the Kāvyādarśa are, generally, the same as those we met with in the work of Bhāmaha. We find indeed (the suggestion mentioned above corroborated in the Kāvyādarśa. Daṇḍin mentions (I, 32) Sanskrit, Prakrit, and Apabhraṃśa, which three idioms are employed for literary purposes. Let us see what Daṇḍin has to say concerning Apabhraṃśa.

I, 36: ābhirādigiraḥ kāvyeṣv apabhraṃśa iti smṛtāḥ, sāstreṣu sanskṛtād anyad apabhraṃśatayoditam.

"In the kāvyas the idioms of the ābhīras etc. are considered as apabhraṃśa. But in the śāstras that is called Apabhraṃśa which differs from Sanskrit"7.

According to this statement, the word apabhraṃśa has a double meaning. In the śāstras, and especially in the books on poetical theories, and thus in the work of Bhāmaha all languages which are different from Sanskrit are united under the name of apabhraṃśa. There is little doubt that, that was in the older period the meaning of the word, and that only in the course of time the word assumed a more pregnant and particular sense, perhaps then, when the 'best' Prakrit, māhārāṣṭrī, was used as a language fit for literary compositions. Then the scholars retained the older meaning of the word, whilst in common use the term prākṛta, once the designation of a determinate language (māhārāṣṭrī),

7 Tarkavāgiśa thinks the term ābhīra not a nomen proprium, but rather used in a professional sense: cowherd. He says: kāvyeṣu nātakādiṣu ābhīrādigiraḥ ābīri-prabhṛtayo gopacāṇḍālaśakārādīnāṃ vyavaharaṇīyā bhāṣā apabhraṃśa iti smṛtāḥ apabhraṃśanāmnā bhāṣānirūpakair nīrūpitāḥ. But this is probably not the case.

134 THE FOUNDATIONS OF INDIAN POETRY

took the place of ‘apabhramśa8. It ‘apabhramśa’ was mentioned in. the kāvyas something different from the apabhramśa of the śāstras was meant, viz. such dialects as were a good deal less important and lower than the vernacular of the ābhīras etc., in which idioms scarcely any literary works were composed in the period of Bhāmaha and Daṇḍin. The home of the ābhīra-language was, as is stated by Taruṇavācaspati, one of the commentators on the Kāvyādarśa, in Western India9.

With respect to the various dialects Daṇḍin says (I, 33):

samskṛtaṃ nāma daivī vāg anvākhyātā maharṣibhiḥ, tadbhavas tatsamo deśīty anekaḥ prākṛtakramaḥ.

"Sanskrit is the divine language, taught by the great sages. The variations of Prakrit are numerous, being tadbhava (come forth from that, i. e. Sanskrit); tatsama (similar to that), or belonging to certain regions."

From the term deśī we may gather that the word Prākṛta is used here in a wider sense including also Apabhraṃśa. Taruṇavācaspati, however, thinks that the Māgadhī language is meant (deśī iti māgadhī grhyate). The author of Hṛdayañgama comments: deśī povalli iti kramukanāmā, cassīmsī iti kañcanasya nāma, dogghaṭa iti gajasya nāma.

8 Prof. Pischel in his famous Prakrit Grammar quotes, (p. I) an observation by Śaṃkara to Śakuntalā 9, 10, saying: saṃskṛtāt prākṛtaṃ śreṣṭhaṃ tato’ pabraṃśabhāṣaṇam.

9 The commentary of Hṛdayaṅgama has another interpretation: ābhiro nāma paścimārṇavattire vartamāno deśaviśeṣaḥ, athavā ābhirādayo gopālādayaḥ.

the word's wider sense is in contrast to Prakrit in the narrower sense, (I, 34:)

mahārāṣtrāśrayāṃ bhāṣāṃ prakṛtam prākṛtaṃ viduḥ, sāgarah sūktiratnānāṃ setubandhādi yanmayam.

"As the most excellent Prakrit the language of Mahārāṣṭra is considered. The Setubandha etc., the ocean of jewels of good sayings, is composed therein."

Other variations of Prakrit are mentioned in I, 35 :

śaurasenī ca gauḍī ca lāṭī cānyā ca tādrśī, yāti prākṛtam ity evaṃ vyavahāreṣu saṃnidhim.

"Saurasenī, Gauḍī, Lāṭī and other dialects of this kind are as 'Prakrit-variations' met with in the dialogues (in nāṭakas etc.)"

As to the division caused by the difference of the languages Daṇḍin states that from this point of view four kinds of kāvya must be distinguished (not three, as Bhāmaha had said) :

tad etad vāṅmayam bhūyaḥ saṃskṛtaṃ prākṛtaṃ tathā, apabhraṃśaś ca miśraṃ cety āhur āptāś10 caturvidham.

"This (body of the kāvya) formed by the languages is further taken by the authorities as being fourfold, as far as it is Sanskrit, Prakrit (Māharāṣṭrī), Apabhraṃśa, or mixed (dialects)11."

In the next stanzas the question is answered, in which manner the various idioms are peculiar to the different sorts of poetical compositions :

saṃskṛtam sargabandhādi prākṛtaṃ skandhakādikam, āsārādin̄y apabhraṃśo nāṭakādi tu miśrakam.

10 I prefer this reading to āryāś. The Tibetan translation has a word which means 'clever, expert'. With respect to āpta compare Kāvyādarśa II, 23f.

11 tad evaṃ vāṅmayam refers to śarīra in I, 9. Bhīyas goes back to stanza 10.

136 THE FOUNDATIONS OF INDIAN POETRY

"A sargabandha etc. is in Sanskrit ; compositions in the metre skandhaka etc. in Prakrit ; compositions in the metre āsāra etc. in Apabhramśa, but nātakas etc. are in mixed dialects".

By the word tu the importance of the new kind 'miśra' (which is not found in Bhāmaha's) is, as it were, underlined.

Kathāpi sarvabhāṣābhiḥ saṃskrtena ca badhyate, bhūtabhāṣāmayīm prāhur adbhutārthāṃ bṛhatkathām.

"The kathā, too, is composed in all languages and in Sanskrit. The Bṛhatkathā12 with its wonderful tenor is said to have been composed in the idiom of the ghosts".

Vāmana and also, be it mentioned here already, Mammata had said nothing about the languages. We turn therefore to Rudraṭa. Bhāmaha and Daṇḍin had pointed out (though differing from each other in details) that there were three groups of idioms : a kāvya may be composed in Sanskrit, Prakrit, and Apabhraṃśa. Rudraṭa mentions six languages, saying (Kāvyālaṃkāra II, 11, 12) :

12 The term āhuh seems to indicate that the Bṛhatkathā was already unknown even to Daṇḍin. I do not believe that the plural sarvabhāṣābhiḥ is of a peculiar significance (Prof. Jacobi, Bhavisattakahā, p. 42). One must bear in mind that the above quoted stanza refers directly to Bhāmaha I, 28. Bhāmaha had taught that the kathā (with regard to which more will be said later on) is composed in Sanskrit and, partly, in Apabhraṃśa. Daṇḍin replies : no, that is not correct ; it is composed in all languages and in Sanskrit. The term saṃskrtena ca is nothing more than an appended note with a delicately hinted irony against Bhāmaha.

THE DIVISION OF POETRY

bhūyo'pi,

bhāṣābhedanimittạḥ sodhā bhedo'sya sambhavati.

prākṛtasaṃskṛtamāgadhapiśācabhāṣāś ca sūrasenī ca,

ṣaṣṭho'tra bhūribhedo deśaviśeṣād apabhraṃśaḥ.

"Further, there is a sixfold division of the (vākya) according to the languages ; these are : Prakrit, Sanskrit, Māgadha, Piśāca, and Sūrasenī ; the sixth, Apabhraṃśa is divided into many kinds according to the various countries".

From this it is evident that 'Prākṛta' does not denote a group of languages, but rather one language, the same one as mentioned by Daṇḍin, which appears to have frequently been used in poetical compositions, the Māhārāṣṭrī.

For if that were not the case it would be impossible to explain, why Māhārāṣṭrī is not mentioned by Rudraṭa. Māgadha was not named by Daṇḍin, perhaps because it was of less importance as a literary idiom in his time.

Apabhraṃśa is, on the whole, probably the same as that Apabhraṃśa in the Kāvyādarśa, when it is spoken of in kāvyas (not in śāstras).

The older Vāgbhaṭa has the following remarks about the languages (Vāgbhaṭālamkāra II, 1-3) :

"Sanskrit, Prakrit, Apabhraṃśa, and the language of the Bhūtas : these four languages become the body of the kāvya.

Sanskrit is the language of the gods ; it is settled in the grammars.

Prakrit is variously divided, as far as it originates in (Sanskrit) is similar to it, differing with respect to dialects etc.

Apabhraṃśa, in its pure shape, is spoken in various regions.

What is spoken by the Bhūtas (the Piśācas) is called Bhūta'(Paiśācika)".

xviii

Thus Vāgbhaṭa comprises by the term Prākṛta several idioms, which, however, are in some way or other connected with Sanskrit. According to the commentary of Simhadevagaṇi Saurasenī and Māgadī is meant by ādi. Prakrit is set in contrast to Apabhraṃśa. In its pure form it is found, says the commentary, in the countries of the Karnāṭas and Pāñcālas. With regard to the last group, the Bhāṇṭa or Paisācī is little known. Vāgbhaṭa was perhaps thinking of the Bṛhutkathā, which was referred to already by Daṇḍin.

In a supplementary way (because they have nothing to do with the division of poetry) some very interesting observations of Rājaśekhara concerning the different kinds of recitation may be mentioned here13. After dealing with the mode of recitation (pāṭha) generally the author goes on to point out what sorts of recitation are peculiar to the different countries. People from Magadha and others from the country lying eastward of Benares have a good pronunciation of Sanskrit, but are 'blunt (kuṇṭha)' when they speak Prakrit. Of the Gauḍas, Rājaśekhara does not speak well. Their recitation of Prakrit is, according to him, very bad. The Gauḍa Brahmans recite neither very distinctly (atispasṭa), nor 'legato (āśliṣṭa)', neither roughly (rūkṣa) nor very tenderly (atikomala), neither with a high (atitāra) nor with a deep (mandra) voice. Karnāṭas recite, whichsoever the rasa, the ṛti, or the guṇa may be, with an elevated, and at the end, with a twanging voice (ṭaṃkāra). All Dravidian poets in verse as well as in prose compositions use a

13 Kāvyamīmāṃsā, 7th adhyāya.

THE DIVISION OF POETRY

musical mode of recital (geyagarḍhe sthitah pāthe sarvo'pi draviḍaḥ kaviḥ). The author is very pleased with the beautiful Prākṛta-pronunciation of the Lāṭas, who hate Sanskrit¹ ⁴. The Surāṣṭras, Travaṇas, and others recite Prakrit well, but their way of speaking Sanskrit has something of an Apabhraṃśa note¹⁵. The people of Kashmir, says Rājaśekhara, are good poets but the author ridicules their pronunciation, which sounds, he says, as if they had the mouth full of Guḍūcīs¹⁶. The poets who dwell westward of them, in Uttarāpatha, though well educated, speak with a nasal twang (sānunāsikapāthinaḥ). Finally, Rājaśekhara praises the beautiful (subhaga) recitation of the Pāñcālas, for whom he shows great sympathy also elsewhere.

3. The Subject matter.

With respect to the division based upon the subject matter, Bhāmaha says (I, 17) :

"The kāvya is also fourfold, as it praises the past deeds of gods etc., or the subject matter is to be invented (by the poet), or the subject matter is formed by the arts, or (finally) by the śāstras.

Though the alaṃkāraśāstra, after the time of Bhāmaha, has not conserved this division of poetry, we shall come back to these ideas more than once later on.

14 Paṭhanti latabhaṃ lāṭāḥ prākṛtaṃ saṃskṛtadviṣaḥ, jihvayā lalitollāpabdhasaundaryamudrayā.

15 Apabhraṃśāvadaṃśāni te saṃskṛtavacāṃsy api.

16 Śāradāyāḥ prasādena kaśmiraḥ sukavir janaḥ karṇe guḍūcigaṇḍūṣas teṣāṃ pāṭhakramaḥ kimu.

139

140 THE FOUNDATIONS OF INDIAN POETRY

4. Sargabandha, Abhidhānārtha, Ākhyāyikā, Kathā, Anibaddha.

This division of poetry is much more important than the previous ones. Bhāmaha says (I, 18) :

"Kārya etc. is said to be fivefold : sargabandha, abhidhānārtha, ākhyāyikā, kathā, and anibaddha".

A. Sargabandha, a composition which is divided into sargas.

I, 19-23 :

"Sargabandha is a mahākāvya, dealing with great (persons)17, large, the words of which are not vulgar18, which has good sense, adorned with figures of speech, based on real events19 ; it is endowed with the description of a counsel-meeting, of (the sending of) a messenger, of a campaign, of a battle, and of the (final) triumph of the hero ; it is furnished with the five sandhis20, needs no extensive explanation.

17 In a stanza mentioned below, note 30, on p. 143 there is said :

mahākāvyam prayoktavyam mahāpuruşakīrtiyuk.

18 Grāmya is well explained by Vāmana II, 1, 7 : lokamātraprayuktam grāmyam.

19 As 'real' must of course also be considered such tales met in the itihāsas, which bear the stamp of reality on account of the holiness of tradition. Itihāsas are in contrast to such tales which have originated in the poet's imagination.

20 The term sandhi is taken from the drama. Compare Daśarūpa I, 34 ff. The five sandhis are : mukha, pratimukha, garbha, avamarṣa (Bharata : vimarsa), upasaṃhrti (Bharata : nirvahana). See also Bharata, Nāṭyaśāstra XIX, 35ff. The question cannot be discussed here, whether all the five sandhis are necessary for a kāvya. At any rate we see from the words of Bhāmaha to which degree later poetics have been influenced by the older doctrine of the drama.

THE DIVISION OF POETRY

tion, is rich (with poetical value); though it has to do with all the four vargas (dharma, artha, kāma, mokṣa), it gives, above all, instruction in artha. It describes things as they happen in the world, and it is endowed with all rasas (but) separately. The poet, after having described the hero's race, energy, fame etc., should not describe his destruction in order to praise thereby the high qualities of another person. If the hero is not described in such a way that he fills the body of the kāvya throughout, then, indeed, the mention of him in the eulogy in the beginning (of the kāvya) is in vain".

The last two stanzas do not touch on the definition of the mahākāvya (sargabandha). They have been caused by some differences of views, which will be clearer only after having learned Dandin's account concerning the matter.

Dandin's treatment of it is based on Bhāmaha's. He says (Kāvyādarśa I, 14–19):

alamkṛtam asaṅkṣiptam rasabhāvanirantaram, sargair anativistṛtaiḥ śravyavṛttaiḥ susandhibhīḥ. sarvatra bhinnavṛttāntaiḥ upetaṃ lokarañjakam, kāvyam kalpāntarāsthāyi jāyeta sadalamkṛti.

"Sargabandhaḥ is a mahakāvya¹. These are its peculiarities: It begins with an āśis, a namaskriyā, or an indication of the contents²². It is based on

21 Tarkavāgiśa: mahākāvyam sargabandhaḥ, tasya sargair nibadhyamānātvāt.

22 Āśis is the benediction (Tarkavāgiśa: āśīḥ iṣṭajana-sya śubhāśaṃsanam. Compare Kāvyādarśa II, 357, where āśis is defined as an alamkāra. Namaskriyā is the respectful salutation). Tarkavāgiśa mentions that the Kṛcakavadha

a story from the itihāsas23, or on another theme, based on a fact24; it is endowed with the success of the four vargas25. The hero is clever and exalted26. It is adorned with descriptions of a town, of the ocean, of a mountain, of a season, of the rising of sun and moon27, further, of sport in park or in water, of drinking, of love-feasts, of separations28, of weddings, of the developinent of a son, with descriptions, of a counsel-meeting, of (the sending) of a

opens with an āśis, the Raghuvamśa with a namaskriyā, and the Śiśupālavadha with a vastunirdeśa.

23 Above all Mahābhārata and Rāmāyaṇa as Tarka-vāgiśa and Vidyāsāgara say.

24 Tarkavāgiśa says that the telling of an invented story in the mahākāvyaś is prohibited (sadāśrayam ity anena kalpitavṛttāntāsya mahākāvye varṇanạṃ pratiṣiddham); and Vidyāsāgara: itarad rāmāyaṇamahābhāratādivyatiriktam api etena asatyavṛttạṃ mahākāvye na varṇanīyam iti pradarśitam,

25 The commentary adds that it is not possible that all the four vargas succeed in one place. Thus the poet should describe all the four vargas indeed, but the fruit of only one among them (tathā caturnāṃ dharmārtha kāmamokṣāṇāṃ vargaś caturvargah, tadrūpeṇa phalena prayojanopetam ekarta caturnāṃ phalatvāsambhavāt sarve punar varṇanīyāḥ parạṃ tv anyatāṃṃ eva phalaṃ iti mantavyyaṃ, uktaṃ ca catvāras tatra vargāḥ syus teṣv ekaṃ ca phalaṃ bhaved iti).

26 TarkavāgIśa: tathā caturo vyavahārakuśala udā-tto dhiṣodatto nāyakah kathāvyāpipraadhānāpuruṣo yatra tat.

27 Of course the description of sun and moon-set, too, though it is not mentioned expressly by Daṇḍin.

28 The plural is used (says Tarkavāgiśa) in order to indicate that there are several kinds of separation (vipra-lambhair iti bahuvacanena tasya pūrvarāgamānāṃpravāsa-karuṇāmakatayā catuvidhatvaṃ sūcitam,

messenger, of a campaign, of a battle, and of the final triumph of the hero29; it is not too compressed (with respect to the description), is penetrated thoroughly by the rasas (sentiments) and bhāvas (emotions), divided into sargas which are not too long30, endowed with metres agreeable to the ear31, at the end of which the metre changes32, and the sandhis of which are good33: such a kāvya that has

29 Tarkavāgīśa: mantrādipañcakaṇ kramikaṇ tathā hi prathamạṁ mantranā tato dūtapreṣaṇaṁ tataḥ prayāṇaṁ tato yuddhaṃ tataś ca ripujayādirūpābhyudayaḥ.

30 Vidyāsāgara thinks that by anativistīrṇaiḥ the number of the sargas (not the length) is meant: śāstraniyam-itāṣṭādibhiḥ tad uktam Īśānasamhitāyāṃ :—

astasargān natu nyūnaṃ triṃśatsargać ca nādhikam, mahākāvyạṃ prayoktavyaṃ mahāpurụsakīrtiyuk.

31 śravyavṛttair iti hatavṛttatādidoṣaparityāgena mādhuryādiguṇasadbhāvena ca vairasyānāvahaśrutisukhadavṛttair ity arthaḥ, says Tarkavāgīśa.

32 Tarkavāgīśa has a twofold interpetation, the first being: bhinnavṛttāntaiḥ pṛthakpṛthig avāntarakathā-prakāśakaiḥ. According to this, the poet should insert some episodical stories, but that is certainly not correct. The second interpretation alone is the correct one :

yadā bhinnạṃ vṛttạṃ pṛthakchandonibaddhaḥ śloko 'nte' avasāne yesāṃ taiḥ, ekena chandasā sargaṃ nirmāya chando 'ntarena samāpayed ity arthaḥ, uktaṃ ca ekavṛttamayaiḥ padyair avasāne 'nyavṛttakaiḥ.

33 One might be inclined to translate susandhibhiḥ by "the connections of which are good", i.e., the connection between the end of one and the beginning of the next sarga should be pleasing, unconstrained, and natural (as Tarkavāgīśa comments: susandhibhiḥ suśliṣṭasandhānaiḥ bhāvis-argasāpekṣair ity arthaḥ. Vidyāsāgara is of the same opinion). There can be no doubt, however, that Daṇḍin takes up Bhā-

good alamkāras and that (on account of its being as described) delights the world, may last till the end of the kalpa."

Comparing these words with Bhāmaha's description of the mahākāvya, Daṇḍin's dependance on Bhāmaha is evident. Several terms are borrowed verbatim, as sargabandho, mahākāvyaṃ, sadāśrayam, mantradūtaprayānājinākābhyudayaīḥ.

As a matter of course, Daṇḍin, on the one hand, is anxious to complete the definition of Bhāmaha and on the other, to omit what in his opinion are no characteristic features of this kind of composition, as the terms ayrāyaśabda and arthya.

These are indeed superfluous as they are characteristic not only of the sargabandha but also of the other sorts of poetry.

There are new points, Daṇḍin thinks, necessary for the definition of mahākāvya: so that the poet should begin with an āśis, a namaskriyā, or a vastuniṅdeśa.

Whether bhāvas and rasas are of importance even for sargabandha-compositions or not, is a question which is perhaps not answered in the affirmative for the bhāva has its proper place in the drama, on account of which the doctrine of bhāva is developed in the books dealing with the dramas.

As to the rasas they belong originally to the drama too, but they have been transferred to the kāvya already

maha's pañcabhiḥ sandhibhir yuktam again and that it must be translated as above; and in this sense is Tarkavāgīśa's second interpretation: yad vā sandhayo nāṭakā lakṣaṇoktā mukha-pratimukha-garbha-vimarṣa-nirvahaṇākhyāḥ pañca.

See also Sāhityadarpana VI, 317, where the author says expressly: nāṭakasandhayaḥ.

THE DIVISIÓN OF POETRY

very early. Some three 'sentiments' are even considered, but certainly erroneously, as alamkāras84. Regarding the things the poet should describe in the mahākāvya, Daṇḍin goes into more detail than Bhāmaha. As a matter of fact all the famous mahākāvya contain such episodic descriptions as are mentioned in the Kāvyādarśa. They are typical of the sargabandha and ought not to be wanting therein. Therefore they must be named, as is done by Daṇḍin. Bhāmaha has mentioned only those descriptions which form, so to speak, the political content of the kāvya. That this must be considered as the foundation may be gathered from the fact that Daṇḍin simply quotes the exact words of Bhā­maha concerning this matter. The four varṇas among which artha is preferred by Bhāmaha, are of similar significance. Bhāmaha says nothing about the metre. The postulation, that at the end of every sarga the metre has to change, is not quite new. There are some cases in the Vedic hymns where the poets like to wind up with one or more stanzas the metre of which is not the same as that of the preceding verses35. In the later literature we find the same method almost everywhere. This change of metre appears to have taken place to suggest to the hearer the

34 Preyas, Rasavat, Ūrjasvin; Kāvyādarśa II, 275.

35 Compare H. Oldenberg: Die Hymnen des Rigveda. Vol. i, Metrische und textgeschichtliche Prolegomena, pp. 441ff.—There is, to mention it by the way, a second method of indicating the end of a chapter, viz., to repeat the last words. Thus, we find it in the Upaniṣads. One must, however, bear in mind that these are, for the greater part, prose texts.

xix

146 THE FOUNDATIONS OF INDIAN POETRY

coming end of the sarga36. The want of punctuation is supplied as it were in that way.

According to Dandin the poet has to draw attention to and to be aware of very many things, so that a young poet may become discouraged, if he has to fulfil conscientiously all requirements. But he need not be so. Dandin continues comforting the poet by the words (I, 20).

nyūnam apy atra yaiḥ kaiścid aṅgaiḥ kāvyam na duṣyati, yady upātteṣu sampattir ārādhayati tadvidaḥ.

" When some of these components are wanting, a kāvya does not become bad, if only the success in those things which are described satisfies the learned."

Up to this point we find no great difference between Dandin's teaching and that of Bhāmaha, but as to the nāyaka, the hero of the kāvya, the views of both scholars diverge from each other. The way Bhāmaha deals with the hero leaves little doubt that he opposes an opinion of some other ālamkārikas before or contemporary with him. The poem, says Bhāmaha, should end with the hero's triumph ; it would be absurd, if the poet praised the excellences of some other person in the kāvya, through whom the chief-nāyaka would perish finally. Dandin, always desirous to go against Bhāmaha, says (I, 21, 22):

gunataḥ prāg upanyasya nāyakaṃ tena vidviṣām, nirākaraṇam ity eṣa mārgaḥ prakṛtisundaraḥ.

36 Though this need not be proved I may draw attention to the fact that even in earlier times (e.g. by the author of the Sāhityadarpana) the whole domain of kāvya is divided into śravya and dṛśya, poetry to be heard and poetry to be seen.

vamśavīryaśrutādini varṇayitvā ripor api, tajjayān nāyakotkarṣavarnanāṁ ca dhinoti naḥ.

"It is a naturally pleasant way, if the hero, after his excellent qualities have been described, conquers his enemies. But we are, on the other hand, pleased, too, when the poet, after he has set forth the race, heroism, fame, etc. even of the enemy, describes the brilliant qualities of the nāyaka, so that he vanquishes the (so described) enemy."

The partly literal congruence of these two stanzas with Bhāmaha's I, 2237 makes it evident that Dandin blames no one else but Bhāmaha, though in this case, the words of Dandin do not bear the stamp of open aggression38. There Dandin's hidden attack shows again a peculiar form, twisting the meaning of Bhāmaha's words, for, according to theKāvyādarsa we would assume that Bhāmaha had stated that the poet should not make the enemy's (pratināyaka) excellent qualities the subject of a detailed description. But this is not so. For, though in the opinion of Bhāmaha the nāyaka ought to wholly 'fill' the body of the kāvya, the chief point is this that the poet should not describe the hero's fall in order to

be seen. The mahākāvyaṣ etc. belong to the former, the nāṭakas to the latter sort of poetry.

37 Dandin: gunataḥ prāg upanyasya nāyakam, Bhāmaha: nāyakam prāg upanyasya. Dandin: vamśavīryaśrutādīni, Bhāmaha: vamśavīryaśrutādibhiḥ, this last is comprised in Dandin's gunataḥ. Even Dandin's nāyakotkarṣavarnanam refers obviously to Bhāmaha's anyotkarṣābhidhītsayā.

38 Thus Tarkavāgiśa, who did not know the work of Bhāmaha, could not see any polemical tendency in the above mentioned stanzas of Dandin.

148 THE FOUNDATIONS OF INDIAN POETRY

praise thereby the excellent qualities of his enemy. That this is the standpoint of Bhāmaha is proved by the tenor of stanza I, 23.

While Vāmana has not given a definition of mahākāvya, because in his opinion the division of kāvya into kathā, ākhyāyikā, and mahākāvya is of no special interest39, Rudrata presents a very exhaustive description of it. This scholar considers the poetical compositions as twofold, utpādya and anut-pādya as far as the subject of the poem is invented by the poet himself or is taken from the itihāsas40. The mahākāvya is an utpādya composition, and described by the stanzas (XVI, 7–18):

"As to the utpādya composition one may in the mahākāvyas begin with the description of a beautiful town. Then, (the poet) should praise the descent of the hero therein (in that town). He must set forth a hero who follows the three vargas (dharma, artha, kāma), who possesses the three abilities (prabhuśakti, mantraśakti, utsāhaśakti41), and all the excellent qualities, whose subjects are devoted to him42, who

39 Vritti to I, 3, 22: yad uta kathākhyāyike mahākāvyam iti, tallakṣaṇam ca nātiva hṛdayaiṅgamam ity upekṣitam asmābhiḥ, tad anyato grāhyam.

40 Kāvyālaṅkāra XVI, 2, 3. A second division is given in XVI, 2–6: mahat and laghu. These compositions are taken as mahat, where all the four vargas and all rasas are to be found. This division and many others of the later ālaṅkārikas too are of small importance and interest. In their endeavours to discover new ideas scholars could not be lucky in every case.

41 See e. g. Raghuvanśa III, 13; VI, 33; XVII, 63; Śiśupālavadha II, 26.

42 Or : whose ministers are devoted to him.

is desirous of victory. He should describe the whole

kingdom of the hero, who rules in due form, and

his behaviour as a king; if opportunity offers, he

should (describe) a season, autumn, etc., in connexion

(with the hero). Out of the high families he should

paint an adversary of the hero, who is anxious to

promote the dharma both for himself and for his

friends, (an adversary) who is endowed with excellent

qualities. Hearing from his spy, who has been sent

out by him, or from another side of enemy's inten-

tions, (the hero) may cause excitement in the assem-

bly of the kings (so that) their thoughts and speeches

are inflamed by anger. After having consulted

together with his ministers and having resolved that

(this enemy) ought to be punished, then (the poet)

should let the hero unertake a campaign or let him

send a loquacious messenger. Then, when a cam-

paign is undertaken, (the poet) should paint the ex-

citement of the women in the town, (further) the

mountains and rivers of the countries, the woods,

the forests, the lakes, the deserts, the oceans, the

dvlpas and the continents43. (The poet should fur-

ther describe) the encampments of the army and,

according to the circumstances, the sport of young

people therein44, the sunset, the dawn, the darkness,

the rising of the moon, the night, and, therein,

meetings, concerts, drinking, and love. In due time,

the poet may resume (the thread of) his tale. In the

same way he may describe the adversary, as he approa-

43 The other worlds, say Namisādhu.

44 Viz. in the woods, rivers, etc., which are enumerated

in the previous lines.

ches, as he arrives, or as he, forced by circumstances, besieges a town. Then the poet should let the warriors, who fear death, send messages in the night, which they spend wholly in drinking, thinking that they have to die the next day, to their wives. When both (the hero and his enemy), after having prepared for the struggle, fight hard with each other, the poet should finally well describe the triumph of the hero.

If one compares this 'definition' with those of Bhāmaha and Daṇḍin, Rudraṭa's dependance on both becomes evident. However, though it is very long, Rudraṭa's definition can scarcely be called a progressive one. On the contrary, we miss various points in it, which are important marks of the composition called mahākāvya, as the occurence of alamkāras, of rasas, the division into sargas, the change of the metre, while, on the other hand, the author introduces some points which are not inherent or necessary parts of the division. This is, however, sometimes a peculiarity of definition of the Indian theorists, whatever their speciality be, that they draw some points of a more incidental and accessory character into their definitions and do not render the real essential features prominent enough, though they are very seldom entirely left out.

In spite of the similarity of the definition of both Daṇḍin (and Bhāmaha) and Rudraṭa, there is, really, a small divergence, though, perhaps, more theoretically than practically. It has already been mentioned, that, according to Rudraṭa, the mahākāvya belongs to utpādya-poetry, that is, that the poet

THE DIVISION OF POETRY

himself is the inventor of the subject-matter of the kāvyasatīra. But Bhāmaha and Dandin have said that the kāvyas should be based on facts, to which also such stories as are sanctioned by tradition and are told of e.g. in the Rāmāyana and Mahābhārata belong. It is true, however, that these older scholars already thought of a more stereotyped form, when they speak of mantradūtaprayānājināyakābhyudaya.

These are more general outlines allowing the poet a wider scope in the development of the story as a whole. All the other points mentioned by Bhāmaha and, still more in detail by Dandin, are meant to give a poetical charm to the description of the story. The impression we gain from Rudrata's statements is different. He prescribes so to speak, a norm for the development of the story. Thereby the story itself is pushed into the background and the poetical ornament is considered the main point of the mahā-kāvya. The action ceases to be interesting compared with the way it is told in. Thus, the working poet has to concentrate his whole attention on the form, to which he should give his own individual stamp.

The definition of mahākāvya has scarcely altered in course of time. Some five hundered years after Rudrata the Sāhityadarpana45 says as follows:

The author of the Sāhityadarpana divides the whole domain of kāvya into many parts. Though this subdivision is only of smaller interest, it may be given here shortly, because Viśvanātha appears to have borrowed from older works. The kāvya is (I) drśya (to be seen) and (II) śravya (to be heard). The drama belongs to the first kind, and is divided in many ways. The śravya-kāvya is (i) padya (in verse) and (ii) gadya (in prose). The description of padya is introduced by an account

152 THE FOUNDATIONS OF INDIAN POETRY

" An arrangement of sargas is called mahākāvya. The one hero therein is a god or a ksatriya of noble family, firm, exalted, endowed with good qualities46, or there may be several heroes: princes sprung from one 47 race, of a noble origin. Among the of the names given to one detached stanza and collections of stanzas: muktaka (one stanza), yugmaka (two), sandānitaka (three,) kalāpaka (four), and kulaka (five stanzas, which form one whole). Then padya is subdivided: (1) sargabandha = mahākāvya (example: Raghuvaṃśa, Śiśupālavadha, Naiṣadha, "my Rūghavavilāsa"); (2) ākhyāna (when the sargas are posed by a rṣi: (Mahābhārata); (3) āśvāsa (when the sargas are written in Prakrit, the metre being mostly āśvāsa, galitaka: (Setubandha, "my Kuvalayāśvacarita"); (4) kadavala (when the sargas are written in Apabhraṃśa: (Karpūrāśkarama). (5) Kāvya (written in many languages and not divided into sargas: (Bhikṣātana, Āryāvilāsa); (6) khandakāvya (when the conditions of the mahākāvya definition are only partly fulfill-ed : (Meghadūta); (7) kośa (a collection of stanzas which do not depend on each other: (Muktāvatī). As to gadya, the author remarks that there are four kinds of prose: muktaka (without compounds), vṛttandhi (containing metrical parts), utkalikāprāya (long compounds), and cūrṇaka (short compounds). With regard to that one may compare Vāmana's treatment of gadya, (see above p. 127). Gadya is divided into : 1 kathā, 2 ākhyāyikā, 3 campū (in prose and in verse, as the Daśarājaka-rita), 4 biruḍa (praise of a king, in prose and in verse: Biru-damaṇinmālā), 5 kasambhaka (in different languages: "my Praśastiratnāvalī, composed in sixteen languages").

46. Pramadādasa Mitra translates: "characterized by firmness and generosity of heart" but we must separate udātta from guṇa, as is suggested also by the corresponding words of Daṇḍin.

47 Though eka can also be rendered by "excellent", the above given translation seems to be better. Of course, the

rasas, śṛṅgāra (love), vīra (heroism) and śānta (calmness) one should be the chief rasa and all (the other) rasas should be subordinate to it. (The mahākāvya) contains all sandhis of the drama. The story is founded upon the itihāsas or another one, (but of such a kind that) it is connected with good persons. It contains the four vargas (dharma, artha, kāma, mokṣa), and of those the poet should describe one as being the goal48. In the beginning there is a namaskṛiyā, āśis, or vastunirdeśa49. Sometimes it opens with the reproof of bad men etc. and the praise of the excellent qualities of the good men. It consists ot sargas, which are neither too short nor too long, more than eight50 (sargas) everyone ot which is nāyakas must be connected in some way or other in order to save the unity of the action. It would be possible to compose a kāvya the heroes of which are both Rāma and Lakṣmaṇa.

48 Pramadādāsa Mitra gives the following translation: "It has for its fruits (i.e. the final objects obtained by the hero or the like) all the four of the class consisting of the great objects of human desire, viz. Merit, Wealth, Enjoyment, and Liberation, or it has only one of them." Although the mahākāvya deals with all the four vargas, the fruit of only one of them is described.

49 Pramadādāsa Mitra renders vastunirdeśa by "or simply with the mention of a matter (leading into the main story of the poem)", but vastu is nothing more than the subject-matter of any poetical composition (Apte, Dictionary, s. v.] ; vastunirdeśa, then, means contents. The Raghuvamśa begins with a vastunirdcśa.

50 From this interesting remark it appears that at the time of Viśvanātha, Kālidāsa's Kumārasambhava was already

xx

154 THE FOUNDATIONS OF INDIAN POETRY

written in one metre, but there should be a change of the metre at the end (of the sargas). In some cases, however, a sarga composed in various kinds of metre is to be seen. At the end of the sarga there should stand a hint of what will happen in the next sarga. Dawn, sun, moon, night, evening, darkness, day, morning, midday, hunting, mountain, sea-son, wood, joys of love, saparation, muni, heaven, town, sky51, battle, campaign, wedding, counsel, birth of a son, etc., all that should be described together with aṅgas and upāṅgas52, according to circumstances. The mahākāvya should receive its title after the poet, or some other, and the sargas should be named after what is described therein.

It is evident that this definition, which contains scarcely anything new, is dependent on Daṇḍin. Like Daṇḍin, Viśvanātha also begins with the words sarga-bandho mahākāvyam. Ādau namaskriyā vastunirdeśa eva vā is the same as: āśīr namaskriyā vastunirdeśo vāpi tanmukham of the Kāvyādarśa; itihāsodbhavam vṛttam anyad vā sajjanāśrayam repeats only Daṇ-din’s itihāsakatodbhūtam itarad vā sadāśrayam53. Viśvanātha’s catvāras tasya varṇāḥ syus tesv ekam ca

enlarged and had more than eight sargas, for it has never been doubted that the Kumārasambhava must be considered as a mahākāvya.

51 Pramadādāsa Mitra renders adhvara by sacrifice.

52 i. e. everything that is connected with the just mentioned subjects either directly or indirectly.

53 The author, however, interprets the former part of sad-āśraya by "good." This is not the opinion of Daṇḍin.

THE DIVISION OF POETRY

phalam bhavet corresponds to Dandin's caturvagaph-alopetam, and sambhogavipralambhau is only the inversion of vipralambhair vivāhais ca54.

B. Abhineyārtha.

The second kind of literary composition is 'abhi-neyārtha', i. e. to be acted or dramatically represented. It is the nāṭaka. The writers on alamkāra agree that it should not be treated in the books on poetics but in those which deal with the dramatic art. Thus Bhāmaha and Dandin refer the reader to the literature thereon. Bhāmaha says (I, 24):

" As regards the nāṭaka, which is divided into dvipadī, samyā, rāsaka, skandhaka, etc., one says that it should be acted. Other scholars have treated (the nāṭaka) in detail."

And Dandin, (Kāvyādarśa I, 31):

"Mixed (i. e. in prose and in verse) are the nāṭakas etc. A detailed description of them is to be found elsewhere"55.

The same author notes in I, 37 that naṭakas are written in various languages.

54 In the Pratāparudrīya (p. 96) the mahākāvya is described by the following words.

nagarārṇavaśailartucandrārkodayavarnaṇam, udyānasalilakrīḍāmadhu-pā-naratotsavāḥ, vipralambho vivāhaś ca kumārodayavarnaṇam, mantradūtaprayāṇājin-āyakā-bhyudayā api, etāni yatra varṇyante tan mahākāvyaṃ ucyate.

That is nothing else than the old definition, but a good deal more superficial.

55 It is worth noting that both Bhāmaha and Dandin use exactly the same term 'vistara.

156 THE FOUNDATIONS OF INDIAN POETRY

Vāmana (I, 3, 30) only says that dramas (daśari-rūpaka) are the best of all literary compositions56.

C. Ākhyāyikā and Kathā.

According to Bhāmaha, both ākhyāyikā and kāthā are the third and fourth kind of composition. The author has (I, 25–29) :

prakṛtānākulaśravyāśabdārthapadavṛttinā, gadyena yuktodāttā socchvāsākhyāyikā matā. vṛttam ākhyāyate tasyāṃ nāyakena svaceṣṭitam57 vaktram cāparavaktraṃ ca kāvya-58 bhāvyarthasaṃsi ca, kaver abhiprāyakṛtair aiṅkanaih kaiścid aiṅkitā59, kanyāharaṇasaṃgrāmāvipralambhodayānvitā. na vaktrāparavaktrābhyāṃ yuktā nocchvāsavatī api, saṃskṛtasaṃskṛtā60 ceṣṭā kathāpabhraṃśabhāk tathā. anyaiñ svacaritāni tasyāṃ nāyakena tu nocyate, svaguṇāviṣkṛtiṃ kuryād abhijñātaḥ kathaṃ janāḥ.

"That kind of literary composition is called (ākhyā-yikā), which is composed in prose, the words of which

56 Daśarūpaka is, in the opinion of Vāmana, the wider, the first group of which is the nāṭakas.

57 Quoted by Tarkavāgiśa in his commentary to Kāvyā-daśa I, 24; the second line (with the v.l. yaspiṃ for tasyāṃ) also by Vidyāsāgara commenting on the same stanza of Daṇḍin's.

58 Thus we may read instead of kāle. Śaṅkara, commenting on the tenth introductory stanza of the Haracarita, quotes the verse in this way: etasmin (viz. in the metre vaktra) ākhyāyikādbhir bhāvivasṭusaṃsūcanāya vṛg vira-cyate, tathā ca Bhāmahaḥ, vaktraṃ cāparavaktraṃ ca kāvye kūvyārthasaṃsini iti. From this sentence the correct reading must be concluded.

59. Trivedī: 'kṛtaiḥ kathā (tha ?) naiḥ kaiścid aiṅkitā. Tarkavāgiśa (commentary to Kāvyādarśa I, 30) quotes the line as follows : kaver abhiprāvakṛtair aiṅkanair aiṅkitā kathā.

60 Trivedī. has saṃskṛtāṃ saṃskṛtā.

(as the bearers of ideas) express the matter intended,

the meaning of which is not confused, and the words

of which (as combinations of sounds) are pleasant

to the ear, and which is divided into ucchvāsas61.

In the (ākhyāyikā) the hero himself narrates

his deeds. (Stanzas written in the metre) vaktra

and aparavaktra indicate what will happen next

in the (course of the) kāvya.

The kathā shows characteristics which spring from

the poet's imagination. It contains the description

of the seizing of a girl, (the description) of a struggle,

of the separation and the (final) triumph of the hero.62

It contains no (stanzas in the metre called) vaktra

and aparavaktra, nor is it divided into ucchvāsas.

It is written in Sanskrit, and also partly in Apa-

bramśa.

In the kathā other (persons) recount the deeds

of the hero. For how should a well-educated man

set forth his own excellent qualities !"

61 As to the meaning of the first line, one can be of a

different opinion. I bring the adjectives prakrta, anākula,

and śravya together with śabda, artha, and pada respectively

(yathāsamkhyam); then vrtti must be connected with pada

as well as with śabda and artha. Prakrta in the sense of

"belonging to the matter in question" is used very often

by later writers on poetics, mostly however, in connection .

with artha (prastuta has the same meaning). Prakrta

does not occur elsewhere in Bhāmaha's work. I confess,

however, that I am not wholly satisfied with the translation

given above. Manuscript T gives prasrta for prakrta which

does not render the meaning clearer.

62 Tarkavāgīśa (commentary to Kāvyādarśa I, 29)

connects line 29d wrongly with the ākhyāyikā.

It is not possible to gain a thoroughly clear idea of ākhyāyikā and kathā from these stanzas, but what we may gather from this short account, is, probably, the following : Both ākhyāyikā and kathā are written in prose (gadyā). As to their tenor, these two kinds of compositions differ from each other on this point that the subject-matter of the ākhyāyikā is what is experienced by the hero himself. This being the case, there is no restraint concerning the matter of the composition. The tenor of the kathā, on the other hand, is much more stereotyped in its outlines. The theme, or rather the disposition, is given : the seizing of a girl, a struggle caused by it, the separation of the two lovers, and finally and as a matter of course, the triumph of the hero, i.e. the happy union of the hero with his beloved wife. From this it is obvious that the poet's imagination takes a prominent part as regards the formation of the plot within the prescribed main points of the whole ; and this is apparently meant by the author's words : kaver abhiprāyakrtain an̉kanaiḥ kaiścid an̉kitā. In the ākhyāyikā we have only to do with events which have really happened. Another remarkable distinction is the fact that in the ākhyāyikā the hero himself is the narrator or, if we dare say so, the reporter of his own deeds and experiences, while in the kathā the speaker is another person. Further, the ākhyāyikā is divided into several chapters which are called ucchvāsas. These ucchvāsas open with some verses in the metre named vaktra and aparavaktra, perhaps a couple of stanzas83, the former in vaktra, the latter

63 I may be allowed to assume this from the terms

THE DIVISION OF POETRY

in aparavaktra, which in some way or other point to what will be the tenor of the ucchvāsa in question. The kathā is not divided into chapters (ucchvāsas), nor are there stanzas in vaktra and aparavaktra. As regards the language there is an interesting remark of Bhāmaha to the effect that the kathā is written in Sanskrit and, besides that, in apabhramśa. Taking into consideration what has been gathered from Bhāmaha I, 16, he does not mean a single language here by the term apabhramśa but the Prakrit idioms in general. It is possible, though it cannot be proved, that Bhāmaha's prototype of kathā was the famous Brhatkathā, which according to Dandin was composed in the Paisācī language.

Bhāmaha's treatment of ākhyāyikā and kathā as two different kinds of gadya poetry is sharply criticized by Dandin, who says (Kāvyādarśa I, 23, 24) :

iti tasya prabhedau dvau tayor ākhyāyikā kila, nāyakenāiva vācyā 'yā nāyakenetarena vā, svaguṇāviṣkriyā doṣo nātra bhūtārthaśaṃsinah.

"Gadya is a series of words without metre. Ākhyāyikā and Kathā are considered as two different kinds of it (gadya). Of these two the ākhyāyikā is to be narrated only by the hero alone, the other (kathā) by the hero or another person. Therein the setting forth of one's own excellent qualities cannot be taken as a fault of the nāyaka, if he recounts real things".

Dandin first deals with the interesting question regarding the person that by means of the poet's words narrates the story. The opinion of Bhāmaha, that vaktra and aparavaktra, and from the words: . . . vaktrāp-ravaktṛe papātha (Bāṇa's Harsacarita p. 115).

160

THE FOUNDATIONS OF INDIAN POETRY

in the kathā the narrator is not the hero, is rejected by Dandin, who states nāyakena itarena vā. In the same way Dandin refuses to accept what Bhāmaha says in order to justify his idea that in the kathā another person than the hero should be the narrator: it is no fault when a person describes his own excellent qualities, supposing that he says the truth. There is one point, however, in which Dandin agrees with Bhāmaha, viz. that in the ākhyāyikā the hero himself is the speaker, for if this were not the case, Dandin must have mentioned it.

Jn the following three stanzas (I, 25-27) the author goes still farther in refusing to accept Bhāmaha's theories :

api tv aniyamo dṛśyas tatrāpy anyair udīraṇāt, anyo vaktā svayaṃ veti kīdṛg vā bhedalakṣaṇam. vaktraṇ cāparavaktran vā socchvāsatvañ ca bhedakam, cihnam ākhyāyikāś cet prasaṅgeṣu kathāsv api. āryādivat praveśaḥ kiṃ na vaktrāparavaktrayoḥ. bhedaś ca dṛṣṭo lambhadir ucchvāso vāstu kiṃ tataḥ.

"But farther, because also in the ākhyāyikā other persons (and not exclusively the hero himself) are speaking, no restriction can be seen regarding this point also. How can that be a distinctive mark whether another or himself is the speaker ?

If the employment of the metre called vaktra and aparavaktra or the division into ucchvāsas should be considered as distinctive attributes, I may reply that even in the kathā, the āryā etc. occur occasionally, why should not vaktra and aparavaktra occur also ? And the chapters may in this case be called lambha etc. in the second ucchvāsa, but what can be concluded from that ?"

As Dandin gave a false interpretation of the words of Bhāmaha in the case of the sargabandha, so also here Dandin intentionally misinterprets the passage in question. Bhāmaha does not deny that there are occasionally other speakers than the hero himself in the ākhyāyikā, but they have to be, of course, introduced by the nāyaka.

With respect to vaktra and aparavaktra as well as to the names of the chapters (ucchvāsa), Dandin seems to be more correct. For, what have these quite subordinate things to do with the characteristic features of ākhyāyikā and kathā? An examination of the question, however, shows that Dandin also in this case takes another standpoint than that which is the correct one for Bhāmaha's treatment of the matter. Bhāmaha's words refer not to some stanzas in vaktra and aparavaktra, which occasionally occur in the story, but to those stanzas which form the headings of the chapters (ucchvāsa). The name of the metre (vaktra and aparavaktra) is by no means of the same importance, nor is the name of the chapters ucchvāsa; but there we meet with questions which will be examined later on.

Thus the author of the Kāvyādarśa insists that neither the person of the narrator, nor the kind of metre, nor the heading of the chapter is are essential peculiarity of ākhyāyikā on the one hand, and of kathā on the other. And so he says (I, 28a):

"Thus kathā and ākhyāyikā are only one sort of literary composition called by different names".

And, as Dandin's commentator Tarkavāgiśa re- xxi

162 THE FOUNDATIONS OF INDIAN POETRY

marks, "with an overbearing manner", the author looks down rather contemptuously on the master and says (I. 28 b):

atraivāntarbhaviṣyanti śeṣāś cākhyānajātayaḥ.

"And there will be contained also the other kinds of stories" 64.

After having shown that the "peculiarities" of Bhāmaha's ākhyāyikā have no distinctive character in reality, Dandin goes on to prove that also what Bhāmaha teaches with respect to the kuthā ought not to be accepted. According to Bhāmaha the description of the seizing of a girl etc. as well as some other "tokens" sprung from the author's imagination are characteristic features of the kathā. Dandin replies (I, 29, 30):

kanyāharanasamgrāmavipralambhodayādayah,

sargabandhasamā eva naitc vaiśeṣikā guṇāḥ.

kavibhāvakṛtam cihnam anyatrāpi na duṣyati,

mukham iṣṭārthasamsiddhau kiṃ hi na syāt kṛtāmanām.

"The description of the seizing of a girl, of a struggle, of the hero's triumph, etc, met with in sargabandhas, too thus these are no distinctive qualities.

A special mark sprung from the poet's imagination is no fault also elsewhere. For what should

64 Tarkavāg I ś a: atra atraivāntarbhāviṣyantiti bhā-viprayogāt prauḍhivādenābhedapratipādanam granthakṛto na tu vastutah prāmāṇikatamaír munibhir api tattadbhedābh-yupagamāt, yathā, āgneyé,—

"ākhyāyikā kathā khaṇḍakathā parikathā tathā kathāllketi manyante gadyakāvyam ca pañcadhā" iti, evaṃ kathākhyāyikayor apy abhinnatvapratipādanam etan-mūlakam cveti mantavyam.

not be a means for the learned ones respecting the complete attainment of the wished-for object"?

Finally Dandin rejects Bhāmaha's view concerning the languages (I, 38):

kathāpi sarvabhāsābhih sanskrtena ca badhyate, bhūtabhāsāmayīm prāhur adbhutārthām brhatkathām.

The kathā, too, is composed in all languages and in Sanskrit. The Brhatkathā, which contains wonderful things, is said to be written in the language of the ghosts"65.

On a previous occasion it has been pointed out that the word apabhramsa is employed by Dandin in a narrower sense than by Bhāmaha. This being the case Bhāmaha's words get, of course, another meaning ; and this modified meaning of Bhāmaha's is rejected by Dandin, when he uses the word sarvabhāsābhih.

The words sanskrtena ca of Dandin can only be fully understood when they are compared with those of Bhāmaha. There is a shadow of irony in them ; and Dandin rejects Bhāmaha's theory concerning this point giving an example from literature by referring the reader to the Brhatkathā. Unfortunately we know only little about this work, which must have been well-known in the period of the Kāvyādarsa.

That this story was composed in the Paisācī language is also stated by other authors66.

Vāmana, though mentioning the two kinds of gadya ākhyāyikā and kathā, gives no further definition.

.. 65 . Paisācī,

66 See Hall in the introductory pages of the edition of the Vāsavadattā (Calcutta 1859, Bibliotheca Indica).

164 THE FOUNDATIONS OF INDIAN POETRY

tion of them, because this division of kāvya is in his eyes of no interest.

In spite of Dandin's bitter attacks the younger writers on alankāra were not willing to give up the theory of the old masters that two sorts of prose must be assumed. In fact the power of tradition was too strong to admit of alterations. Rudrata deals very exhaustively with the matter in question and gives the following account (XVI, 20-23) :

"The poet, after having expressed his reverence for the desired gods and gurus in verse in the great kathā, should (also in verse) shortly67 describe his own family and his authorship68. Then he should arrange in prose, which is endowed with anuprāsas and which has light syllables69, the body of the kathā as before (in the utpādyakāvyas), viz. descriptions of towns, etc. In the kathā he should give in the beginning either another story, well developed, and then, in an easy manner the connecting link in order to come to the main tale, or he should (without giving another story before) arrange the kathā in Sanskrit and in another (language), but, in the latter case, not in prose, (the kathā) the end of which is the winning of a girl, wherein the whole śṛṅgāra-rasa (the sentiment of love) is rightly developed".

The account concerning the ākhyāyikā is still more detailed (Kāvyālankāra XVI, 24-30) :

"The poet after having, as before (v. 20), worshipped

67 Which is done at length in the ākhyāyikā.

68 Namisādhu remarks: tena sujanakhalastutinindādi-kam cābhidadhyād iti sūcyate.

69 Syllables with not too many consonants.

THE DIVISION OF POETRY

the gods and gurus, should praise the (former) poets in the ākhyāyikā, thinking that he is not able to compose a kāvya, after those (great poets) have existed. Then (as being the reason, why, notwithstanding, he proceeds to compose a kāvya) he may describe his devotion to the princes, or his fruitless effort to praise the excellent qualities of another, or he may give another, but not a trivial, reason, why he is composing the ākhyāyikā. Thereafter he should arrange the ākhyāyikā in prose as well as the kathā. And the poet may describe his own descent, but not in verse. As (the mahākāvyas) are divided into sargas, so he should divide (the ākhyāyikā) into ucchvāsas. In the opening of every (ucchvāsa), except the first70, he may for the sake of the (main subject) mention two (verses in the) āryā-metre, which are connected with each other and the tenor of which is a general idea.

Should there be the occasion to utter a doubt concerning a present or a past object, which is hidden, or concerning a future object, though it is manifest, the poet, in order to dispel the doubt, should in the presence of the doubting person let some body quote one or two of the alamkāras anyokti, samāsokti, or śleṣa. In this case he should use one of the metres āryā, aparavaktra, or puṣpitāgrā, or according to circumstances any other, mostly mālinī."

In spite of the very detailed description the definitions of both kathā and ākhyāyikā contain scarcely any new idea. As to the introductory stanzas, however, Rudraṭa gives some interesting remarks. Bhā-

70 Where the just mentioned rules are to be observed.

maha and Dandin mentioned nothing on this matter, so that it is doubtful whether for the ākhyāyikā and kathā of that older period the quoting of such introductory stanzas was prescribed or not.

On the other hand a very important question is not touched on by Rudraṭa, viz. whether the hero himself or another person is the narrator. If this point had been of interest still in the time of Rudraṭa, he would have mentioned it, Thus we cannot but assume that the question concerning the narrator was no longer considered as a characteristic feature of one of the gadyā-compositions. With this we have, however, arrived at a point regarding which more will be said later on.

Other important writers on alamkāra consider kathā and ākhyāyikā as two different kinds of prose-poetry. The author of the Dhvanyālokalocana mentions them occasionally, without giving, however, any further description. Dealing with the arrangement of words (saṅghatanā) Ānandavardhana states that the various sorts of saṅghatanā depend on the length of compounds. There are (the author says on p. 134) three kinds of saṅghatanā, whether no compounds, short compounds, or long compounds are found. Which of these three sorts should be used is determined by the sort of composition (p. 141). Under these literary compositions saṅgabandha, abhineyārtha, ākhyāyikā, and kathā are mentioned. These are the very same groups and in the very same arrangement as are found in the work of Bhaṭṭa-maha. The rules regarding the prose compositions are of course different from those of the verse compo-

sitions. As to the prose all depends on the condition of the compounds. Thus the ākhyāyikā is characterised by short and long compounds. Though with respect to the kathā it is similar in general, it should be borne in mind that in this sort of composition the rasas are described (sūtra III, 8). Under these circumstances, those rules are valid for the suitableness (aucitya) of compounds here which are taught with respect to compositions dealing with the rasas. So, in the case of śṛṅgāra-rasa long compounds should be employed (p. 135), in the case of raudra-rasa compounds are wanting. Some-times sentiments as love, sadness, anger, heroism occur also in the ākhyāyikā. Then, as a matter of course, the poet has to work according to the same rules, the arrangement of words depending always on the difference of rasas.

Viśvanātha gives the following account (Sāhityadarpaṇa VI, 332-336a) : kathāyāṃ sarasam vastu gadyair1 eva virnirmitam. kvacid atra bhaved āryā kvacid vaktrapavaktrake, ādau padyair namaskāraḥ khalāder vṛttakīrtanām. Yathā kādambaryādiḥ.

71 We must read thus with the Nirṇaya Sāgara Press edition instead of padyair, as the reading is in the edition of the Bibliotheca Indica. The stanza is quoted with the correct reading by Krishnamachariar in the introduction to Subandhu's Vāsavadattā (Srirangam 1906). Pramadā-dāsa Mitra translates as follows: "In the kathā (tale) which is one of the species of poetical composition in prose, a poetical matter is represented in verse, and, sometimes, the Āryā and, sometime, the Vaktra and Aparavaktra are the metres employed in.it." This gives a wrong idea of

168

THE FOUNDATIONS OF INDIAN POETRY

Ākhyāyikā kathāvat syāt kaver vaiśānukirtanam, asyām anyakvināṃ ca vrttam padyam kvacit kvacit. kathāṃśānāṃ vyavaccheda āśvāsa iti badhyate, āryāvaktrāpavaktrāṇāṃ chandasā yena kenacit. anyāpadeśenāśvāsamukhe bhāvyarthasūcanam.

Yathā harṣacarita'uddhṛtau.

'api tv anīyamo dṛśyas tatrāpy anyair udīraṇāt' iti Daṇḍ-yācāryavacanāt kecit 'ākhyāyikā nāyakenaiva nibaddhavyā' ity āhuḥ, tad ayuktam, ākhyānādayaś ca kathākhyāyikayor evāntarbhāvān na prthag uktāḥ, yad uktam Daṇḍināiva: atrai-vāntarbhaviṣyanti śeṣāś cākhyāyajātayaḥ'. eṣām udāhara-ṇaṃ pauñcatantraūdi.

kathā, as this translation is based on the incorrect reading padyair. Peterson (edition of Kādambarī, Intruduction-Part II, p. 69) rejects the translation of Pramādāsa Mitra and thinks that the meaning is: "When... we turn back to the description of kathā, we find that species of composition distinguished as a narration in prose, with here and there a stray verse or two, of matter already existing in a metrical form." The last words are the rendering of padyair eva viniṛmitam. Even supposing that the reading padyair is correct I think it impossible to translate as Peterson does. As a matter of course, all the hypotheses Peterson founds on this rendering of his, need not be discussed. The same is the case with respect to what Miss Ridding says in the translation of the Kādam-barī, p. XII.

72 Though in Bāṇa's Harṣacarita the chapters are not called āśvāsa, but ucchvāsa, I cannot alter āśvāsa to ucchvāsa. In an older dictionary, Uttaratantra by name, from which some passages are quoted by Śivarāma in his commenary to Subandhu's Vāsavadattā, we find the following remark: ākhyāyikāpariccheda ūśvāsocchvūsakūv api. These words are quoted by L. H. Gray, Journal of the American Oriental Society, Vol. 24, first half, 1903, p. 60. But L. H. Gray separates incorrectly ākhyāyikā pariccheda.

THE DIVISION OF POETRY

169

"In the kathā a rasa-containing73 subject is described in prose. Sometimes a stanza in the kathā may appear in the āryā-metre, sometimes stanzas in vaktra and aparavaktra. In the beginning there should stand in verse a namaskriyā and a description of the behaviour of bad people and so on.

Example: Kādambarī etc.

The ākhyāyikā is similar to the kathā. An account of the poet's race, and, sometimes, a description of the deeds of their poets is presented therein. The division of the separate parts of the story is made by chapters, the names of which are āśvāsa. The coming subject is indicated in any of the metres āryā, vaktra, or aparavaktra in the beginning, but in such a way that the poet takes another matter as pretext.

Example: Harṣacarita etc.

It is not right what some scholars say viz. that the ākhyāyika should be narrated by the hero. For Dandin has taught: 'But farther because also in the ākhyāyikā other persons (and not exclusively the hero himself) are speaking, no restriction can be seen regarding also this point' (Kāvyādarśa I, 25). Because the other kinds of prose compositions ākhyāna etc. are contained in kathā and ākhyāyikā, these are not described separately. For Dandin has said: And there will be contained also the 'other kinds of stories.

Example: Pañcatantra etc."

73 Pramadādāsa Mitra translates sarasam vastu by "a poetical matter", but the well-known "sentiments" śṛṅgāra etc. are meant:

xxii

170 THE FOUNDATIONS OF INDIAN POETRY

In these last words Viśvanātha states his agreement with Dandin, but, in spite of this fact, he does not revolve upon identifying ākhyāyikā and kathā. He goes only so far that he touches the theory of the narrator 74 in the definition no longer.

As a whole, Viśvanātha's definitions are not satisfactory. Nothing is said concerning the subject-matter, a very important point in the view of the older masters. What we may gather from this is that the real and old distinctions of both ākhyāyikā and kathā were forgotten very long ago.

After we have seen in which way the two main kinds of prose have been defined by the writers on alamkāra, we shall make an attempt whether it is possible or not to obtain an idea of the real state of kathā and ākhyāyikā, what both have in common and in which points they differ from each other.

Unfortunately we have no older authorities than Bhāmaha, for from Bhāmaha's treatment of the subject it appears that there had been discussions on the matter in question long before his time.

74 The author of the Pratāparudrīya does not know kathā. After having defined mahākārya he mentions that there are gadya and padya-kāvyas. An example for the former is the Kādambarī, for the latter the Raghuvaṃśa. Then he continues :

asargabandham abi yad upakāravam udīryate, asargabandharūpam—sūryaśatākādi.

gadyapadyamayam kāvyam campūr ity abhidhīyate, vaktram aparavaktram ca socchvāsatvaṃ ca bhedakam,

varnyate yatra kāvyajñair asāv ākhyāyikā matā.

The Harṣacarita is considered as an example. It is evident that the author is influenced by Dandin.

THE DIVISION OF POETRY

need not be proved that prose works must have existed before the period of Bhāmaha and Dandin, which differed from each other on very important points, so that scholars were in a way right in calling these compositions by different names. The definitions given by the Indian scholars are, however, very often not of such a kind as to show what the real condition of things might have been, putting aside the chief characteristic features for points of less importance.

The ākhyāyikā seems to have been the older kind, out of which the kathā was developed as a peculiar form. In the opinion of the older masters four points are of particular importance; they are :

1 The subject matter;

2 The narrator;

3 The division into chapters called ucchvāsa, and

4 Stanzas in the metre called vaktra and aparavaktra, which open every ucchvāsa (but not the first) and indicate its content.

We will first consider the last two points. Apparently it is not right to see the essential part of the definition in the name of the chapter (ucch-vāsa) and in the kind of the metre (vaktra and aparavaktra). But this is essential that the ākhyā-yikā was divided into chapters (they were usually called ucchvāsas), and that there were stanzas at the beginning of every chapter, with the exception of the first (it was, however, usual to compose these in the metre called vaktra and aparavaktra), the purpose

pose of which stanzas was to give in one way or other a certain idea of what would happen in the uohvāsa in question. The poet might be allowed, however, to quote some other stanzas; but these are of no importance for the disposition of the whole work, and have by no means anything to do with the characteristic features of the ākhyāyikā. In this incorrect way, however, the words of Bhāmaha are interpreted by Dandin, who, as we have seen and shall see on other occasions, is always endeavouring to find fault with Bhāmaha.

In dealing with the introductory stanzas something must be said concerning the metres called vaktra and aparavaktra. Even their names show that both are to be closely connected with each other, though we are not able to find out their origin. Vaktra is probably the simple śloka. This we gather from the rules Pingala (v. 9) presents concerning this metre. The term śloka itself is not used by Pingala. In the older language the word śloka meant not a certain metre but was the term for metre generally. Then vaktra became later on the designation of a certain form of epic śloka. The vaktra verse we find e.g. in Bāna's Harṣacarita (Nirnaya Sāgara Press, p. 125) differs from the usual śloka only in so far that the penultimate syllable of the second and of the fourth pāda are long. Except this difference the form of the second and fourth pāda is pathyā. Thus, the scheme of the vaktra is as follows :

~_~~_~/~~_~~_~

THE DIVISION OF POETRY

I am inclind to assume that this metre rather is a younger invention, made when the term śloka was already employed instead of the old name. Now, the term vaktra is to be found in the old books on alamkāra, when ākhyāyikā and kathā are spoken of. To do justice to these authoritative doctrines, the new metre vaktra was invented. The above mentioned stanza of Bāṇa's is combined with a second one in the metre called aparavaktra; both verses are introduced by the author's works : vaktrāparavaktre papātha, from which we conclude that both vaktra and aparavaktra were considered by Bāṇa as being connected in some way. In older times the form of aparavaktra was, however, not the same as later on, when in the Chandraḷśāstra the scheme was given as ~~~—~—~~~ — —.

As to the narrator of the story Bhāmaha had said that in the ākhyāyikā the hero himself describes what he has done (vrttam ...... svaceṣṭitam). Regarding this point there seems to be a difficulty. For the fact that in the kathā another person, but not the nāyaka, is the narrator, is explained by Bhāmaha, who says, that a well educated man is never inclined to sing his own praises. Is this not, however, the very same case as regards the ākhyāyikā ? Though here the hero himself recounts his deeds he is not blamed for doing so. But the case is far from being the very same. For, in the ākhyāyikā the hero relates real facts, which he has experienced himself, his own personal adventures (vrtta). By doing so he gives, so to speak, an autobiography. Theoreti-

cally, he may describe not only his good but also his less good deeds and qualities. Thus we have not only to do with gunāviṣkriyā here.

Matters are quite different in the kathā. The subject-matter of the kathā is invented, as we shall see afterwards. On account of this peculiarity the hero appears in quite a different light. As a matter of course he was presented as being endowed with as many excellent qualities as possible. When he had to be the speaker himself, then he could only praise his own good qualities. This was thought a fault (doṣa), because we have not to do with facts (vṛtta) here. It is in agreement with Dandin's statement, that the praise of one's own good qualities is no fault, if the related story contains the truth.

The ākhyāyikā, then, is, as it were, an autobiography. Considering the matter from this point of view, and bearing in mind that the kāvya in all its forms was intended only to be heard, the first of the mentioned points, viz. that concerning the division into chapters called ucchvāsas, appears in a new light. There is a connection between the ucchvāsas called chapters and the narrator, the hero. The term ucchvāsa means verbatim breathing out. Referring to our case, it appears to be quite natural that the whole story can scarcely have been told by the narrator in one breath. Pauses were necessary, and so the whole story was divided into parts of such a length that the narrating person, the hero, could relate his story in an easy way. The name ucchvāsa, which meant originally pause, was later on transferred to the part from one pause to another, and

became finally the term for chapter. Every uchvāsa, set forth in a poetical way, was with regard to its tenor, in some way or other complete in itself, and introduced by that couple of stanzas which has been spoken of above.

In the ākhyāyikā the hero himself narrates his own deeds. Bhāmaha's words regarding this are quite clear: vṛttam ākhyāyate tasyām nāyakena svaceṣtitam. Every word is of importance here: the subject-matter is not invented, but is based on real facts, it is recounted in a series of uchvāsas; ākhyāyate is nothing more than the simple act of narrating, recounting, informing, as well as its derivations as ākhyāna etc. Because the hero recounts what he has experienced, this sort of prose composition is called ākhyāyikā.

The kathā is of quite a different nature. The subject-matter of the kathā is a more or less invented story. Kathā is the novel. The peculiarity of this kind of literary composition favoured the rising of more regular forms. In India this development was the easier as there always has been an inclination to create fixed schemes for all domains of human understanding. Thus, with respect to the kathā the case became similar to that of the mahākāvya (saṅrgabandha): The tenor losing its individuality was placed into the background, while the poetic embellishment took the first place. What was to be described in the kathā? No longer svaceṣṭita, which was shown to the hearer in all its natural variety and in all its individual features, but the seizing of girl, a struggle, separation, and finally,

as a matter of course, the triumph of the hero. As in the nātaka, the happy ending of the plot is remarkable and characteristic for all compositions. Free play was given to the poet's fancy. Bhāmaha mentions some "signs or tokens," aṅkana. It is, however, not easy to say what those aṅkanas were. Perhaps the poet only wished to say, that the kathā is characterized by descriptions which spring from the kavi's imagination. Daṇḍin replies (and some later writers have borrowed these objections) that such description as kanyāharana etc. are seen also in the mahā-kāvya, so that they ought not to be considered as peculiarities of the kathā. This is not right in this form, as it is again a false interpretation of Bhāmaha's words. It may be that in the sayar-bandha, too, similar descriptions are met with, but they take another rank therein. For, while those descriptions as we have seen are of a subordinate character in the mahākāvya, the main content of which is not a love-story but rather a poetical account of the deeds of a god or a king, in the kathā they are the very basis of the composition. Besides that, the mode of expression is so widely different in these two kinds of composition, that this fact alone would justify calling both by different names.

The kathā is, if we may be allowed to say so, a novel. If this is right, then śṛṅgāra-rasa, the sentiment of love, is above all developed therein. There are many intimations iddeed that this is the case. As we have seen, the Dhvanyāloka says a few words regarding the sorts of poetic compositions. As to the kuthā, the author says, that, on the one hand,

it shows the same style as the ākhyāyikā, but that, on the other hand, attention must be drawn to a distinctive feature peculiar to that sort of poetry ; and this peculiarity consists in the fact that in the composition in question the description or rather development of the śṛṅgāra-rasa takes place. Rudraṭa characterizes the kathā, still more distinctly, pointing out that this composition winds up with the winning of a girl, and that on account of this fact the rasa of śṛṅgāra becomes fully developed ; and therefore we read in the Sāhityadurpaṇa : kathāyām sarasam vast u gadyair eva virimitam.

The character of the kathā being as described, we are not surprised to find that its linguistic form differs from that of the ākhyāyikā. As the hero is not himself the narrator, and as we have a continuous narration here, the conditions of which are different from those of the ākhyāyikā, it can be understood that a division into chapters (called ucchvāsa) is wanting, by which fact the appearance of the couple of vaktra and apara-vaktra stanzas is excluded ; but, on the other hand, there some stanzas may occur in the course of the story in the kathā as well as in the ākhyāyikā. The character of the whole is not changed thereby. It has been pointed out before that Daṇḍin's remark regarding this point is based on an intentionally false interpretation of Bhāmaha's words. The same must be said concerning the matter in question here. Daṇḍin states (I, 24) that the kathā in contrast to the ākhyāyikā is to be narrated (nāyakenetarena vā) by the nāyaka or another person. If we had not the book of Bhāmaha, we could not

xxiii

but assume that Dandin states the opinion of his predecessor correctly, This is not the case. Through the supplementary itarena vā an important point of distinction is almost effaced. Dandin intends to make the question of the narrator quite subordinate, and to point out that ākhyāyikā and kathā have no essential differences.

These have probably been the characteristic features of and also the connection between both ākhyā-yikā and kathā during the period of Bhāmaha and Dandin. But, as other subjects out of the wide domain of alamkāra have undergone many changes in course of time, so, in the present case, the old views could not remain unaltered for ever. While the kathā was less touched by such changes, (probably because it was younger), the ākhyāyikā had lost its original character already in the time of the Kāvyā-darśa.

Among the literary compositions there may have been many types of ākhyāyikā. That, however, is certain : Had Bāṇa's Harṣacarita and Kādambarī, which two compositions are called ākhyāyikā and kathā by the poet himself, been written before Dandin's time, then Dandin's treatment of these sorts of composition would have been totally different from the one we now find in the Kāvyādarśa.

The oldest example of ākhyāyikā we know of is the Harṣacarita by Bāṇa, who lived about the year 620 A. D. We shall analyse this poem from the standpoint of the śāstra, in order to get an idea what the form of ākhyāyikā was in that period.

The composition opens with twenty-one introductory stanzas, the metre of which is the śloka.

THE DIVISION OF POETRY

beginning is a namaskriyā, first to Siva and Pārvatī (1, 2), secondly to Vyāsa (3). Then follows a long discourse about poet and poetry in general. We are told of bad kāvys and plagiaries, of the difference of style, and of the difficulty for the good poet to combine all necessary qualities (4-8). In the highly interesting and important stanzas 10-18 the famous poets and poems are mentioned, eg. the poets Hari-candra, Sātavāhana, Pravarasena, Bhāsa, Kālidāsa, and the poems Vāsavadattā, Bhatkathā, and, finally, Ādhyarāja's Utsāha. In spite of these "stars" Bāṇa has decided to write an ākhyāyikā in honour of the king Harṣa (19). Stanza 20 contains a praise of the ākhyāyikā, and with a jagatī verse this metrical part closes.

We will stop here for a moment in order to compare what Rudraṭa has said regarding the ākhyā-yikā with these introductory stanzas. According to Rudraṭa the introduction is in verse. It should contain first a namaskriyā to the gods and to the gurus or the guru. Secondly the introduction contains the praise of the former poets and the confession of one's own inability, and, finally, the expression of the poet's devotion to the king, which devotion is the cause (as we may add to complete Rudraṭa's idea) that the poet notwithstanding his want of ability is willing to write the poem. All these peculiarities are found strictly observed in the introductory verses of Bāṇa's. What else can be concluded from this strict agreement of theory and practice than that the prototype for Rudraṭa's ākhyāyikā was just the Harṣacarita?

Bhāmaha and Dandin say nothing concerning these opening stanzas, so that it is doubtful whether according to the older opinion the ākhyāyikā was introduced by a metrical passage or not. Probably it was; for as according to Bhāmaha at the beginning of the first ucchvāsa there were not these two stanzas spoken of above, it must have been something else that was placed before the first chapter.

We return to the Harṣacarita. After the described metrical part comes the prose story. As the poem has not been brought to an end by the author we possess only eight ucchvāsas. That the chapters were called ucchvāsas by Bāṇa himself is proved by what he says in the tenth introductory stanza : ucchvāsānte 'py akhinnās te yeṣāṁ vaktre sarsvatī, katham ākhyāyikākāraṁ na te vandyāḥ kavī-svarāḥ. Here three terms : ucchvāsa, vaktra, and aparavaktra are alluded to excepting the first ucchvāsa, which begins with a 'scholarly' tract, every ucchvāsa opens with a pair of stanzas. The reader will very easily find out that these stanzas contain an indication of what is to happen in the chapter in question.

As to the metre we have the following scheme : second ucchvāsa : two āryās, third ucchvāsa : one śloka and one āryā, fourth ucchvāsa : the same ; fifth, sixth, seventh, and eighth ucchvāsa : two āryās each.

The prose part gives first a very detailed account of the poet's family which subject extends to the third ucchvāsa : the poet speaks of his ancestors and his youth (ucchvāsa 1), his introduction to the court by king Harṣa's stepbrother, Though Bāṇa was

received here rather coolly he soon grew the prince's favourite (2). After the poet had returned to his own country he narrates to his relatives the history of King Harṣa. In the same (3rd) ucchvāsa this chief story begins and is treated up to the end of the whole tale, where it breaks off.

Sometimes there are stanzas in the ucchvāsas which are, however, of no great importance to the construction and the development of the story. In the first ucchvāsa we meet a stanza called aparavaktra by the author himself (p. 18), the second ucchvāsa contains three stanzas in vasantatilaka (p. 54), śārdūlavikrīdita (p. 69), and aparavaktra (p. 78) ; the third two pairs of stanzas : āryā (p. 86), sragdharā (p. 93) ; the fourth a couple of verses in vaktra and aparavaktra (p. 125 : “vaktrāparavaktre papāṭha") and besides that, one verse in āryā (p. 140) ; the fifth a stanza in the śloka-metre (p. 153) and another in aparavaktra (p. 159), the sixth only one stanza in āryā ; the two last ucchvāsa contain no verses at all, which is somewhat remarkable. Thus, the usual metres are the vaktra, aparavaktra, and āryā, which may perhaps be accounted for by the fact that in the older authoritative books on alaṃkāra these metres were spoken of chiefly.

Comparing these points also with what Rudraṭa says, a full agreement appears again. According to this scholar the poet should describe his own race in prose. Can this be considered as a typical feature of every ākhyāyikā ? As that is not very probable, Rudraṭa has apparently generalised the case of the

Harsacarita. Still something else may be gathered from this peculiar treatment of the matter by Rudraṭa, viz., that in the time of Rudraṭa the old type of ākhyāyikā was but a matter of past ages and was only spoken of in a traditional way out of respect to the great masters. Further : at the head of every ucchvāsa the poet should give two stanzas in āryā. A couple of stanzas, indeed, opens every chapter ; and the metre is āryā with the exception of the third and fourth ucchvāsa, where the former of these stanzas is a śloka. This is, of course, only a very slight variation, showing that this rule of Rudraṭa's should not be taken too strictly. Finally there is agreement concerning the metric form of the stanzas within the ucchvāsas themselves : According to Rudraṭa the metre is not prescribed, though vaktra, aparavaktra, and āryā are employed with preference, and this is the fact with the Harsacarita.

It is of still greater interest to compare the ākhyāyika, as it is represented by Bāṇā's Harsacarita with the rules given by Bhāmaha. There are some points of manifest uniformity : first, the Harṣacarita is divided into chapters called ucchvāsas, secondly, these ucchvāsas, but not the first, are opened by stanzas indicating the tenors of the chapter in question, and thirdly the subject matter of the poem is not an invented story but a true account of something which was experienced (vrtta) by the hero or whosoever it may be.

Considering this last point, we see, however, a highly important difference. In the Harsacarita the hero himself is apparently not the narrator, but

Harṣadeva's biography is told by another person, who witnessed the described events. Though Bāṇa gives a very minute account of his own race and though his personal affairs are closely connected with Harṣadeva, he can by no means be considered as the nāyaka ; the title of the work, too, is Harṣacarita. As the poem is but a fragment, the length of the author's own biography appears to be of a mere casual character. A second difference (but far less important than the one just mentioned) concerns the metre of the stanzas which introduce every ucchvāsa : according to Bhāmaha the metre is vaktra and aparavaktra, while in the Harṣacarita it is never of this kind but mostly āryā, vaktra and aparavaktra occur in Bāṇa's work only in the ucchvāsas themselves. This difference is, however, not so important as might be thought at first sight, because it has been pointed out on a previous occasion that, in spite of Daṇḍin's interpretation, not the form of the metre but rather the indicative character of this couple of stanzas was in the opinion of Bhāmaha the essential matter.

These considerations lead to an important result : the prototype of Bhāmah's ākhyāyikā was not the Harṣacarita by Bāṇa, which does not agree with the theory, but some older work, which has not come to us. This fact makes it very improbable that Bhāmaha should have lived after Bāṇa, as some scholars are inclined to believe.

Thus we have a younger form of the ākhyāyikā, which in the seventh century A. D. was considered as a separate kind, in spite of Daṇḍin's denial of the

184 THE FOUNDATIONS OF INDIAN POETRY

difference. The following points may be mentioned as its chief characteristics :

1 The subject-matter gives historical facts.

2 It is not necessary that the hero himself is the narrator.

3 There are chapters called ucchvāsas.

4 Every ucchvāsa, but not the first, opens with two stanzas (metre usually āryā), which indicate what will happen in the ucchvāsa in question.

5 The whole story begins with a metric introduction of a literary character.

The ākhyāyikā was in contrast to the kathā ; and the kathā is, as we have seen, the Indian novel. We possess two examples of kathā from the older period : Subandhu's Vāsavadattā and Bāna's Kādambarī. Though as regards this case matters are less complicated, we shall analyze both stories shortly in order to compare them with the doctrine of the śāstra.

Subandhu's Vāsavadattā begins with twelve introductory āryā stanzas, that commence with a namaskriyā to Sarasvatī, Krṣṇa, and Śiva. After that the sinfulness of man and the inferiority of the whole world since king Vikramāditya's death, is spoken of. Then the good poets are praised, and with the statement of the poet's authorship the metric introduction closes. The contents of the prose need not be given here in detail.

It is a well-known fact that we do not find this form of the story as it is presented in Subandhu's work any where else in Indian literature, so that

we may venture to assume that, at least as regards particulars, the story is invented by the poet, and this is the very fact which according to the opinion of the older ālaṃkārikas is a mark of difference between ākhyāyikā and kathā. To this point the words kaver abhiprāyakrtir aṅkanaiḥ kaiścid aṅkitā used in the definition of Bhāmaha must evidently refer. We have in this case the peculiar feature of a novel, which fact finds expression by Bhāmaha's term kanyāharaṇasamgrāmāvipralambhodayā.

Mentioning this we come to a point in which the Vāsavadattā does not agree wholly with the theories of the old scholars, for in our novel the events are peaceful on the whole. The struggle we hear of in the last pages of the story has nothing to do with the development of the plot. It seems to be, on the contrary, totally unnecessary here. We will, however, abstain from pursuing the question further here. Conceding that the way Vāsavadattā was carried off from her native town to the Vindhya mountains can be correctly called a kanyāharaṇa (which is not probable), we hear not one word about a struggle caused thereby, and the word samgrāma can have no other meaning than "struggle", especially not here on account of the preceding āharaṇa which means "taking by force."

Rudraṭa presents, as we have seen, some differences concerning the matter in question. Instead of kanyāharaṇa he speaks of kanyālābha, which has not quite the same meaning as the former term, and what appears to be more important, he does not mention any samgrāma. Considering this there can be no

xxiv

doubt that the prototype of Bhāmaha's kathā was by no means Subandhu's Vāsavadattā, but another story, in which the struggle caused by the seizing of a girl was described.

In other less important points the Vāsavadattā is in accordance with the opinion of Bhāmaha : the narrator is not the hero, nor is there a division into chapters (called ucchāsas), the tale being told without interruption up to the end.

It need not be mentioned that there are no stanzas (vaktra and aparavaktra) indicating the tenor of the following parts.

Within the story, however, the poet gives stanzas sometimes, but their metre is never vaktra and aparavaktra ; verses occur three times : three stanzas : āryā, sārdūlavikrīdita, sārdūlavikrīdita : three stanzas : śikharinī, śikharinī, sragdhrā ; one stanza : āryā.

Another example of the kathā is the Kadambari by Bāṇa.

It does not matter that this story has not been completed by the author himself, but by his son Bhūṣaṇabhaṭṭa.

Though the tale is more intricate than the Vāsavadattā, the form is similar to that of the composition just described.

The prose-story is introduced by some stanzas in the vaṃśastha metre.

They begin with a namas-kriyā to Brahman, Śiva, and Viṣṇu, and to the author's guru Bhatsu.

After that the poet speaks of the sinfulness of men and of the effect of good poetry.

Then follows an account of Bāṇa's race and family, and with the statement of the authorship the introduction comes to an end.

Then follows the prose-story, the tenor of which need not be detailed.

THE DIVISION OF POETRY

It can easily be pointed out that, as in the case of the Harsacarita, Bāṇa's Kādambarī was the model for Rudrata's kathā. Concerning the introductory stanzas (śloka, in the word's wider sense) Rudrata says ; iṣṭān devān gurūn namaskṛtya. Further the poet should, according to this author, describe his own race in verse shortly. The prose-story.begins with another tale and ends with the winning of the beloved girl.

Thus the Kādambarī is a novel quite similar to the Vāsavadattā. The subject-matter is not based on a known itiḥāsa, but is sprung from the poet's own imagination. As well as in the Vāsavadattā nothing is said of a samgrāma.

It is very obvious here just as in the case of the Hārṣacarita that the prototype of Bhāmaha's kathā cannot have been the Kādambarī (nor the Vāsavadattā), but an older work, which we do not know yet. Combining this fact with some other considerations I can scarcely believe that Bāṇabhaṭṭa should have written before Bhāmaha.

The newer kathā, then, shows the following features :

1 The subject-matter is a story for the most part invented by the poet, a love story ending with the union of the lovers.

2 The narrator is not the hero himself.

3 The story is not divided into chapters.

4 At the beginning there is a literary introduction in verse.

RAMKRISHNA MISSION INSTITUTE 1431 UIC OF CALCUTTA THE LIBRARY * CALCUTTA *

ABDOL JABBER & PROS

BOOK-BINDER

91, Bayaulane Lane