Pramāṇa · Search · About · API

Studies in the History of Sanskrit Poetics v1.c6: II Bharata

Studies in the History of Sanskrit Poetics v1.c6

Language
English

II Bharata

v1.c6.1

Although Indian tradition glorifies Bharata, the reputed author of the *Nāṭya-śāstra*, with the title of *muni* and places him in a mythic age, the widest possible divergence of opinion exists among scholars as to his actual date; and he has been variously assigned to periods ranging from the 2nd century B.C. to the 2nd century A.D.¹ That he is the oldest writer on dramaturgy, music and kindred subjects, whose work has survived, is generally admitted; but at the same time the question arises as to how far the extant version of his work represents his original text. As there are indications that it has been subjected to considerable rehandlings in later times before it assumed its present shape, this question has an important bearing on the date of the supposed author.

v1.c6.2

¹ Regnaud in *Annales du Mus. Guimet* ii p. 66, also introd. to Grosset's ed. ; Pischel in *GgA*, 1885, pp. 763 f ; P. R. Bhandarkar in *IA* xli, pp. 157 f ; H. P. Śāstrī in *JASB* v p. 352 f ; Sten Konow in *Ind. Drama* p. 2 ; Rapson in art. on Indian Drama in *Encyclop. of Religion and Ethics*, vol. v p. 886.

v1.c6.3

There are several passages in the present-day text which probably throw some light on this process of gradual interpolation and recasting. The curious colophon at the end in the Kāvyamālā edition, which appears to have puzzled its editor, designates the latter portion of the work as Nandi-bharata.¹ Rice mentions² a work called Nandi-bharata on music; while a chapter, apparently from a work on dramatic gesture, is referred to as nandibharatōkta-saṃkara-hastādhyāya in a manuscript of a treatise on music and abhinaya, noticed in Madras Catalogue no. 13009. These works, probably late compilations, refer to Nandi or Nandikeśvara, whom tradition acknowledges as an ancient authority on music and histrionic art³. This designation, therefore, of the latter part of Bharata's text, a part of which deals, among other things, with music, probably implies that it was compiled or recast at some later period in accordance with the views of Nandikeśvara. The last chapter, to which this colophon is appended, also contains a prediction that the rest of the topic will be treated in detail by Kohala¹ (who apparently belonged to the same school²), plainly showing that the rewriting of the portion in question was done some time after Kohala, as well as Nandikeśvara, had spoken on the subject. Nandikeśvara's date is unknown; but Kohala, side by side with Bharata, is recognised as an ancient authority, as early as the end of the 8th century A. D., in Dāmodaragupta's *Kuṭṭinī-mata* (*śl.* 81). It is interesting to note in this connexion that Abhinavagupta, commenting on Bharata vi 10, says³ that although *nāṭya* is usually said to consist of five aṅgas, the enumeration of eleven aṅgas in the text is in accordance with the view of Kohala and others, to whose opinions the commentator makes many other incidental references. It is likely, therefore, that between Bharata’s original work and its existing version, there came “Kohala and others” whose views found their way into the compendium, which goes by the name of Bharata and which indiscriminating posterity took as genuine and unquestionable.

v1.c6.4

¹ samāptaś cāyaṃ [granthah] nandibharata-saṃgīta-pustakam.

v1.c6.5

² Mysore and Coorg Catalogue, p. 292.

v1.c6.6

³ See above p. 2 fn. Similarly we hear of a work called Matangabharata by Lakṣmaṇa Bhāskara, which apparently sets forth the views of Matanga, who is another old authority cited by Śārṅgadeva and his commentator, and by Śiṅgabhūpāla (i 51). It appears that the term bharata in course of time came to mean the dramatic art generally, as it also came to mean the actor; and Rāghavabhaṭṭa on Śakuntalā expressly refers to ādi-bharata, by whom he means our author, in contra-distinction to these later bharatas.

v1.c6.7

¹ *śeṣam prastāra-tantreṇa kohalaḥ kathayiṣyati* (emended for *kolāhalaḥ kathiṣayti* in the text), xxxvii 18.

v1.c6.8

² see xxxvii 24. Kohala is cited as one of his sources by Śārṅgadeva (I. i. 15); and Hemacandra, with reference to the classification of the different species of the drama, says: *prapañcas tu bharata-kohalādi-śāstrebhyo'vagantavyaḥ* (p. 329; also p. 325). Śiṅgabhūpāla (i 51) acknowledges him as an authority on the drama and allied arts. A work on music called *Tāla-lakṣaṇa*, probably a late compilation, is attributed to him (Aufrecht i 130b, *IOC* 3025,3089; *Madras Cat.* 12992 with a Telegu comm.), and a *Kohalīya Abhinaya-śāstra*, purporting to embody his views, is also entered in the last named catalogue (with a Telegu comm. no. 12989). Another work on music, entitled *Dattila-kohalīya*, is mentioned by Burnell p. 60b, apparently a compendium of the opinions of Kohala and Dattila (or Dantila), who is mentioned by Damodaragupta (*śl.* 123), and cited as an old authority by Śārṅgadeva (I. i. 16) and his commentator Kallinātha (p. 49), by Śiṅgabhūpāla (i 51), as well as in various works on music (Peterson iv p. 43 extr.; *ABod* 199b; 201a).

v1.c6.9

³ *abhinaya-trayam gītātodye ceti pañcāṅgam nāṭyam* ........*anena tu ślokena kohalādi-matenaikādaśāṅgatvam ucyate*.

v1.c6.10

The process of incorporation must have occurred very early and was apparently complete by the end of the 8th century, when the work assumed more or less its present shape. Udbhaṭa, about this time, actually appropriates (iv 4) the first-half of the verse vi 15 of the Nāṭya-śāstra, and makes only enough verbal change in its second-half to admit śānta as the ninth rasa in the category of eight recognised by Bharata.¹ Abhinavagupta, who commented on the existing text at the end of the 10th century, himself mentions several other previous commentators, of whom Lollaṭa and Śaṅkuka in all probability belonged to the 8th and the 9th centuries. These indications will make it clear at any rate that the text existed in its present form in the 8th century A. D., if not earlier.

v1.c6.11

¹ I It may be noted that Abhinavagupta, immediately after the passage cited above, goes on to say: *nanu bharate tat-saṃgṛhitasyûpi punar atroddeśānirdeśe caitat krama-vyatyāsānûdityudbhaṭaḥ, neti lo llaṭaḥ . . . . . vayaṃ tvatra tattvam agre vitaniṣyāma ityûstām tāvat* (on Bharata vi 10). This difference of opinion between Udbhaṭa and Lollaṭa on a question of textual interpretation supports our conclusion that Udbhaṭa was probably familiar with the text of Bharata, as Abhinava knew it and as it has come down to us. It is not clear, however, if this evidence is sufficient to confirm the tradition, recorded by Śārṅgadeva (I. i. 19), that Udbhaṭa was one of the early commentators on Bharata. Abhinava himself refers to Lollaṭa, Śaṅkuka aud Bhaṭṭa Nāyaka as

v1.c6.12

We have, on the other hand, the tradition as well as the statement of Bhavabhūti, who refers to to the mythical Bharata as the *tauryatrika-sūtra-kāra*¹, that Bharata's work existed originally in the *sūtra*-form, which is also made likely by Pāṇini's early reference to such *naṭa-sūtras* in his own time. Reminiscences of the *sūtra*-style may indeed be presumed in the *Nāṭya-śāstra* vi and vii, which deal with *rasa* and *bhāva*; for in ch. vi we find a

v1.c6.13

¹ *Uttara-carita*, ad iv 22 (ed N. S. P., 1906, p. 120)

v1.c6.14

the principal commentators whose views he sometimes thought worth refuting in detail. Śārṅgadeva omits from his list Bhaṭṭa Nāyaka but adds Udbhaṭa, as well as one Kirtidhara who apparently flourished after Abhinava (*vyākhyātāro bhāratīye lollaṭodbhaṭa-śaṅkukāḥ | bhaṭṭābhinava-guptas ca śrīmat-kīrtidharo'paraḥ*). Abhinava also cites Rāhula (or Rāhala) and Bhaṭṭa Yantra, but it is not clear whether they were also commentators on Bharata. Śārṅgadeva mentions Rāhula elsewhere as one of his authorities, apparently on music (cf Hemacandra p. 316).

v1.c6.15

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

v1.c6.16

¹ *tatra vibhāvānubhāva-vyabhicāri-saṃyogād rasa-niṣpattiḥ*, ed. Grosset p. 87, l. 8 ; ed. Kāvyamālā p. 62, l. 6. This dictum is cited as a *sūtra* by all later writers, including Abhinavagupta, and presumably by his predecessors Loṭṭaṭa and others.

v1.c6.17

If we get the lower limit to the date of Bharata’s work at 8th century A.D., the other limit is very difficult to settle, when we consider that there were apparently two versions, one based on the other. But it is not clear what weight should be placed on the testimony of Bhavabhūti; for if in the first quarter of the 8th century Bharata was known to him as a *sūtra-kāra*, it is not intelligible how at the end of the same century, Udbhaṭa makes use of Bharata’s *kārikā*, and Lollaṭa and others, immediately following, apparently comment on the same text. The short space of less than half a century is not enough to obliterate all signs of the older version and replace it entirely with a new *kārikā*-text which, to all intents and purposes, is taken as the only authoritative version in later times, and in which, strangely enough, we find still lingering traces of the earlier *sūtra*-text. The only possible explanation of Bhavabhūti’s reference is that the historical Bharata, who was the *sūtra-kāra* on the three arts of dancing, singing and instrumental music, had already, in Bhavabhūti’s time, become identified with the mythical Bharata; for the passage in the *Uttara-carita* gives an obviously mythical account, through the mouth of Lava, that *bhagavān* Vālmīki, having composed his story of *Rāma*, gave it to *bhagavān* Bharata (the *sūtra-kāra* on the three arts and apparently the *nāṭyācārya* of the gods) who revised it and got it acted through celestial nymphs.

v1.c6.18

But this does not exhaust all our textual difficulties. Independent prose-passages also survive, in the midst of kārikās, in chs. xvii, xxviii, xxix, xxxi and xxxiv, which, forming an integral part of the text, cannot be taken as mere vṛtti, but which resemble, in some respects, the prose smṛti-fragments or, more closely, the prose-fragments in the recently published Bhela-saṃhitā. Again, the anubaddha or anuvamśya ślokas, referred to above, correspond to the parikara- or saṃgraha-ślokas in later writings, and certainly indicate the probability of earlier speculations on the subject. These verses are generally taken from two distinct sources, for some of them are in āryā while others are in anuṣṭup metre.

v1.c6.19

From the facts adduced above, we are confronted with the problem of the inter-relation of these apparent survivals in our text, which contains vestiges of (1) independent prose-fragments (2) anuvamśya ślokas in āryā and anuṣṭup metres and (3) sūtra-bhāṣya style, as well as ( 4 ) the present systematic kārikās. We have no room to dilate upon this point, but an examination of these passages will reveal that these different styles do not belong to the same period, but they probably indicate several stages in the growth of particular forms of composition of dramaturgic works in general, each stage betraying its own partiality towards a particular form. Taking the present kārikā-text as the starting point, we find in it traces of an earlier sūtra-bhāṣya style, of which it is presumably a recast. In the sūtra-bhāṣya, again, there are fragments of metrical quotations, which indicate, in their turn, another and still earlier *kārikā*-stage; while the independent prose-fragments perhaps represent the earliest form taken by such technical treatises. We can, therefore, distinguish in their order of development (1) a stage of prose-treatises (2) a tentative period of *kārikā*-writing (3) the *sūtra-bhāṣya* stage and (4) the final period of compilation of compendiums, which reverts again to the *kārikā*-form.¹ This conclusion perhaps finds some support in the repetition, more or less, of a similar phenomenon in the sphere of the Dharma-śāstra, the Artha-śāstra, the Vaidya-śāstra and probably the Kāma-śāstra. The loss of earlier tentative treatises makes it difficult to dogmatise; but if this conclusion is correct generally, then our text may be supposed to contain remnants of all these styles and forms. It is not argued here that Bharata's work itself passed through all these stages or forms, from a rudimentary prose-version into a systematic metrical manual; but our text contains enough to betray the existence of previous speculations in prose and in verse, as well as indicate the fact that it might have itself been once written in the *sūtra-bhāṣya* form, which was recast, with considerable additions from other sources, into a convenient metrical compendium.

v1.c6.20

¹ This conclusion does not apply to the more or less imitative periods after the 10th century in which we find the *kārikā* and the *sūtra*-style, existing almost side by side e. g. we find in the respective works of Mammaṭa and Ruyyaka.

v1.c6.21

Taking the substance of the work, apart from the vexed question of different versions, the portion of the *Nāṭya-śāstra*, which deals principally with music, has been conjectured¹ on internal evidence to have been compiled about the 4th century A. D. ; and it may appear likely that the other portions were also put into their present shape about the same time. Pischel's argument, however, on the date of the work, derived from the reference to Pahlavas in a text of such composite character is of doubtful value in determining the question finally ; but it perhaps makes it probable that the upper limit of its date cannot be put too early. On the other hand, Rāghavabhaṭṭa on *Śakuntalā*² and Vāsudeva on *Karpūra-mañjarī*³ quote one Mātṛguptâcārya as a writer on dramaturgy ; and Sundaramiśra in his *Nāṭya-pradīpa* (dated 1613 A.D.), commenting on Bharata's remarks on *nāndī*, says *asya vyākhyāne mātṛguptâcāryaiḥ........iyam udāhṛtā*⁴. This has been taken by Lévi to imply that Mātṛgupta wrote a *vyākhyāna* or commentary on Bharata, and that, assuming him to be same as the poet who lived under Harṣa-vikramâditya (*Rāja-taraṅg*.

v1.c6.22

¹ *IA* xii p. 158 f.

v1.c6.23

² ed. N. S. P., 1922, pp. 5, 6, 7, 8, 9, 13, 15, 20, 57, 62, 74, 110, 123, 126, 151, 154, 199 etc.

v1.c6.24

³ ed. N. S. P., 1900, p. 5 : cf Aufrecht i 448a.

v1.c6.25

⁴ quoted in *IOC* iii p. 347. This Mātṛgupta is also cited by Ranganātha on *Vikramôrvaśī* (dated 1656 A. D., see *AFL* 144) ed. N. S. P., 1914, p. 5.

v1.c6.26

iii 125-252), we get in him a very early commentator on Bharata. But this evidence does not appear to be conclusive; for our Mātrgupta, as Rāghavabhatta’s profuse quotations shew, probably wrote an independent metrical work on dramaturgy, in which he might have in the usual course commented on Bharata’s precepts; and the word *vyākhyāna* need not be construed to mean a commentary. This author, cited only by late commentators, is therefore to be distinguished from the poet of that name; and probably he belongs to comparatively recent times, being certainly unknown to Abhinavagupta, Dhanika and other older writers. But he must be earlier than Śārngadeva (beginning of the 13th century) who mentions him in his work as an authority on music.

v1.c6.27

We are in a position, however, to infer that the substance of Bharata’s work is probably much older than that of Bhāmaha¹, who may be assigned to the last quarter of the 7th century. Bhāmaha, in his treatment of poetic figures (*kāvyālaṃkāras*), groups them in a curious but suggestive way, which probably indicates the different periods in the growth and multiplication of such figures². At the outset, he names and defines only five poetic figures (ii 4), recognised, as he says, by other writers, viz., *anuprāsa*, *yamaka*, *rūpaka*, *dīpaka* and *upamā*. This represents the first stage; but in course of time, six other figures appear to have been added, and Bhāmaha mentions and deals with them next in ii 66. Then he goes on to enumerate, two (or three, including *svabhāvókti*) more figures admitted by writers like Medhāvin (ii 88), who also appears to have dealt with *upamā* etc. (ii 40). Finally, Bhāmaha defines and illustrates a further long list of twenty-three more figures in a separate chapter (iii 1-4). The differentiation and multiplication of poetic figures with the progress of speculation is a familiar fact in Alaṃkāra literature, and the way in which Bhāmaha successively enumerates and groups these figures probably shows that to the original five mentioned by him at the outset, others were added in course of time as the study itself advanced. Now Bharata, in his treatment of *alaṃkāras*, names (xvi 41) only four such figures known to him, viz., *yamaka*, *rūpaka*, *dīpaka* and *upamā*. These four in reality correspond to the five mentioned by Bhāmaha; for *anuprāsa* may be taken as falling in the same class as *yamaka*, the one being *varṇábhyāsa* and the other *padábhyāsa*. At the same time the very fact *anuprāsa* is thus differentiated from *yamaka* may indicate further refinement in the analysis of these poetic figures. It is clear, therefore, that Bharata's work belonged to a period when the number of figures had not yet multiplied; and one, if not more, stages must have intervened between it and Bhāmaha's *Kāvyálaṃkāra* in which the number had already swelled into nearly forty in all.¹ To this intermediate stage belonged Madhāvin and others, whom Bhāmaha cites, and the loss of whose work makes it difficult for us to trace the development thus indicated by Bhāmaha.

v1.c6.28

¹ Besides showing himself conversant with some theory of *rasa* (ii 281, 283 f), Dandin mentions the dramaturgic technical terms *sandhi*, *anga*, *vrtti* and *laksana* and refers to *āgamāntara* for their discussion (ii 366).

v1.c6.29

² Cf Jacobi in *Sb. der Preuss. Akad.* xxiv (1922) pp. 320 f.

v1.c6.30

¹ By the end of the 6th and beginning of the 7th century, Bhaṭṭi illustrates as many as 38 different species of poetic figures, indicating that the process of refinement had proceeded very far indeed even in his time.

v1.c6.31

There are also indications that Bharata’s teachings are probably older than Kālidāsa, who generally adheres to Bharata’s dramaturgic prescriptions². Kālidāsa refers to Bharata as the mythical nāṭyācārya; but apart from this, it may be pointed out that while Bharata ignores Māhārāṣṭrī as one of Prakrits used in the drama, it is well known that in Sanskrit dramatic works, including those of Kālidāsa, Prakrit verses are in Māhārāṣṭrī and the prose-parts are in Śaurasenī and other allied Prakrits. In Raghu xix 36, again, Kālidāsa speaks of aṅgasattva-vacanāśraya nṛtya, which, as Mallinātha rightly points out, agrees with Bharata’s dictum sāmānyābhinayo nāma jñeyo vāgaṅga-sattvajaḥ³.

v1.c6.32

² The discovery of the dramas, ascribed to Bhāsa, does not invalidate this argument for it is possible that they follow a tradition or a system of opinion of which all traces are now lost, and the date of Bhāsa itself is uncertain.

v1.c6.33

³ The Amara-koṣa only mentions āṅgika and sāttvika abhinaya. This lexicon, which is said to have been translated into Chinese about 561-66 A. D. (see Nandargikar's introd. to Meghadūta, 1894, p. 73), admits after Bharata the eight rasas, and explains some of the dramaturgic technicalities, giving three synonyms of the actor (śailālin

v1.c6.34

The lower limit of the date of Bharata's work, therefore, can be provisionally shifted back to the fourth or fifth century A.D., while it is almost certain that it existed in its present shape in the 8th century A.D. The upper limit cannot be put too early, because of the mention of *śakas*, *yavanas*, *pahlavas* and other tribes, and probably does not go beyond the commencement of the Christian era. It is difficult to settle the relative age of the *sūtra* and the *kārikā*-texts; but if the tendency towards *sūtra-bhāṣya* style may be presumed to have been generally prevelant in the last few centuries B.C., then the suspected *sūtra*-text of Bharata belongs apparently to this period¹. It was certainly much earlier than the present *kārikā*-text, in which Bharata already appears as a mythical sage as an expounder of the *nāṭya-veda*.

v1.c6.35

¹ It will be shown later that the tradition that Bharata was the author of a *Kāvya-lakṣaṇa*, which forms in substance of the *kārikā*-verses of Mammata's *Kav. prak.*, is entirely erroneous, as is also the statement made use of by Lévi

v1.c6.36

*kṛśāśvin* and *bharata*) from the names of the three well-known teachers on dramaturgy. Pāṇini refers to the formation of the first two terms, but not that of the third; but this silence of the grammarian does not prove anything. The Jaina *Aṇuogadārasutta* (ed. N. S. P. 1915, fol. 134-145, also quoted in Weber ii 2, pp. 701-02) which, Winternitz thinks, was probably put together by the middle of the 5th century, mentions nine *rasas*, which, however, have hardly any reference poetic or dramatic *rasas*; but the enumeration is interesting from the inclusion of *praśānta* (not mentioned by Bharata), apparently from religious motives.

Linked texts (16)