Pramāṇa · Search · About · API

A History of Sanskrit Literature: Classical Period, Vol. I c4.s7: 3. The Literature of Tale and Fable

A History of Sanskrit Literature: Classical Period, Vol. I c4.s7

Language
English

Contents

3. The Literature of Tale and Fable

c4.s7.1

The Buddhist anecdotal literature perhaps reflects an aspect of the literary, as well as popular, taste of the time, which liked the telling of tales in a simple and unadorned, but distinctly elegant, manner; for the origin of the Sanskrit *Pañcatantra* and the Prakrit *Bṛhatkathā*, which represent story-telling from another point of view, is perhaps synchronous, although the various extant versions of the two works belong to a much later period. The Avadāna, the didactic beast-fable and the popular tale are indeed not synonymous. While the Avadāna, closely related to the Jātaka, is clearly distinguishable as a Buddhist *gest*, which has a definite religious significance, the other two species are purely secular in object and character. The method of story-telling is also different; for in the Jātaka or Avadāna, we have generally the application of a past legend to a tale of to-day. In the Jātaka the Bodhisattva tells a tale of his past experience, but it is not narrated in the first person ; the device of first-hand narrative as well as of enclosing a tale, is a feature which characterises the classical method. The Sanskrit poetic theory ignores the Jātaka and Avadāna, presumably because they have a religious objective and seldom rises to the level of art, but it does not also clearly define and discriminate between the fable and the tale. The elaborate attempt to distinguish between the Kathā and the Ākhyāyikā,¹ as the invented story and the traditional legend respectively, is more or less academic, and has hardly any application to the present case. Some of the stories of the *Pañcatantra* are indeed called Kathās, but one of the versions of the entire work is styled *Tantrākhyāyikā*, while Guṇāḍhya's work is designated as the Great Kathā. Possibly no fine distinction is meant, and the terms Kathā and Ākhyāyikā are employed here in the general sense of story. A rigid differentiation, however, cannot perhaps be made in practice between the fable and the tale ; for the different elements in each are not entirely excluded in the other, nor isolated. The beast-fable, as typified by the *Pañcatantra*, is not seldom enriched by folk-tale and spicy stories of human adventure, while the tale, as represented by *Bṛhatkathā*, sometimes becomes complex by absorbing some of the elements of the fable and its didactic motive. Both these types, again, should be distinguished from the prose romance, the so-called Kathā and Ākhyāyikā, such as the *Harṣa-carita* and the *Kādambarī*, in which all the graces and refinements of the Kāvya are transferred from verse to prose, either to create an exuberantly fanciful story or to vivify and transform a legend or folk-tale.

c4.s7.2

¹ See S. K. De, The Kathā and the Ākhyāyikā in Classical Sanskrit in *BSOS*, III, p. 307f.—Daṇḍin (i-28) speaks of Ākhyāna as a general species, in which collections of tales like the *Pañcatantra* were probably included.

c4.s7.3

The currency of tales and fables of all kinds may be presumed from remote antiquity, but they were perhaps not used for a definite purpose, nor reduced to a literary form, until at a comparatively late period. The ancestor of the popular tale may have been such Vedic Ākhyānas as are preserved, for instance, in the Ṛgvedic dialogue-hymn of Purūravas and Urvaśī, or in such Brāhmaṇic legends as that of Śunaḥśepa ; but it is futile to seek the origin of the beast-fable in the Ṛgvedic hymn of frogs (*vii.* 103), which panegyrises the frogs more from a magical than didactic motive, or in the Upaniṣadic parable of dogs (*Ch. Up.* i. 12), which represents the dogs as searching out a leader to howl food for them, but which may have been either a satire or an allegory. Nor is there any clear recognition of the fable in the Epics as a distinct literary *genre*, although the motifs of the clever jackal, the naughty cat and the greedy vulture are employed for the purpose of moral instruction. But all these, as well as the Jātaka device of illustrating the virtues of Buddhism by means of beast-stories,¹ may have suggested the material out of which the full-fledged beast-fable developed in the *Pañcatantra*. In its perfected form, it differed from the simple parable or the mere tale about beasts, in having the latent didactic motive clearly and deliberately brought out and artistically conveyed in a definite framework and a connected grouping of clever stories, in which the thoughts and deeds of men are ascribed to animals. There is nothing simple or popular in such a form ; and the beast-fable as an independent literary creation diverged considerably in this respect from the popular tale, which is free from didactic presentation and in which the more or less simple ideas of the people and their belief in myth and magic, as well as racy stories of human life, find a direct expression. In the case of beast-fable, again, the connexion with the courts of princes is cleared. The popular tale, no doubt, speaks of romantic prince and princess of a fairly-land ; but the framework of collection of beast-fables like the *Pañcatantra*, which is delivered in the form of instruction to tender-minded young princes in statecraft and practical morality, leaves no doubt about one form of its emyloyment. It is thus closely related to the Nīti-śāstra and Arthaśāstra,¹ but it is not directly opposed to the Dharma-śāstra. The fact is important; for even if the beast-fable inculcates political wisdom or expediency in the practical affairs of life, rather than a strict code of uprightness, it seldom teaches cleverness at the expense of morality.²

c4.s7.4

¹ The Barhut Stūpa reliefs, depicting some of the stories, establish the currency of the beast-fable at least in the 2nd Century B.C.

c4.s7.5

¹ No direct influence of Kauṭily's *Artha-śāstra* can be traced in the *Pañcatantra*.

c4.s7.6

² F. Edgerton in *JAOS*, XL, p. 271 f.

a. The Pañcatantra

c4.s7.t1.1

The only collection of beast-fable and the solitary surviving work of this kind in Sanskrit is the *Pañcatantra*, which has come down to us in various forms; but it is a work which has perhaps a more interesting history than any in world-literature.³ There can be little doubt that from the very beginning it had a deliberate literary form. Each of its five parts, dealing respectively with the themes of separation of friends (Mitra-bheda), winning of friends (Mitra-prāpti), war and peace (Saṃdhi-vigraha), loss of one's gains (Labdha-nāśa) and hasty action (Aparīkṣita-kāritva), is a narrative unit in itself; but all together they form a perfect whole fitted into the frame of the introduction.

c4.s7.t1.2

³ J. Hertel (*Das Pañcatantra, seine Geschichte und seine Verbreitung*, Leipzig and Berlin, 1914, Index, p. 451 f.) records over 200 different versions of the work known to exist in more than 50 languages (three-fourths of the languages being extra-Indian) and spreading over a region extending from Java to Iceland. For a brief résumé of this history, as well as for brief summary of the work, see Winternitz, *GIL*, III, pp. 294-311; Keith, *HSL*, pp. 248 f, 357 f.—The question whether the individual tales or the Indian fable itself as a species, were borrowed in their origin, from Greece is much complicated. Chronology is in favour of the priority of Greece, but the suggestion that India consciously borrowed from Greece is not proved. Some points of similarity may be admitted, but they may occur without borrowing on either side. At any rate, if reciprocal influences and exchanges occurred, India seems to have given more than it took. Benfey's position that the tale is entirely Indian, while the fable came from Greece, need not be discussed, for folklorists to-day no longer seek to find the birthplace of all tales and fables in any one country.

c4.s7.t1.3

The stories are told, as in the case of the popular tale, in simple but elegant prose, and there is no attempt at descriptive or sentimental excursions or elaborate stylistic effects. The combining of a number of fables is also a characteristic which it shares with the popular tale, but they are not merely emboxed; there is, in the weaving of disjointed stories, considerable skill in achieving unity and completeness of effect. The insertion of a number of general gnomic stanzas in the prose narrative is a feature which is dictated by its didactic motive; but the tradition is current from the time of the Brāhmaṇas and the Jatakas. More interesting and novel, if not altogether original, is the device of conveniently summing up the moral of the various stories in pointed memorial stanzas, which are not general maxims but special labels to distinguish the points of individual fables. The suggestion¹ of a hypothetical prose-poetic Vedic Ākhyāna, in which the verse remained fixed but the prose mysteriously dropped out, is not applicable to the case of the blend of prose and verse in the fable literature; for the prose here can never drop out, and the essential nature of the stanzas is gnomic or recapitulatory, and not dramatic or interlocutory. There must have existed a great deal of floating gnomic literature in Sanskrit since the time of the Brāhmaṇas, which might have been utilised for these passages of didactic wisdom.

c4.s7.t1.4

¹ H Oldenberg in *ZDMG*, XXXVII, p. 54 f. XXXIX, p. 52 f; also in his *Zur Geschichte d. altindischen Prosa*, Berlin 1917, p. 53 f and *Lit. d. alten Indien*, cited above, pp. 44 f, 125 f, 153 f.

c4.s7.t1.5

The *Pañcatantra*, however, is not a single text, but a sequence of texts; it exists in more versions than one, worked out at different times and places, but all diverging from a single original text. The original,² which must have existed long before 570 A.D. when the Pahlavi version was made, is now lost; but neither its date nor its title nor provenance, is known with certainty. The character and extent of the transformation, to which the work was subjected in course of time, make the problem of reconstruction one of great intricacy, but the labours of Hertel¹ and Edgerton² have succeeded in a great measure in going back to the primary *Pañcatantra* by a close and detailed examination of the various existing versions. That it originally contained five books with a brief introduction and was called *Pañcatantra*, is now made fairly certain, but there is a considerable discussion of the meaning of the word *Tantra*. It may denote nothing more than a book or its subject-matter, but since it occurs in the title *Tantrākhyāyikā* of one of the versions,³ it may indicate a text of polity as an art. There is no evidence at all of authorship; for the name Viṣṇuśarman, applied in the introduction to the wise Brahman who instructs, with these stories, the ignorant sons of king Amaraśakti of Mahilaropya in Deccan, is obviously as fictitious as the names of the king and the place. Hertel thinks that the work was composed in Kashmir, but his arguments are inadequate; while nothing can be confidently inferred from the mention of Gauḍa or Ṛṣyamūka or of well known places of pilgrimage like Puṣkara, Vārāṇasī, Prayāga and Gaṅgādvāra.

c4.s7.t1.6

² The idea of a Prakrit original is discredited both by Hertel and Edgerton. The literature on the Pañcatantra is vast and scattered, but the results of the various studies will be found summarised in the works, cited below, of these two scholars.

c4.s7.t1.7

¹ *Das Pañcatantra*, cited above, as well as works and editions cited below.

c4.s7.t1.8

² *The Pañcatantra Reconstructed*, Text, Critical Apparatus, Indroduction and Translation 2 vols., American Orient. Soc., New Haven, Conn., 1924.

c4.s7.t1.9

³ Jacobi, however, would translate it apparently as a collection of ākhyāyikā in tantras, 'die in bücher eingeteilte Erzählungssammlung.' See F. W. Thomas in *JRAS.* 1910, p. 1347.

c4.s7.t1.10

The various important recensions of the *Pañcatantra* have been classified into four main groups,⁴ which represent diversity of tradition but all of which emanate from the lost original. The first is the lost Pahlavi version,⁵ from which were derived the old Syriac¹ and Arabic² versions ; and it was through this source that the *Pañcatantra*, in a somewhat modified form, was introduced into the fable literature of Europe. The second is a lost North-western recension, from which the text was incorporated into the two North-western (Kashmirian) Sanskrit versions of Guṇāḍhya's *Bṛhatkathā*, made respectively by Kṣemendra and Somadeva (11th century A.D.).³ The third is the common lost source of the Kashmirian version, entitled *Tantrākhyāyikā*,⁴ and of the two Jaina versions, namely, the Simplicior Text, well known from Bühler and Kielhorn's not very critical edition,⁵ and the much amplified Ornatior Text, called *Pañcākhyāna*, of Pūrṇabhadra (1199 A.D.).⁶ The fourth is similarly the common lost source of the Southern Pañcatantra,⁷ the Nepalese version and the Bengali *Hitopadeśa*.¹ A detailed study of the character and interrelation of the various recensions and versions is not possible here, but some of their general characteristics may be briefly noted. The *Tantrākhyāyikā* is perhaps the oldest Sanskrit version, and preserves the original text better and more extensively than any other version. But none of the recensions—not even the *Tantrākhyāyikā*, the claims of which have been much exaggerated by Hertel—represents in its entirety the primitive text. The North-western original of Kṣemendra and Somadeva must have been a version made much later in Kashmir. Kṣemendra's fairly faithful, but dry, abstract suffers from its brevity, but Somadeva's narrative, in spite of a few omissions and some interruption of sequence by the introduction of extraneous tales, is normally clear and attractive. There is a great deal of reshuffling of stories, as well as intrusion of additional matter, in both the Simplicior and Ornatior Texts, the former adding seven and the latter twenty-one new stories. The Southern recension exists in several sub-versions ; it is much abbreviated, but nothing essential appears to have been omitted, and only one complete story (The Shepherdess and her Lovers) is added. The *Hitopadeśa*,² which has currency mostly in Bengal, is practically an independent work, containing only four and not five books, by one Nārāyaṇa, whose patron was Dhavalacandra and who must have lived before 1373 A.D., which is the date of one of the manuscripts of the work. The compiler amplifies the stories derived in the main from the *Pañcatantra*, by drawing upon an unknown source, considerably omits, alters, remodels the sequence of books and stories, and inserts large selections of didactic matter from *Kāmandakīya Nīti-sāra*.

c4.s7.t1.11

⁴ Hertel, however, believes in two versions of one Kashmirian recension only as the archetype of the other three recensions, namely, the *Tantrākhyāyikā* and what he calls '**K**'.—For a short genealogical table, setting forth the relationship of the four main recensions or groups, see Edgerton, *op. cit.*, II, p. 48, and for a full and detailed table of all known versions see Penzer's *Ocean of Story*, Vol. V. p. 242 (also by Edgerton).

c4.s7.t1.12

⁵ Made by the physician Burzoē under the patronage of Chosroes Anūshīrwān (531-79 A.D.) under the title Karataka and Damanaka.

c4.s7.t1.13

¹ Made by Būd, a Persian Christian, about 570 A.D. under the title *Kalilag wa Damnag*. Ed Schulthess, Berlin 1911.

c4.s7.t1.14

² Made by 'Abdallāh Ibnu'l-Muquffa about 750 A.D. under the style *Kalila wa Dimna*. Ed. L Cheikho, 2nd Ed., Beyrouth 1923.

c4.s7.t1.15

³ *Bṛhatkathā-mañjarī* xvi. 255 f ; *Kathā-sarit-sāgara* lx-lxiv. Leo von Mankowski has *edited, with trans. etc.*, (from only one imperfect MS), Kṣemendra's version separately in *Der Auszug aus dem Pañcatantra in Kṣemendras Bṛhatkathāmañjarī*, Leipzig 1892. Lacôte, Hertel and Edgerton make it probable that the original *Bṛhatkathā* of Guṇāḍhya did not contain the *Pañcatantra*.—Somadeva's version of the *Pañcatantra* (according to Emenau's computation in *JAOS*, LIII, 1933, p. 125) contains 539 Ślokas, while Kṣemendra's in Mankowski's edition, has 306 ; but deducting tne stories not found in Somadeva, Kṣemendra's total would be about 270 only.

c4.s7.t1.16

⁴ Ed, J. Hertel, Berlin 1910, containing two sub-versions ; also ed. J. Hertel in Harvard O. S., Cambridge Mass. 1915 ; trs J. Hertel, 2 vols., Leipzig and Berlin 1909.

c4.s7.t1.17

⁵ Bomby Skt. Ser. 1868-69 ; also ed. L. Kosengarten Bonn 1848 ; ed. K. P. Parab, NSP, Bombay 1896 (revised Parab and V. L. Panshikar 1912). J. Hertel, Über die Jaina Recensionen des Pañcatantra in *BSGW*, LIV, 1902, pp. 23-134, gives selections of text and translation.

c4.s7.t1.18

⁶ Ed. J. Hertel, Harvard Orient Ser., Cambridge Mass., 1908-12 ; trs into German by Schmidt, Leipzig 1901 ; into English by A. W. Ryder, Chicago 1925.—Pūrṇabhadra uses both the *Tantrākhyāyikā* and the Simplicior text.

c4.s7.t1.19

⁷ Ed. J. Hertel (Text of recension β, with variants from recension α), Leipzing 1906 ; Text of recension α, ed. Heinrich Blatt, Leipzig 1930. See also J. Hertel, Über einen südlichen textus amplior des Pañcatantra in *ZDMG*, 1906-07 (containing translation of text). Of the Nepalese version' Bk. i-iii are included in Hertel's ed. mentioned above, while Bk. iv-v in his ed. of *Tantrākhyāyikā*, introd., p. xxvii. Selections from the Nepalese version published with trs. by Bendall in *JRAS*, 1888, pp. 465-501. See Hertel in *ZDMG*, LXIV, 1910, p. 58 f and *Das Pañcatantra*, pp. 37 f, 313 f.

c4.s7.t1.20

¹ Repeatedly printed in India, but not yet critically edited. The better known ed. is by P. Peterson, Bomb. Skt. Ser., 1887 ; also *Hitopadeśa nach Nepalischen Handschrift*. ed. H. Blatt, Berlin 1930 (Roman characters). The earliest ed. is that of A. Hamilton, London 1810, and the earliest trs. by C. Wilkins, London, 1787.

c4.s7.t1.21

² See J. Hertel, *Über Text und Verfasser des Hitopadeśa nach* (Diss.) Leipzig 1897, p. 37, and *Das Pañcatantra*, p. 38 f. In spite of omissions and alterations, the *Hitopadeśa* preserves over half the entire sub-stories of the Pañcatantra, and follows closely the archetype which it shares with the Southern recension.

c4.s7.t1.22

Although Hertel is right in believing that the *Pañcatantra* was originally conceived as a work for teaching political wisdom, yet the fact should not make us forget that it is also essentially a story-book, in which the story-teller and the political teacher are unified, most often successfully, in one personality. There are instances where the professed practical object intrudes itself, and tedious exposition of polity prevails over simple and vivid narration; but these instances are happily not too numerous, and the character of the work as a political text-book is never glaring. Inequalities doubtless appear in the stories existing in the different versions, but most of them being secondary; it can be said without exaggeration that the stories, free from descriptive and ornamental digressions, are generally very well and amusingly told. They show the author as a master of narrative, as well as a perfect man of the world, never departing from an attitude of detached observation and often possessed of a considerable fund of wit and humour veiled under his pedagogic seriousness. If he makes his animals talk, he makes them talk well and the frankly fictitious disguise of the *fabliau* eminently suits his wise and amusing manner. With a few exceptions, the individual stories are cleverly fitted together into a complex but well planned form. The language is elegantly simple, and the author shows taste and judgment in never saying a word too much, except for a touch of the mock-heroic, and in realising that over-elaboration is out of place. The gnomic stanzas, if not the title-verses, are not always demanded by the narrative, but they are meant to give sententious summary of worldly wisdom and impressive utterance to very ordinary, but essential, facts of life and conduct. We do not know how far these stanzas are original, for some of them occur in the Epics and elsewhere; but they are generally phrased with epigrammatic terseness, and form an interesting feature, in spite of the tendency to over-accumulate them. It is not without reason, therefore, that the work enjoyed, and still enjoys such unrivalled popularity as a great story-book in so many different times and lands. /

b. The Bṛhatkathā of Guṇāḍhya

c4.s7.t2.1

The popular tale is represented by a number of works in Sanskrit, but the earliest appears to have been the *Bṛhatkathā*, or 'the Great Story', of Guṇāḍhya, the Prakrit original of which is lost, but which is now known from three comparatively late Sanskrit adaptations. Its exact date¹ cannot be determind, but that it already received recognition before 600 A.D. is clear from the references to its importance by Bāṇa² and Subandhu³ ; and there is nothing to show that it cannot be placed much earlier. If it belongs to a period after Christian era, it is not improbable that the work took shape at about the same time as the lost original of the *Pañcatantra*; and to assign it to the fourth century A.D. would not be an unjust conjecture.⁴ The recorded tradition informs us that the original *Bṛhatkathā* was composed in Paiśācī Prakrit; and it is noteworthy that the literary form which the popular tale first assumed was one in Prakrit. Like the *Pañcatantra*, the work of Guṇāḍhya was undoubtedly a new literary creation, but the medium of expression perhaps indicates a difference in method and outlook.

c4.s7.t2.2

¹ On the question of date and author, see J. S. Speyer, *Studies about Kathāsaritsāgara*, Amsterdam 1908, p. 44 f. Bühler in his *Kashmir Report* summarily places the work in the first century A.D., with which F. Lacôte (*Mélanges Lévi*, p. 270) appears to agree ; but S. Lévi (*Théâtre indien*, 1891, p. 317) cautiously adjusts it to the 3rd century. See Keith in *JRAS*, 1909, p. 145f. Both Daṇḍin's *Daśa-kumāra-carita* and Subandhu's *Vāsavadattā* refer to the story of Naravāhanadatta.

c4.s7.t2.3

² *Harṣa-carita*, Introductory st. 17.

c4.s7.t2.4

³ Ed. F. E. Hall, p. 110.

c4.s7.t2.5

⁴ The alleged Sanskrit version of Durvinīta of the 6th century (R. Narasimhachar in *IA*, LXII, 1913, p. 204 and *JRAS*, 1913, p. 489 f; Fleet in *JRAS*, 1911, pp. 186 f) and the supposed Tamil version of the 2nd century A.D. (S. K. Aiyangar in *JRAS*, 1906, p. 689 f; and *Ancient India*, London 1911, pp. 328, 337) are too doubtful to be of any use for chronological purposes. See Lacôte, *Essai sur Guṇāḍhya et la Bṛhatkathā*, Paris 1908, p. 198 f.

c4.s7.t2.6

An obviously legendary account of the origin of the work and the personality of the author is given, with some variations, in the introductory account of the two Kashmirian Sanskrit versions and in the apocryphal *Nepāla-māhātmya*¹ of a pseudo-Purāṇic character. It makes Guṇāḍhya an incarnation of a Gaṇa of Śiva, who under a curse is born at Pratiṣṭhāna on the Godāvarī and becomes a favourite of king Śātavāhana; but the king has another learned favourite in Śarvavarman, the reputed author of the *Kātantra* grammar. Having lost a rash wager with Śarvavarman, with regard to the teaching of Sanskrit to the king, who had been put to shame by the queen for his ignorance of the language, Guṇāḍhya abjures the use of Sanskrit and society, and retires to the wild regions of the Vindhya hills. There, having learnt from another incarnated Gaṇa of Śiva the story of the *Bṛhatkathā*, originally narrated by Śiva to Pārvatī, he records it in the newly picked up local Paiśācī dialect, in 700,000 Ślokas, of which only one-seventh was saved from destruction and preserved in the work as we have it! The Nepalese version of the legend, howeved, places Guṇāḍhya's birth at Mathurā and makes king Madana of Ujjayinī his patron; it knows nothing of the wager but makes Guṇāḍhya, on being vanquished by Śarvavarman, write the story in Paiśācī for no other explicit reason than the advice of a sage named Pulastya. The legend is obviously a pious Śaiva invention modified in different ways in Kashmir and Nepal²; from the reference in the *Harṣa-carita*, one may infer that it was known in some form to Bāṇabhaṭṭa; but the value of biographical and other details is not beyond question. If Śarvavarman is introduced, Pāṇini, Vyāḍi and Vararuci-Kātyāyana also figure in the legend as contemporaries, although the Nepalese compiler does not appreciate the grammatical interest, nor the use of Prakrit. The association with Satavahana recalls one of the brilliant periods of Prakrit literature, and probably suggests that the employment of Sanskrit by the Kṣatrapa rulers probably found a counter-movement in favour of the patronage of Prakrit literature; but Satavahana being a dynastic name, which may denote any of several kings, it does not help to solve the chronological problem.¹

c4.s7.t2.7

¹ Given in Lacôte, *op. cit.*, Appendix p. 291 f.

c4.s7.t2.8

² It is as a saint of Śaivism that Guṇāḍhya figures in the Nepalese work, as well as in a Cambodian inscription of about 875 A.D., which is of Śaivite inspiration (S. Lévi in *JA*, 1885 p. 412).

c4.s7.t2.9

¹ On the alleged Greek influence on Guṇḍhya's work, see Lacôte, *op. cit.*, pp. 284-86 who argues the opposite way to show that the Greek romance was influenced by the Indian. See Keith, *HSL*, p. 366 f.

c4.s7.t2.10

But much controversy has naturally centred round the value of the Guṇāḍhya legend regarding its testimony on the form of the lost work and its language. The legend speaks of Guṇāḍhya's work being written in Śloka and in the dialect of the wild people of the Vindhya regions, which is called the dialect of the Piśācas or Paiśācī. Daṇḍin, in his *Kāvyādarśa* (i. 38), appears to know the legend in some form, and states that the work was written in the Bhūta-bhāṣā; but he thinks that it was a type of the prose romance known as Kathā, in which, of course, verse was allowed to be inserted. The three existing Sanskrit versions are all metrical, but this need not invalidate Daṇḍin's statement, if Daṇḍin can be presumed to have possessed a direct knowledge of the work already famous in his time. More inconclusive is the evidence regarding the nature and location of the dialect in which the work was composed. In accordance with the legend, the Paiśācī Prakrit is localised² as the dialect of the Vindhya regions lying near about Ujjayinī, but it is also maintained³ that it was a North-western Prakrit of Kekaya and eastern Gandhāra, which is regarded as the ancestor of the group of Dardic dialects now spoken in Kafirstan, Swat valley, Citral and adjacent places. The difficulty of arriving at a final conclusion¹ lies in the fact that the statements of fairly late Prakrit grammarians about Paiśācī Prakrit, as well as the doubtful fragments cited by them as specimens,² are meagre and uncertain. It is also not safe to argue back from the character and location of present-day dialects to those of a hypothetical Prakrit. The designation Paiśācī was perhaps meant to indicate that it was an inferior and barbarous dialect, and the sanction of a vow was required for its employment; but what we know about it from Prakrit grammarians and other sources makes it probable that it was an artificial from of speech nearer in some respects to Sanskrit than the average Prakrit. If it hardened *t* and *d* alone, it is a characteristic which may be equally applicable to a Vindhya dialect influenced by Dravidian and to a dialect of the North-west. The question, therefore, does not admit of an easy solution, although greater plausibility may be attached to the linguistic facts adduced from the Dardic dialects.

c4.s7.t2.11

² Sten Konow in *ZDMG*, LXIV, 1910 p. 95 f and *JRAS*, 1921, p. 244 f; Keith, *HSL* P. 269. Rājaśekhara (*Kāvya-mīmāṁsā*, p. 51) apparently holds the same view. Sten Konow's view, in brief, is that the Paiśācī was an Indo-Aryan language spoken by Dravidians in Central India.

c4.s7.t2.12

³ G. Grierson in *JRAS*, 1905, p. 285 f, *ZDMG*, LXVI, 1912, pp. 49 f, at pp. 73-86, *JRAS*, 1921, p. 424 f, as well as in his *Linguistic Survey*, 1919, Vol, III, pt. 2 and in Hastings *ERE*, under Paiśāca, Vol. X (1918), p. 43 f.

c4.s7.t2.13

¹ Lacòte, *op. cit.*, p. 51 f. Lacòte believes the Paiśācī to be based upon the Indo-Aryan language of the North-west, but spoken by non-Aryan people. He suggests a *via media* by stating that Guṇāḍhya picked up the idea of the dialect from travellers from the North-west, but his sphere of work lay around Ujjayinī! *Cf.* F. W. Thomas, Foreword to Penzer's ed. of *Ocean of Story*, Vol. IV pp. ix-x.

c4.s7.t2.14

² Hemacandra's Prakrit Grammar, ed. Pischel, iv. 303-24; for Mārkaṇḍeya, see Grierson in *JRAS*, 1913, p. 391. For a discussion of the Passages, see Lacòte, *op. cit.*, p. 201 f. Vararuci speaks of one Paiśācī dialect; Hemacandra appears to distinguish three varieties; Mārkaṇḍeya increases the number to thirteen! Different localities are mentioned, but one locality is agreed upon, viz., Kekaya or N. W. Punjab.

c4.s7.t2.15

The exact content and bulk of the original *Bṛhatkathā* cannot also be determined, even to the extent to which we can approximate to those of the original *Pañcatantra*. We have two main sources of knowledge, derived from Kashmir and Nepal respectively, but both of them employ a different medium of expression, and are neither early nor absolutely authentic. The first is given by two metrical Sanskrit adaptations of Kashmir, namely, the *Bṛhatkathā-mañjarī*,³ 'the Bouquet of Great Tale,' of the polymath Kṣemendra, and the *Kathā-sarit-sāgara*,¹ 'the Ocean of Rivers of Tales,' of Somadeva, the latter written between 1063 and 1082 A.D. and the former about a quarter of a century earlier.² Like Somadeva's work, that of Kṣemendra is divided into eighteen Lambhakas,³ but it is of the nature of a condensed abstract, industriously and perhaps (as his other Mañjarīs shows) faithfully compiled. It consists of about 7,500 ślokas, as against more than 21,000 of Somadeva's work; but Kṣemendra makes up for the brevity and dreariness of his narrative by a number of elegant, but mannered, descriptive and erotic passages.⁴ Somadeva, on the other hand, is not anxious to abridge; but he shows considerable restraint in avoiding useless elaboration, and tells his stories with evident zest and in a clear and attractive manner. At one time it was thought that these two Kashmirian versions drew directly from the Prakrit original, but the idea has now been discarded, not only from the comparative evidence of their contents but also in view of the discovery in Nepal in 1893 of the second important source, namely, the *Bṛhatkathā-śloka-saṃgraha* of Budhasvāmin,⁵ which is also in Śloka, but unfortunately incomplete. Its date is unknown, but it is assigned, mainly on the probable date and tradition of the manuscript, to the 8th or 9th century A.D. Although this work is a fragment of 28 Sargas and 4,539 stanzas and also, as its name implies, an abbreviated abstract, its evidence is highly important regarding the existence of two distinct traditions of the text, which show considerable and remarkable divergences.¹

c4.s7.t2.16

³ Ed. Śivadatta and Parab, NSP, Bombay, 1901. Parts of it (introduction and first two stories), translated with the Roman text, by S. Lévi in *JA*, 1885-86.

c4.s7.t2.17

¹ Ed. Durgaprasad and Parab, NSP, Bombay 1889 (reprinted 1903, 1915 etc.). H. Brokhaus edited i-v (with trs), 2 vols. Leipzig 1843, and vi-viii, iv-xviii (text only) in Abh für die Kunde d. Morgenlandes, II and IV, Leipzig 1862 and 1866. The work is well known from its Eng. trs. by C. H. Tawney under the title *Ocean of Story* in Bibl. Ind., Calcutta 1880-87, reprinted with notes and essays, etc., by N. M. Penzer in 10 vols., London 1924-28.

c4.s7.t2.18

² See Bühler, *Über das Zeitalter des kaśmirischen Dichters Somadeva*, Wien 1885. Somadeva wrote the work to please Sūryamatī, princess of Jalandhara, wife of Ananta and mother of Kalaśa. Kṣemendra also wrote most of his works under king Kalaśa of Kashmir.

c4.s7.t2.19

³ The division does not seem to be original, being missing in Budhasvāmin's version, which has Sarga division. The sections are called Gucchakas 'clusters' in Kṣemendra, and Taraṅgas 'billows' in Somadeva, according to the respective titles of their works.

c4.s7.t2.20

⁴ On these descriptive passages, see Speyer, *op. cit.*, p. 17 f. Speyer estimates that Kṣemendra's work contains 7, 561 ślokas, Somadeva's 21, 388.

c4.s7.t2.21

⁵ Ed. F. Lacôte, with trs., Paris 1908-29 (i-xxviii). The work was first discovered by Haraprasad Sastri in Nepal, but its importance was not realised till Lacôte edited the work and published the results of his investigations. The MS is from Nepal, but otherwise there is no sign of the Nepalese origin of the work.

c4.s7.t2.22

¹ See Lacôte, *Essai* cited above, for a discussion of the Kashmirian versions, pp. 61-145, the Nepalese version, pp. 146-195, comparison of the two versions, pp. 207-18, and of the original *Bṛhatkathā*, pp. 1-59.

c4.s7.t2.23

The main theme of both the recensions appears to be the adventures of Naravāhanadatta, son of the gay and amorous Udayana, famed in Sanskrit literature, and his final attainment of Madanamañjukā as his bride and the land of the Vidyadharas as his empire; but in the course of the achievement, he visits many lands and contracts a large number of marriages with beautiful maidens of all kinds and ranks. A vital difference, however, occurs in the treatment of the theme. While the Nepalese recension concentrates upon the main theme and gives a simple and connected narrative, comparatively free from extraneous matters, the Kashmirian recension is encumbered by a stupendous mass of episodic stories, indiscriminately accumulated and remotely connected, regardless of the constant break and obscuration of the original theme. The Nepalese recension, for instance, omits the introductory Guṇāḍhya legend, which occurs in the Kashmirian, and plunges at once into the story of Gopāla and Pālaka and of the love of Gopāla's son for Suratamañjarī, connecting it with the story of Naravāhanadatta, who is made the narrator of the tale of his twenty-six marriages. The Kashmirian authors are apparently aware of this beginning, but the necessity of commencing with the Guṇāḍhya legend and making Guṇāḍhya the narrator of the tale makes them shift the story of Gopāla, Pālaka and Suratamañjarī, and place it, unconnectedly, as a kind of appendix at the end. The Nepalese recension omits also the unnecessary tale of Udayana's winning of Padmāvatī, and does not think it desirable to provide royal ancestry for the courtesan Kaliṅgasenā, mother of Madanamañjukā, in order to conceal the questionable origin of the heroine. In the Kashmirian recension, the hero Naravāhanadatta does not even make his appearance till his birth in Bk. IV (in both versions), but the narrative of the hero is interrupted for two more books by the stories of Śaktivega and Sūryaprabha, who, recognising in the infant the destined emperor of the Vidyādharas, relate their own adventures as aspirants to the same rank. In this way, the main theme is constantly interrupted by a vast cycle of legends, although Kṣemendra and Somadeva are not in perfect agreement, after Bk. IV, regarding the sequence and arrangement of the extra mass of material. It is clear that both the Kashmirian versions do not, in their zeal for collection, succeed in producing a unified or well-constructed work, although the narrative of Somadeva, who is a consummate story-teller, is marked, in spite of its bulk, by greater coherence and desire to preserve, however strenuously, the effect of the main story. The accretions, for example, not only bring in entirely irrelevent stories of Mṛgāṅkadatta and Muktāphalaketu, of expedition to the Camphor Land and the White Island for the winning of Ratnaprabhā and Alaṃkāravatī respectively, but also incorporate the Vikramāditya cycle of legends and interpolate versions of the entire *Pañcatantra* and the *Vetāla-pañcaviṃśati*. All this, with the addition of countless number of small tales, legends and witty stories, would justify the quaint, but appropriate, name of Somadeva's largest collection as the ocean of the streams of stories, and which in their rich mass would make the overwhelmed reader exclaim that here is indeed God's plenty !

c4.s7.t2.24

How far these episodes and legend-cycles belonged to the original *Bṛhatkathā* cannot be precisely determined, but it is clear that much of them is remotely and sometimes confusedly connected with the main theme, and is entirely missing in the Nepalese recension. It is true that Budhasvāmin's work is specially styled a compendium (Saṃgraha) and that his omissions may have been dictated by a desire for abbreviation; it is also possible¹ that Budhasvāmin is an independent writer rather than a mere epitomator, although he may have adhered to Guṇāḍya's narrative in the main. But it is clear, from the way in which the thread of the main story of Naravāhanadatta is kept from being lost in an interminable maze of loosly gathered episodes, that these interruptions or deviations from the predominant interest could not have occurred on a large scale in the original, if we are to presume from its reputation that it was a work of no small literary merit. It seems, therefore, that Budhasvāmin follows the original with greater fidelity² than Kṣemendra and Somadeva, who, apart from minor stories which they individually insert, are following a recension refashioned and much enlarged in Kashmir. In this recension the central theme appears to occupy, after the fashion of Kāvya-poets, a subordinate interest; their essentials are often abridged and throughout sacrificed to the elaboration of subsidiary adventures, as well as to a somewhat confused insertion of tales derived from other sources. Whether this Kashmirian recension was in Paiśācī or in Sanskrit is not known; but Somadeva distinctly speaks of having altered the language, and there are not enough verbal similarities³ between Somadeva and Kṣemendra to warrant the supposition of a common Sanskrit original.

c4.s7.t2.25

¹ Winternitz, *GIL*, III, pp. 315-17.

c4.s7.t2.26

² Lacôte, *Essai*, p. 207 f. Lacôte believes that the Kashmir recension is far removed form the original *Bṛhatkathā*, and was compiled about the 7th century A.D.

c4.s7.t2.27

³ Speyer, *op. cit.*, p. 27 f.

c4.s7.t2.28

In the absence of the original work of Guṇāḍhya, an estimate of its literary merit would be futile. Each of the three adaptations have their own characteristics, which may or may not have been inherited from the original. Kṣemendra's abridged compilation is rapid, dreary and uninspiring, except in ornamental passages, which doubtless show the influence of the Kāvya. Somadeva's larger and more popular masterpiece has been rightly praised for its immensely superior quality of vivid story-telling and its elegantly clear, moderate and appropriate style. Budhasvāmin's abstract, considered nearer to the original, is marked by a sense of proportion both in matter and manner, rapid narration, power of characterisation and simple description, as well as by a more bourgeois spirit and outlook suiting the popular tale; but, in spite of these qualities, it is of a somewhat prosaic cast. It is difficult to say how far all the praiseworthy qualities, if not the blemishes, of these late versions, produced under different conditions, were present in the primary *Bṛhatkathā*, a verbal or even a confident substantial reconstruction of which is wellnigh impossible. To judge, however, from the principal theme, stories and characters, as well as from the general method and outlook, it is possible to assert that Guṇāḍhya must have been a master at weaving into his simple story of romantic adventure all the marvels of myth, magic and fairy tale, as well as a kaleidoscopic view of varied and well-conceived characters and situations. Although Naravāhanadatta is a prince, the story is not one of court life or courtly adventure, nor even of heroic ideals; it is essentially a picture consonant with the middle class view of life and sublimated with the romance of strange adventure in fairy lands of fancy. It is certainly a work of larger and more varied appeal, containing a gallery of sketches from life, romantic as well as real; and Keith is perhaps just in characterising it as a kind of bourgeois epic. The loves of the much-married Naravāhanadatta are perhaps too numerous and too lighthearted, like those of his famed father Udayana, but his chief and best love, Madanamañjukā, has only one parallel in Vasantasenā of the *Mṛcchakaṭika*; while in Gomukha we have a fine example of an energetic, resourceful and wise courtier and friend. It cannot be determined with certainty if the numerous tales of fools, rogues and naughty women existed in the original; but they form an unparalleled store-house of racy and amusing stories, which evince a wide and intimate experience of human life and are in keeping with the humour and robust good sense of people at large.

Linked texts (25)