Pramāṇa · Search · About · API

A History of Sanskrit Literature c19: XII. The Bṛhatkathā and its Descendants

A History of Sanskrit Literature c19

Language
English

Contents

XII. The Bṛhatkathā and its Descendants

1. Guṇāḍhya and the Bṛhatkathā

c19.s1.1

THERE is no doubt that one of our really serious losses in Indian literature is the disappearance of the *Bṛhatkathā* of Guṇāḍhya,¹ a work which ranked beside the *Mahābhārata* and the *Rāmāyaṇa* as one of the great storehouses of Indian literary art. Its existence is asserted first definitely by name in the seventh century when Subandhu, Bāṇa in both his romances, and Daṇḍin in his *Kāvyādarça* attest its fundamental importance. Later references are not rare; we have both the text and the comment of the *Daçarūpa* of Dhanaṁjaya as evidence of its existence, it is mentioned by Trivikrama in his *Campū*, and by Somadeva Sūri in his *Yaçastilaka*, both works of kindred type, while Govardhana celebrates it in his *Saptaçatī*. A Cambodian inscription (*c.* 875) expressly mentions Guṇāḍhya and his aversion to Prākrit, and we cannot doubt for a moment the existence before A. D. 600 of a romantic work by Guṇāḍhya.

c19.s1.2

¹ F. Lacôte, *Essai sur Guṇāḍhya et la Bṛhatkathā* (1908).

c19.s1.3

Of his personality we have an account, reproduced with little variation in three Kashmirian sources, the *Bṛhatkathāmañjarī* of Kṣemendra, the *Kathāsaritsāgara* of Somadeva, and the *Haracaritacintāmaṇi* of Jayaratha. Çiva one day, asked by Pārvatī for a new tale, related to her the substance of the *Bṛhatkathā*, *inter alia*. A Gaṇa Puṣpadanta overhearing it told it to his wife Jayā, who repeated it to Pārvatī. She in great anger cursed Puṣpadanta to lose his rank, which he was not to regain until he had met and related to a Yakṣa Kāṇabhūti the tale he had overheard, the Yakṣa being another unfortunate under a curse. Further, Mālyavant, a comrade of Puṣpadanta who had intervened to intercede for him, was cursed to leave heaven until he should have met and heard from Kāṇabhūti the tale. In course of time Puṣpadanta was born as Vararuci-Kātyāyana at Kauçāmbī; becoming the minister of Nanda, he finally retired to the Vindhya and there told to Kāṇabhūti the tale of the seven emperors of the Vidyādharas, and attained release. Guṇāḍhya meantime had been born at Pratiṣṭhita or Pratiṣṭhāna on the Godāvarī as a reincarnation of Mālyavant; he is in high favour with Sātavāhana, but the latter suffers a severe mortification when during the water play with his wives he is told by his queen not to throw any more on her (*modakaiḥ*), which he in his ignorance of the laws of verbal combination misunderstands as a request to be pelted with sweetmeats—an appalling request if the ancient Indian sweets were like the modern. Dejected, he refuses to be comforted unless he can learn Sanskrit. Guṇāḍhya offers to teach him in six years, but when Çarvavarman the author of the *Kātantra* laughs at this offer, and suggests that he can do it in six months, Guṇāḍhya vows to use neither Sanskrit, Prākrit, nor the vernacular if the deed is done. It is accomplished and Guṇāḍhya wanders disconsolate in the Vindhya, where Kāṇabhūti meets him and relates the tales learned from Vararuci. Guṇāḍhya would record them, but must write in Paiçācī, the language of the goblins, as he is debarred from use of any other speech by his vow. His disciples take the vast work to the king Sātavāhana, who rejects it. Guṇāḍhya recites it to the beasts and birds, burning the manuscript as he proceeds; the beasts, intent on the sweet poetry, become thin, and the cooks in the royal kitchen no longer serve good soup. Hence the marvel is revealed and the king saves one-seventh of the 700,000 Çlokas of the original, the tale preserved in the *Bṛhatkathā*. The Nepalese version contained in the *Nepālamāhātmya* is different. We hear nothing of Vararuci-Kātyāyana, there is but one sinner Bhṛṅgin who enters the private room of Çiva and Pārvatī in bee form; he is reborn at Mathurā as Guṇāḍhya, becomes a Paṇḍit of king Madana of Ujjain, is vanquished by Çarvavarman, and is advised to write in Paiçācī by a seer Pulastya. Nothing is said of the pledge as to language, naturally enough, for Nepal lay outside the interest on this point of India proper.

c19.s1.4

The legend seems to have been known already in some form to Bāṇa, and therefore must be moderately old; how far and in what form it goes back to Guṇāḍhya, it is idle to say. The location of Guṇāḍhya is clearly different in the two sources, for it is vain to seek to make out that there has been confusion between Pratiṣṭhāna on the Godāvarī and a place of like name at the junction of the Ganges and the Yamunā. What is clear is that Ujjain or Kauçāmbī was the scene whence Guṇāḍhya derived much of his inspiration, which is a very different thing from the place where he was in royal honour and composed his work. The cennexion with Sātavāhana, which the Kashmirian recensions suggest, is borne out to some extent by certain facts. In the first place, the Sātavāhanas were at one time patrons of Prākrit as opposed to Sanskrit literature; the evidence of the inscriptions¹ shows that Sanskrit was used by their Kṣatrapa rivals before they adopted it, and the Māhārāṣṭrī lyric flourished under them. Secondly, the mention of the study of Sanskrit in this connexion does suggest that there was a tradition regarding the time when the Sātavāhanas determined to copy the Kṣatrapas and Sanskrit became popular in court. Further we cannot go.

c19.s1.5

¹ Bloch, *Mélanges Lévi*, pp. 15 f.; Lévi, JA. 1902, i. 109 ff.

c19.s1.6

Nor can we say anything definite of the date of Guṇāḍhya. The connexion with the Sātavāhanas after all means nothing definite even if real, and the most important evidence we could have would be a clear² allusion in literature to, or employment of, the *Bṛhatkathā* before Daṇḍin or Bāṇa. It may be³ that Bhāsa's dramas drew some inspiration from this source, but we have no strict proof. We can fairly claim that Guṇāḍhya is not later than A. D. 500, but to place him in the first century A. D. is quite conjectural, nor in reality is any other later date more assured.

c19.s1.7

² The supposed Tamil version of the 2nd cent. A. D. (S. K. Aiyangar, *Ancient India*, pp. 328, 337) is too dubious in date to be evidence. The alleged version into Sanskrit by Durvinīta (? 6th cent.) is quite dubious (R. Narasimhachar, JRAS. 1913, pp. 389 f.); see Fleet, JRAS. 1911, pp. 186-8.

c19.s1.8

³ Denied by Hertel, *Pāla und Gopāla*, pp. 153 f.; cf. P. D. Gune, *Ann. Bhand. Inst.*, ii. 1 ff.

c19.s1.9

Obscure also is the question of the form of the work. The Kashmirian version suggests that what Guṇāḍhya produced was a work in Çlokas, but that may be quite misleading, and on the other hand we have the express statement of Daṇḍin that a Kathā to which type he refers the *Bṛhatkathā* was written in prose. Verses may have been interwoven as in the case of the *Jātakamālā*, but this must remain a mere hypothesis, and there is no other evidence to invalidate the impression given by Daṇḍin. A prose citation by Hemacandra may conceivably be from the *Bṛhatkathā*, but it would be quite idle to assert that it was; it may have come from a later recension or from some other source.

c19.s1.10

The dialect used was Paiçācī, and over this term a controversy has raged, accentuated by the fact that we really cannot be sure that we have a single relic of the *Bṛhatkathā*, still less that so late a grammarian as Mārkaṇḍeya (17th cent.) actually¹ had the text before him. A further confusion has arisen from Sir G. Grierson's decision to group a certain number of north-western dialects, spoken in Kafiristan, the Swat valley, Chitral, and Gilgit, as Piçāca languages, claiming both that they have a true relation to the ancient Paiçācī dialect, and were so called because the speakers were cannibals, and thus styled Piçācas, eaters of raw flesh, by their neighbours. The assertions of the grammarians are confused and unsatisfactory, nor is the matter improved by the existence of two schools of Prākrit grammar with divergent traditions and views, especially as these are represented by comparatively late texts. But, as we have seen, there is more probability that Paiçācī was a dialect rather of the Vindhyas than of the north-west; the hardening of *d* to *t* or of other soft letters is not, as Grierson's theory requires to make it plausible, solely a feature of the north-west, but occurs in other dialects including Pāli, and the fact that Paiçācī has but one sibilant prejudices its claim to be akin to the north-west dialects which in Açoka's time and later preserve distinctions.² Lacôte, however, while accepting connexion with the north-west, agrees with the view that the phenomenon of hardening is a sign of the use of an Aryan speech among a non-Aryan people, and holds that Guṇāḍhya adapted this dialect to literary purposes, avoiding any too serious deviations from Sanskrit, and, if we substitute a Vindhyan dialect spoken in a Dravidian area, we probably approach the truth. At least for the connexion with the Vindhya we have the clear assertions of the Kashmirian recension, which had no special motive for misrepresentation of the facts, and the testimony of Rājaçekhara¹ is clear in favour of the actual use of Paiçācī in a wide region, including the Vindhya area. This view is much more plausible than Lacôte's suggestion that Guṇāḍhya picked up the idea of the dialect from some visitors from the north-west, his sphere of work lying round Kauçāmbī and Ujjain, and Grierson admits that, even if originally a north-western dialect, Paiçācī might have been carried to the Vindhya.

c19.s1.11

¹ As Grierson asserts, AMJV. i. 121; JRAS. 1913, p. 391. All that is said is *Bṛhatkathāyām*, and common sense forbids us to assume that Mārkaṇḍeya used it, or that the quotation is really from Guṇāḍhya's own text and not, for instance, the Kashmirian version.

c19.s1.12

² Chap. i, § 4.

c19.s1.13

¹ *Kāvyamīmāṅsā*, p. 51.

c19.s1.14

It is impossible to determine with precision the content of the *Bṛhatkathā*; our sources are too slight, but we can gather a general impression of the task accomplished by Guṇāḍhya. The sources on which he drew were, it is clear, three in number. The *Rāmāyaṇa* gave him the *motif* of the search of a husband for a wife cruelly stolen from him soon after a happy marriage; from Buddhist legends and other traditions of Ujjain and Kauçāmbī he was deeply familiar with the tales of Pradyota or Mahāsena and the gallant and dashing hero Udayana,² whose love-adventures were famed for their number and variety; he was also in touch with the many tales of sea-voyages and strange adventures in far lands which were current in the busy centres of Indian trade, and with the abundant fairy-tales and legends of magic current in India. From the latter source and from Buddhist legend he derived the conception of the emperor, Cakravartin, who is the secular counterpart of the Buddha; Naravāhanadatta, his hero, is born with the thirty-two auspicious signs which assure him Buddhahood if he enter the ascetic life, universal dominion if he remain in the affairs of the world. But the empire is not of this earth; it is essentially a fairy land, the realm of the Vidyādharas, who dwell beyond the formidable defences of the Himālayas and who by reason of their magic powers have semi-divine attributes. The Vidyādharas do not appear early in Indian religion, but we can recognize easily enough in them a contamination of the old ideas of the Gandharvas with notions derived from the mysterious powers of Hindu seers and ascetics and Buddhist saints. The hero is a son of Udayana, and in effect is Udayana revised and remodelled for his new destiny, while the *Rāmāyaṇa* lends the decisive element of the plot, the rape of Madanamañcukā or Madanamañjukā by Mānasavega, and the efforts of her husband to discover her, in which he has the aid of his faithful minister Gomukha. His success is accomplished simultaneously with his winning the empire of the Vidyādharas, just as the recovery of Sītā is followed forthwith by the royal consecration of Rāma. But there must have been a vital difference in the tales, for Guṇāḍhya clearly was the poet not of kings so much as of the merchants,¹ the traders, the seafarers, and even the handicraftsmen of his day; his epic was a bourgeois epic, and in lieu of the stainless purity of Rāma we have as hero a son of Udayana, even lighter in love, despite his affection for Madanamañcukā, than his father. Hence we find that certainly even in the original there must have been much said of Naravāhanadatta's other loves and many a tale of adventurous journeying as well as Märchen and fairy lore. In Gomukha we have a picture of a minister such as is Yaugandharāyaṇa in the dramas of Bhāsa, bold, energetic, courageous, if slightly devious in modern views as regards choice of expedients. The portrait of Madanamañcukā was clearly definite; she was, like Vasantasenā in the *Cārudatta* of Bhāsa and still more clearly in the *Mṛcchakaṭikā*, a hetaira who hated her position, and whose great aim was to be recognized as a woman of family (*kulastrī*), and thus be permitted legitimate marriage in lieu of compulsory polyandry. We have here perhaps a valuable chronological hint, if we could be sure that it was from the *Bṛhatkathā* that Bhāsa really drew the picture. It is striking at least, however, that the description of the eight courts and the garden of the palace of Vasantasenā in the *Mṛcchakaṭikā*, but not in the *Cārudatta*, should correspond minutely with the description of the place of Kaliṅgasenā given in the *Bṛhatkathāçlokasaṁgraha* of Budhasvāmin.

c19.s1.15

² Cf. Przyluski, *La légende de l'empereur Açoka*, pp. 74 ff.; J. Hertel, BSGW. lxix. 4 (1917); Lacôte, *JA.* 1919, i. 493 ff.; P. D. Gune, *Ann. Bhand. Inst.*, ii. 1 ff.; Burlingame, HOS. xxviii. 51, 62 f., 247-93.

c19.s1.16

¹ Cf. Foucher, *L'Art Gréco-Bouddhique du Gandhāra*, ii. 102 ff.

c19.s1.17

Guṇāḍhya's influence is seen also in Daṇḍin who borrowed from him, we may fairly assume, the conception of placing his kings' sons, fallen in station by the action of misfortune among vagabonds, in positions where a series of adventures drawn from low life is allied to marvellous happenings of every kind. The arrangement indeed of the story may be due to the same cause, for it resembles the scene in which Naravāhanadatta and his friends, reunited after separation, recount their adventures to one another. The fantasy of Guṇāḍhya lives on also in the *Yaçastilaka* of Somadeva Sūri and in the *Tilakamañjarī* of Dhanapāla, both of whom recognize the importance of Guṇāḍhya.¹ Moreover, the name of his hero seems from his use of it to have won acceptance in royal usage as a suitable title for a prince as well as in literature. But his enduring memorial is furnished by the versions of the *Bṛhatkathā* which have reached us.

c19.s1.18

¹ The degree of his originality may, of course, be questioned, and no poet is without some predecessor; but his success points to a very real creative power, which permits us justly to ascribe to him the creation of the genre.

2. The Bṛhatkathāçlokasaṁgraha of Budhasvāmin

c19.s2.1

Budhasvāmin, the author of the *Çlokasaṁgraha*, abbreviation in Çlokas of the *Bṛhatkathā*, is no more than a name to us.² The manuscripts of his work are from Nepal, but there is no mark otherwise of his Nepalese origin, which must remain merely a conjecture. The form of name is not modern; but, as it is attested from early days down to the twelfth century, the probable date of one of the manuscripts, we reach no satisfactory result from that. If he is assigned to the eighth or ninth century, it is without any special ground save that the manuscript tradition suggests that a long time has elapsed before the extant manuscripts came into being.

c19.s2.2

² Ed. and trans. F. Lacôte, 1908 ff.

c19.s2.3

The work preserved is merely a fragment, though there is no adequate reason to hold that it is defective at the beginning or that it ever contained anything as to the origin of the collection of tales comparable to the legend in the Kashmirian versions and the *Nepālamāhātmya*. It is divided into cantos (*sargas*), of which twenty-eight survive, probably a mere fraction of the original, though it extends to 4,539 verses. We are carried at once *in medias res*; Pradyota dies, and is due to be succeeded by Gopāla, but the latter, learning that he is credited with having disposed of his father, insists on his brother Pālaka reigning in lieu (i). Pālaka is a bad ruler, and is induced by what he deems divine suggestion to abandon his throne to Avantivardhana, Gopāla's son (ii). The latter falls in love with the daughter of a Mātaṅga, Surasamañjarī, who, like her father, is really of the race of the Vidyādharas; he marries her, only to be snatched away with his bride by Ipphaka (Ityaka) a jealous Vidyādhara; they are rescued by another of these genii, and the emperor Naravāhana pronounces judgement in favour of the marriage (iii). The seers so admire the emperor's judgement that they demand from him the account of his achievement of empire; he consents to tell of the twenty-six marriages but only when Gaurī undertakes that his revelations will be kept religiously secret. He then tells the desire of his father Udayana for a son, which ultimately is granted (v, vi). When young Naravāhana grows up, he shows the signs of a Cakravartin, and a Vidyādhara, Amitagati, recognizes them and attaches himself to him; finally he wins the hand of Madanamañjukā, daughter of Kaliṅgasenā who, however, is a hetaira, thus rendering a true union impossible (vii–xi). One day Madanamañjukā disappears, but is found under an Açoka; she relates that Kubera has demanded that she should be really married to Naravāhanadatta; this desire is conceded, but shortly after the king makes the unpleasant discovery that in lieu of his beloved he is really consorting with Vegavatī. She reveals herself to him as sister of Mānasavega, a Vidyādhara, who has taken Madanamañjukā, but who cannot do her harm, just as Rāvaṇa could not put force on Sītā in her captivity. Naravāhanadatta celebrates a new marriage with her, but immediately after he is carried off by Mānasavega; falling to earth, he finds himself in a well but is rescued (xii–xv). He is now lost, and posing as a student commences a new adventure, ending in marriage with Gandharvadattā, daughter of Sānudāsa whose history is narrated at length (xvi–xviii). Two further marriages are in wait for him, that with Ajināvatī (xix, xx), and that with Priyadarçanā, whose bosom he recognized when it was revealed for a moment when she was posing as a merchant (xxi–xxvii). The next canto gives us only the beginning of a new marriage adventure, and, as so many more were still before him, the extent of the work can be guessed.

c19.s2.4

There is much to prove that Budhasvāmin followed far more faithfully his original than the Kashmirian authors. Assuming that the *Çlokasaṁgraha* was written on the same scale throughout, it may have contained 25,000 verses, certainly an adequate number but not necessarily excessive. On the other hand, comparison of relevant portions of the work with the *Kathāsaritsāgara* shows that the latter is very greatly abridged in the vital parts of the narrative, those intimately connected with Naravāhanadatta. It is a reasonable conclusion, therefore, that the Kashmirian versions contain much added matter, especially the episodes which are merely in nominal connexion with the main story. This impression is certainly strengthened by the fact that the character of Madanamañcukā and her relations with Naravāhanadatta are much more coherently set out in the *Çlokasaṃgraha*; in the Kashmirian versions both her mother and herself are provided with royal connexions, Kaliṅgadatta and Madanavega, in order to spare us the discomfort of seeing a king marry a lady of the demi-monde. The bourgeois character of Gandharvadattā and her merchant father are similarly minimized in the Kashmirian version; they spare Ajināvatī, because she was a princess, but omit Priyadarçanā as being of middle-class origin. The *Çlokasaṃgraha* again in many details serves to explain obscurities in the Kashmirian version and to motive adequately incoherent episodes. On the other hand, it is fair to note that Budhasvāmin assumes that we know the tale of Udayana, and that we need not doubt that in the original *Bṛhatkathā* it was recognized, though Budhasvāmin preferred to confine his work to the adventures of Naravāhanadatta. From the paucity of his episodes we may fairly conclude that these were not over-numerous in the original, though it is impossible to stress this point.

c19.s2.5

Budhasvāmin is unquestionably worthy of praise for his art. Admitting his debt to Guṇāḍhya does not diminish the pleasure afforded by his lively outlook on life, the complex picture of adventure and marvels which he paints, or the romance of his well-conceived characters and the kaleidoscope of the swiftly altering scenes in which they are placed by fate or their own action. He restrains his desire for mannered description of which he doubtless felt competent by the necessity of getting on with the tale, and displays his virtuosity, partly by his large vocabulary with its not rare Sanskritizations of Prākrit terms which are doubtless sometimes derived from him by the lexicographers, and partly by the revival of obsolescent forms such as aorists. As a rule, he is simple, clear, fluent without verbosity, and if he seems on the whole rather devoid of ornament the magnitude of his undertaking may be deemed excuse enough for a very venial fault.

3. The Kashmirian Bṛhatkathā

c19.s3.1

The older view that the *Kathāsaritsāgara* and the *Bṛhatkathāmañjarī* were directly drawn from the *Bṛhatkathā* cannot be retained¹ in view of the discovery of the *Çlokasaṁgraha*. The Kashmirian recensions show themselves at once as vitally similar in contrast with the Nepalese and leave no option but to assume that they are derived from one source, and that not the original *Bṛhatkathā*. The date of this form of the *Bṛhatkathā* is clearly impossible to decide beyond that it must have been considerably before A.D. 1000. Nor can we say who the author was, or by what process the work assumed form. It may have been the outcome of a continued process of change if the story was regarded as specially attractive. All that can be conjectured is that the work received its final form through two main processes. In the first place, the essentials of the legend of Naravāhanadatta, including his parentage, were extracted from the original of Guṇāḍhya, and abbreviated. Then, secondly, the account was expanded and completed by inserting as satisfactorily as was possible other great legend-complexes which were popular in Kashmir, making a work essentially different from the original *Bṛhatkathā* because the original theme, the adventures of Naravāhanadatta, had fallen into a position of subordinate interest and the episodes had become of predominant importance. Which the additions were it is, of course, frankly impossible² to say on the strength of the present evidence; the absence of the rest of the *Çlokasaṁgraha* deprives us of the one useful control. But we may reasonably hold that the additions included both the version of the *Pañcatantra* and that of the *Vetālapañcaviṁçatika* which are found in both Kṣemendra and Somadeva, but which have plainly no real or original connexion with the legend of Naravāhanadatta.

c19.s3.2

¹ Despite F. D. K. Bosch, *De legende van Jīmūtavāhana* (1914), pp. 85 ff.

c19.s3.3

² Subandhu may have known the Vikramāditya legends (cf. *Vāsavadattā*, p. 110).

c19.s3.4

The language and form of the new text do not permit of precise determination. It is possible that the references to Paiçācī forms and citations in the case of Hemacandra are derived from this Kashmirian text, and if so they would show that the work was handed down in a form of Paiçācī. Nor, of course, is there anything implausible in such a proceeding. It is not rare for a dialect once established to remain in use for a certain work after the original has been changed. We have the perfectly clear statement of Somadeva that the language was altered, and this can hardly mean anything less than a translation. If the original had been in Sanskrit, it seems incredible that it would not have influenced both Kṣemendra and Somadeva sufficiently to cause frequent verbal similarities, and this is not the case. The similarities which do occur, as for instance in the stories of the *Pañcatantra*, can easily be explained by the fact that both authors were dealing with a work in a dialect which admittedly was considerably more Sanskritic than the ordinary Prākrit; indeed, on one list of the relative position of Prākrits Paiçācī is ranked after Sanskrit in honour.

4. Kṣemendra's Bṛhatkathāmañjarī

c19.s4.1

The work of Kṣemendra¹ was probably produced in his youth like the Mañjarīs of the *Mahābhārata* and the *Rāmāyaṇa* which he composed, perhaps in accordance with his own doctrine that the would-be poet ought to undertake exercises of this kind. The character of these abridgements is well known; they are dry and sober, reproducing faithfully, though with much omission and curtailment often to obscurity, their originals, but depriving them of all life and attraction. Kṣemendra has, moreover, in lieu of seeking to write interesting summaries, thought it enough to relieve the barrenness of his versions by interpolating elegant descriptions at intervals, a procedure not to be commended, as it merely adds to the bulk of the works without serving any useful purpose. But his accuracy, which we can test for the epics, is assured, and therefore we may *a priori* assume that his account of the contents of the *Bṛhatkathā* of Kashmir accords with reality.

c19.s4.2

¹ Ed. KM. 69, 1901. Cf. Bühler, IA. i. 302 ff.; Lévi, JA. 1885, ii. 397 ff.; 1886, i. 216 ff.; Speyer, *Studies about the Kathāsaritsāgara*, pp. 9 ff.

c19.s4.3.1

It appears from the coincidence of the two recensions that the original was divided into eighteen Lambhakas as its main divisions, and it is a plausible conjecture that the term applies to the victories of the hero, each section dealing with some achievement of his. As we have it in both our sources, the work begins with the Kathāpīṭha, which gives as an introduction to the tale the legend of Guṇāḍhya already noted. In Book ii, the basis of the story is furnished in the adventures of Udayana, which are carried in iii to his winning of Padmāvatī, the book taking its style Lāvānaka from the place where the first queen, Vāsavadattā, was reported to have perished, a necessary preliminary to the second espousals. In iv we have the birth of the hero, Naravāhanadatta, who is to be the emperor of the Vidyādharas. The next book, Caturdārikā, is decidedly episodical. The Vidyādhara Çaktivega comes to visit the future sovereign, and relates how he himself has reached the wonderful city of the Vidyādharas and won the four beauteous maidens whence the title of the book is derived. From this point Kṣemendra and Somadeva diverge vitally. Kṣemendra continues with the legend of Sūryaprabha (vi), a strange and remarkable tale, of how that hero rose from royal rank to becoming emperor of the Vidyādharas after a desperate struggle against his foe Çrutaçarman, who was finally induced to content himself with a minor kingdom, thanks to the direct personal intervention of Çiva himself.

c19.s4.3.2

The tale is remarkable in its obvious blending of mythology involving Vedic and epic beliefs, Buddhist legends, and popular story matter; but in Kṣemendra's hands it suffers greatly from excessive condensation. The two books, it will be seen, have a certain relevance to each other and to the work as a whole, despite their episodic character; they deal with the career of other aspirants to emperorship over the Vidyādharas. In vii we return to the main story a little more clearly. The essence of the book is a long account of Kaliṅgadatta, father of Kaliṅgasenā, who serves merely to give his daughter a royal ancestry; Udayana is sought in marriage by her and he would gladly wed her, but Yaugandharāyaṇa resists the match, lest the king should become too much enamoured of his wife and neglect his duties, a ludicrous excuse seeing that he had already arranged two marriages for the prince. Doubtless in a more original form it was Kaliṅgasenā's character as hetaira which motived the objections. At any rate Udayana is induced to abandon the project, but he determines to allow her daughter to wed Naravāhanadatta, and the book carries us to his consent to a formal marriage. Book viii, which is very short, is styled Velā after the name of the character of whom and her husband a legend is narrated, quite episodically, but it ends with the vital statement that Madanamañcukā has been abducted by the Vidyādhara Mān his beloved he has to be the hero of four episodic books (ix–xii).

c19.s4.3.3

In the first he is carried off in sleep and ends by espousing another Vidyādhara maiden, Lalitalocanā, with whom he spends time on mount Malaya, but is saddened by longing for Madanamañcukā; Lalitalocanā disappears, but a hermit, Piçaṅgajaṭa, comforts him by telling him the tale of Mṛgāṅkadatta, a prince of Ayodhyā, who won in marriage Çaçāṅkavatī, daughter of his enemy Karmasena of Ujjain, who gives the book its name (ix). The next consolation is administered by Kaṇva and consists in the narration of a vast cycle of legends of the emperor Vikramāditya, though it is inconceivable that Guṇāḍhya himself could have been guilty of so flagrant an anachronism; the title is Viṣamaçīla (x). In xi, Madirāvatī, the prince is encouraged to persevere by the tale of two Brahmins who by manly effort (*puruṣakāra*) succeeded in defying the decrees of fate (*karman*) and achieving their desire, and he also recovers the missing and apparently not much regretted, Lalitalocanā. Yet another episode follows: Gomukha tells the tale of the emperor Muktāphalaketu and his beloved Padmāvatī, who gives the book its name (xii).

c19.s4.4.1

After this long interlude action is resumed in Book xiii, Pañca, so called because in it the prince wins five more brides, Vidyādhara maidens who are determined to espouse him. The main business, however, of the book is the effort to attain Madanamañcukā. With the aid of Prabhāvatī, a Vidyādhara, the prince penetrates to her place of confinement, using a woman’s form lent by Prabhāvatī; as she, however, has to resume it, suddenly he is discovered and Mānasavega has him tried by the court of the Vidyādharas, but will not accept their decision in his favour. Prabhāvatī takes him in safety away from the Vidyādharas; ultimately he reaches Kauçāmbī, and many Vidyādharas join him for an attack on his foes. After great efforts, he attains Çiva’s favour, and in a great battle slays Gaurīmuṇḍa and Mānasavega in single combat. He prepares to attack his remaining foe Mandaradeva in the north of Kailāsa, and marries the five damsels who seek his love. The next step ought obviously to be the attack on Mandaradeva, as it is in Somadeva, but there now occurs a long series of episodes which doubtless had been inserted here in the Kashmirian *Bṛhatkathā*. In Book xiv he marries Ratnaprabhā, whose name the book bears, and pays an important visit to the land of camphor, returning in a flying machine of the kind which the Yavanas, Greeks, were experts at constructing. In Book xv we have a sort of duplication of this adventure;

c19.s4.4.2

he marries Alaṁkāravatī, and proceeds to an expedition to a White Island or Continent¹ where he worships Nārāyaṇa with an elaborate prayer written in the most finished Kāvya style; the parallel to the famous episode of the *Mahābhārata* in which sages seek the Çvetadvīpa and take part in the worship of a wonderful deity—which has been deemed a reference to actual experience of Nestorian rites or even of Alexandrian Christianity—is complete, and suggests very strongly that the Kashmirian or the original *Bṛhatkathā* borrowed the episode from the epic as we know it. The next book (xvi) is much more banal; it gives the prince another wife, Çaktiyaças, and imparts a number of unimportant episodes. We resume now in Book xvii the lost thread. Before he can attack Mandaradeva, Naravāhanadatta must receive from the sage Vāmadeva on mount Malaya the seven jewels, emblematic of sovereignty. He then reaches the north by passing under a great tunnel, and by his offer of his own head induces the dread Kālarātri, who guards the exit, to permit his passage. Mandaradeva falls, five more maidens are wed—a repetition of the *motif* in Book xiii, and the great consecration, Mahābhiṣeka, whence the book is named, is duly celebrated, the emperor insisting on his father being present. The work is now finished, but very inconveniently a further book (xviii) is necessary;

c19.s4.4.3

under the style Suratamañjarī it tells how, after the death of Pradyota and Udayana, Gopāla and Pālaka resigned their tenure of the kingship of Ujjain, how Avantivardhana wedded the heroine, and how the two were protected against a jealous Vidyādhara by the emperor. The only excuse for this absurd position of the tale is the fact of the existence of Book i with its account of the telling of the tale by Guṇāḍhya. In the original, as the Nepalese version shows, the episode of Suratamañjarī led up to the telling by Naravāhanadatta himself of his adventure, which would have clashed with the version of Book i, and the old exordium was, therefore, relegated to an appendix. This view is confirmed by the fact that Somadeva in his Book vi expressly tells us that Naravāhanadatta is relating his adventures in the third person, an admission that he knew that the tale of Suratamañjarī had originally been placed at the beginning of the work. On this point Kṣemendra is silent in his corresponding Madanamañcukā book (vii), but he reveals the fact in the summary (*upasaṁhāra*) with which he ends his poem, for he tells us, for the first time, that the work is supposed to be set forth by Naravāhanadatta to the sage Kaçyapa on a visit.

c19.s4.5

¹ Cf. W. E. Clark, JAOS. xxxix. 209-42; Garbe, *Indien und das Christentum*, pp. 193 ff.; Grierson, IA. xxxvii. 251 ff., 373 ff.

c19.s4.6

Two other points at once stand out revealing the defect of the original Kashmirian recension. The break between the end of Velā (viii) and the continuation in Pañca (xiii) is lamentable; but its harshness is concealed in some measure by making the intervening books recognize the plight of the prince and the endeavour to console him during his search. Evidently it was thus that the compilers of the Kashmirian recension hoped to work in not too awkwardly their extra matter, and in a sense they succeeded. The same thing cannot be said regarding the interpolation of Books xiv–xvi between Pañca and the book of triumph and consecration. The break is ludicrous; Naravāhanadatta, who is left at the end of Pañca as accepted as lord by the great majority of the Vidyādharas, but who has Mandaradeva still to overcome, is now treated for three books as a prince in the home of his father, without any consciousness of his great adventures or his imperial dignity in the land of the Vidyādharas. Here the compiler had evidently not the skill to make even a passable transition, and Kṣemendra loyally followed his incoherence. This is conclusive evidence against the original *Bṛhatkathā* ever having contained this material; no author would permit himself such confusion, while a compiler could easily slip into it when he desired to knit together varying cycles of legend.

5. Somadeva's Kathāsaritsāgara

c19.s5.1

Somadeva, a Brahmin of Kashmir, son of Rāma, wrote the *Kathāsaritsāgara*¹ between 1063 and 1081 in order to divert the troubled mind of Sūryamatī, a princess of Jalandhara, wife of Ananta and mother of Kalaça, his work falling, therefore, a considerable period after that of Kṣemendra. In addition to the division into Lambhakas Somadeva has one of his own composition into Taraṅgas, 124 in all, the name, ‘billows’, being chosen obviously in relation to the title of the work, which is most naturally analysed as ‘Ocean of the Rivers of Stories’, rather than with Lacôte as ‘(Bṛhat-) Kathā, an Ocean of Rivers (of Stories)’. These divisions are not original; Kṣemendra, indeed, has subdivisions for some of the longer books which he calls Gucchas, ‘clusters’, in the older manner. Kalhaṇa apparently was influenced in his choice of title for his chronicle by Somadeva.

c19.s5.2

¹ Ed. Durgāprasād, NSP. 1903; trans. C. H. Tawney, BI. 1880–4. Cf. J. S. Speyer, *Studies about the Kathāsaritsāgara* (1908).

c19.s5.3

Somadeva sets out by telling his purpose, and one stanza of his has caused trouble, evoking different renderings from Hall, Lévi, Tawney, Speyer, and Lacôte:

c19.s5.4

*aucityānvayarakṣā ca yathāçakti vidhīyate*
*kathārasāvighātena kāvyāṅçasya ca yojanā.*

c19.s5.5

The sense of this stanza appears to me clear: ‘Literary convention and the connexion of topics have been presented as best I could, as well as the arrangement of a part of the poem so as not to offend against the sentiment of the story (or the story and its sentiment).’ We have, it seems, a recognition of the fact that there has been change of order, and that it was made in order to preserve the sentiment in the tale. This accords exactly with what we find in the arrangement. In the first five books there is no change. But for the rest Somadeva was dominated by his desire to preserve the effect of the poem, and obviously this compelled the breach of the gap between Pañca and Mahābhiṣeka; in his text the transition is perfect; the former book ends with the decision of the prince to obtain the jewels necessary for the coronation of a would-be emperor, and the next book carries on the proposal, though in a slightly casual manner which Somadeva has not altogether obliterated. This left him, however, with three books to fit in, Ratnaprabhā, Alaṁkāravatī, and Çaktiyaças, and obviously necessitated a complete overhauling of the earlier part of the poem in order not to overburden it. The solution adopted was to fit these three books, which all deal with adventures of the prince before he became emperor, in the space before Pañca and to eliminate from the earlier matter two books, which could, as not dealing with the hero's own adventures but merely being stories told to him, be fitted in as an appendix, that is the books Padmāvatī and Viṣamaçīla. The arrangement of the material before Pañca is carried out artistically in so far as an effort is made to interpose books mainly episodic with those giving important if incidental acts of the hero. Thus after the fifth book which is episodic we have the important book Madanamañcukā (vi); this is followed by the Ratnaprabhā (vii); the Sūryaprabha (viii), which intervenes before Alaṁkāravatī (ix), is essentially merely episodic; Çaktiyaças (x) runs naturally on from Alaṁkāravatī as containing incidental stories; then follow Velā (xi), Çaçāṅkavatī (xii), Madirāvatī (xiii), the all-important Pañca and Mahābhiṣeka (xiv and xv), and, by way of appendix, Suratamañjarī, Padmāvatī, and Viṣamaçīla (xvi–xviii). One change in the actual contents of a book was necessary. In Kṣemendra and probably in the original Velā was not merely episodic; it contained at the close the vital element of the disappearance of Madanamañcukā, which explains the grief of the king alluded to in the following books. Nothing of this sort accorded with Somadeva's plan of working in the books Ratnaprabhā, Alaṁkāravatī, and Çaktiyaças, and therefore the allusion had to disappear, although it was not possible for Somadeva to avoid leaving occasional traces in the books before Pañca in his order that Madanamañcukā had already been lost.

c19.s5.6

We may admit at once that despite his efforts Somadeva has not succeeded in producing a unified work. But the merit of the *Kathāsaritsāgara* does not rest on construction. It stands on the solid fact that Somadeva has presented in an attractive and elegant if simple and unpretentious form a very large number of stories which have for us a very varied appeal, either as amusing or gruesome or romantic or as appealing to our love of wonders on sea and land, or as affording parallels to tales familiar from childhood. Kṣemendra’s example shows how, by undue condensation and obscurity, tales can lose all point and interest; Somadeva, how by care the point can be fully expressed without fatigue to the reader. We meet with the old but still amusing tales of fools, scattered in the *Kathāsaritsāgara* among the tales of its version of the *Pañcatantra*, but collected together after it by Kṣemendra; chance proves that half at least go back to a collection made before A. D. 450, used in a work by a monk Ārya Saṅghasena, and rendered into Chinese by his pupil Guṇavṛddhi in 492.¹ We hear once more of the foolish servants who, bidden protect the leather of the new trunks, take out the clothes in them and thus protect them against the rain, of the fool who insists that his father never violated chastity and that he must have been a mind-born son, of the fellow who filled himself with seven cakes and then bitterly lamented that he had not eaten the seventh first and saved the rest, and we may, if young enough, still laugh with the stones at these japes. Rogues who prosper lend another series of tales; one is ingenious; dressed as a rich merchant, he craved an interview with the king, to whom he promised for the honour of a daily repetition of the audience a gift each time of 500 *dīnāras*; the king accedes, and the courtiers, thinking that he is all-powerful with their master, bribe him until he has fifty million gold pieces, which he has the good sense to share with the king to whom he reveals his effective ruse. Much is told also of the thief, gambler, roué, but always brilliant Mūladeva,² who is in Indian literature the *beau idéal* of a perfect cheat but who has a son even wilier than himself. Another rascal is so clever that we may forgive him his evil deeds; he is to suffer after death an age in hell by reason of his misdeeds, but a single gift to a pious person entitles him to one day’s life as Indra. Of this he takes opportunity to gather all his friends and to traverse with them the sacred places of India, thus acquiring such merit as to remain Indra. But yet this god also is lightly treated; still more often are religious ascetics denounced; one of them who in order to get into his possession a pretty girl frightens her father into exposing her in a chest Danaë-like, finds the wrong chest, and has his nose and ears bitten off by an ape, while the girl is rescued by a prince.

c19.s5.7

¹ Hertel, *Ein altindisches Narrenbuch*, BSGW. 64, 1912. Cf. the story of the foolish monkeys (Jāt. 46 and a Bharhut scene, GIL. ii. 108).

c19.s5.8

² PAPS. lii (1913).

c19.s5.9

A book of tales about women seems to have been used by the compilers of the Kashmir recension, to judge from the mass of stories, unhappily often to their disadvantage; we hear of murderous women, of one who mutilates her husband in revenge for a beating, of one who regularly betrays him but insists on burning herself on his pyre, of the woman who got rid of ten husbands, and apparently met her match in the man who had disposed of ten wives, but defeated him also and became so unpleasantly notorious that she turned into an ascetic. Full of reminiscences of various Märchen *motifs* is the tale of the king whose white elephant can be healed only by the touch of a chaste woman; none of 80,000 in the kingdom can help it, until a poor young wife succeeds; the king marries her sister, immures her in a palace, and is after all betrayed. But Somadeva gives us also tales of faith and truth among women. Devasmitā revenges herself on her would-be lovers by giving them assignations, but merely in order to brand them; charming is the picture of an Indian Philemon and Baukis.¹ It is death to tell another what one has remembered of existence in a former birth; nevertheless the queen of Dharmadatta and her husband are alike seized with the feeling that they must tell each other of their suddenly aroused memories. The story is pretty; the lady was a faithful servant in the house of a Brahmin, while her husband was the loyal retainer of a merchant; they lived together in poverty, eating the little they had over when gods, ancestors, and guests had taken their share. In time of famine a Brahmin comes, the husband gives him the little they had, and then his life leaves him, indignant that he had preferred the Brahmin to himself. His wife follows him in death, and the same fate again meets them when they have exchanged these memories of a faithful love.

c19.s5.10

¹ J. S. Speyer, *Die indische Theosophie*, pp. 97 .

c19.s5.11

The religious world of Somadeva reminds us of the superstitious nature of the people of Kashmir; we can hardly doubt that the Kashmir recension added readily anything that seemed interesting in this regard, even if Somadeva himself is rather inclined to rationalizing Märchen. Çiva and Pārvatī in her dread forms are the great deities, though Viṣṇu inevitably appears in the episode of Naravāhana's visit to the Çvetadvīpa. Human offerings are specially frequent, the Pulindas, the Bhillas, are regarded as ever on the outlook for victims for the goddess to whom Jīmūtavāhana is prepared to offer homage before his act of self-sacrifice. Witchcraft is taken as a matter of course, and many details are given of the dreadful deeds of the witches and of the horrible scenes enacted nightly at the places where the dead are burned or flung out as prey for beasts, birds, and the ghouls who haunt these cemeteries; in his eeriness of description Somadeva is a match for the author of the *Mālatīmādhava*. Buddhistic traits are not rare, though only sporadic; it must be remembered that, as we know from Kalhaṇa, Buddhism had in a degraded form a strong hold in Kashmir. A number of tales are told to relate the action of Karman in determining man’s life; we have a legend of a prince who tears out an eye because women loved so deeply his beauty, a parallel to the *Mittavindaka Jātaka*, and the legend of Jīmūtavāhana, though the Buddhist origin of that has been questioned.¹ The *Vetālapañcaviñçatikā* legends show distinct Buddhist traits. On the other hand, we have frequent mention of the worship of the Liṅga, Çiva’s phallic symbol, and of the Mothers, and popular superstition is everywhere abundant. The gods and minor spirits mingle freely in ordinary life, innumerable apparent mortals are merely beings driven from heaven by curses who can be restored to their former estate by some act of cruelty or kindness. The love of the marvellous is fully satisfied by tales of adventures at sea, with shipwrecks and subterranean palaces, or not less marvellous wanderings on land to strange places like camphor-land where princesses can easily be won. The loves of Naravāhanadatta are too numerous and too inevitable—for they are all fore-ordained even if we are only told so at the end—to be exciting, but there are many others recounted in episodes, and a picture or a dream often proves the starting-point for a deep if transient affection. Nor can we ignore the interest lent by the inclusion of effective versions of the Vetāla cycle, of the *Pañcatantra*, of anecdotes of Vikramāditya, as well as those in the less satisfactory book Padmāvatī.

c19.s5.12

¹ Bosch, *De legende van Jīmūtavāhana*, pp. viii, 143 ff.

c19.s5.13

Somadeva's taste is shown by the fact that, though he likes to conclude a tale with a different metre, only 761 of his 21,388 verses are in more elaborate metres, and he resists the temptation to indulge himself in word-plays, contenting himself with the swift easy flow of the simple narrative. He permits himself in his metre a certain lightness of touch exhibited in minor negligences, which in no way make it inaccurate, but save it from the pedantry of following in absolute strictness the rules regarding caesuras and Sandhi rigorously adopted by the great Kāvya writers. His abnegation is the more remarkable because he obviously could have won repute as a poet in the elaborate style. As it is, we owe him many happy passages in which simplicity is not inconsistent with ornament. Thus we have the description, brief but effective, of a storm at sea:

c19.s5.14

*aho vāyur apūrvo 'yam ity āçcaryavaçād iva*
*vyāghūrṇante sma jaladhes taṭeṣu vanarājayaḥ.*
*vyatyastāç ca muhur vātād adharottaratāṁ yayuḥ*
*vāridher vārinicayā bhāvāḥ kālakramād iva.*

c19.s5.15

'The forests on the banks of the sea shook to and fro as though amazed at the wondrous force of the gale, and inverted by the wind the waves went up and down as do men's hopes through the force of fate.' The good deed of the Gandharva, who saves the prince from the well into which he had fallen, is summed up in an admirable line:

c19.s5.16

*parārthaphalajanmāno na syur mārgadrumā iva*
*tapacchido mahāntaç cej jīrṇāraṇyaṁ jagad bhavet.*

c19.s5.17

'Were there not high-souled men born to do good to others, like wayside trees which dispel the heat, this world were nothing but a worn-out forest.' There is a very pathetic picture of the death of Çūrasena; he was a Rājput and had to obey his king's summons, despite his love for his wife Suṣeṇā; she awaits his promised return and, when he comes not, her breath leaves her body as if consumed by the forest fire of love. Her husband meantime, scarce able to leave his lord, is hastening to her on a swift camel:

c19.s5.18

*tatrāpaçyad gataprāṇāṁ priyāṁ tāṁ kṛtamaṇḍanām*
*latām utphullakusumāṁ vātenonmūlitām iva.*
*dṛṣṭvaiva vihvalasyaitāṁ kurvato ’ṅge viniryayuḥ*
*pralāpaiḥ saha tasyāpi prāṇā virahiṇaḥ kṣaṇāt.*

c19.s5.19

‘There saw he his wife lying dead in all her finery like a creeper in full bloom that the wind hath uprooted; seeing her he grasped her in his arms, beside himself with grief of separation, and his breath straightway departed with his lamentations.’ There is a brilliant description of summer:

c19.s5.20

*bhrāmyataç ca jagāmāsya bhīmo grīṣmartukesarī*
*pracaṇḍādityavadano dīptatadraçmikesaraḥ.*
*priyāvirahasaṁtaptapānthanihçvāsamārutaiḥ*
*nyastoṣmāṇa ivātyuṣṇā vānti sma ca sakīraṇāḥ.*
*çusyadvidīrṇapaṅkāç ca hṛdayaiḥ sphuṭitair iva*
*jalāçayā dadṛçire gharmaluptāmbusampadaḥ.*
*cīrīcītkāramukharās tāpamlānadaladharāḥ*
*madhuçrīvirahān mārgeṣv arudann iva pādapāḥ.*

c19.s5.21

‘And as he wandered there came on him the dread hot season in lion shape with the blazing sun for mouth and his fiery rays for mane. The winds blew with cruel heat as though warmed by the dolorous sighs of travellers parted from their loved ones. The tanks, their waters wasted by the heat, with their drying white mud seemed to show their broken hearts. The trees bewailed the departure of the glory of spring with the shrill moaning of their bark, their lips of leaves being parched by the heat.’

Linked texts (23)