वासवदत्ता (महाकवि सुबन्धु - श्लेष-प्रधान शास्त्रीय गद्यकाव्य सान्वय सटीक)

ABBREVIATIONS xiii

• = when prefixed to a word or meaning in the lexicographical appendix, such word or meaning is cited by the St. Petersburg lexicons only on the authority of native Sanskrit lexicographers. ° = when prefixed to a word or meaning in the lexicographical appendix, such word or meaning is entirely omitted by the St. Petersburg lexicons ; elsewhere it denotes the omission of words or parts of words to be supplied from the context. ‹ › = single paronomasia. « » = double paronomasia. ⋘ ⋙ = triple paronomasia. ( ) = when placed around or in words of the transcribed text, the enclosed portions differ from the text of Hall ; when placed around numerals in the translation and transcription, these refer to the pages of the Madras edition of 1862. [ ] = when placed around or in words of the transcribed text, the enclosed portions, though contained in Hall’s edition, are entirely omitted by the Madras edition of 1862 ; when placed around numerals in the translation and transcription, these refer to the pages of Hall’s edition.

This image contains a blank page with no readable text.

INTRODUCTION

Title. The title of the Vāsavadattā of Subandhu, the oldest romantic novel in India, seems to be derived from that of a long lost drama by Bhāsa,¹ the Svapnavāsavadattā, or ‘Dream-Vāsa- vadattā’ (for compounds of this type cf. Wackernagel, Altindische Grammatik, 2. 1. 244–245, 250–253, Göttingen, 1905). The dream as a novelistic device in India first occurs in Subandhu (see below, p. 28); though in the drama it is found in the first act of the Viddhaśālabhañjikā and the third of the Karpūrama- ñjarī (both written by Rājaśēkhara, who was acquainted with Bhāsa’s work), as well as in the first of Viśvanāthabhaṭṭa’s Śr̥ṅgāravāṭikā (Catalogue of the Sanskrit Manuscripts in the Library of the India Office, 7, 1618, London, 1904). In the fifth act of Bhāsa’s Svapnavāsavadattā the hero, King Vatsarāja, sleeping, dreams of his love Vāsavadattā, who enters, disguised as an attendant of the queen, but who, he thinks, has been burned to death at Lāvānaka (cf. svapnavāsavadattasya dāhakō, ‘the conflagration of the “Dream-Vāsavadattā”’ [Rājaśēkhara, cited in the Sūktimuktāvalī (see Peterson and Durgāprasāda, Subhāṣitāvali of Vallabhadēva, Introd., p. 81, Bombay, 1886)], and Bhāsa’s epithet jalaṇamitta, ‘friend of fire,’ in Gaüḍavaha, v. 800), this being employed both in the famous fire-scene in the fourth act of the Ratnāvalī (first half of the seventh century) and in the Tāpasavatsarāja (before the second half of the ninth century; see the analysis by Hultzsch, in Nachrichten von der königlichen Gesellschaft der Wissenschaften zu Göttingen, 1886, pp. 224–241). Not only was the fire-scene thus borrowed from Bhāsa by later dramatists, but from him, it may be conjectured, came, at least in literary form, the entire story of Vāsavadattā and Udayana, or Vatsarāja, as given in the Ratnāvalī, Priya-

¹ On Bhāsa, see, in general, Hall, ‘Fragments of Three Early Hindu Dramatists,’ in JASBe. 28. 28–29; Lévi, Théâtre indien, 1. 157–160, 2. 31–32, Paris, 1890. B

2 VĀSAVADATTĀ darśikā, and Tāpasavatsarāja,¹ the ultimate source probably being the lost Bṛhatkathā. With the Vāsavadattā of these latter works Subandhu's heroine has only her name in common, nor is any other story concerning her known to exist in Sanskrit literature (cf. Krishnamachariar, Introd., pp. 48–50; Lacôte, Essai sur Guṇāḍhya et la Bṛhatkathā, pp. 15–16, Paris, 1908). Though sleeping on the stage is forbidden by Sanskrit drama- turgy, the hero of the Svapnavāsavadattā dreams of the heroine, an episode imitated, I would suggest, by Rājaśēkhara in his Viddhaśālabhañjikā and Karpūramañjarī. In similar fashion Subandhu seems to have derived from this play ² both the dream- episode and the name of his heroine, who was indeed a 'dream- Vāsavadattā' ; and it would then appear that he invented the remainder of the romance. This is also the opinion of Krishna- machariar (Introd., pp. 48, 50), who suggests, however, that Subandhu may have adapted some old wives' tale. Author. The author of the Vāsavadattā was the only Suban- dhu to win for himself a name in Sanskrit literature, unless exception be made in favour of Subandhu, son of Gōpāyana or Lōpāyana, who, according to the Sarvānukramaṇī (ed. Macdonell, p. 19, Oxford, 1886), was one of the four authors of Ṛg-Veda 5. 24 (cf. the legend concerning him in Bṛhaddēvatā, 7. 84–102, ed. and tr. Macdonell, Cambridge, Mass., 1904). The name, however, occurs with tolerable frequency in Sanskrit (cf. Böhtlingk and Roth, Sanskrit-Wörterbuch, 7. 1086, St. Petersburg, 1875), and it would even seem to be found, under the form Šu-ba-an-di, or Šu-ba-an-du, in one of the Tell-el-Amarna Tablets, dating approxi- mately from the fourteenth or fifteenth century B.C. (Winckler, Thontafeln von Tell-el-Amarna, Nos. 224–229, Berlin, 1896). ¹ To this list Krishnamachariar (Introd., p. 37) adds the lost drama Udayanacarita (cf. Lévi, Théatre indien, 1. 92; 2. 39, Paris, 1890; Schuyler, Bibliography of the Sanskrit Drama, p. 90, New York, 1906). ² In May, 1910, the Svapnavāsavadattā and nine other dramas of Bhāsa were dis- covered near Padmanābhapura by Gaṇapati Śāstrī, who later found another manu- script containing, among other plays, a second copy of the Svapnavāsavadattā. These dramas were edited, after this introduction was already in type, in the Trivandrum Sanskrit Series.

INTRODUCTION 3 Sanskrit References to Subandhu. References in Sanskrit literature and inscriptions to our Subandhu (whose date is dis- cussed below, pp. 8–11) are but scanty. By far the most important allusion is contained in the eleventh stanza of Bāṇa’s introduction to his Harṣacarita (seventh century): kavīnām agalad darpō nūnaṁ vāsavadattayā śaktyē "va pāṇḍuputrāṇāṁ gatayā karṇagōcaram, ‘verily, the pride of <poets> melted away through the ≪Vāsava- dattā coming to their ears≫ even as the pride of the <sages> melted away through the Pāṇḍavas’ ≪Indra-given spear coming nigh Karṇa≫.’¹ Bāṇa is also supposed to allude to the Vāsava- dattā when, in the twentieth stanza of his introduction to his Kādambarī, he declares his new work to be iyam atidvayī kathā, ‘this story surpassing the two,’ these being, according to the commentator, the Vāsavadattā and the Bṛhatkathā. About a century later Vākpati, the author of the Prakrit historical poem Gaüḍavaha, wrote (v. 800), in describing himself: bhāsammi jalaṇamittē kantīdēvē a jassa rahuārē sōbandhavē a bandhammi hāriyandē a āṇandō, ‘in Bhāsa—the friend of fire—in the author of the Raghu (vaṁśa)—that lord of beauty—in Subandhu’s work, and in that of Haricandra is his delight.’ In Kavirāja’s Rāghavapāṇḍavīya, which dates from about 1200 A. D., occurs the stanza (I. 41): subandhur bāṇabhaṭṭaś ca kavirāja iti trayaḥ vakrōktimārganipuṇāś caturthō vidyatē na vā, ‘Subandhu, Bāṇabhaṭṭa, and Kavirāja—these three be skilful in the path of ambiguity²; a fourth there is not found.’ The ¹ This verse is interpolated at the end of the Vāsavadattā by the Telugu and Grantha editions, and Krishnamachariar (Introd., pp. 38–39) implies that it may have been written by Subandhu and later have found its way into the Harṣacarita. His theory is to me untenable. On the use of the signs < >, etc., see p. 17. ² On vakrōkti see Sāhityadarpaṇa, No. 641; Appayyadīkṣita, Kuvalayānandakārikā,

1. 158–159; Kāvyaprakāśa, tr. Jhā, pp. 181–182, Benares, 1898; Bernheimer and Jacobi, in ZDMG. 63. 797–821; 64. 130–139, 586–590, 751–759; 65. 308–312. Subandhu is also mentioned immediately before Bāṇa in the Sarasvatīkaṇṭhābharaṇa according to Müller, India, What can it Teach us?, p. 331, note 5, London, 1883, but I have not been able to find the reference. B 2

4　　　　　　　　　VĀSAVADATTĀ twelfth-century Śrīkaṇṭhacarita of Maṅkha contains the stanza (2. 53) : mēṇṭhē svardviradādhirōhiṇi vaśaṁ yātē subandhāu vidhēḥ śāntē hanta ca bhāravāu vighaṭitē bāṇē viṣādaspr̥śaḥ vāgdēvyā viramantu mantuvidhurā drāg dr̥ṣṭayaś cēṣṭatē śiṣṭaḥ kaścana sa prasādayati tāṁ yadvāṇisadvāṇinī, ‘Mēṇṭha having mounted the elephant of the sky [i.e., having died], Subandhu having yielded to the will of destiny, Bhāravi, alas, being at rest, and Bāṇa being broken, let the reason-reft glances of the sorrow-stricken goddess of speech [Sarasvatī] quickly find repose; for any one left that bestirreth himself doth win her, to whose voice she is a goodly dancer.’ Two centuries later, Śārṅgadhara, quoting Rājaśēkhara (cf. above, p. 1), who flourished about 900 A. D., in his Paddhati (cf. Aufrecht, ‘Ueber die Paddhati von Çārṅgadhara,’ in ZDMG. 27.77; Śārṅgadha- rapaddhati, ed. Peterson, 1. No. 188, Bombay, 1888), made the citation : bhāsō rāmilasāumilāu vararuciḥ śrīsāhasāṅkaḥ kavir mēṇṭhō bhāravikālidāsataralāḥ skandhaḥ subandhuś ca yaḥ daṇḍī bāṇadivākarāu gaṇapatiḥ kāntaś ca ratnākaraḥ siddhā yasya sarasvatī bhagavatī kē tasya sarvē ’pi tē, ‘Bhāsa, Rāmila, Sāumila, Vararuci, the poet Sāhasāṅka, Mēṇṭha, Bhāravi, Kālidāsa, Tarala, Skandha, and Subandhu, Daṇḍin, Bāṇa, Divākara, Gaṇapati, and the charming Ratnākara,—what are all they to him by whom the exalted Sarasvatī is possessed’? Rājaśēkhara also refers to Subandhu in the following stanza quoted by Aufrecht (ZDMG. 36. 366) from the Saduktikarṇā- mr̥ta (cf. also Peterson and Durgāprasāda, Subhāṣitāvali of Vallabhadeva, Introd., p. 57, Bombay, 1886; Krishnamachariar, Introd., p. 41) : subandhāu bhaktir naḥ ka iha raghukārē na ramatē dhr̥tir dākṣīputrē harati haricandrō ’pi hr̥dayam viśuddhōktiḥ śūraḥ prakr̥timadhurā bhāravigiras tathā ’py antarmōdaṁ kam api bhavabhūtir vitanutē,

INTRODUCTION 5 ' in Subandhu is our delight; who rejoiceth not in the author of the _Raghu(vaṃśa)_¹? satisfaction delighteth in the son of Dākṣī²; even Haricandra joyeth the heart; of faultless diction is Śūra³; the words of Bhāravi are delightsome in theme; Bhavabhūti doth infuse an inward pleasure.' Krishnamachariar also cites an allusion to the author of the Vāsavadattā in Abhi- navabhaṭṭabāṇa's Vīranārāyaṇacarita (l.c.) : pratikavibhēdanabāṇaḥ kavitātarugahanaviharaṇamayūraḥ sahṛdayalōkasubandhur jayati śrībhaṭṭabāṇakavirājaḥ, ' victorious is the noble Bhaṭṭabāṇa, king of poets, an arrow [or, “ a Bāṇa ”] for piercing hostile poets; a peacock [or, “ a Mayūra ”] for wandering through the forest of the trees of poetry; a goodly kinsman [or, “ a Subandhu ”] for all connoisseurs.' An anony- mous citation in the modern anthology Subhāṣitaratnabhāṇḍāgāra (ed. Parab, 3 ed., p. 56, Bombay, 1891) runs: māghaś cōrō mayūrō muraripur aparō bhāraviḥ sāravidyaḥ śrīharṣaḥ kālidāsaḥ kavir atha bhavabhūtyāhvayō bhōjarājaḥ śrīdaṇḍī ḍiṇḍimākhyaḥ śrutimukuṭagurur bhallaṭō bhaṭṭabāṇaḥ khyātāś cā 'nyē subandhvādaya iha kṛtibhir viśvam āhlādayanti, ' Māgha, Cōra, Mayūra, Mura's second foe (Murāri), Bhāravi in climax learned, Harṣa, Kālidāsa, and also the poet named Bhavabhūti, Bhōjarāja, Daṇḍin (hight “ the Drum ”), Bhallaṭa weighty with the diadem of fame, Bhaṭṭabāṇa, and other renowned ones, such as Subandhu, here on earth rejoice the universe with their compositions.' In the latter part of the sixteenth century, Ballāla, in his Bhōjaprabandha, which he set forth as a history of Bhōja, who ruled at Dhārā (the modern Dhār) in the eleventh century, mentioned Subandhu, according to some manuscripts, as one of the thirteen principal members of the host of five hundred literati who graced the royal court (cf. Wilson, Works, 5. 174, London, 1865; Hall, Introd., p. 7, note 1); but the list varies so extremely in the different manuscripts of the Bhōjaprabandha ¹ Kālidāsa. ² Pāṇini. ³ See Aufrecht, Catalogus Catalogorum, 1. 660, Leipzig, 1891.

6 VĀSAVADATTĀ that little stress can be laid upon it, especially in view of the legendary character of the work as a whole. Finally, mention should also be made, for the sake of completeness, of an apparent allusion to the Vāsavadattā in Daṇḍin's Daśakumāracarita (ed. Godabole and Parab, p. 110, lines 11–12, Bombay, 1898): anurūpabhartṛgāminīnāṁ ca vāsavadattādīnāṁ varṇanēna grā- hayā 'nuśayam, ‘and make her repent by a description of Vāsavadattā and others who gained suitable husbands.’ This clearly refers, however, to the well-known story of Vāsavadattā and Udayana (see above, p. 2); and it is equally impossible that the vāsavadattām adhikṛtya kṛtō granthaḥ mentioned in the Vārttika (probably third century B.C.) on Pāṇini, 4. 3. 87, should be connected in any way with Subandhu’s romance. It should also be noted that Narasiṁha Vaidya, one of the glossators of the Vāsavadattā, says: kavir ayam vikramāditya- sabhyaḥ. tasmin rājñi lōkāntaram prāptē ētan nibandham kṛtavān, ‘this poet [Subandhu] was a retainer of Vikramāditya. When this king attained the other world, he [Subandhu] composed this work’ (Hall, Introd., p. 6, note). Hall’s manuscript D, more- over, which belongs to what I may tentatively call the ‘South Indian recension’ of the Vāsavadattā (see below, p. 38), terms Subandhu ‘the son of Vararuci’s sister’ (śrīvararucibhāginēya), Vararuci himself being, as is well known, one of the ‘nine gems’ of Vikramāditya’s court, flourishing at least later than the fifth century (Bloch, Vararuci und Hemacandra, p. 13, Gütersloh, 1893; cf. Macdonell, History of Sanskrit Literature, p. 324, London, 1900). Hall denies that Subandhu was Vararuci’s nephew (Introd., pp. 6–7), but it is possible that the tradition contains a larger element of truth than is often supposed (cf. Wilson, Works, 5. 177, London, 1865). It is, at all events, echoed by the Bhōjaprabandha,—whatever be the value assigned to such testimony—which associates Subandhu and Vararuci in the passage already referred to. Inscriptions of India mention Subandhu only once to my knowledge. This single instance is a Canarese record of 1168 A.D., found at Balagāmi (Rice, Mysore Inscriptions, p. 111,

Bangalore, 1879), which contains the words: ‘In śabda a Pāṇini paṇḍita, in nīti Bhūṣaṇācārya, in nāṭya and other bharata śāstras Bharatamuni, in kāvya Subandhu, in siddhānta Lakulīśvara, at the feet of Śiva a Skanda adorning the world, thus is Vāma Śaktiyati truly described.’ The only additional information thus gained is that by the twelfth century his fame had spread to southern India. Subandhu's Allusions to Sanskrit Literature. If Subandhu is thus recognised but sparingly and indefinitely in the literature and epigraphy of his native land, he is himself most generous in alluding to the productions of other authors. The majority of his references, however, cast little light upon his date, for no real conclusions can be drawn from his mention of the Bṛhatkathā (ed. Hall, pp. 110, 147), the Kāmasūtra (ed. Hall, p. 89), the Chandōviciti section of the Bhāratīyanāṭyaśāstra (ed. Hall, pp. 94, note, 119, 235), the Mahābhārata and Harivaṁśa (ed. Hall, pp. 21, 27, 93, 234, 254; on Subandhu's indebtedness to the Mahābhārata see Cartellieri, ‘Das Mahābhārata bei Subandhu und Bāṇa,’ in WZKM. 13. 57–74), the Rāmāyaṇa (ed. Hall, p. 234), the Upaniṣads (ed. Hall, p. 235), and the Mīmāṁsā and Nyāya philosophies (ed. Hall, pp. 93, 235, 297), any more than we can derive any definite conclusions from his general and hostile mention of the Buddhists (ed. Hall, pp. 144, 179, 235, 255, 297 bis) and Jains (ed. Hall, pp. 93, 187, 297; cf. on these allusions to the Buddhists and the Jains Telang, ‘Subandhu and Kumārila,’ in FRASBo. 18. 150–159).¹ It has been held, on the basis of Śivarāma's commentary, that the words bāuddhasaṅgatim ivā 'laṁkārabhūṣitām, ‘decked with ⟨adornments⟩ as an assembly of Buddhists is decked with the ⟨Alaṁkāra⟩’ (ed. Hall, p. 235), refer to the Bāuddhasaṅgatyalaṁ- kāra of Dharmakīrti (cf., in general, on Dharmakīrti, Pathak, ‘Dharmakīrti and Śaṁkarāchārya,’ in FRASBo. 18. 88–96, and

¹ Reference should also be made, in this connexion, to the allusions collected in Krishnamachariar's Introduction, received after these lines were written, to other Sanskrit literature, especially the Mahābhārata and Rāmāyaṇa, as well as to religion and philosophy (pp. 22-24, 27-28).

8 VĀSAVADATTĀ the criticism of Telang, ib. 18. 148–150). Since, however, Dharmakīrti is described by I-Tsing, who travelled in India in 671–695 A. D., as among those ‘of late years’ (Record of the Buddhist Religion, tr. Takakusu, p. 181, cf. p. lviii, Oxford, 1896); and since Tāranātha (Geschichte des Buddhismus in Indien, tr. Schiefner, pp. 184–185, St. Petersburg, 1869) makes him a con- temporary of the Tibetan king, Srong-btsan-sgam-po, who died about 650 A. D. (Duff, Chronology of India, p. 53, Westminster, 1899), Lévi (‘La Date de Candragomin,’ in Bulletin de l'École d'Extrême-Orient, 1903, p. 18; cf. Kern, Manual of Indian Buddhism, p. 130, note 11, Strassburg, 1896) is doubtless right in denying that Subandhu makes any allusion to Dharmakīrti’s activity (for an opposing view see Krishnamachariar, Introd., p. 32). This leaves but a single literary allusion in the Vāsavadattā which can in any way be construed as casting light on the date of the romance. The reference in question is nyāyasthitim ivō 'ddyōtakarasvarūpām, ‘⟨revealing her beauty⟩ as the permanence of the Nyāya system has its ⟨form from Uddyōtakara⟩’ (ed. Hall, p. 235). Since we know that Uddyōtakara wrote his Nyāya- vārttika to refute the heterodox (i.e., Buddhist) views of Dignāga, who flourished between 520 and 600 A. D. (Kern, op. cit. p. 129; Müller, Six Systems of Indian Philosophy, p. 477, London, 1899), it is obvious that Uddyōtakara, to whom Subandhu so unmis- takably refers, can not have lived before the latter part of the sixth century. It is, therefore, certain that the Vāsavadattā can not be prior to the late sixth century of our era. The Date of Subandhu. There is but one allusion in Subandhu's romance itself which can be interpreted as referring to a historical event. This is the tenth introductory stanza: sā rasavattā vihatā navakā vilasanti carati nō kaṃ kaḥ¹ sarasī ’va kīrtiśēṣaṃ gatavati bhuvi vikramādityē, ‘⟨moisture⟩ is destroyed, ≪cranes sport not≫, ⋘the heron fares not forth⋙; yea, ⟨eloquence⟩ is destroyed, ≪new-comers make ¹ The theory of Mazumdar (JRAS. 1907, pp. 406–408), that the kaṃ kaḥ of this stanza involves an allusion to a Kaṅka dynasty, must be regarded as not proven.

INTRODUCTION 9

disport», ««who devours not whom»»?—for Vikramāditya, like a lake, hath passed away on earth.’ This has been taken, particularly by Hall (Introd., p. 6), to imply that Subandhu ‘lived long posterior to the great Vikramāditya of Ujjayinī.’ Although some deny that this monarch, about whom cluster so many legends, ever existed (see, for example, Macdonell, History of Sanskrit Literature, pp. 323–324, London, 1900), such a view can scarcely be supported, and there is good reason to believe him to be identical with Candragupta II, who reigned from about 374 to 413 (Smith, Early History of India from 600 B.C. to the Muhammadan Conquest, 2 ed., pp. 275–283, Oxford, 1908). It would be most tempting, so far as the special problems of the Vāsa- vadattā are concerned, could one accept the view, argued with great learning by Hoernle (‘Some Problems of Ancient Indian History,’ in JRAS. 1903, pp. 545–570, and ‘The Identity of Yaśodharman and Vikramāditya, and some Corollaries,’ ib. 1909, pp. 89–144 ; against this Fleet, ‘Dr. Hoernle’s Article on Some Problems of Ancient Indian History,’ ib. 1904, pp. 164–166, and Smith, ‘The Indian Kings named Śīlāditya, and the Kingdom of Mo-la-p’o,’ in ZDMG. 58. 787–796), that Yaśōdharman (on whom see also Smith, op. cit. pp. 301–302), whom he identifies with the great Vikramāditya, ‘founded his Mālava empire about 533 A.D., and reigned up to about 583 A.D.’ Hoernle accord- ingly dates Subandhu in the second half of the sixth century, and holds that the Vāsavadattā was written before 606–612, the latter year being that of Harṣa’s coronation. Attractive as is this hypothesis, I am compelled to admit that it can scarcely be used to determine the date of Subandhu, the whole evidence of Indian history being against it. The period following the death of Vikramāditya is described in the stanza quoted above as one of degeneration, and there may be a covert allusion to the same (or a similar) evil state of affairs in the phrase navanṛpaticittavṛttibhir iva kulyāpamānakāriṇībhir, ‘as the disposition of new monarchs causes dishonour to the honourable’ (ed. Hall, p. 220). Hoernle, holding that Vikra- māditya’s successor was his son Śīlāditya, who was dethroned

10 VĀSAVADATTĀ

by his enemies (probably about 593 A. D.), being 'replaced in the kingdom of his father' (probably before 604 A. D.) only by the aid of the Hun, Pravarasēna II of Kashmir (Rājataraṅgiṇī 3. 330), has evolved a most ingenious theory which I was long inclined to adopt. Śīlāditya is, on this hypothesis, described as not only unfortunate, but cruel, as evidenced by his execution of the Māukharī Grahavarman, king of Kanauj, and the brutal fettering and imprisonment of the dead monarch’s young wife, Rājyaśrī (Harṣacarita, tr. Cowell and Thomas, p. 173, London, 1897); and as unpatriotic, this being shown by his acceptance of assistance from non-Aryan Huns. Despite his restoration by Pravarasēna, the reign of Śīlāditya, who, Hoernle maintains, succeeded his father, Vikramāditya, about 583 A. D., came to a disastrous end in 606 (or 605), when he was utterly defeated by Rājyavardhana II, the brother of the famous Harṣavardhana who is the hero of the Harṣacarita. Harṣavardhana himself succeeded to the throne of Thāṇēsar in 606, when Rājyavardhana was treacherously slain by the Gāuḍa king, Śaśāṅka, and reigned until 648 (on Harṣavardhana, in general, cf. Ettinghausen, Harṣa Vardhana, empereur et poète de l'Inde septentrionale, Paris, 1906). While holding this theory, I gave to it the pleasing embellish- ment of an hypothesis, without real basis, that the dynasties to which Vikramāditya and Harṣavardhana belonged were rivals, and that Bāṇa was the faithful eulogist of Harṣavardhana exactly as Subandhu was loyal to Vikramāditya. Since, moreover, Bāṇa's monarch had been victorious over the degenerate son of Subandhu's royal patron, I deemed that Bāṇa had deliberately set out to surpass Subandhu, so that Harṣavardhana's court might excel Vikramāditya's in literature as well as in arms. Thus, there would have been a deeper motive for Bāṇa to write the Harṣacarita than the mere incentive of literary emulation which is generally ascribed to him. History does not sustain this elaborate figment, which I have recorded mainly to keep others from possible pursuit of a false clue. Not only was Vikramāditya not identical with Yaśōdharman,

INTRODUCTION II as already noted, but Śīlāditya was the very reverse of a cruel monarch (Smith, Early History of India from 600 B.C. to the Muhammadan Conquest, 2 ed., p. 306, Oxford, 1908). If one were to stress the theory of rivalry both in letters and in war, one might suppose that Subandhu was a courtier either of Śaśāṅka of Gauḍa or of Dēvagupta of Eastern Mālava (cf. Ettinghausen, op. cit. pp. 36–38, 148), both of whom were ignoble in character. But of this there is not the slightest evidence ; and even if the name of the father of either of them was Vikramāditya (a most impro- bable hypothesis), that would give little point to Subandhu's stanza, which plainly alludes to the famous Vikramāditya, and is, therefore, only a conventional harking back to happy times long past. In determining the date of the Vāsavadattā I am forced to consider the lines under discussion as utterly valueless. While the sole known basis for assigning a terminus a quo to the composition of Subandhu's romance is, as we have seen, the allusion to Uddyōtakara, who probably flourished in the latter half of the sixth century, the terminus ad quem is almost cer- tainly the date of Bāṇa's Harṣacarita. This romance, which was left unfinished by its author, ends abruptly with the rescue of Rājyaśrī, the sister of Harṣavardhana and widow of Graha- varman (i. e. 607, or 606), though Harṣa had reigned several years when Bāṇa wrote (Harṣacarita, tr. Cowell and Thomas, pp. 75–76, London, 1897). The precise date of composition of Bāṇa's second romance, the Kādambarī, is unknown; but, as Bāṇa died before completing it (Kādambarī, tr. Ridding, p. 182, London, 1896), it must have been written considerably after the Harṣacarita. It may also be regarded as certain that Subandhu lived later, probably by at least a century (cf. p. 12), than Daṇḍin, the author of the picaresque Daśakumāracarita (Weber, Indische Streifen, 1. 311–315, 353, 372, Berlin, 1868 ; Daśakumāracarita, tr. Meyer, pp. 120–127, Leipzig, 1902 ; Collins, The Geographical Data of the Raghuvamśa and Daśakumāracarita, p. 46, Leipzig, 1907, places Daṇḍin's literary activity before 585 A. D.). The Place of Composition of the Vāsavadattā. The question next arises as to the place of composition of the Vāsavadattā.

12 VĀSAVADATTĀ Here the answer must be still more vague.¹ It is obviously impossible that the romance was written at the court of Bhōja, as some manuscripts of the Bhōjaprabandha would imply, for that ruler did not reign at Dhārā until the eleventh century; nor does there seem to have been any Bhōja reigning in the latter part of the sixth century at whose court Subandhu might have been, thus being confusedly located by Ballāla in the train of the famous Bhōja of Dhārā. One might, indeed, by reckless theorising, allege that Subandhu, thus being placed at Dhārā in Mālava, had actually been a courtier either of Dēvagupta of Eastern Mālava or of Śīlāditya of Mo-la-p'o (Western Mālava); but the real reason for this wild statement by the author of the Bhōja- prabandha (or, more probably, by one of his interpolators) was obviously the identification, occasionally made by Sanskrit authors (cf. Weber, Akademische Vorlesungen über indische Literaturgeschichte, 2 ed., pp. 218–219, Berlin, 1876; Rajendralala Mitra, 'Bhoja Rájá of Dhár and his Homonyms,' in JASBe. 32. 93), of Bhōja and Vikramāditya, an equation too absurd to require refutation. To sum up the discussion, we can say with reasonable certainty only that the Vāsavadattā was written by Subandhu at a place unknown, probably between 550 and somewhat after 606 A. D., the terminus a quo being the circumstance that Uddyōtakara cannot have flourished until at least the middle of the sixth century, and the terminus ad quem by the date of composition of the Harṣacarita, early in the seventh century.² ¹ Absolutely no clue is given by the purely conventional geography of the romance, on which see Weber, Indische Streifen, 1. 385, Berlin, 1868. ² Krishnamachariar devotes a long section of his Introduction (pp. 30–48) to a discus- sion of Subandhu's date, which he places after Bāṇa and before Vāmana, the author of the Kāvyālaṁkāravṛtti, whom tradition makes a minister of Jayāpīḍa of Kashmir (779–813; cf. Duff, Chronology of India, pp. 68, 70–71, Westminster, 1899). He rightly argues that the various references in Sanskrit authors to Subandhu and Bāṇa allow of no conclusion as to the priority of the Vāsavadattā; but some of his hypotheses, as that Subandhu's dislike of Buddhism proves him to be later than Bāṇa (p. 45), as well as his general implication that the difference between the two writers is due to degeneration of style (cf. pp. 14–18), are, in my judgment, certainly untenable; nor does he touch with sufficient depth upon what evidence may be drawn from Indian history.

INTRODUCTION 13 Data Concerning Subandhu’s Life. Our knowledge of Subandhu is most meagre. In the thirteenth stanza of his intro- duction to the Vāsavadattā he terms himself sujanāikabandhu, which Hall (Introd., p. 24), following the commentator, Śivarāma, renders ‘an intimate of none but the virtuous,’ although the word should rather be translated ‘Sujana’s only brother.’¹ The tradition that Subandhu was the nephew of the Prakrit grammarian, Vararuci, has already been mentioned (see above, p. 6), though with disapproval; and there seems also to have been a legend that he was, by birth, a Kashmirian Brāhman (Weber, Indische Streifen, 1. 371, Berlin, 1868, quoting Cunning- ham, in JASBe. 17. 98–99).² Subandhu is not known to have written anything besides the Vāsavadattā. Citations are made from him in the Śārṅga- dharapaddhati, Subhāṣitāvalī, Padyāvalī, and Sūktikarṇāmṛta (Aufrecht, Catalogus Catalogorum, 1. 726, Leipzig, 1891); but the quotations in the first two anthologies, which alone are thus far edited [by Peterson (Bombay, 1888) and by Peterson and Durgāprasāda (Bombay, 1886), respectively], are drawn exclu- sively from the Vāsavadattā. Hall (Introd., p. 48, note), it is true, found in the still unedited Padyaveṇī of Veṇīdatta, com- piled about the reign of Shāh Jahān (early 17th century), the fol- lowing distich then supposed to have been written by Subandhu: akṣamālāpavṛttijñā kuśāsanaparigrahā brāhmī "va dāurjanī saṃsad vandanīyā samēkhalā, ‘an assembly of scoundrels, knowing ‹how to live by disparaging speeches›, accepting «evil teachings», and «wicked to the just», should be honoured even as an assembly of Brāhmans knowing ¹ Cartellieri, ‘Das Mahābhārata bei Subandhu und Bāṇa,’ in WZKM. 13. 72, trans- lates the stanza thus: ‘Durch eine Gnadengabe, die Sarasvatī ihm verliehen, hat Subandhu — d. h. der edle Freunde hat — dieses Buch gemacht; obzwar Subandhu — d. h. der hundert Freunde hat — hat er doch nur den Edlen zum einzigen Freund; eine wahre Schatzkammer ist er in der Kunst, Silbe für Silbe doppelsinnige Dicht- ungen zu verfertigen.’ On Sujana as a proper name see Aufrecht, op. cit. 3. 149, Leipzig, 1903. ² It is interesting to note, in this connexion, that Krishnamachariar holds that Subandhu was a Vāiṣṇavite and an adherent of the Mīmāṃsā philosophy (Introd., pp. 23, 28).

14 VĀSAVADATTĀ ‹the end of their rosaries›, accepting «seats of kuśa-grass», and «girt with their girdles».' The distich was not, however, written by Subandhu, but by Trivikrama Bhaṭṭa, the author of the Damayantīkathā, or Nalacampū (1. 7; cf. Böhtlingk, Indische Sprüche, 2 ed., No. 52, St. Petersburg, 1870–1873), who flourished about 915 A.D. (Duff, Chronology of India, p. 85, Westminster, 1899).¹ The Vāsavadattā a Kathā. The Vāsavadattā is expressly stated by many manuscripts (cf. Hall’s ed., p. 300, note 7, and Śivarāma ad loc.) to be an ākhyāyikā, or ‘tale,’ this being very possibly influenced by the reference to some work entitled, from the name of its heroine, vāsavadattākhyāyikā in the Vārttika on Pāṇini 4. 3. 87 (cf. also the Vārttika on 4. 2. 60, and see Krishnamachariar, Introd., pp. 36–37). The ākhyāyikā, according to Sanskrit rhetoricians (cf. Regnaud, Rhétorique sanskrite, pp. 76– 77, Paris, 1884), is a division of gadya, or poetical prose; and the classical example is the Harṣacarita of Bāṇa, who himself seems to intimate that the Vāsavadattā likewise belongs to this cate- gory by using the term ākhyāyikākārā, ‘authors of ākhyāyikās,’ immediately before his allusion to Subandhu’s romance, in the tenth stanza of his introduction to the Harṣacarita. The classic description of the ākhyāyikā is given in the following passage of the Sāhityadarpaṇa (ed. Roer, No. 568, Calcutta, 1851): ākhyāyikā kathāvat syāt kavēr vaṁśādikīrtanam asyāṁ anyakavīnāṁ ca vṛttaṁ gadyaṁ kvacit kvacit kathāṁśānāṁ vyavacchēda āśvāsa iti badhyatē āryāvaktrapavaktrāṇāṁ chandasā yēna kēnacit anyāpadēśēnā ’’śvāsamukhē bhāvyarthasūcanam, ‘the ākhyāyikā should be as the kathā. (There should be) in it an account of the lineage of the poet and of other poets; poetry ¹ Krishnamachariar (Introd., pp. 39-40) calls attention to a number of passages in the Nalacampū (ed. Bombay, 1885; new ed., 1903) in which he holds that Trivikrama Bhaṭṭa imitated Subandhu. He likewise notes parallels between the Vāsavadattā and the Jīvandharacampū of Haricandra (p. 52), who wrote after 897 A.D. (p. 44), Śiśupālavadha (p. 53), Rāmāyaṇa (p. 64), Mēghadūta (p. 54), Vikramōrvaśī (pp. 62, 64), and Mālatīmādhava (pp. 61–62), as well as the Harṣacarita (pp. 53–57), and Kādambarī (pp. 52, 53, 55, 57, 63).

INTRODUCTION 15 in some places (and) prose in others (should be employed); divisions, called “sighs,” are used for the divisions of the story; at the beginning of the “sighs” (there should be) an intimation of the theme, under the guise of something else, by any metre whatsoever of the āryā, vaktra, or apavaktra (classes).’ The kathā, or ‘story,’ best represented by Bāṇa's Kādambarī, is described by the Sāhityadarpaṇa (No. 567) as follows : kathāyāṁ sarasaṁ vastu padyāir ēva vinirmitam kvacid atra bhavēd āryā kvacid vaktrāpavaktrakē ādāu padyāir namaskāraḥ khalādēr vṛttakīrtanam, ‘in the kathā a theme with poetic sentiments is represented even with poetry ; in it there should be the āryā metre in some places, (and) the vaktra and apavaktra metres in other places ; at the beginning (there should be) homage in verse (to a divinity, also) a description of the character of knaves and the like.’ The older, and in my judgment the better, definition of this type of Sanskrit literature, however, is given by Daṇḍin, the author of the pica- resque Daśakumāracarita, who says (Kāvyādarśa I. 23–25, 28) : apādaḥ padasantānō gadyam ākhyāyikā kathā iti tasya prabhēdāu dvāu tayōr ākhyāyikā kila nāyakēnāi "va vācyā "nyā nāyakēnē 'tarēṇa vā svaguṇāviṣkriyā dōṣō nā 'tra bhūtārthaśaṁsinaḥ api tv aniyamō dṛṣṭas tatrā 'py anyāir udīraṇāt anyō vaktrā svayaṁ vē "ti kīdṛg vā bhēdalakṣaṇam ............................................................................ tat kathākhyāyikē "ty ēkā jātiḥ sañjñādvayāṅkitā atrāi "vā 'ntarbhaviṣyanti śēṣāś cā "khyānajātayaḥ, ‘ prose is a series of words without strophes ; its two classes are the ākhyāyikā (and) the kathā. Now, the ākhyāyikā should be spoken by the hero, the other (the kathā) by the hero or another. A revelation of one’s own personality, if he narrates facts, is no fault here. Nevertheless, the lack of fixed distinction is seen from the story being told by others even there (in the ākhyāyikā). Whether another (is) the speaker, or one’s self, is a sorry standard

16 VĀSAVADATTĀ of discrimination........ Therefore the kathā (and) ākhyāyikā are one category marked with a double name ; and here, too, will be comprised the other categories of stories.' ¹ In support of this statement of Daṇḍin, it may be noted that the Vāsavadattā, though termed, as we have seen, an ākhyāyikā lacks the necessary divisions into 'sighs' ; in its opening stanzas it (like a kathā) describes 'the character of knaves and the like' (introductory stanzas 6–9) ; and it contains a long episode spoken by another than the hero—the conversation of the maina with his mate concerning the heroine of the story. The manifest resem- blance of the Vāsavadattā to the Kādambarī, which is considered to be a kathā, together with its unlikeness to the Harṣacarita, whose technique it should share, were it really an ākhyāyikā, also serves to confirm the views of Daṇḍin rather than those of the Sāhityadarpaṇa. One need have little hesitation, therefore, in regarding the Vāsavadattā as technically a kathā.² The 'Style' and Rhetorical Embellishments of the Vāsava- dattā. The rīti, or 'style,' of the Vāsavadattā is the Gāuḍī, which the Sāhityadarpaṇa (No. 627) defines as follows (cf., in general, Regnaud, Rhétorique sanskrite, pp. 253–255, Paris, 1884): ōjaḥprakāśakāir varṇāir bandha āḍambaraḥ punaḥ samāsabahulā gāuḍī, 'the Gāuḍī, moreover, is a resonant arrangement (of words) with sounds expressing strength, (and) abounds in compounds.' Vā- mana, in his Kāvyālaṁkāravṛtti (1. 2. 12), describes this 'style' as 'consisting of strength (ōjas) and grace (kānti)', while avoiding 'sweetness' (mādhurya) and 'softness' (sāukumārya). Accord- ing to the Kāvyādarśa (1. 44a, 46a, 54a, 92a), moreover, the Gāuḍī especially affects alliteration, etymologising, and hyper- bole.³ When it is added that, as the Kāvyādarśa (1. 14–29) also ¹ It may be mentioned in passing that Ānandavardhana's Dhvanyālōka, 3. 8 (tr. Jacobi, in ZDMG. 56. 789), states that compound words are longer in the ākhyāyikā than in the kathā. ² I am glad to note that my conclusion in this respect is confirmed by Krishna- machariar (Introd., pp. 8–9). ³ Krishnamachariar (Introd., pp. 28–29) notes the prevalence in the Vāsavadattā

INTRODUCTION 17 states, an ākhyāyikā, kathā, or other form of narrative should, like poetry in general, include descriptions of battles, cities, oceans, mountains, seasons, sunrise, moonrise, and the like (each and all of which may be exemplified from the Vāsavadattā), we see at once how closely Subandhu was restricted in the composi- tion of his romance, and how faithfully and minutely he discharged his self-imposed task. The slender thread of narrative in the Vāsavadattā is em- bellished with many forms of literary adornment, which, indeed, constitute by far the major portion of the work. First and fore- most among these embellishments stands the śleṣa, or 'parono- masia,' and with good reason Subandhu declares himself to be 'a repository of cunning skill in arranging a series of parono- masias in every syllable' (pratyakṣaraśleṣamayaprabandhavinyā- savāidagdhyanidhir, introductory stanza 13). The śleṣa is well defined by Daṇḍin, in his Kāvyādarśa (2. 363; cf. Regnaud, Rhétorique sanskrite, pp. 227–229, Paris, 1884; Sāhityadarpaṇa, No. 705; Kāvyaprakāśa, tr. Jhā, pp. 188–197, 217–218, Benares, 1898; Kuvalayānandakārikā, 1. 62), as follows : śleṣaḥ sarvāsu puṣṇāti prāyō vakrōktiṣu śriyam bhinnaṁ dvidhā svabhāvōktir vakrōktiś cē 'ti vāṅmayam 'the paronomasia generally enhances the beauty in all equivoca- tions; the phraseology (is) divided in two parts: the natural meaning and the equivocal meaning.' Examples of the śleṣa, usually intimated in the Vāsavadattā by iva, 'as' (and indicated in this translation by < > or, when double and triple, by « », «« »»), abound in Subandhu's romance. As a single specimen may be cited vānarasēnām iva sugrīvāṅgadōpaśōbhitām, 'adorned with a of utkalikāprāya, or style of long compounds and words containing alliteration (Regnaud, Rhétorique sanskrite, p. 75, Paris, 1884), and of the vṛtti ārabhaṭī, or 'violent manner' of scenes of awe and conflict (Lévi, Théatre indien, 1. 92–93, Paris, 1890). The 'manner' is also sometimes madhyamakāiśikī (according to Vidyānātha, the author of the Pratāparudrayaśōbhūṣaṇa [cf. Regnaud, op. cit., pp. 377–378], quoted by Krishnamachariar, mṛdvarthē 'py anatiprauḍhabandhā madhyamakāiśikī, 'not conjoined with excessive dignity in a gentle theme'), and the style is mostly nārikēlapāka (according to Vidyānātha, sa nārikērapākaḥ syād antargūdharasōdayaḥ, 'the rising of hidden flavour'), although sometimes āmrapāka (for which no definition is given). C

18 VĀSAVADATTĀ ‹beautiful throat (sugrīva) and with armlets (aṅgada)› as the army of monkeys was adorned by ‹Sugrīva and Aṅgada›' (ed. Hall, pp. 63–64).¹ The figure next in frequence to the śleṣa in the Vāsavadattā is the virōdha, or ‘antithesis,’ where the superficial meaning is self- contradictory, while the paronomasiac reading renders the phrase consistent, and even intensifies it. This rhetorical embellishment is defined as follows in the Kāvyādarśa (2. 333; cf. Sāhityadar- paṇa, No. 718; Kāvyaprakāśa, tr. Jhā, pp. 233–235, Benares, 1898; Kuvalayānandakārikā, 1. 74) : viruddhānāṁ padārthānāṁ yatra saṁsargadarśanam viśēṣadarśanāyāi ”va sa virōdhaḥ smṛtō yathā, ‘when there is an apparent union of antithetical objects simply to show the distinction (between them), it is called virōdha.’ The conventional sign of the virōdha in the Vāsavadattā is api, as iva is indicative of the śleṣa. As an example of the countless in- stances of the virōdha in Subandhu’s romance, mention may be made of agrahēṇā ’pi kāvyajīvajñēna, ‘which ‹has no planets (a-graha)› yet knows «Venus (kāvya = Śukra = the planet Venus) and Jupiter (jīva = Bṛhaspati = the planet Jupiter)», for it is ‹free from theft (a-graha)› and knows «the essence (jīva) of poetry (kāvya)»’ (ed. Hall, pp. 113–114). Besides these two rhetorical devices, Śivarāma, in his com- mentary on the Vāsavadattā, enumerates a long series of alaṁ- kāras, or ‘adornments,’ which will now briefly be considered. The parisaṅkhyā, or ‘special mention,’ usually combined with the śleṣa in the Vāsavadattā, is an affirmative statement with the implied negation of the paronomasiac meaning of the phrase, and is thus defined by the Sāhityadarpaṇa (No. 735; cf. Kāvyapra- kāśa, tr. Jhā, pp. 245–246, Benares, 1898 ; Kuvalayānandakārikā,

1. 112) : ¹ On Subandhu’s fondness for paronomasia see, further, Krishnamachariar, Introd., pp. 18–20, who also calls attention to repetitions of paronomasia on the same word (p. 27) as well as to the frequent repetition of the same phrase in the romance (pp. 25–26).