Pramāṇa · Search · About · API

A History of Indian Literature v2.c19.s6.t1: Remnants of the Sanskrit Canon

A History of Indian Literature v2.c19.s6.t1

Language
English

Remnants of the Sanskrit Canon

v2.c19.s6.t1.1

In the first place the Sarvāstivāda school² of the Hīnayāna, which had its adherents more especially in Kashmir and Gandhāra, and spread thence to Central Asia, Tibet and China, had a Sanskrit Canon of its own. Though no complete copy of this canon has come down to us, we know it firstly, from many fragments, large and small, which have been discovered among the manuscripts and block-prints brought from Eastern Turkestan by M.A. (Sir Aurel) Stein, A. Grünwedel, A. von Le Coq, P. Pelliot, and others; ¹ secondly, from quotations in other Buddhist Sanskrit works, such as Mahāvastu, Divyāvadāna and Lalita-Vistara, and lastly, from Chinese and Tibetan translations.² The principal texts of the Mūla-Sarvāstivāda Canon were translated from Sanskrit into Chinese by the Chinese pilgrim *I-tsing* in the years 700-712.³ There are, however, Chinese translation of single texts dating from the middle of the 2nd century onwards, and there were adherents of the Sarvāstivāda in India as early as the 2nd century B.C.⁴ In wording and in the arrangement of the texts, the Sanskrit Canon evinces great similarity to the Pāli Canon, but on the other hand, there are many points of difference too. A feasible explanation of this is that both canons had a common source, probably the lost Māgadhī Canon, from which first the Pāli Canon branched off in one part of India, and then, later on, the Sanskrit Canon in another district.

v2.c19.s6.t1.2

² ) The literal meaning of Sarvāstivāda is "the all-is doctrine," the doctrine that everything exists permanently. *I-tsing* (A Record of the Buddhist Religion, transl. by J. *Takakusu*, Oxford, 1896, pp. XXIV, 8 ff., 20) calls it Āryamūlasarvāstivāda, and he mentions the schools of the Mūla-Sarvastivādas, the Dharmaguptas, the Mahīśāsakas and the Kāśyapīyas, as its sub-divisions. The relationship between the Sarvāstivāda and the Mūla-Sarvāstivāda is, however, by no means clear. *Cf.* *Takakusu*, ERE XI, p. 198 ff.; *Lévi* in T'oung Pao, sér. 2, t. VIII, 1907, p. 114 f., N. *Dutt*, Early History of the Spread of Buddhism and the Buddhist Schools, p. 271 n. and table to p. 224. In the Buddhist dictionary Mahāvyutpatti, para. 275, a list of the schools begins with: Āryasarvāstivādāḥ,

v2.c19.s6.t1.3

¹ Cf. *W. Geiger*, Die archaeologischen und literarischen Funde in Chinesisch Turkestan und ihre Bedeutung für die orientalistische Wissenschaft, Erlangen (Rede beim Antritt des Prorektorats) 1912; *H. Oldenburg* in NGGW 1912, 171 ff.; *H. Lüders*, Über die literarischen Funde von Ostturkestan, in SBA 1914, p. 85 ff.; *A. F. Rud. Hoernle*, Manuscript Remains of Buddhist Literature found in Eastern Turkestan, I, Oxford, 1916.

v2.c19.s6.t1.4

² Cf. *Oldenburg* in ZDMG 52, 1898, 654 ff., 662 ff.; *M. Anesaki* in OO XIII, Hamburg, 1902, p. 61; Le Muséon, N. S. VI, 1905, p. 23 ff.; VII, 1906, p. 33 ff.; JRAS 1901, p. 895 ff.

v2.c19.s6.t1.5

³ *J. Takakusu*, A Record of the Buddhist Religion by I-tsing translated, p. XXXVII. Cf. *Ed. Huber* in BEFEO VI, 1906, pp. 1 ff., 36; *S. Lévi* in T'oung Pao, sér. 2, t. V, 1904, p. 297 ff.; t. VIII, 1907, p. 110 ff. The Chinese "Tripiṭaka," though also divided into a Vinaya, Sūtra and Abhidharma, does not, however, mean the same as the Pāli Tipitaka, for it contains many non-canonical texts besides the canonical ones, in fact even Brahmanical philosophical treatises (*s. Takakusu* in JRAS 1896, p. 415). In the Tibetan Kanjur, too, which is also designated as a "Tripiṭaka" and consists of a Vinaya, Sūtra and Abhidharma, much is included which does not admit of comparison with the Tipitaka, and was certainly not a part of the Canon.

v2.c19.s6.t1.6

⁴ A Sarvāstivādin ācārya Buddhila is mentioned in inscriptions on the lion capitol in Mathurā dating from the 2nd century A.D., *s. F.W. Thomas*, Ep. Ind. 9, 1907-8, p. 135 ff.

v2.c19.s6.t1.7

Fragments of the Prātimokṣa-Sūtra of the Sarvāstivādins, as well as other texts of the Vinayapiṭaka of the Sanskrit Canon, have been found in Central Asia (Eastern Turkestan), and a few in Nepal too. It is also possible to reconstruct the Prātimokṣa-Sūtra from Chinese and Tibetan translations. Both the Sanskrit Canon of the Sarvāstivādins and the Vinayas of the Mahīśāsakas, Dharmaguptas and Mahā-Saṅghikas, show manifold divergences in separate details from the Pāli Canon and from one another, not only as regards the rules of the Prātimokṣa, but also as regards the rules of the Vinaya in general. Nevertheless, the original stock of rules is one and the same.¹ It is in the stories prefaced to the rules in order to relate on what occasion the Buddha proclaimed the rules in question, that the texts show greater divergences. The Mūla-Sarvāstivāda Vinaya contained many legends having reference to the conversion of Kashmir and North-western India to Buddhism.² Some of these legends are also to be found in the Divyāvadāna, which borrowed a large portion of its tales from the Vinayapiṭaka of the Mūla-Sarvāstivāda.¹ The Vinayapiṭaka also includes the Śikṣās and Saṅghakarmans corresponding to the Pāli Sikkhās and Kammavācās, and fragments of these have been brought to light in Central Asia.²

v2.c19.s6.t1.8

¹ The Sanskrit text of the Prātimokṣa-Sūtra has been edited by *L. Finot* on the basis of the fragments of manuscripts found by *Pelliot* in Central Asia, and has been translated into French, with the Chinese version of Kumārajīva, by *Ed. Huber*, in JA ser. 11, t. II, 1913, pp. 465-558 ; a fragment from M. A. Stein's collection, by *La Vallée Poussin* in JRAS 1913, p. 843 ; a fragment in the Kuchean language, by *S. Lévi* in JA s. 10, t. XIX, 1912, 101 ff.; JRAS 1913, p. 109 ff.; Hoernle, Manuscript Remains, p. 357 ff. *E. Waldschmidt*, Bruchstücke des Bhikṣuṇī-Prātimokṣa der Sarvāstivādins, Leipzig 1926, has edited the, fragments of manuscripts of the Prātimokṣa of the nuns, which were found at Turfan, and compared them with the corresponding texts of all the other schools. For the Tibetan version of the Prātimokṣa, see *G. Huth*, Die tibetische Version der Naiḥsargika-prāyaścittika-dharmāḥ, Strassburg 1891. See also *Oldenberg* in ZDMG 52, 1898, p. 644 ff.

v2.c19.s6.t1.9

² Cf. *J. Przyluski* and *S. Lévi* in JA s. 11, t. IV, 1914, p. 493 ff. In addition, the reign of Kaniṣka is prophesied, the seven-day week is mentioned, among other things, from which *Lévi* (T'oung Pao, s. 2, Vol. VIII, 1907, p. 115 f.) concludes that the Mūla-Sarvāstivāda-Vinaya was not completed before the 3rd century A.D. However, as there were already Sarvāstivādins at a much earlier period (see above, p. 232), it is most likely that there was an earlier Sarvāstivāda-Vinaya too, and possibly the Mūla-Sarvāsti-vāda-Vinaya was merely a later version of that work.

v2.c19.s6.t1.10

¹ Fragments corresponding to the Mahā-Vagga of the Vinayapiṭaka, found in the Stein collection, have been edited by *La Vallée Poussin* in JRAS 1913, p. 550 ff. A fragment of a Buddhist ritual of ordination, corresponding to the Culla-Vagga X, 17, in Sanskrit, was discovered by *C. Bendall* in Nepal and edited in Album Kern, p. 373 ff. Various fragments of the Vinaya, ed. by *Hoernle*, Manuscript Remains, pp. 4-16. A fragment of the Bhikṣuṇikā-dharma-vacana, ed. by Miss *C. M. Ridding* and *La Vallée Poussin* in BSOS I, 1920, p. 123 ff. See also *S. Lévi* in T'oung Pao, s. 2, Vol. VIII, 1907, p. 110 ff. ; *N. Dutt*, Early History of the Spread of Buddhism and the Buddhist Schools, p. 280 ff.

v2.c19.s6.t1.11

² Cf. *La Vallée Poussin* in JRAS 1913, 843 ff.; *Lüders* in SBA 1922, p. 243.

v2.c19.s6.t1.12

The Nikāyas of the Pāli Canon find their parallel in the Āgamas of the Sanskrit Canon; the Dīrghāgama corresponding to the Dīghanikāya, the Madhyamāgama to the Majjhimanikāya, the Saṃyuktāgama to the Saṃyuttanikāya, and the Ekottarāgama to the Aṅguttaranikāya. Fragments of various Sūtras of the Dīrghāgama (Saṃgīti-Sūtra, Āṭanātiya-Sūtra), of the Madhyamāgama (Upāli-Sūtra, Śuka-Sūtra), of the Saṃyuktāgama and the Ekottarāgama, have been found in Central Asia.³ These fragments are not always in agreement with the corresponding Pāli texts. The comparison of the Chinese Āgamas with the corresponding Pāli Nikāyas, has also shown both agreement to a considerable extent, and notable divergences.¹ In the first place there is no agreement as regards the order of the Sūtras in the separate Āgamas. In the Saṃyuktāgama the division into Vargas and Saṃyuktasa is totally different, and there are Saṃyuktas in the Pāli which are missing in the Chinese version; on the other hand, there are some in the Chinese which do not occur in the Pāli.² The greatest differences are those which we find between the Ekottarāgama and the Aṅguttaranikāya. On the whole it may be said that the points of agreement and the divergences prove that the Sanskrit Āgamas and the Pāli Nikāyas were compiled from the same materials, but were arranged in different ways in the different schools. Now in the Chinese Tripiṭaka there are also texts which, though bearing the same titles as the corresponding Pāli texts, are nevertheless essentially different. Thus there is a Brahmajāla-Sūtra, which was translated into Chinese by Kumārajīva, and is held in high esteem in China and Japan but which teaches Buddhist ethics in the spirit of the Mahāyāna. On the other hand, there are in the Abhidharma-Kośa-Vyākhyā, several quotations from a Brahmajāla-Sūtra, which must have corresponded to the Pāli text.³ The Chinese Tripiṭaka contains ten different translations of the Mahā-Parinirvāṇa-Sūtra. Three of these translations belong to the Hīnayāna and seven to the Mahāyāna; and the only thing they have in common is the external form, inasmuch as they contain speeches which the Buddha is supposed to have uttered prior to his death (parinirvāṇa).¹

v2.c19.s6.t1.13

³ Fragments of the Āgamas of the Sanskrit Canon were discovered amongst the remains of MSS. found in Eastern Turkestan, and edited by *R. Pischel* in SBA 1904, pp. 808 f., 1138 ff.; *S. Lévi* in T'oung Pao, s. 2, Vol. V, 1904, p. 297 ff.; JA s. 10, t. XVI, 1910, pp. 433 ff., 450 ff.; JRAS 1911, p. 764 ff. ; *L. de La Vallée Poussin* in JRAS 1911, p. 772 ; 1912, p. 1063 ff.; 1913, p. 569 ff.; *Hoernle*, Manuscript Remains, pp. 16-52. A fragment of the Śroṇa-Sūtras of the Saṃyukta-Āgama (cf. Saṃyuttanikāya 22, 49 f. = Vol. III, p. 48 ff.) has been edited by *La Vallée Poussin* from MSS. of Cecil Bendall's collection, in JRAS 1907, 375 ff. Quotations from the Saṃyukta and Ekottara Āgamas have been traced by *S. Lévi* in the commentary on Asaṅga's Mahāyāna-Sūtrālaṃkāra (ed. *Lévi*, on XVI, 21 f., 75). *V. A. Smith* and *W. Hoey* found Buddhist Sūtras in Sanskrit, written on bricks, in the ruins of Gopālpur, together with inscriptions which are dated between 250 and 400 A.D. (JASB, proceedings 1896, p. 99 ff.).

v2.c19.s6.t1.14

¹ See especially M. *Anesaki*, The Four Buddhist Āgamas in Chinese, in Transactions of the Asiatic Society of Japan, Vol. 35, Part 3, 1908, and above, p. 232, note 3.

v2.c19.s6.t1.15

² The Sagāthavarga in the three Chinese versions of the Saṃyuktāgama contains 318 Sūtras, out of which 244 coincide with those of the Pāli text. It is impossible to determine whether the quotations traced by *Anesaki* (Le Muséon, N. S. VII, 1906, 38 ff.) in the Chinese Mahā-Prajñā-Pāramitā-Śāstra, are derived from a Pāli or Prākrit text of the Saṃyutta and Suttanipāta. For the Saṃyukta Āgama, see also *Lévi* in T'oung Pao, s. 2, Vol. V, 1904, p. 297 ff.

v2.c19.s6.t1.16

³ Cf. *La Vallée Poussin* in JRAS 1908, 359 ff., *Anesaki* in ERE, Vol. V, p. 452.

v2.c19.s6.t1.17

¹ Cf. F. Max Müller in SBE, Vol. 10 (I), p. xxxi note; J. *Edkins* in JRAS 1881, p. 66 ff.; R. *Kimura*, in JDL 4, 1921, 188 f. note, and Historical Study of the Terms Hīnayāna and Mahāyāna, p. 94 note. The Hīnayāna includes the translation of Po-fâ-tsu (290-306 A.D.) and that of Fa-hien (between 317 and 420 A.D.); translated into Italian, after the last-named, by Carlo *Puini* in GSAI 21, 1908, p. 59 ff.; 22, 1909, p. 1 ff. The earliest Vaipulya-Parinirvāṇa-Sūtra was translated into Chinese by Dharmarakṣa (between 205 and 316 A.D.). Specimens of it are given by S. *Beal* in Ind. Ant. 5, 1876, p. 222 ff. and SBE Vol. 19, p. 365 ff.

v2.c19.s6.t1.18

The Sanskrit Canon also contained a "Kṣudraka" corresponding to the Khuddakanikāya.² We do not know whether this included all those texts which in the Pāli Canon are counted as belonging to the fifth Nikāya; but we do know that the Sanskrit Canon also contained the Sanskrit texts Udāna, Dharmapada, Sthavira-Gāthā, Vimānavastu and Buddha-Vaṃśa, corresponding to the Pāli texts Udāna, Dhammapada, Thera-Gāthā, Vimāna-Vatthu and Buddha-Vaṃsa. In the Chinese Tripiṭaka, it is true, there is no single text to correspond to the Suttanipāta; but more than half of the texts included in the Suttanipāta occur in the Āgamas, for instance the texts corresponding to the Aṭṭhaka-Vagga and the Pārāyana. Fragments of the Sanskrit version of the two last-named have been found in Eastern Turkestan too.³ A collection corresponding to the Itivuttakas was translated into Chinese by Hsüan Tsang in about 650 A.D.⁴ Considerable quotations, forming a parallel to passages from the Khuddakapāṭha, the Vimāna-Vatthu, Buddha-Vaṃsa and Dhammapada, are to be found in the Mahāvastu.¹

v2.c19.s6.t1.19

² A Kṣudraka is mentioned by Hsüan Tsang as the fifth Āgama of the Śrāvakapiṭaka, s. *Lévi* in JA s. 11, t. VIII, 1916, p. 20. Nevertheless, the "four Āgamas" are especially emphasized as the real Āgamas, as for instance, when the Divyāvadāna makes frequent mention (pp. 17, 331, 333) of the *āgamacatuṣṭayam*, cf. *Oldenberg* in ZDMG 52, 654 f.

v2.c19.s6.t1.20

³ Cf. M. *Anesaki* in JPTS 1906-7, p. 50 f.; *Hoernle* in JRAS 1916, p. 709 ff.; 1917, p. 134. In the Sanskrit version of the Aṭṭhaka-Vagga (it is mentioned as the arthavargīyāṇi Sūtrāṇi in the Divyāvadāna, p. 20) prose narratives precede the verses, but there are none such in the Pāli. There are also other divergences from the Pāli text. The Sanskrit text is shorter.

v2.c19.s6.t1.21

⁴ K. *Watanabe* in JPTS 1906-7, p. 44 ff. In the Chinese Tripiṭaka there is a work Ṣaṭpāramitā-saṃnipāta-Sūtra (translated in 222 A.D.) which is a kind of Cariyāpiṭaka; s. R. *Kimura*, A Historical Study of the Terms Hīnayāna and Mahāyāna, p. 129, note 1.

v2.c19.s6.t1.22

¹ Mahāvastu, ed. *Senart*, I, 290 ff.; II, 191 ff.; I, 250 ff.; Quotations from "the Dharmapada" (sing) or "the Dharmapadas" (plur.): II, 212; III, 91, 156, 434 ff.

v2.c19.s6.t1.23

We glean from one of the Mahāvastu quotations,² that there was a Dharmapada in the Sanskrit Canon, divided into Vargas, after the style of the Pāli Dharmapada. This Sanskrit version was not identical with that of the Pāli Canon, but both versions had an original stock of verses in common, and were arranged on one and the same principle. The collection of verses which belonged to the original Canon, also served as a prototype and as a foundation for more comprehensive anthologies, which were compiled in later times, and were widely propagated in Central Asia, Tibet and China. Thus, for instance, there is a Chinese anthology with the same 26 sections (Vargas) as the Pāli Dhammapada though it has 13 additional sections.³ Fragments of a Sanskrit anthology have been found in Central Asia, which were at first regarded as passages from recensions of the Dharmapada, but which in reality belong to the Udāna-Varga, which is known to us from a Tibetan translation. This anthology was compiled by Dharmatrāta, who, according to Tāranātha, is supposed to have lived in King Kaniṣka's day. In the Tibetan version, it contains not only 375 verses, corresponding to verses of the Dhammapada, but also the greater part of the verses of the Udāna, and numerous verses which have parallels in the Suttanipāta and other portions of the Pāli Canon.⁴ Another anthology was written in a Prākrit dialect, which is a later form of the North-west Indian dialect familiar to us from the Shāhbāzgarhī rock edicts of Aśoka, and compiled after the pattern, and with the help of the Dhammapada or some similar verse collection of the Sanskrit Canon, or the original Canon. Fragments of this anthology have come down to us in a few leaves of a manuscript written in the Kharoṣṭhī script, which M. Petroffsky and J. L. Dutreuil de Rhins had found at Khotan, and brought to Europe.[^1)

v2.c19.s6.t1.24

² III, 434 ff. Here a Sahasravarga is quoted from "the Dharmapadas;" it contains 24 verses, whilst the Sahassa-Vagga in the Pāli consists of only 16 verses. On the Dharmapada in Nepal, Tibet and China, s. *Oldenberg* in ZDMG 52, 1898, p. 662 ff.

v2.c19.s6.t1.25

³ Cf. *Max Müller* in SBE, Vol. 10 (1), p. 2 ff.; S. *Beal*, Texts from the Buddhist Canon commonly known as Dhammapada, London 1902; and the important study on the Dharmapada recensions by *S. Lévi* in JA. s. 10, t. XX, 1912, pp. 203-294.

v2.c19.s6.t1.26

⁴ Fragments of the Udānavarga, from finds in Central Asia, have been published by *R. Pischel*, Die Turfan-Rezensionen des Dhammapada, SBA 1908, p. 968 ff.; *S. Lévi* and *La Vallée Poussin* in JA s. 10, t. XVI, 1910, p. 444 ff.; t. XVII, 1911, 431 ff.; t. XIX, 1912, p. 311 ff.; *JRAS* 1911, p. 758 ff ; 1912. p. 355 ff. The Tibetan version of the Udānavarga has been translated into English by W. W. *Rockhill*, London 1883, and the Tibetan text was published by H. *Beckh*, Berlin 1911. Cf. A. *Schiefnen* Über Vasubandhus Gāthāsaṃgraha, in Mélanges Asiatiques VIII (Bulletin XXV, 1878), pp. 559 f., 590 ff.; *La Vallée Poussin* in GGA 1912, 191 f.; *Lüders* in SBA 1914, p. 102; and *Seidenstücker*, Das Udāna, I, 1913, pp. 37, 71 ff. A voluminous anthology, about five times as large as the Pāli Dhammapada, is the *Dharmasamuccaya* by Avalokitasimha, who has also drawn from Mahāyāna sources; on a MS. of this work, found in Nepal, see Dharma Aditya *Dharmācārya* in Ind. Hist. Qu. 1, 1925, 422 ff., 677 ff. [^1): ¹) The old Kharoṣṭhī manuscript, important from the points of view both of language and palaeography, has been edited by *Senart* in JA s. 9, t. XII, 1898, pp. 193 ff., 515 ff. This edition was the foundation for the new edition by Benimadhab *Barua* and Sailendranath *Mitra*, Calcutta 1921. The fragments found by *Petroffsky* were identified by S. *d'Oldenburg*, St. Petersburg 1907. On the "MS. Dutreuil de Rhins," cf. Comptes rendus de l'académie des inscriptions, 14. mai 1895 et 15. avril 1898; *Senart*, OC XI Paris 1897, I, 1 ff.; G. *Bühler*, On the Origin of the Indian Brahma Alphabet, 2nd Ed., Strassburg 1898, p. 122 ff.; *Lüders* in NGGW 1899, p. 474 ff.; T. W. *Rhys Davids* in JRAS 1899, p. 426 ff.; R. O. *Franke* in ZDMG 60, 1906, p. 477 ff.; M. A. *Stein*, Ancient Khotan, Oxford 1907, I, 188; Jules *Bloch*, JA ser. 10, t. XIX, 1912, p. 331 ff.; *Konow* (Festschrift Windisch, p. 85 ff.) is of opinion that, though this anthology is composed in a dialect originating in the North-west of India, it was written in the neighbourhood of Khotan. Like *Bühler* (l.c.) he ascribes it to the 1st century A.D., whilst *Lüders* (SBA 1914, p. 101) thinks it was written in the 3rd century A.D. It is probably the earliest Indian manuscript which has been discovered so far.

v2.c19.s6.t1.27

It is a certainty that there were Jātakas too, in the Sanskrit Canon; but it must remain an open question whether this Canon ever included a Jātaka Book or a large collection of Jātakas of its own. It is doubtful, too, whether the collection of "seven Abhidharmas," which is translated in the Chinese Tripiṭaka, is also derived from the old Canon, for these Abhidharmas have nothing in common with the Abhidhammapiṭaka of the Pāli Canon save the number seven and a few titles.¹

v2.c19.s6.t1.28

¹ J. *Takakusu*, JRAS 1905, p. 3 ff. and JPTS 1905, p. 67 ff. See above, p. 173.

v2.c19.s6.t1.29

The Canon of the Mūla-Sarvāstivādins has thus only come down to us in an incomplete way, but as regards the other Buddhist sects, we do not know whether they had any set Canon at all. As far as we can see, each sect merely has one or several texts which pass as especially sacred, being regarded as a kind of "Bible," and which partly assimilated and partly displaced the old texts of a Tripiṭaka which they, too, acknowledged in principle.

Linked texts (19)