国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Sec. iv] MSS. FROM THE HIDDEN LIBRARY, IN BRAHMY AND CHINESE 915

illustrate, it may suffice here to mention Ch. ii. 002, 003, which contain extensive portions, counting sixty-five and seventy-one folios respectively, of medical texts translated or extracted from Sanskrit originals, and Ch.00274, a Buddhist text in thirty-nine folios, apparently complete but as yet unidentified.

Among the Khotanese rolls, written almost exclusively in Cursive Gupta, we find Buddhist texts, some of them of considerable length, statements of an apparently documentary nature, and also medical formulae.'`= Their number and the way in which the blank reverses of old Chinese manuscript rolls have been utilized for them leave little doubt about their having been written locally. That there were settled at Tun-huang Buddhist monks familiar with the language and script prevalent in the Khotan region and elsewhere in the south of the Târim Basin may thus be safely assumed, and various indications point to these Khotanese texts having been produced at a relatively late period.13 But still more conclusive evidence that the Khotanese language was locally studied is supplied by the numerous alphabetic tables and syllabaries for the Cursive Gupta script which are contained among these rolls. As shown by Dr. Hoernle, who has very fully discussed these tables corresponding to the siddham-chang of Chinese Buddhist writers, they possess considerable value for determining the palaeography of a script rendered difficult by its generally very cursive, and often slovenly, character."

Sanskrit and Khotanese are not the only languages represented among our Brahmi manuscripts from Chien-fo-tung. Three folios, belonging to two different Pbthis, Ch. 00316. a, b (Plate CLII), contain portions of text in that newly discovered Indo-European language which, first designated as the ' Language I ', then identified with ` Tokhâri', has by a brilliant and convincing demonstration of M. Sylvain Lévi been proved to have had its principal home in the Kuchâ region and can, in consequence, be justly called by the name of ' Kuchean '.16 M. Sylvain Lévi, who at my request kindly examined our two Pôthi fragments, both written in Slanting Gupta, has identified them as belonging, one to a medical text, the other to a Buddhist poem bearing on the Udeinavarga. Extracts from them have since been published and utilized by him and M. Meillet in a paper on the grammatical forms of Kuchean.16 The disproportion in numbers between these few Kuchean leaves and the relatively plentiful Pôthis and rolls in Khotanese which I gathered from the Chienfo-tung hoard is certainly striking. It may be premature to draw any definite conclusion from it until it is known what Kuchean materials, besides the three leaves of a bilingual medical text in Kuchean and Sanskrit specified by M. Sylvain Lévi," rewarded M. Pelliot's search of the bundles

Khotanese MS. rolls in Cursive Gupta script.

Alphabetic tables.

Kuchean manuscript remains.

Idfyuh Sûtra, the Old Khotanese Version together with the Sanskrit Text and the Tibetan Translation, etc. in MS. Remains of Buddhist Literature, ed. Hoernle, i. pp. 2x 4-88, 289-356, resp.

12 For specimens of such rolls, see PI. CXLVII, CXLVIII. For Buddhist texts of great extent, cf. e.g. Ch. 0041, 0026569.

13 There is close resemblance both in style of writing and in language between the documentary rolls of Chien-fo-tung and the Khotanese records brought to light by me at sites like Dandan-oilik, Khâdalik, Mazar-tagh, all occupied down to the end of the eighth century or later. On the other hand, indications derived from Chinese palaeography have similarly led M. Pelliot to assume a relatively late date, eighth to tenth century, for the numerous ' Eastern Iranian', or Khotanese, manuscripts brought away by him from Chien-fo-tung; cf. Un fragment du Suvarnaprabhasasûtra, loc. cit., p. 3.

" Cf. Hoernle, The ' Unknown Languages' of Eastern Turkestan, II, J.R.A.S., 191r, pp. 45o sqq.; Pl. I—Iv. Such

tables and syllabaries form the sole or principal contents of the rolls Ch. 0042, 46, 271, 273, 327 ; i. oorg; xl. 002, 003; lviii. 007 (Pl. CXLV); C. 002.

15 Cf. Sylvain Lévi, Le ' Tokharien B ', langue de Koutcha, J. Asiat., sept.—octobre 1913, pp. 312 sqq. For a brief but lucid review of the researches bearing on this ' unknown' Central-Asian language, first rendered accessible for study by Dr. Hoernle's publication of the Weber-Macartney Manuscript (Igor) and recognized in its true linguistic character by Professors Sieg and Siegling 0908), cf. Sylvain Lévi, Étude des documents tokhariens de la Mission Pelliot, J. Asia., mai juin I9II, pp. 43r sqq.

'e Cf. S. Lévi and A. Meillet, Remarques sur les formes grammaticales de quelques lexies en tokharien B, in Mémoires de la Société de Linguistique de Paris, xviii. (reprint) pp. 2, 17, 21 sq.

17 See S. Lévi, J. Asia., mai juin IgII, p. 433. The absence of local ' site-marks' in the references made to other Kuchean materials from M. Pelliot's collection in MM. S. Lévi

6A2