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. ii]   FINDS IN A POLYGLOT LIBRARY   819

limits of the use of these Sogdian texts in Buddhist communities of Eastern Turkestan, nor the extent of the territories into which this use had been carried. But the fact that some of these Sogdian texts from the Chien-fo-tung hoard are written on the back of Chinese rolls which manifestly belong to Tang times and were put to fresh use locally is instructive by itself. It proves that Buddhist monks of Iranian origin must have penetrated to Tun-huang. This perfectly accords with what Professor Pelliot has been able to prove, partly from texts among the Chinese manuscripts I brought back from the ` Thousand Buddhas ', about a Sogdian settlement under a chief from Samarkand having been established in the Lop tract from about the middle of the seventh century.I"

It was from one of those ` mixed ' bundles (Ch. xxxix) that there emerged on the third day of my search a remarkable manuscript, exhibiting a third variety of the Syriac script transplanted to Central Asia, and the one which discoveries at the ruined sites of Turfan first revealed as peculiar to Manichaean writings. It was an excellently preserved narrow roll of paper, close on 15 feet long (see Plate CLXII for specimen portion), containing the beautifully written and almost complete text of what Professor A. von Lecoq has since recognized as the Turkish version of the Khuastuanift, a Manichaean confession prayer.17 The discovery amidst Chinese Buddhist texts and monastic records of this Manichaean relic was interesting, indeed, but could scarcely cause surprise ; for the Turfan excavations and finds had previously shown us how easy it evidently was for Mani's church in Central Asia to share the same sacred site with Buddhist cult, and with Christian worship, too, though remaining for centuries a formidable rival to them both.

We shall see further on that the presence of Turkish-speaking Manichaeans at Tun-huang is also attested by another important find from the walled-up chapel. I mean the perfectly preserved small book in Turkish ' Runic ' script seen in Plate CLX.1H But this did not come to light until the solid bundles of Chinese Sutra rolls subsequently acquired from the hoard could be searched in London. During the time of my actual stay at the caves only some torn fragments of a text in Turkish Runic script, Ch. 0014 (Plate CLXI),19 afforded proof that this earliest known form of Turkish writing had also been familiar to some of those who visited the sacred site before its great deposit of relics was hidden.

Discovery of Manichaean manuscript roll.

Manuscript book in Runic' Turkish.

SECTION III.—ACOUISITION OF MAINPUSCRIPTS AND ART RELICS

However interesting such stray finds were as illustrations of the remarkable polyglot aspect Miscel-

laneous which this place of Buddhist worship must have once presented, and whatever their philological monastic

value might prove hereafter, I could not expect them to offer much help towards settling the records, etc.

question, which archaeologically was of special importance, as to the date at which the chapel was walled up. I realized from the first that for this purpose the miscellaneous records in Chinese; such as monastic documents, letters, memoranda, and accounts, which filled those bundles in abundance, were bound to prove far more useful. Guided by Chiang Ssti-yeh's rapid examination of their contents and by peculiarities of their shape and paper (for specimens, see Ch. 365, 1283, Plate CLXVIII), I soon learned to distinguish them and to pick them out where they were embedded among packets of Chinese Buddhist texts. They were likely to throw instructive light, not only on

16 Cf. Pelliot, Journal Asial., janv.—févr. 1916, pp. 120 sqq.; also above, pp. 653 sq.

" For an edition and annotated translation of this text, see Prof. von Lecoq's paper Dr. Stein's Turkish Khuasluanifl from Tin:-huang, J.R.A.S., 191 r, pp. 277-314.

19 Cf. Professor V. Thomsen's paper, J.R.A.S., 1912, pp. 190 sqq.; also below, p. 921.

19 See for Prof. Thomsen's analysis of these fragments, J.R.A.S., 1912, pp. 215 sqq.

5M2