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

explains that yār means « earth » and kyāng, « broad » (Hsi-yü t'u-chih, 18, 1 a; Hsi-yü t'ung-wên
chih, 3, 18-19). It is true that yār means « earth » in Turki; as to kyāng, the word intended is the
« kīng », « wide », « broad », of SHAW, Vocabulary, 177, the real pronunciation of which, as I heard
it in Chinese Turkestan, is kāng and keng (cf. RADLOV, II, 1067). The etymology is of course
absurd, and the form itself is artificial, being based on a nasalized pronunciation of Yārkān, the
Kalmuk form of Yārkānd. « Yārkāng » never occurs in Turkish texts, except perhaps in some local
documents purposely written for the use of Chinese officials. The Tibetan 'Jigs-med nam-mkha,
writing in 1818, gives Yer-khen (misread « Yed-khen » in HIRTH, Gesch. des Buddhismus, I, 18;
II, 28; cf. also Sarat CHANDRA DAS's Dictionary, 1143; the would-be Sanskrit name « Arghan »
quoted in WATTERS, On Yuan Chwang, II, 294, from a paper by Sarat CHANDRA DAS is valueless).
The present note has so far dealt with Yārkānd only from the time when the name « Yārkānd »
appears in the 11th cent. But the place existed at a much earlier date.
All scholars, both Chinese and European, agree that the kingdom of 疏 勒 So-chü of Han
times must be identified with Yārkānd. On the basis of this equivalence, the old name was officially
revived for Yārkānd when the « province » of Chinese Turkestan was created in 1882 : Yārkānd
became the « independent chou » of So-chü, was later promoted to a « prefecture » (fu), and became
the « district » (hsien) of So-chü when all « prefectures » were suppressed by the Republican Govern-
ment. HERRMANN adds that So-chü also is Ptolemy's Σοῖτα (VI, 15, 4), which he believes to be
corrupt for Σοΐγα; Σοΐγα and even Σάγχα occur among the readings of the Mss. (Southern Tibet,
VIII, 448; HERRMANN, Das Land der Seide und Tibet, III, 144). The geographical identification
is possible, but so far unproved. As to the names, the phonetic similitude, though real, is not
so clear as HERRMANN's words would seem to imply. Towards A. D. 600, So-chü sounded *Suâ-kịwo.
The first character, so (*suâ), always stand for sā (or sæd) in early transcriptions. The second one,
chü (*kịwo), is also pronounced ch'ê (*t'ś'ja), and it is with the latter pronunciation that it is used
in post-Han times in Buddhist transcriptions, for instance to render the last syllable of mleccha
or the second one of Licchavi. But in the case of the old name of Yārkānd, the traditional pronun-
ciation is So-chü, and the use of the second character in transcriptions of Han times with the chü
(*kịwo) pronunciation is confirmed by the double form of the native name of the ancient kingdom of
Turfan, which is 姑 師 Ku-shih (*Kuo-şi) in Ssü-ma Ch'ien, but 車 師 Chü-shih (*Kịwo-şi) in the
Chien-Han shu. Although I have no doubt that there is an etymological connection between
the two pronunciations chü (*kịwo) and ch'e (*t'ś'ja), I think that in Han times the final vowel of
chü must have been a labial vowel. In other words, the theoretical original of So-chü probably
was *Sāku (or *Sākā?), which is fairly far removed from *Σοΐγα. I must add that the readings
Σοΐγα and Σάγχα seem to be of small authority, the only important variant of Σοῖτα being Σότα.
Saga is, however, the form given by Ammianus Marcellinus (cf. CŒDÈS, Textes d'auteurs grecs et
latins, 94).
For a period which must be the 5th cent. (c. A. D. 437 ?), the Pei shih (97, 3 b) mentions a king-
dom of 渠 沙 Ch'ü-so (*C'ịwo-suâ), which had its centre at the «ancient city of So-chü». Here again
Chinese scholars and HERRMANN agree that Ch'ü-so must be Yārkānd (cf. Hsi-yü t'u-chih, 18, 6 a;
Southern Tibet, VIII, 19, 440); but this is because they leave out of consideration all sources except
the Pei shih. As is well known, this section of the Pei shih is based to a great extent on the original
17 a.