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

is RUGE's misreading). In other words, we would have here another example of -g- (-gh- before i
not to give -j-) used to mark the hiatus stop (see « Coigangiu »).

The hsien of Pao-ying received that name in 676, became a chou in 1227, and almost immedia-
tely (?) thereafter was changed into a military area (chün); it is under the name of Pao-ying-chün
that the diarist of 1276 registers his passage there (cf. TP, 1915, 396, 413-414). In 1279, it was
raised to An-i-fu, but degraded to Pao-ying-hsien in 1283 (cf. YS, 59, 11 a; TP, 1915, 414; Ta-Ch'ing
i-t'ung chih, 66, map, 3 a; text, 2 a). The name used by Polo is thus quite correct for the time when
he wrote.


312. PEM

pain, paines, pera FB pem Ft, FA(?), Z piem, pien VA
pein F, Fr, FA, L, V(cor.) peym TA¹, TA²; R pin V
peiti, terchin VB peyn LT, P; G poyn VL
peiu F

This is certainly the correct form, despite BENEDETTO's hesitation (B¹, 446); we might only
suppose that « Pein » of F is Peim > Peï > Pein, and that the « Pem » of Z is this same last form Pein
misread *Pem (hence *Pê > Pen in Fra Mauro, Zu, 36; HALLBERG, 409). R has « Peym ». There is
no doubt that Polo dictated a form with final -m. « Pem » is Hsüan-tsang's P'i-mo (*P'ici-muâ);
this last name has been restored into Bhīmā by Stan. JULIEN, but the p'i of P'i-mo is an ancient
aspirate (p'-), and the theoretical reading is *Phema, or possibly *Phima. The « Pim » of the ancient
Turkish map mentioned in Y, I, 192, is interesting as a survival, but gives no clue to the proper
pronunciation, as Pim can also be read Bim, Pêm, Bêm. More stress may be laid on the Phye-ma
of ancient Tibetan texts referring to the region of Khotan, and F. W. THOMAS is probably right in
seeing in that Phye-ma Hsüan-tsang's P'i-mo (Zeitschr. für Buddhismus, September 1924; Notes
relating to... Ancient Khotan, p. 2 of the reprint).

« Pem » has been placed by HUNTINGTON at Keriya, by STEIN at Uzun-tati (better Uzun-tatïr,
« Long Tatïr »; tatïr means a stretch of hard barren ground), by CHARIGNON at Endereh. But
HUNTINGTON's view belongs to an early stage of the archaeological study of Chinese Turkestan;
CHARIGNON's theory (Ch, I, 104-107) is vitiated from the start because he denies the identity of P'i-
mo and Pem and keeps « Pein » so as to identify it phonetically (!) with the T'ang station of Po-
hsien. In the present state of our knowledge, the identification of P'i-mo and Pem with Uzun-
tatïr is the only plausible one, and it is at least probable.

I leave aside Abū Dulaf's بيم Bahā, corrected to بيم Bimā by the last commentators (cf. Fe,
217-218). The form is uncertain, the location unknown, and the whole account most untrust-
worthy.