183. COTTON 441
pei (*kivp pudi; cf. ibid., vi, 59 a; vII, 58 a; ix, 147 a). The chieh- [ ]chü or chieh-chü-so in SOOTHILL and HoDOUS's Dictionary of Chinese Buddhist terms, 232, is a misprint; and so is «kiupei », chü pei, instead of « kie-pei », chieh-pei, in PRZYLUSKI's paper, JA, 1919. i, 407; outside of Buddhist texts, chieh-pei occurs in a verse of the T'ang poet P'i Jih-hsiu, doctor of 867. A text
in Tripit., < , III, 99 b, speaks of Ag 51 Chia-shih chieh-pei, « Kàsi cotton », i. e. Benares
muslin (kdsi or kdsika is not « silk », as said by COWELL, Divydvaddna, 678; the same is true of Pali kiisika). One is immediately struck by the fact that none of these transcriptions is made directly from karpdsa; all are based on Prâkrit forms beginning like the Pali kappdsa. As a matter of fact, if we except Li Yen's Sino-Sanskrit Vocabulary published by BAGCHI, karpdsa never occurs as such in Chinese Buddhist texts, where we find only kdrpdsika (cf. supra, p. 433), and the Mandvyupatti registers only keirpdsakam (No. 9164). In the middle of the 7th cent., Hsüan-ying, commenting on the above transcriptions which he declares to be erroneous, states that the true
Sanskrit word was A NI chia-po-lo (*ka puâ-lâ), and Hui-lin copies it on his own account; but
chia-po-lo can only render kapeila, « skull ». ODA Tokuno has suggested in his Dictionary (p. 470') that the second and third characters may have to be reversed; but the result, *chia-lo po, would only give karpd-, without -sa. I rather think that Hsüan-ying misunderstood some information imparted to him by Hsüan-tsang, a fact which of course would sadly reflect upon the Sanskrit attainments of both Hsüan-ying and Hui-lin. As to the chia-lo- [ ]p'o-chieh given in STUART'S Materia Medica, 198, as a Sanskrit name of cotton, this is a mistake of the Pen-ts'ao kang-mu (36, 71 b; it has passed into T'u-shu chi-ch'êng, ts'ao-mu tien, 303, 1 a, 11 a), due to a partial juxtaposition of chic-lo po and chieh-pei, and simply does not exist. Chieh-pei-so seems to be an artificial, half-scholarly transcription, in which an attempt has been made to add the final -sa of Pali kappa-sa to the transcription already current chieh-pei; but, as will soon be seen, chieh-pei is not merely kappa-. Chieh po-yü would suppose an original like *kappdiuk; it probably represents a Prâkrit form *kappdik" corresponding to the adjectival form kappdsika, « made of cotton », « cotton stuff », of the Pali. This last form, which is the more usual, is the most interesting, since it supposes a Prâkrit *kappei , corresponding to the feminine Pali form kappeisi, « cotton ». Now this is exactly the stage of the final element in ku-pei, which would suppose *kupdi, or chi-pei, which in its turn would suppose *kirpdi. The Cheng-tzû t'ung (s. v. ~j; mien) says that chi-pei is more probable than kupei, because it is « nearer » chieh-pei. As to Yü Chêng-hsieh, he boldly states (Kuel-ssd lei-kao, 7, 21 a) that ku-pei is merely a misprint of the Sung editions, and, in his special section on cotton (Mu-mien k'ao; ibid. 14, 4-6), always writes chi-pei, without mentioning any other reading. But we must not forget that we have no editions older than the Sung; on the other hand, one cannot see why an almost general corrupt reading ku-pei should have crept into the Sung editions of the dynastic histories and of the Tripitaka, when chi-pei was in fact the only form alive in Sung times. The opinion of the authors of the Chêng-tzei t'ung and of Yü Chêng-hsieh seems to have been determined by the fact that both chi and chieh have an -i-, which ku never had, and that both arc ju-shêng words, while ku is pronounced with shang-shêng; but they were not aware of the ancient final consonants, one being -t in chi (*kiét), and the other -p in chieh (*kill)). As we have seen, chi-pei (*kiétpuâi) would suppose *kirpdi *Skr. kdrpdsi, but the vowel of the first syllable would be abnormal, and moreover it seems impossible that a Prâkrit form in which °past had been
IIIFカラー高解像度
