n Fu-tu ». The character 51 has two pronunciations, to (*tuo) and she (*d"z'ia), but it occurs in transcriptions thousands of times, and it has never been found to be used in these transcriptions with any other value than shê. Its regular value in transcription is Ja (or fa) ; Ltvi was absolutely right on this point. There is something more. In the name of one of the Khotanese kings, the texts give as the first two characters not Fu-shê, but fC gjj Fu-shih (*B'juk-si), which in all likelihood is the rendering of a dialectical pronunciation of the same name or title (cf. CHAVANNES, Doc. sur les Tou-kiue, 127, 207, 273; TP, 1904, 48); but, although shih (*,si) has no to reading, THOMAS adds to between brackets as an alternative sound after shih (Tibetan Texts and Documents, I, 162); this is clearly impossible.
Although I agree with LEvi that the second character must be read shê, with a transcription value Ja, I am in great doubt as to his correction of the first character as f fei (*b'iwvi). It is true that this character occurs in transcriptions, but in rendering ve or vai rather than vi. What is more important is that Fu-shê, and for one king Fu-shih, occur in a number of Khotanese royal names, in both dynastic T'ang histories, in the Ts'ê fu yüan-kuei, and in the T'ang hui yao (73, 23 a), and there is not one instance in which fei is used instead of fu. My conclusion is that we must retain Fu-shê. If this is, as I think, another transcription of Vijaya, Vi a or the like, we may imagine different solutions. One would be that the vowel of the first syllable was slurred and that the use of the first character with a final consonant chosen by a Court interpreter who was not, like Hsüan-tsang, a trained phonetician, was intended to render something like *Vila or *Nisi. Another hypothesis would be to read fu with its other and less
usual pronunciation *b'i2u. In such a case, *B'jau-d'i'ja or *Blau-si would render *Vu)ä or
*Vusi, both forms in which the initial labial consonant has developed after it a labial vowel as in Skr. visatya Arab. bu. an. At the present state of our knowledge it would be premature to try to choose between the two explanations.
But, while I believe that Fu-shê and Fu-shih are only other transcriptions of the name Vijaya adapted somewhat earlier as Yü-ch'ih, a serious difficulty remains which LEvi, KoNOw and THOMAS have overlooked. Since both Yti-ch'ih and Fu-shê or Fu-shih represent Vijaya, we ought not to find names in which both forms occur simultaneously. We find, however, Yüch'ih Fu-shih (without the last character of the name, the one which is really personal) in Chiu T'ang shu, 198, 8 a, and in Ts'ê fu yüan-kuei (cf. TP, 1904, 48), Yü-ch'ih Fu-shih Chan in Hsin T'ang shu (CHAVANNES, Doc. sur les Tou-kiue, 127), Yü-ch'ih Fu-shê Ta and Yü-ch'ih Fu-shê Hsiung in Ts'ê fu yüan-kuei (cf. TP, 1904, 23, 24; p. 56, CHAVANNES'S index is misleading as it gives Fu-t'u [J as if it were the actual reading in both passages; but this t'u is only a clerical error for jj shê in the text translated, p. 23; in that of p. 24, the text has shê as usual; moreover, the Ts'ê fu yiian-kuei correctly gives twice Fu-shê Hsiung in ch. 999, 8 b, 9 a, for two
embassies of 674 and 687 respectively, overlooked by CHAVANNES). If we had only the case of Yü-ch'ih Fu-shih or Yü-ch'ih Fu-shih Chan, we might suppose that, in spite of what seems
natural, Fu-shih is not the same as Fu-shê, and that Fu-shih or Fu-shih-chan really is the personal name of that king. But the solution would not be possible for the Yü-ch'ih Fu-shê Ta and the Yü-ch'ih Fu-shê Hsiung of the Ts'ê fu yiian-kuei. I incline to the following explanation : since the current notices of Khotan give, from the beginning of the T'ang dynasty, the family
IIIF Color HighRes
