the would-be original Syriac text, and no mention of the novel itself occurs before the 11th cent. The result of PEETERS'S investigations, with which I fully agree, is that the legendary biography of Sâkyamuni translated into Pahivi about A. D. 500 (?) was later retranslated into Arabic, then into Georgian; the Georgian (or possibly the Arabic) version was used as a basis c. A. D. 1000 by the Greek author of the novel of Barlaam and Ioasaf, itself soon translated into Latin as the history of Barlaam and Josaphat. A Persian version of the biography, possibly translated from the Arabic, but more probably independent of it and based on the Pahivi, has been partly published by OL'DENBURG (ZVOIRAO, Iv, 229-265); a fragment in Turkish found by LE COQ (Türkische Manichaica aus Chotscho, I, [1912], 5-7) has been identified by OL'DENBURG (in RADLOV, Alttürk. Studien VI, Izv. Ak. Nauk., 1912, 779-782) as belonging to the same recension of the biography as the Persian text.
Under «Avenir », I have called attention to the final -r in the Arabic and Greek forms of that name, and not -s as in Georgian, as an indication in favour of the dependence of the Greek upon the Arabic instead of the Georgian. Of course, it is the Arabic writing which accounts for the alternation , b- and , i- (y-) at the beginning of Bodisaf > Yodisaf; but the form is also altered in the Georgian Iodasaf, so that, in so far as the initial is concerned, either the Georgian or the Arabic may account for the i- of 'Ico&aap. But in the rest of the Greek name, there is perhaps another indication of an Arabic, and not Georgian prototype. KUHN, as I said above, explained 'Iw&aacp through a
misreading in Greek for which there is no support in the Mss. On the contrary, a form ~ .1. Yoasaf
is actually given in one case by the Fihrist (cf. HOMMEL, in Verhandl. d. VII. Intern. Orient.- Congresses... 1886, Vienna, 1888, Semit. Sekt., 119) ; if the author of the Greek novel had an Arabic prototype with the same reading, he could find no better rendering than 'Iox ocp.
In 1888, SACHAU (Alberuni's India, I, xxxlll) still expressed some surprise at the equivalence Bodhisattva > Bôdâsâf, because « there is no law in Indian phonetics which admits the change of sattva to saf ». The texts discovered in Central Asia provide us now with many instances of Bodhisattva occurring in Buddhist Turkish texts as bodistv, bodastv and bodisavt (cf. BANG and VON GABAIN, Analyt. Index, in SPA W, 1931, 474). Probably Bodastv is a scribe's mistake; I hesitate to connect it with the Arabic -a- form Bôdâsâf. As for bodistv (= bodisäty) > bodisvt (= bodisävt), it is a metathesis of a type quite common in Iranian and Turkish languages; and it is of course bodisävt, not bodisäty, which has given Bbdâsâf. As far as I can remember, the form in -säv (> -säf ), not -sävt or -säty, has only been noticed once in a Turkish text, and the text is of Manichaean origin (cf. LE COQ, Ein christl. and ein manich. Ms.-Fragment, in SPA W, 1909, 1204-1205).
This leads me to speak of the only point on which I feel inclined to dissent from PEETERS'S views, that is the part possibly played by the Manichaeans in spreading the legend of Sakyamuni's life through Western and Central Asia. PEETERS'S attitude is negative. On the other hand, in 1917-1918, ALFARIC not only maintained that the Manichaeans had helped in the diffusion of Buddha's legend, but that Mani had used and transmitted to his disciples a gnostic life of Sakyamuni (JA, 1917, II, 269-288; Rev. d'hist. des Relig., Vol. 78 [1918], 227-235; reprinted in Les écritures manichéennes, II, 211-219).
Without sharing ALFARIC'S extreme opinions, I think that the part played by the Manichaeans cannot be dismissed so easily in this connection. While the same tale occurs in practically the same words in the old Persian version of Bilöhar and Bôdâsâf and in a manuscript discovered at Turfan by
IIIF Color HighRes
