science [of the Indians] [and he accordingly sends the Jew to Arin, whence he brings back the Indian numerals and several important astronomical works].”
There are two ways of explaining this account, supposing it to be substantially true. Either Al-Mokaffa employed the Jew as a “ghost” or “devil,” or there were two Arabic versions, one made from the Pehlevi, the other from the Sanskrit. In the former case it would not be surprising to receive different accounts from the “devil” and the advocate. But it would be difficult to account for the biography of the Persian Barzoye in a translation from the Sanskrit, and I am therefore inclined to think that Ibn Ezra’s account points to an independent translation by a Jew from the Sanskrit direct into Arabic. I am confirmed in this belief by the remarkable variations in the Arabic MSS., which clearly indicate two prototypes (Guidi, Studij sul testo arabo del libro di Calila e Dimna, Rome, 1873), but must reserve details for another place. And in this connection it is interesting to observe the reference to illustrations in the Indian book in Ibn Ezra’s account. We have seen that some
