not to mention other copies he was consulting and collating. Further testimony, apart from this internal evidence, is provided by Rashi's references to texts parallel to the Talmud, among them the Tosefta. Sometimes he records two readings without giving either the preference, though as a rule the reasoning or the context shows that he leans one way or the other, so that his alterations, which are usually correct, do not necessarily represent the early text. When Rashi has good cause for deciding a point in a certain way, he does not pay attention to possible errors or contradictions on the part of the Talmudists. In other words, though his text may be the most rational, it is not always the most authentic. Rashi exercised this criticism of the text to a wide extent, yet prudently. I have already mentioned what Isaac of Vienna said concerning the numerous erasures that covered an autograph manuscript of his.[96] Many readings that Rashi rejected might have been kept - in fact they sometimes were kept - by force of finesse and subtlety. His method affords a striking contrast to that of the Talmudist Hananel,[97] who either eliminates the phrases unacceptable to him or preserves them only by doing violence to the sense. Rashi, on the contrary, compared the different versions of difficult or suspicious passages and prefers the one not requiring a subtle explanation. It is only when no reading satisfies him that he assumes an interpolation or an error, in this event frequently resorting to the Responsa of the Geonim. Needless to say, he also paid heed to the revision of Gershom; but since he deemed that Gershom had himself preserved faulty readings, he took up the work again, despite Gershom's prohibition. He realized that this careful and detailed critical revision of his predecessor, however ungrateful the soil might appear, was nevertheless fertile ground, and might serve as the solid basis of a thorough commentary. He acquitted himself of the task with such success that his has become the official text, the "Vulgate," of the Talmud. In fact, his disciples inserted into the body of the Gemara the greater part of his corrections or restitutions (but not all; and one does not always comprehend the reasons for their choice), which have now become an integral part of the text. Thus a single, definite, and official text was established - a thing of great value in assuring the stability of rabbinical tradition in France and Germany.