VII

PREFACE

It remains for me to set forth what critical resources were available to me in preparing this edition and what method I adopted in using them. Therefore, the following codices were collated for our use and noted by us in such a way that from silence, one can always conclude agreement: I. Q (p. 1, 1—85, 33), P (p. 35, 7—118, 5), EF (p. 52, 4—64, 7). II. V (p. 1, 1—191, 31), B (p. 119, 1—485, 14), for which—as it is missing—p. 433, 8—452, 31 have been noted as GT, U (p. 186, 27—314, 24 and p. 387, 1—485, 14), Y (p. 315, 1—386, 6), L (p. 293, 17—297, 13; 324, 15—331, 14; 348, 11—355, 12), M (p. 301, 10—316, 32; 416, 12—422, 6; 452, 31—457, 10). Furthermore, I have noted the full variety of the Trincavelli edition (t), excluding obvious printer's errors. 1 Therefore, where QPEF are available (p. 1, 1—118, 5), two witnesses (Vt) were used; where they are missing, from p. 119, 1 onwards, in addition to LM, three witnesses of the common class were always used (BVt, BUt, BYt, B(GT)Ut). From the paraphrase (S), however, so as not to overburden the apparatus with things that anyone could know from the published edition, I considered it sufficient to note those things which either contributed to correcting the record of our codices, or, where they differ from each other, to verifying the reliability of individual ones, or were otherwise worthy of note, so that it can never be inferred from silence that the paraphraser agrees with our Philoponus or follows his traces at all.

Equipped with these resources, I performed the task of recension in such a way that, as far as possible, I followed the record of the second class, yet I did not fear to prefer the common version (Vt) where they seemed in some way preferable. At the beginning of the commentary (p. 1, 1—35, 7), however, where QVt were present, since Q deviated into error no less often than it alone preserved the truth, one might sometimes doubt which record to follow. Regarding the manuscripts of the common class BUVY(LM), since none surpasses the others in value to the extent that it should be preferred by itself, I always followed that record which seemed to be commended in some way by the usage of language, the order of thoughts, the authority of Aristotle, or confirmed by the number of witnesses. In these matters, I place not only the paraphrase (S), which I have said above also has some value for correction, but also the Trincavelli edition. For even if one may be cautious about what t alone provides, 2

1) Cf. vol. II 2 p. V and my commentary 'Die griech. Ausleger der Arist. Topik' (Scientific Appendix to the Program of the Sophien-Gymnasium in Berlin, 1891, Nr. 65).
2) We have refrained from printing or mentioning in the apparatus the tables and syllogistic figures with which the margins of the codices and especially the editio princeps are adorned, especially since the books by no means agree with each other, for example V lacks these tables and figures entirely. Examples of this are what we proposed in the Supplementary Preface, p. XXXVII-XXXVIII.