It is quite understandable that Silvia Fazzo wished to defend her edition of book Lambda of Aristotle’s Metaphysics against what she has taken to be a ‘dismissive’ review of her book. To leave no puzzle unsolved, I should make clear that my aim was neither to dismiss her edition – which I did commend for its following the manuscripts E and J – nor to ‘impress’ the readers, as she also surmises, but to provide them with tools and substantial information regarding the textual transmission of the Metaphysics, so as to allow a right appreciation of Fazzo’s edition of book Lambda or, for that matter, of any edition of the Metaphysics. Indeed, I consider Fazzo’s edition to be an essential part of her project, and a sound stemma codicum to be the most essential step towards a sound edition of a text. Since Fazzo’s response is quite likely to have put the readers of BMCR into perplexity, I must now assume the inelegant task to reply to her response.
Fazzo has apparently not yet understood my main objection to her edition; or, else, she is committed to a piece of sophistry when she writes that “[Golitsis] recognizes”, that is, in accordance with her, “that Ab splits away from the M-C subfamily in Lambda 1-7”. Nobody would disagree that this is so, not only for Lambda 1-7 but for the whole Metaphysics, since it is plain that these manuscripts are copies of different exemplars. Our divergence consists in that (1) where I take A b to be, either directly or indirectly, a more or less faithful copy of the lost hyparchetype β, Fazzo says that it is a copy of the lost manuscript δ, which in its turn goes back to the lost hyparchetype α; and (2) where Fazzo believes that M and C are faithful (and not ‘direct’, as she has me saying) copies of β – it is true that Fazzo does not explicitly says so, but the stemma codicum she provides at p. 117 of her book implicitly says so; see also p. 138, where she qualifies M as “eminente rappresentante della famiglia β” –, I take them to be, either directly or indirectly, copies of ε, a lost manuscript that has been collated and corrected according to a surviving manuscript of the α-family, namely H a. This is confirmed not only by some cases in Lambda but by further cases in the whole Metaphysics, which of course could have no place in the limited space of a review. Fazzo insists on the thesis expressed in her publication of 2010, namely that A b “moves from β to α” at Kappa 8, 1065a25. As this thesis does not immediately concern Lambda, I contented myself with referring in note 5 of my review to a forthcoming article, in which I extensively deal with it. For the time being, let it suffice to say that Fazzo has failed to recognize in Kappa not only ε’s partial dependence on H a but also ε’s revision according to the text of the Physics, given that Kappa 8-12 is more or less a collection of extracts from books II, III and V of the Physics. These are the two main reasons for which A b and ε(ΜC) diverge considerably in that part of book Kappa.
I regret to say that Fazzo adduces in her response some further weak arguments in support of her thesis. She appeals to the extracts of Lambda found by Stefan Alexandru in V k, a manuscript copied by George Scholarios around the middle of the fifteenth century, in order to corroborate her reconstruction of the lost readings of β. In principle, however, V k alone can only confirm the reading of the lost manuscript ζ, of which V k and C are copies; accordance of ζ, thus reconstructed, with M can further confirm the reading of ε; and accordance of ε, thus reconstructed, with A b can further confirm the reading of β. Given that it was a sort of principle in Byzantium to compare at least two copies of a text before producing a new one, it is evident that V k ’s testimony for reconstructing β is rather weak: its model, namely ζ, has been checked against another surviving manuscript of the α-family, namely the Escorialensis Y.III.18 (as is suggested by the extensive collations that I have made of several manuscripts of the Metaphysics, and which I will not cite here), while ζ’s model, namely ε, has been corrected according to H a, as said repeatedly. Ιn the same vein, Fazzo hastily dismisses the crucial importance for reconstructing β of some supplementary readings found in E (= E Σ) and in V d (= V d2), which I have pointed out in my review. Fazzo appeals now to the mysterious existence of a Byzantine vulgate of the Metaphysics influenced by A b, but, to the best of my knowledge, no such vulgate existed – nor can I imagine how it could have existed. Byzantine scholars rationally checked the texts they were reading against other manuscripts that were readily accessible to them. This has manifestly been the case of several scholars engaged with the Metaphysics in Byzantium: V d2 has thoroughly corrected V d with the help of a manuscript that was stemmatically close but different from A b, as is suggested by the slight disagreement of his readings with the readings of A b; E Σ has thoroughly added variant readings onto E taken from a manuscript that was stemmatically close but different from A b, as is suggested by the slight disagreement of his readings with the readings of A b; E Σ and V d2 have used different manuscripts, as is suggested by their mutual disagreements. Similarly, the scribe labeled J 2 by Fazzo, whom she now seems to consider at least contemporary with J,1 has used two models for his new copy of the Metaphysics. These remarks bring me anew to the main problem in Fazzo’s work towards a new constitution of the Greek text: she has based her edition on a very limited assessment of the available sources; she ventured to revise Dieter Harlfinger’s stemma without taking any notice of the hints that Harlfinger provides to the “open” branches of the manuscript tradition of the Metaphysics, which we have been (wrongly) used to associate with the so-called ‘contamination’ of manuscripts. Had Fazzo realized that Byzantine scholars themselves were interested in producing editions of ancient texts (and not in ‘contaminating’ manuscripts), she would have produced a far more accurate edition of Lambda.
I finally turn to the chronological issues raised by Fazzo in her response. I am eager to declare that I completely share with her the concern for accuracy. Fazzo says that I claim the fragment Y to be the most ancient manuscript of the Metaphysics; but this is not accurate. I say in my review that Y is perhaps the most ancient manuscript of the Metaphysics; perhaps it is not. As any trained paleographer knows, it is impossible to date precisely manuscripts written in ancient minuscule, unless they are dated by the copyists themselves. Thus J’s pretty precise dating to the middle of the ninth century, to which Fazzo appeals in her response, is an estimation made by Jean Irigoin in 1957.2 I have recently argued in a colloquium organized by Filippo Ronconi and Daniele Bianconi in Paris that J was copied around 900 and that it is slightly younger than E, which has to be dated earlier than usually assumed; this seems to have already found acceptance among some paleographers. E Σ has been dated to the 13th/14th century, not by Myriam Hecquet-Devienne, as Fazzo says, but by Silvio Bernardinello in 1982;3 I believe to have shown elsewhere that this dating is no longer to be retained.4 Finally, A b ’s dating to the end of the 11th century is not mine but Guglielmo Cavallo’s.5 The dates 1118-1138 for Michael of Ephesus’ exegetical activity, accepted by Fazzo, depend upon acceptance of two unsafe assumptions: (1) that the princess Anna Comnena conceived the idea of having scholars produce commentaries on Aristotle only after she first conceived a plot against her brother John in 1118; and (2) that Michael, born around 1060, would comment on Aristotle only because Anna has asked him to do so.6
1. I quote Fazzo’s words as to a possible dating of J 2 (p. 151): “Gli editori non hanno registrato alcuna differenza di epoca fra J et J 2. . . Pertanto essi non hanno identificato precisamente il testo di J prima delle correzioni, il quale, risalendo al secolo IX, presenta il singolarissimo interesse di essere il piu antico documento che ci sia conservato . . . Le correzioni J 2 invece potrebbero essere posteriori alla data di E,” traditionally dated to the 10th century. Fazzo observes in her response some graphic differences between between J and J 2; they all can be explained by the very nature of “J 2 ”’s interventions, which are not those of a copyist stricto sensu but those of a corrector.
2. J. Irigoin, ‘L’Aristote de Vienne’, Jahrbuch der Österreichischen Byzantinistik 6 (1957), 5-10.
3. See S. Bernardinello, ‘Gli scoli alla Metafisica di Aristotele nel f. 234r del Parisinus graecus 1853 (E)’, Elenchos 1 (1982), 39-54.
4. P. Golitsis, ‘Trois annotations de manuscrits aristotéliciens au XIIe siècle: les Parisini gr. 1901 et 1853 et l’ Oxoniensis Corporis Christi 108’, in D. Bianconi (ed.), Paleografia e oltre (Rome 2013), 33-52.
5. See G. Cavallo, ‘Scritture informali, cambio grafico e pratiche librarie a Bisanzio tra i secoli XI e XII’, in G. Prato (ed.), I manoscritti greci tra riflessione e dibattito, Florence 2000, p. 219-238.
6. For more details on Michael’s exegetical activity, I refer the reader to my entry ‘Michel d’Ephèse’ in R. Goulet (ed.), Dictionnaire des philosophes antiques vol. 6 (Paris, due 2013).