On alternative proofs and other substitutions of proofs in Euclid's \textit{Elements} (Q1762835)

Relying on the research work carried out in the 20th century on the medieval translations of Euclid's \textit{Elements} into Arabic and then Latin, \textit{Wilbur Knorr} [``The wrong text of Euclid: on Heiberg's text and its alternatives'', Centaurus 38, 208--276 (1996; Zbl 0861.01014)] reopened the 19th century discussion between Heiberg and Klamroth, which had seemed for a long time to have been definitively settled in favor of Heiberg's view. The question at stake in the debate related to how the extant textual evidence regarding the \textit{Elements} had to be taken into account in a critical edition of the text. Heiberg, who produced the edition of the \textit{Elements} still considered a reference today, clung to the view that only the Greek manuscripts were worth considering, holding that the medieval translations -- the indirect tradition -- were unreliable witnesses for it. By contrast, Klamroth considered that the medieval translations gave evidence of a state of the text closer to the Greek original than the extant Greek manuscripts, a view in favor of which Knorr offered new arguments. The article under review relies on research work that developed within the framework of this discussion, and it provides ingredients that may allow clarifying the issues as well as formulating projects meaningful to work on them. It addresses the question of alternative proofs to be found in the various versions of Euclid's \textit{Elements}. Why and when did they develop? What motivation led to their formulation? How can we account for the difference between alternative proofs? On the basis of which arguments can we determine the one that may have been the closer to Euclid's original one, if any? In his philological work evoked above, Heiberg gave preeminence to a manuscript (P: ms. vat.gr. 190) and worked under the assumption that Euclid gave only one proof for each proposition. He regularly assumed that, when the manuscript P contained two proofs, the one to be considered the original one was the one occurring in the first place in manuscript P. Bernard Vitrac reopens this question, by considering systematically the corpus of all alternative proofs. By this, he does not mean only those to be found in the Greek manuscripts considered by Heiberg in his critical edition of the \textit{Elements}, but also those attested to in the Arabic and then the Arabo-Latin translations, which doubles the size of the corpus to be considered. Moreover, he offers a typology of alterations that may be used to distinguish various kinds of alternative proofs. The question he has in mind is to analyze whether, for each case, Heiberg in his edition made the right choice. The quantitative analysis Vitrac offers shows that the phenomenon is far from being marginal. The article reviews the various cases for which we have more than a proof (dissatisfaction with logical dimensions, with the formulation or with mathematical features of a proof or a set of proofs). Each proof is then analyzed in relation to the types of motivations that can be identified as having led to the production of a second proof. In each case, Vitrac discusses the history of the text and the question of the proof that could be thought as Euclid's own demonstration. His analysis leads Vitrac to formulate his conviction that most of the structural modifications in the proofs of the \textit{Elements} originated with Greek authors. On the one hand, Vitrac argues in favor of the importance to be granted to the commentary ascribed to Hero of Alexandria in the production of a set of alternative proofs that were used in some editions to replace, or complement, Euclid's original proofs. On the other hand, he sees in the school use of the \textit{Elements} in late Antiquity a second key source for the production of modifications in the proofs.