The concept of existence and the role of constructions in Euclid's \textit{Elements} (Q1862870)

Since Zeuthen's days it has been regularly suggested that the constructions in Euclid's \textit{Elements} were meant as existence proofs, ascertaining that mathematical objects corresponding to specific definitions or prescriptions are constructible and thus ``exist'' (in the sense given to that word in modern constructivist philosophy of mathematics). Part of the foundation of this view is the persuasion that the Grundlagenkrisis presumed to follow from the discovery of irrationality could not but create the need for such existence proofs; the other is derived from some of Aristotle's discussions of definitions and existence of objects. In the first part of the present paper, Orna Harari shows (against frequent views) that Aristotle's texts are not pertinent to the question. Firstly, the only passage where Aristotle speaks of mathematical objects ``existing'' (i.e., ``being'' \textit{simpliciter}, not ``being something''; namely Metaphysics Z. 10, 1035b31-1036a12) denies existence to individual circles as soon as we do not think of them, and reserves it for the universal, the circle corresponding to the definition; but Euclid's arguments concern located, that is, individual circles. Secondly, the passages in the \textit{Posterior Analytics} where Aristotle refers to particular geometrical proofs within the framework of syllogistic logic are shown to be fundamentally different from what Euclid does -- which excludes that Euclid followed the model set forward by Aristotle in that work. In the second part, the author proposes an alternative function for the Euclidean constructions, analyzing as an example the proof of \textit{Elements} I.5, where construction is shown to serve ``either as a means of measurement by which quantitative relations are deduced or as a means of exhibiting qualitative relations, i.e., the order or the position of geometrical figures''. One may disagree with a number of interpretations made in the article, but these disagreements hardly affect the overall conclusions; three examples may be mentioned. (1) One might see the references to geometric proofs in the \textit{Posterior Analytics} not as intended full proofs but as references to familiar proofs where Aristotle highlights only those aspects of the proofs that can be fitted into a syllogism, omitting those which \textit{establish} the premises (in \textit{Analytica priora} 65a4-7 he demonstrates to know perfectly well that a proof cannot be made without this second aspect being taken into account). All that is changed by this reading, however, is that even Aristotle would have known that deductive geometry could not be made solely on the basis of syllogistic analytics. (2) Similarly it is unimportant whether the triangle is considered as a species falling under the genus of lines, or a proof involving triangles just depends on the properties of lines, namely as ``καθ' αὑτὸ συμβεβηκότων'', ``attributes necessarily resulting from the notion of a thing, but not entering in the definition thereof'' (Liddell and Scott). (3) Without consequences for the conclusions is finally the question whether even sensible and not only mathematical, intelligible entities have existence solely as universals possessing a definition and not as individuals -- a point where the author appears to be contradicted by several oft-quoted Aristotelian passages.