Eudoxos and Dedekind: On the ancient Greek theory of ratios and its relation to modern mathematics (Q2641282)

For excellent reasons, the Greek (Eudoxian) theory of ratios, better said of proportions (as exemplified in the book V of Euclid's Elements) has exerted a fascination among mathematicians, philosophers and historians. The literature on the subject is enormous and offers all kind of possible styles, from hypotheses about the genesis of the theory to reconstructions of proofs based only on this theory, from explanations of ``modern'' points of view in Euclid (this last operation was even performed by the scrupulous scholar T. Heath when he used Dedekind's cuts to favour a proof by continuity of the fourth proportional) to critical comments on the defects of the theory. Its very influence on the growth of mathematics, from the 14th century to the 17th century was essential, not to mention the treatment in Arabic speaking countries. For the late seventeenth century, at the time of the birth of the calculus, D. T. Whiteside has clearly shown its role (and its hindrance) in a paper published with the first issue of Archive for History of Exact Sciences; E. Giusti has recently examined its importance in the thought of Galileo and his school; P. Duhem, then M. Clagett, E. Grant, H. L. L. Busard, among others, have shown its use by the Oxfordian and Parisian schools using the fourteenth century; the reviewer worked on what remained from the theory during the eighteenth century. At least for the last twenty years, and if we restrict ourselves to the study of the theory during the antiquity, there exist two fundamental books on the subject, one by \textit{W. Knorr} [The evolution of the Euclidean Elements; a study of the theory of incommensurable magnitudes and its significance for early Greek geometry (Dordrecht, Reidel 1975; Zbl 0362.01001)] and the other by \textit{M. Caveing} [La constitution du type mathématique de l'idéalité dans la pensée grecque, 3 tomes (thèse, Lille 1982)]. The paper under review does not mention these works. We certainly should have provided more references, had the author not wished to remain in a sort of no-man's land. He certainly prefers to choose a philosophical bias, using only classical and valuable philological sources (Tannery, Heath, Zeuthen) or a more recent epistemological study (Mueller). His first chapter characteristically bears the title: the philosophical grammar of the category of quantity. But even under such a scope, he could have mentioned, possibly as references to be discussed, the books by \textit{J. L. Gardies}: one at least having a title quite close to the one chosen by the author [Pascal entre Eudoxe et Cantor (Paris, Vrin 1984); L'héritage épistémologique d'Eudoxe de Cnide, un essai de reconstitution (Paris, Vrin 1988)]. Clearly the author wants to take by himself a fresh view on ratios, using the construction of the real field during the last third of the last century as a model for understanding proportion theory. This is certainly not an original approach, and it certainly does not give a good image of history of science if every author builds his own ideas, without relating them to already published and classical papers. This being said, the presentation of the frame provided by book V is well done. Outside this, the paper mainly deals with two interesting questions, and finds a fine way of answering both. The first question (treated last in the paper) concerns the gap between a pursuit of a theory and modifications which can be made in order to reach new and original objectives: Archimedes is then opposed to other authors following Euclid. The second question is a tentative to determine differences between real numbers as they are now commonly used in the form of a totally ordered Cauchy complete field, and the antique ratios. In this respect, two arguments seem new. One is related to the so-called Eudoxos-Archimedes axiom. For which it is claimed that by itself Definition 5 in Book V of Euclid prevents the existence of a one-to-one correspondence of ratios with the set of real numbers. While Archimedes, insisting for the axiom to be valid for all members of the same species (lengths, or areas or volumes, etc.) could get rid of infinitesimals. Nonstandard historians, I mean those interested in a reading of ancient mathematics through the lenses of nonstandard analysis may have comments here to make. Another argument is a clear discussion of an interpretation of the role of homogeneous magnitudes by \textit{E. J. Dijksterhuis} in his Archimedes (1957). The author claims that Dijksterhuis is wrong. He raises a good technical point, but is the composition of ratios validated for nonhomogeneous magnitudes as well? More should be added on the subject. Will that be done if other historians choose the way of the author not even to discuss past comments?