|
Résumé : |
Matthias Cléry, who holds a PhD in the history of science and technology, studies the institutional, educational, and mathematical conditions that led the French mathematician Maurice Fréchet (1878–1973) to propose an axiomatization of probability based on additive set functions in his 1929–1930 course at the Chair of Probability Calculus and Mathematical Physics (CPPM). The aim is not simply to trace the origins of a formal system, but rather to show how, in this case, teaching and research jointly contributed to mathematical work.
Fréchet occasionally taught probability from 1920 onward, but without proposing an axiomatic framework. His appointment in 1928 as a lecturer at the CPPM Chair, in the newly created Institut Henri Poincaré, marked a turning point. The author places this appointment in the broader context of changes in French universities under the Third Republic. Reforms introduced in the late nineteenth and early twentieth centuries gradually gave science faculties a role in conducting research and training students for research, while preserving their role in preparing future secondary-school teachers. In this system, teaching remained an important factor in shaping academic careers. Fréchet’s career illustrates this particularly well: his teaching abilities, regularly praised in his professional records, played an important role in his appointments first in Strasbourg and later in Paris.
The CPPM Chair held a special position. Strengthened by the creation of the Institut Henri Poincaré in 1928, it closely connected advanced teaching, research, and the growing international development of probability theory. Fréchet arrived there after gaining experience in Strasbourg in teaching research-level mathematics and organizing international scientific groups. Cléry thus shows that Fréchet’s axiomatization should be understood within a “range of possibilities” shaped both by his career and by the institutions in which he worked.
The article then examines Fréchet’s views on mathematical teaching and research. In an opening lecture in 1919, he argued that students should first acquire a broad mathematical education before working on a research problem. His 1925 lecture in Bern, Sur une désaxiomatisation des sciences, further developed this position. Fréchet criticized a form of teaching that presented mathematics as a finished sequence of definitions, lemmas, and theorems. He did not reject the axiomatic method, but argued that it should be accompanied by an account of the problems, attempts, and successive approximations that lead to mathematical results.
This view helps explain his 1929–1930 course. Fréchet introduced a family of “probabilizable” events, closed under complements and finite or countable unions, together with a nonnegative and normalized probability function that is countably additive on mutually exclusive events. This construction extended his earlier work on families of abstract sets and additive set functions and placed probability within what he called “general analysis.” At the same time, however, Fréchet retained an empirical definition of probability, understood as a physical constant estimated from observed frequencies. His axiomatization therefore combined mathematical abstraction with a continued concern for applications.
A central issue is the extension of additivity from the finite to the countable case. Fréchet stressed that σ
-additivity does not follow logically from finite additivity. The difficulties involved in this extension, together with his 1930 controversy with Bruno de Finetti, led to several possible positions: rejecting σ
-additivity, checking its validity case by case, or adopting it as an axiom while accepting that some events may not have a probability. Fréchet chose the last option, which he regarded as theoretically convenient, logically consistent, and compatible with experience.
Finally, Cléry examines the place of Fréchet’s axiomatization in the later history of probability theory. Fréchet influenced Kolmogorov, who acknowledged in his 1933 Grundbegriffe the importance of Fréchet’s work on the abstraction of measure and integration. Fréchet did not, however, adopt Kolmogorov’s axiomatic framework, since he rejected a simple identification of probability with measure. The article thus brings attention to a less studied episode in the history of the foundations of probability and, more broadly, shows convincingly how a mathematical construction can emerge from the close interaction of research, teaching, professional career, and institutional context.
|
ORCID: https://orcid.org/0000-0001-7610-4648
(2026).
Du rôle de l’enseignement dans la recherche d’un universitaire mathématicien sous la Troisième République : le cas de l’axiomisation des probabilités de Maurice Fréchet (1929–1930) [compte rendu de l'ouvrage de Cléry, Matthias]. zbMATH Open.
![[thumbnail of 32- zbMATHOpen PDF - 2026 - 10.pdf]](https://r-libre.teluq.ca/style/images/fileicons/pdf.png "32- zbMATHOpen PDF - 2026 - 10.pdf")
