With his book The Nature of Mathematical Knowledge (1983), Ph. Kitcher, that had been doing extensive research in the history of the subject and in the contemporary debates on epistemology, saw clearly the need for a change in philosophy of mathematics. His goal was to replace the dominant, apriorist philosophy of mathematics with an empiricist philosophy. The current philosophies of mathematics all appeared, according to his analysis, not to fit well with how mathematicians actually do mathematics. A shift in orientation should invoke the more general reflection that causal, genetic factors are as significant for epistemology as logical structure. As I am to a large extent sympathetic with Kitcher’s proposal, my aims here will be simple: first I start presenting Kitcher’s argument, and then I try to raise some doubts about his contribution. These doubts come probably from my unskilfulness to follow the torrential flow of Kitcher’s ideas.