Abstract
This paper focuses on the historicity of science, that is, the historicity that is inherent to knowledge itself, to scientific knowledge, to scientific reason or rationality: if scientific discoveries take place over time, in a form of temporal progression, is it legitimate to say that this is also the case with reason itself? Does the term reason, which designates the subjective structures of rationality, refer to a noetic invariant, to some invariable constitution of the knowing mind which would only make its effort in history by applying itself
to successive and changing problems, and which would be presupposed by any history of science and scientific discoveries? Or does scientific rationality, on the contrary, have an intrinsic evolution, essential transformations, even breaks in style or paradigm shifts? We start from the Kantian thesis of the structural invariance of the knowing understanding, as well as from the duhemian thesis of historical continuity, to contrast them with the epistemology of ruptures or discontinuities inaugurated by Koyré, Metzger and Bachelard long before Canguilhem, Kuhn and Foucault; scientific thinking is never directly based on the perceptive world, but on a set of pre-scientific or scientific habits that must be challenged and overcome.