In this answer to “Evolutionism as a Modern Form of Mechanicism”(SiC 2(2): 287–306) I discuss the strange double use the authors make of their reference to Kant in order to deny the relevance of far-from-equilibrium thermodynamics, and more generally, of the physical irreversibility question in the problem of evolution.

On the one hand, the authors quite legitimately use a materialist version of Kantian apriorism in the guise of “means of cognition” presupposed by any physical theory. But on the other hand, they accept a theoretical interpretation of physical irreversibility (as introduced in physics via statistical supplementary conditions) whose precise historical function was to occult the intrinsic difference between the cognitive means of dynamics and those of thermodynamics in order to promote a mechanistic unification of physics.

I thus argue that at stake is not the case of such or such a physicist transgressing the border between philosophy and natural science as defended by von Borzeszkowski and Wahsner, but rather the very interpretation of irreversibility accepted by most physicists (including von Borzeszkowski and Wahsner) since Boltzmann. In contrast, Prigogine's enterprise could be understood as an attempt to take the notion of cognitive means seriously, by approaching the problem of evolution (and not by proposing a theory of evolution) via the problem of the relevance of these means.

I conclude with comments on the strategic role Kant is still made to play today by physicists who wish to transform dissents in physical theory into philosophical ones and by philosophers who wish to judge scientists instead of trying to understand them.