Kantian Accounts of Thought Experiments

The paper outlines a critical history of the concept of TE from a broadly Kantian viewpoint. The thread of our discussion has been Kant's position on the nature of the a priori and how each neo-Kantian theory of TEs can be understood in terms of its own position on this issue. Section 2 examines some aspects of Kant's philosophy that are related to today's debate on TEs. Section 3 is devoted to the precursors of neo-Kantian accounts of TEs, with a special emphasis on Ørsted's account. This author wavers between two positions: sometimes he claims that TEs are merely the hypothetical aspect of REs while other times he assigns to them the capacity of determining a priori some fundamental laws of natural science. The latter position recalls the "material" aspect of the Kantian a priori, a notion which was abandoned by the philosophy of science at the end of the nineteenth century. This is probably the main reason Ørsted's point of view had no real influence on the historical development of the concept of TE and why some recent Kantian or neo-Kantian accounts of TE depart from Ørsted's concept of the a priori. I, on the other hand, have defended an interpretation of the a priori as purely functional by retaining the necessary and universal character of Kant's a priori while rejecting the material a priori, since content can be given only by experience. On this view, TE and RE are complementary in a sense similar to the complementarity of form and matter in Kant: TEs without REs are empty; REs without TEs are blind (Section 4). On the other hand, Yiftach Fehige has defended a relativized and contingent notion of the a priori which has been recently advocated by many authors, notably by Michael Friedman. Unlike myself, Fehige rejects the universality and necessity of Kant's a priori but retains the idea that the a priori, as a constitutive element of experience, is endowed with material content that may be made explicit by thought experimentation (Section 5)