I have posted a draft of the TE paper online now (the title is the same as the title of this post). Here.
Abstract. Brown (1991a,b, 2004, 2008) and Bishop (1999) argue that thought experiments (TE) in science cannot be arguments and cannot even be represented by arguments. They rest their case on examples of TEs which either proceed through a contradiction to each a positive resolution (Brown calls such TEs "platonic") or are used by different people with opposite results. This, supposedly, makes it impossible to represent them as arguments for logical reasons: there is no logic that can adequately model such phenomena. (Brown further argues that this being the case, "platonic" TEs provide us with irreducible insight into the abstract realm of laws of nature). I argue against this approach by describing how "platonic" TEs can be modeled within the logical framework of adaptive proofs for prioritized consequence operations. To show how this mundane apparatus works, I use it to reconstruct one of the key examples used by Brown, Galileo's TE involving falling bodies. I also address Bishop's qualms about the clock-in-the-box TE which Einstein and Bohr employed when they disagreed about the uncertainty principle.
Many thanks to Christian Strasser, Frederik Van De Putte, Erik Weber, Rawad Skaff and Joke Meheus for reading and discussing with me earlier versions of this manuscript and to all the people who discussed this topic with me: Graham Priest, Diderik Batens, Anouk Barberousse, Peter Simons, Margherita Arcangeli, Gillman Payette, and the audiences in Geneva, Paris and Ghent, where I gave talks based on this material.