Abstract
Many thought experiments (TEs) are used to probe theoretical statements. One crucial strategy for doing this, or so I will argue, is the following. A TE reveals an inconsistency in part of our previously held, sometimes empirically well-established, theoretical statements. A TEer or her critic then proposes a resolution in the form of a conjecture, a hypothesis that merits further investigation. To explore this characterisation of the epistemic function of such TEs, I clarify the nature of the inconsistencies revealed by TEs, and how TEs reveal and resolve them. I argue that this can be done without settling the question of which cognitive processes are involved in performing a TE; be they propositional or non-propositional. The upshot is that TEs’ reliability, like real experiments, is to be found, in part, in their replicability by the epistemic community, not in their cognitive underpinnings.
Highlights
Five decades following Kuhn’s (1964) question “[h]ow [...] relying exclusively upon familiar data, can a thought experiment lead to new knowledge or to new understanding of nature?” (p. 241), scientific thought experiments (TEs) remain a hot topic in philosophy of science
While if we analyse the TE as aiming at revealing an internal inconsistency—which would be the case under the stronger reading of the ET, where the experimental details are eliminated in order to transform the TE into a non-TE-argument— Popper would be justified in claiming that this amounts to the strange assertion that QM and Gn, as a set of theoretical statements, are internally inconsistent
An argument that somehow takes us from the observation that an external inconsistency is revealed by a TE to the conclusion that the inconsistency is somehow contained in these general statements
Summary
Five decades following Kuhn’s (1964) question “[h]ow [...] relying exclusively upon familiar data, can a thought experiment lead to new knowledge or to new understanding of nature?” (p. 241), scientific thought experiments (TEs) remain a hot topic in philosophy of science. I will argue throughout the paper, in particular while analysing the nature of the inconsistencies revealed by TEs in 3 This content is taken to be represented purely propositionally in Norton’s “Empirical Psychological Thesis”. Pace Norton’s insistence that the actual conduct of a TE is the mental execution of an argument, proponents of the mental model accounts of TEs (see, for example, Gendler 2004; Miščević 1992; Nersessian 1992), argue that a TE’s content is not always propositional. The upshot is that TEs’ reliability lies, in part, in their replicability by the epistemic community, and this can be investigated without reference to their cognitive underpinnings This said, it remains interesting to investigate this latter subject separately and even draw consequences for TEs (see, for example, Stuart’s (2019) analysis of imagination and its implication for several accounts of TEs).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
