Abstract

The practice of thought experimentation plays a central role in contemporary philosophical methodology. Many philosophers rely on thought experimentation as their primary and even sole procedure for testing theories about the natures of properties and relations. This test procedure involves entertaining hypothetical cases in imaginative thought and then undergoing intuitions about the distribution of properties and relations in them. A theory’s comporting with an intuition is treated as evidence in favour of it; but a clash is treated as evidence against the theory and may even be regarded as falsifying it. The epistemic power of thought experimentation is mysterious. How can experiments carried out within the mind enable us to discover truths about the natures of properties and relations like knowledge, causation, personal identity, reference, meaning, consciousness, beauty, justice, morality, and free will? This epistemological challenge is urgent, but a model of philosophical thought experimentation would seem to be a necessary propaedeutic to any serious discussion of it. An adequate model would make the relevant test procedure explicit, thereby assisting in the identification of points of potential epistemic vulnerability. In this monograph I advance the propaedeutical model-building work already done by Timothy Williamson, Anna-Sara Malmgren, and Jonathan Ichikawa and Benjamin Jarvis. Following the lead of these philosophers, I focus on a single Gettier-style thought experiment and the problem of identifying the real content of the Gettier intuition. My first contribution is to establish the inadequacy of all of the existing models. Each of them, I argue, fails to solve the content problem. It emerges from my discussion, however, that Ichikawa and Jarvis’s truth in fiction approach holds out the prospect of a solution. My second contribution is to develop and defend a new way of implementing the general idea behind the truth in fiction approach. The model I put forward does a better overall job of modelling Gettier-style thought experiments than any of the existing models. It has none of the defects which render those models inadequate and Iam unable to find any major defects peculiar to it. This should make us feel confident that my model is adequate. Moreover, since the Gettier-style thought experiment I focus on is paradigmatic, we should also feel confident that my model will generalise naturally to other philosophical thought experiments.