Exemplary Examples

Normative theory is not peppered but salted with thought experiments and examples. In this paper, we have two objectives. The first is to explore and categorize the various objections that have been made to the deployment of thought experiments. The second is to give an account of the role that thought experiments can and do play in the work of normative theory, and to defend a 'best practice' guide for their use. Thought experiments can be used at many stages of theorizing. One may use them in order to: (i) show that a concept fits our intuitions; (ii) to argue that a theory or principle is correct; (iii) to provide evidence for a theory or principle; or (iv) to justify a theory or principle. We argue that the widespread opposition to thought experiments or hypothetical examples only applies to some elements of normative theorizing, and so cannot be thought of as objectionable across the entire domain. The helps to defuse some of the recent tensions in the relevant literatures. As far as the testing of concept use is concerned, the appeal to thought experiments is a useful device but its application is usually much less contested because thought experiments are typically used to evoke real-world examples to observe and assess how we use natural language. The real disagreement about the use of thought experiments comes to the fore when they are used to support theories. We suggest that thought experiments and hypothetical examples should be thought of as models of real-world situations. Analysing thought experiments as models helps to defuse the most important critique: that thought experiments (sometimes called intuition pumps) are too unrealistic to trigger reliable, unbiased intuitive responses. The critique is beside the point insofar as the inventor of the experiment has thought about how the model can represent real-world examples and gives the audience some guidance about this intended representation. The model-based approach also shows possible limitations in the use of “outlandish” thought experiments: if they cannot be “fleshed out” into realistic real-world cases they are less useful to prompt intuitions, though they may have other uses apart from pumping intuitions. For example, some of the highly abstract thought experiments in political philosophy are not about garnering intuitions at all, but about presenting arguments in a more vivid and accessible way. The representation of arguments in examples requires different criteria of assessment. Finally, we also relate out approach to recent anti-intuitionist moves in analytic political philosophy and defend the current practice of normative theorizing.