A Foundation for Universalisation in Games

Résumé

Revealed preference theory equates choices with preferences over the consequences these choices induce. Nevertheless, if a decision criterion prescribes an act for reasons unrelated to its consequences, the inference drawn regarding preferences can be misleading. I study the behaviour of non-consequentialist individuals who have preferences for universalisation. They choose the action that, in a counterfactual scenario where it is also chosen by everyone else, leads to their preferred consequences. I develop a model for individuals who value their choices in light of the counterfactual consequences they induce. Choices are interpreted as revealing a preference for counterfactual consequences. I impose axioms to single out the most prominent models of universalisation, compare them, highlight and arguably overcome their limitations. I propose a unifying model of universalisation inspired by the equal sacrifice principle.