Abstract

Register to attend in person

Revealed preference approaches to modelling choice in the social sciences face seemingly devastating predictive, explanatory, and normative objections. In this talk, I will focus on predictive and explanatory objections, and offer two defences. First, I argue that when revealed preferences are multiple realizable, revealed preferences can causally explain behaviour well. But, considerations of multiple realizability open the revealed preference theorist to an equally plausible interpretation of these models, that they pick out a coarse grained psychological disposition. Second, I argue that when agential preferences cannot be easily analytically separated from the environment that produces the relevant behaviour, revealed preferences also causally explain, if one adopts a counterfactual dependence account of causal explanation. An upshot of these two arguments is an explanatory argument against a unified dispositional interpretation of ‘preference’.