What, If Anything, Does It Mean that Social Preferences 'Explain' or 'Cause' Cooperation?

Ken Binmore has argued that social preferences interpreted as revealed preferences, i.e. as behavioral dispositions rather than as mental events, are empty constructs. This paper shows that social preference models of behavior are not empty but constitute genuine causal explanations of behavior: They indicate how the environment causes behavior. To make that point, I construct a semi-formal framework to reduce utility functions to what I call "behavior functions." Once I have translated social utility functions into social behavior functions, the emptiness criticism is easily rebutted.