Conditional Preference and Causal Expected Utility

Abstract

I am going to describe how we can make use of the idea of conditional preference to provide a foundation for Brian Skyrms’ version of causal decision theory (K-expectation, or causal expected utility, decision theory). The foundation has the following two virtues: first, it puts the theory on an equal footing with competitors for which foundations have already been given; second, this foundation will provide the basis for a reply to the most serious objections made so far against causal decision theory, and against Skyrms’ version of it in particular (more about this below).¹ I will only say a little about the other versions of causal decision theory developed by Gibbard and Harper (1976), Lewis (1981), and Sobel (1978). There are interesting differences between the theories, but in spirit they are very much alike. The foundation and the replies are immediate payoffs of using conditional preferences, but I think the treatment of conditional preference sketched below is also interesting for its own sake.