FEW 2006 Abstract Katie Steele What is it rational to value? There are many responses to the Allais Paradox, some reconciling it with expected utility (EU) theory, others holding that the two cannot be reconciled (for better or worse for the theory). In this paper, I will focus on a particular aspect of the debate-the extent to which Savage's sure-thing principle constrains the content of (not just the relationship between) our preferences. Many accept that an agent's attitudes towards risk and feelings of regret can legitimately be considered part of an act's outcomes. This claim is nonetheless contentious, because risk and regret do not reside in any particular outcome, but rather arise from the way in which potential act outcomes are related to one another, i.e. they are "global properties" of the outcome space or decision problem. Indeed some hold that incorporating global properties in individual outcomes is outright inconsistent with Savage's (in particular) EU theory. I think this claim is too strong. I argue that the issue rests, rather, on whether we are interested in a descriptive, predictive or normative decision theory. While there may be good descriptive or predictive reasons for assigning value to global properties of the outcome space, I argue that this move severely weakens EU theory as a guide to rational choice. Of course, when we consider examples beyond Allais, in particular some telling choice problems posed by Machina (1982 & 1991), then it seems incorrect to exclude risk and regret sentiments from a model intended to guide choice. I examine the different ways we might respond to this, some defending and others challenging EU theory's dominance as the theory of rational choice. Abstract references: Machina, M. 1982. "Expected Utility" Analysis without the Independence Axiom. Econometrica 50. Machina, Mark J. 1991. Dynamic Consistency and Non-expected Utility. In Foundations of Decision Theory, edited by M. Bacharach and S. Hurley. Oxford and Cambridge, MA: Basil Blackwell.