Towards a Theory of Rational Desire

Objective

The standard framework for rational decision making is expected utility theory. In this framework, decision makers are modelled as having a probability function, which represents their degrees of belief about what is likely to happen, as well as a utility function that represents how desirable they find each possible outcome. Expected utility theory has been influential in many disciplines, including philosophy, but it has important limitations and flaws. In particular, desires are typically treated as subjective and hence not a suitable target of criticism, aside from the requirement that they obey very generic coherence constraints. Agents’ degrees of belief, on the other hand, are held to much more stringent standards. For example, most decision theorists maintain that degrees of belief should be mathematical probabilities and that agents should use a rule called “Bayesian updating” to adjust their degrees of belief in response to evidence. This project will argue that the asymmetric way in which belief and desire are treated in the standard theory is unjustified. Desires can be misaligned with the world, just like beliefs can be false. For example, suppose you live in Alaska and you detest snow and desire tropical weather. Then your desires are clearly misaligned with the world—it would be better for you if you either (somehow) grew to like snow or if you moved to the tropics. In the same way that it is rational to aim to believe the truth, it seems rational to aim to have desires that are aligned with the world. The main goal of the project is to develop a novel theory of rational desire management that takes seriously the idea that our desires can align to a greater or lesser extent with our circumstances and values. In so doing, the project will inaugurate a completely new subfield of philosophy devoted to the formal study of what makes desires rational.