Epistemic Utility Theory and the Aim of Belief

It’s widely accepted that rational belief aims at truth.1 Objectively correct belief is true belief. How should rational believers pursue the aim of truth? Epistemic utility theorists have argued that the framework of decision theory can explain what it means to aim at truth. By pairing decision theory with a distinctively epistemic form of value—gradational accuracy, degrees of proximity to the truth—we can provide arguments for various epistemic norms in terms of truth. These arguments generally use either the notion of expected accuracy or the notion of accuracy dominance. For example: it’s been argued that the reason why we should have probabilistically coherent degrees of belief is that it’s the only way to avoid accuracy domination (Joyce, 1998, 2009), and the reason why we should update by conditionalizing on our evidence is that doing so uniquely maximizes expected accuracy (Greaves and Wallace, 2006; Leitgeb and Pettigrew, 2010b). Caie (2013) and Greaves (2013) show that results depend on notions of dominance and expected utility that are different in important respects from the versions of dominance and expected utility used in standard (practical) decision theory. If we use the more familiar forms of expected utility and dominance, we can’t justify the epistemic norms that epistemic utility theory had hoped to justify. Indeed, the prescriptions of epistemic utility theory conflict with these norms. I argue that the things epistemic utility theorists often call “expected accuracy” and “accuracy dominance” can’t really be the expected accuracy or accuracy dominance of epistemic states in any conventional sense. It’s not clear what they are; so far we don’t have a good philosophical interpretation of these pieces of math. Without a philosophical interpretation, they are ill-equipped to do the epistemological work they were meant to do. For example, just telling us that conditionalization maximizes some particular quantity— whatever it is—doesn’t explain why we should conditionalize on our evidence. In short, those of us who are attracted to the project of epistemic utility theory face a dilemma. We must choose between old and new versions of rules like dominance avoidance and expected utility maximization. Call the familiar, decision-theoretic versions of these rules consequentialist rules and the new versions that appear in epistemic utility theory nonconsequentialist rules.2 If we choose the consequentialist rules, then we can vindicate the idea that rational belief has the aim of accuracy—but at the cost of giving up rules of probabilistic coherence, conditionalization, and other attractive epistemic norms. On the other hand, if we choose the nonconsequentialist rules, we can avoid predicting rational violations of probabilistic coherence, conditionalization, and so on—but at the cost