Epistemic Risk and Picking Posteriors

Abstract

AbstractIn the preceding two chapters, I asked how we should use epistemic utilities to pick our priors. In this chapter, I turn my attention to picking posteriors. In the early section, I consider and reject a range of possible approaches. Then I turn to an argument by Dmitri Gallow (2019) that we should pick our posteriors by maximising expected epistemic utility from the point of view of our priors, which entails that we should obtain our posteriors from our priors by conditioning on our evidence, just as Bayes' Rule demands.