Abstract Epistemic agents care about the truth, but not all truths equally. These preferences are epistemic: they reflect our preferences over which truths are more interesting, cohere better with our existing beliefs, test our tolerance for inconsistency, or are more significant for our aims. Many religious, moral, or political beliefs are shaped by such preferences. Some epistemic utility theories fail to account for our interest in particular truths. I argue they should, and I propose a modification to one prominent model to show how they might. My revised model, compared to other attempts to model content-sensitivity, has the advantage of being (1) question-sensitive and (2) responsive to the agent’s preference over informativity. This revised model also yields a novel theory of the base-rate fallacy: the tendency for people to overestimate the probability of an event relative to its population-level likelihood based on individuating information.
Pacy Yan (Sat,) studied this question.