Dynamic Rationality and Disproportionate Belief

I argue that rationality does not always require proportioning one’s beliefs to one’s evidence. I consider cases in which an agent’s evidence deteriorates over time, revealing less about the agent’s position in the world than their earlier evidence. I argue that the agent should retain beliefs that were supported by the earlier evidence, even if they are no longer supported by the later evidence. Failing to do so would violate an attractive principle of epistemic conservatism; it would foreseeably decrease the accuracy of the agent’s beliefs; it would make the agent susceptible to simple Dutch Books; it would allow them to manipulate their evidence so as to increase their confidence in desirable propositions over which they have no control. I defend the assumption that these dynamic considerations are relevant to epistemic rationality.