Cognitive Decision Theory and Permissive Rationality

Abstract

Cognitive Decision Theory (CDT) says that a rational doxastic state is one with maximum expected epistemic value―one such that your expected epistemic value for how states of the world will turn out, given your adopting it, is at least as high as that of any alternative doxastic state you might adopt. One of the central issues in epistemology is whether epistemic rationality is permissive or not: Some claim that (Uniqueness) for any total evidence, there is a unique doxastic state that any agent with that total evidence should take; Others claim that (Permissivism) for some total evidence, there are multiple doxastic states that an agent with that total evidence can take. How does CDT relate to the debate over permissive rationality? The aim of this paper is to explore this question. In particular, as one way of addressing this question, I present and assess an argument against CDT that goes as follows: On the assumption of Epistemic Conservatism, the correct theory of epistemic rationality will not endorse non-conservative doxastic state shifts from one credence function to another, in the absence of new evidence. In some cases, however, CDT endorses non-conservative doxastic state shifts, in the absence of new evidence. This seems unfortunate. However, I further show that when we clearly distinguish among several types of Permissivism, the argument is not a real threat to any CDTer. Depending on which version of Permissivism or Uniqueness a CDTer endorses, the argument may be avoided in one of two general ways. One response appeals to the stability of beliefs over time, while the other allows that the instability of beliefs over time fits naturally with epistemic rationality.