Abstract

We usually assume that we should be rational; however, it is not clear why. My thesis considers two interpretations of the question "why be rational?" First of all, is it prudent to be epistemically rational? Second, under what conditions would natural selection favor epistemically rational belief and desire formation policies? I formulate and defend both a prudential justification as well as an evolutionary explanation of epistemic rationality. ;In the first part, I examine famous examples of non-epistemic arguments for belief, including work by Immanuel Kant, Blaise Pascal, and William James. Drawing on this earlier work, I develop a precise account of the difference between epistemic and non-epistemic reasons for belief. I defend a truth-linked approach to epistemic reason, and argue that various prima facie problems for such an account can be overcome with a question-relative understanding of epistemic value. ;Many philosophers have made claims about the evolution of rationality. Several philosophers, including Quine and Dennett, have argued that natural selection will favor rationality in one form or another. Others, such as Stich, Plantinga, Godfrey-Smith and Sober, have been more skeptical. I use optimality modelling techniques, adapted from evolutionary biology, to develop a model that represents both belief and desire formation devices. I show how certain kinds of deviations from epistemic rationality can be useful, but demonstrate that the more systematically important a proposition is, the more natural selection will favor organisms with reliable ways of finding out about the world. ;In the final part of my essay, I argue that it is prudent to be epistemically rational only if we make certain substantial assumptions about the nature of an agent's environment, as well as her mind. Once again, the notion of systematic importance plays a crucial role. I show that for propositions of a certain kind, the more systematically important they are, the closer an agent's degree of belief must be to the evidence in order for beliefs about such propositions to be prudent, provided that the evidence indicates the truth.