Abstract

The main thesis of this dissertation is that epistemic rationality is different from justification. While justified beliefs must be likely to be true, a belief with low probability can nevertheless be epistemically rational because informativeness is an important factor to consider in assessing the epistemic rationality of a belief. To argue for this thesis, I first raise the question of why it is reasonable for physicists to either reject or continue to accept Newton's theory before the proposal of Einstein's general theory of relativity. I then examine three major conceptions of justification and argue that if justification is to be truth-conducive, then theories of justification based on these three conceptions cannot account for the reasonableness of the two opposite scientific practices. In chapter two, a Bayesian theory of epistemic rationality is presented and defended against two common objections. It is shown that the Bayesian theory takes into account informativeness and can explicate why both the rejection and the continued acceptance of Newton's theory are reasonable. In chapter three, I compare the Bayesian theory with reliabilism and internalism to further elucidate the difference between epistemic rationality and justification. The Bayesian analyses of cases against reliabilism show that it can be epistemically irrational to hold a reliably formed belief. In comparing the Bayesian theory with internalism, I argue that epistemic rationality is different from epistemic responsibility and point out the differences between the Bayesian theory and various versions of internalism. In chapter four, I discuss Foley's theory of epistemic rationality and argue that the main weakness of Foley's view lies in its not taking into account of informativeness. This weakness not only makes Foley's theory untenable but also precludes his theory from explaining why it is reasonable for the physicists who thought Newton's theory was probably false to continue accepting it