DescriptionMany epistemologists equate the rational and the justified. Those who disagree have done little to explain the difference, leading their opponents to suspect that the distinction is an ad hoc one designed to block counterexamples. The first aim of this dissertation--pursued in the first three chapters--is to improve this situation by providing a detailed, independently motivated account of the distinction. The account is unusual in being inspired by no particular theoretical tradition in epistemology, but rather by ideas in the meta-ethical literature on reasons and rationality. The account is also unusual in proposing that the distinction between rationality and justification can be derived from a reasons-based account of justification. Historically, this is a striking claim. In epistemology, reasons-based accounts of justification are standardly treated as paradigmatically internalist accounts, but this dissertation argues that we should believe the reverse: given the best views about reasons--again drawn from meta-ethics--we should expect reasons-based accounts of justification to be strongly externalist. The first half of the dissertation might leave one wondering why rationality matters from the epistemic point of view. The second aim of the dissertation is to answer this question. The final two chapters argue (1) that we can only explain why rationality matters from the epistemic point of view if we reject the nearly universal assumption that all derivative epistemic value is instrumental value, and (2) that there are powerful reasons to reject this assumption, since it is the true origin of the so-called "swamping problem". It is then argued that if we reject the instrumentalist assumption, we can get a truth-oriented account of epistemic value that provides a unified explanation of how rationality, justification, and knowledge matter from the epistemic point of view. This result is unprecedented: while some epistemologists find room for both internalist and externalist species of epistemic value, virtually all assume that these properties cannot have a common evaluative ground, and especially not a common truth-oriented one. The concluding moral of the dissertation is that once we properly understand what it is to be truth-oriented, we see that this common assumption is mistaken.