We are blessed in the twenty-first century to live in a world in which moral theorists have learned much from epistemologists, and conversely. Reasons First is motivated by the conviction – a conviction that I dub the Core Hypothesis of the book – that this learning can go yet further. Disappointingly and despite its provocative title, the thesis of the book is not that reasons come first. It is, rather, motivated by the striking asymmetry between the prevalence and popularity of the idea that reasons have a central explanatory role in ethics and in epistemology, and by the task of exploring the roots and consequences of this asymmetry. There are, I argue, deep and central challenges to how deep the explanatory role of reasons can go that are more visible and central in epistemology than in ethics, which together rationalize this asymmetry of attitudes. But there are also central puzzles in epistemology that I argue are shaped in unsatisfactory ways by these very same challenges. In each case, I suggest, overcoming these distinctively epistemological challenges to the centrality of reasons instead of rejecting the centrality of reasons on their basis leads to new and, I argue, more satisfactory solutions to some of the central puzzles in epistemology. So epistemology can benefit, I argue, by not being so hasty to reject such a central explanatory role for reasons.

So what do I mean, then, by the hypothesis – sometimes perceived to be hegemonic in moral philosophy – that reasons come first? Just that the reason relation – or at least, some reason relation – has both analytic and explanatory priority over every other normative property and relation. This idea has been advanced in very prominent places in moral philosophy, and it is a generalization of Ross’s idea that your all-things-considered duty is nothing other than a matter of whatever wins the competition among your prima facie duties. Ross supported this claim by noting that what your all-things-considered duty turns out to be is affected in systematic ways by what other things turn out to be the case in your circumstances. This argument, which can be generalized to apply to other moral concepts, I call the classical argument for the priority of reasons.

But historically this idea has played a much smaller role in epistemology – and for good reason. According to the problem of unjustified belief, if reasons are to explain either justification or knowledge, then there must be some prior constraint on which states of mind give you reasons – lest unjustified beliefs justify. So, the problem alleges, we must impose a prior constraint on which beliefs give you reasons – and that prior constraint must look something like justification or knowledge. Hence, reasons cannot come first among reasons, justification, and knowledge. The problem of sufficiency points out that whereas in ethics it may be okay to do something when your reasons to perform it are at least as good as the reasons for any alternative, in epistemology this is not the case, because it is never epistemically rational to believe when the evidence is tied. So, the problem alleges, we cannot extract the conditions of justification from reasons, but must instead know what makes reasons good enough to justify – which puts justification first again, and not reasons.

The book is organized around these two problems. After two introductory chapters in which I lay out the issues and clarify what I take to be central to the notion of a reason for purposes of the book, Part 2 takes up the question of what evidence we have for basic perceptual beliefs and argues that the historical options in answering this question have been strangely and unsatisfactorily constrained by the assumption that evidence must be true, in order to rationalize belief. This assumption is so alien to most historical thinking in moral philosophy that it is the place where the sociological clash between the subdisciplines becomes most obvious. In chapter four I trace what I take to be the most promising motivation for this assumption to the reasons-rejecting response to the problem of unjustified belief.

The picture of basic perceptual justification that emerges in chapter five is what I call the apparent factive attitude view. According to this view, when you have a visual perceptual experience as of p, you are in a state that presents itself to you as a seeing that p. And so this puts you in possession of the following piece of evidence: that this is a seeing that p. This view has a great deal in common with prominent disjunctivist accounts of perceptual justification including those of Williamson and McDowell. Like Williamson, it grants that states like seeing that p have a factive core, but in contrast to Williamson it rejects the claim that this state is knowledge. Like both Williamson and McDowell, it grants that the reasons of subjects in good cases and bad cases are different. But in contrast to both, it distinguishes between the kinds of reasons that make rational, and the kinds of reasons that make correct. Paired good-case and bad-case subjects share the same subjective reasons, and so the same beliefs are rational for each. But the evidence of bad-case subjects is false – they do not, in fact, see that p. And so they lack objective reason. They are in no better position to know.

In Part 3 I turn to the problem of sufficiency. The problem is that once we give up the idea that there is any non-trivial answer to what makes reasons sufficient to justify other than that they beat all comers, we find ourselves without any understanding or explanation for exactly how preponderant the evidence must be, in order to rationalize belief, how belief can be rational at all about philosophy, ancient history, or other topics where evidence is hard to come by, and other related puzzles. But these sorts of puzzles are exactly analogous to the puzzles about different circumstances affecting your all-thing-considered duty that drive the classical argument. So my answer to all of them, in chapter six, is that sufficiency of reasons just is winning the competition with other reasons, in epistemology as in ethics, but what made this invisible was forgetting that there is always a third option in epistemology – we can lack either belief.

So, I say, there are properly epistemic reasons against belief that are not evidence against its content. But what are these reasons, and why should we believe that they are properly epistemic, in the sense that facts about them can make it harder or easier to know? I argue that the latter question is just a special case of a much more general problem of which reasons for any attitude are the “right kind” of reason for that attitude, and which are the “wrong kind”. The solution to this general problem, I argue in chapter seven, must flow from an account of the nature of belief – and I show how different natural accounts of what is distinctive of binary belief lead to different natural accounts of what sorts of thing can be non-evidential epistemic reasons against belief. In chapter eight I offer my own account of the nature of belief – the default reliance account – and argue that it results in a very conservative but very attractive kind of explanation of pragmatic encroachment in epistemology. The package of views advanced in chapters six, seven, and eight I call pragmatic intellectualism. Chapter nine extends pragmatic intellectualism to account for doxastic wrongs and explain how moral considerations can directly raise the standards for knowledge or epistemically rational belief.

Finally, in Part 4 I go on to argue that thinking in terms of reasons gives us a particularly promising way of seeing how to analyze knowledge. This is because knowledge is not so unlike a wide range of what I call well properties from all across normative theory. Aristotle contrasts the person who acts from virtue with the one who acts in accordance with virtue, and similarly we can contrast the person who fears rationally and the one who merely fears what it is rational to fear. Every well property is characterized by its relation to a corresponding thinner normative property that it entails but that is less demanding on the agent, and I argue that the best account of this relationship reveals that these thinner properties must consist facts about the competition between reasons. Because this new argument for an explanatory role of reasons draws on a wider range of the central distinguishing characteristics of reasons, I call this the fundamental argument.

Because knowledge is a well property standing to the thinner property of correct belief, the fundamental argument shows that facts about correct belief consist in facts about the competition between reasons, and that knowledge itself is a matter of believing for the right reasons – which I call the Kantian account. I diagnose which features made this sort of view look like it fell into trouble in the 1970’s, and argue that it has much to offer in explaining not only prominent features of knowledge, but the attractiveness of many of the prominent ways in which smart people have theorized about knowledge.

The three main views that I advance in each of parts 2, 3, and 4 of the book – the apparent factive attitude view, pragmatic intellectualism, and the Kantian account of knowledge, are independent of one another, but mutually supporting. And each of them pushes us to continue to think harder about how ethics and epistemology relate to one another, while acknowledging and respecting their differences.