UCLA

Believing for reasons is ordinary. For example, I believe that irises respond well to heavy watering for the reason that my plants have always responded well in the past. During the last rainstorm, I believed that there was lightning outside for the reason that it was thundering. Believing for reasons is part of our rational endowment, and believing for good reasons is an important part of our attaining knowledge of the world. But what is the relation between a belief and a reason, such that the belief is held for that reason?

One reason we need a theory of believing for reasons is to explain the distinction between, on the one hand, the so-called “normative” or good reasons that there are for a subject to believe something, and on the other hand, the reasons for which a subject believes it. In order to appreciate this distinction, consider that there may be a normative reason for a person to believe that she will win a raffle (perhaps she knows that she has purchased nearly all of the tickets), but she not believe it on the basis of this reason. She might believe it, instead, for the reason that she dreamed of winning. But that she dreamed of winning the raffle is not a normative reason for her to believe it.

One tempting explication of this distinction is that, when a person believes for a reason, the reason must explain her believing, whereas a normative reason to believe something need not explain why the person holds the belief. But this understanding of the distinction is inadequate. Many philosophers have observed that it is possible for a person’s belief to be explained by a reason without the person believing it for that reason. Suppose that I hear thunder. My belief that there is thunder may make me nervous, my nervousness may cause me to trip, and my tripping may cause me to look up and see lightning. In such a case, my belief that there is thunder will be causally explained by my belief that there is lightning, but I will not believe that there is lightning for the reason that there is thunder. I will believe it, instead, because I see that there is lightning. This “problem of deviant causal chains” shows that we need an account of believing for reasons that does not understand the reasons for which we believe to be merely causal or explanatory.

I consider a variety of attempts to give a theory of believing for reasons, and I argue that even the most plausible available accounts fail. In chapter two, I argue against the view, endorsed by Gilbert Harman, that to hold a belief for a reason is to hold a belief that is the result of some historical process of reasoning. This view has difficulties explaining the sorts of control we have over our own believing. It is a necessary feature of our doxastic control that we have the power to change the reasons for which we believe, but we cannot exercise the appropriate sort of control over the history of our beliefs. I argue also against the view, defended by Kieran Setiya, that believing that p for the reason that q is nothing more than to believe that p and separately to believe that the fact that q is evidence that p. This view is subject to counterexamples, it over-intellectualizes the phenomenon, and it leaves unexplicated the nature of the explanatory connection between a belief and the reason for which it is held.

In chapter three, I argue that problems with these theories push us toward adopting a causal sustaining theory of the basing relation, but I hold that the available causal sustaining theories cannot solve the problem of deviant causal chains mentioned above. Philosophers supporting a broadly causal conception have proposed solutions to the problem of causal deviance by appealing variously to direct mental causation (supported by Kevin McCain), cognitive dispositions (supported by David Armstrong and John Turri), and causation “in virtue of” rationalizing (supported by Ralph Wedgwood). I argue that even the most sophisticated of these causal sustaining theories fail to solve the problem of deviant causation. Indeed, one version of the problem, involving neurological breakdown, seems intractable given the resources of these theories.

In chapter four, I give a positive account of the basing relation that explains the difference between believing for reasons and believing something for which there are merely normative reasons to believe. I argue that believing for reasons is constitutively the exercise of a capacity to causally sustain a belief with another belief, where the capacity has the function of sensitivity to normative reasons. Importantly, pace the characterizations of cognitive 'abilities' sometimes found in epistemology, for instance in the work of John Turri, being the exercise of a capacity in this sense is not the same as being the manifestation of a disposition. Instead, I isolate a sense of “capacity” according to which capacities, unlike dispositions, have essential aims, and exercises of capacities have characteristic norms of functioning. I argue that this view avoids the problems that beset the theories discussed in chapters two and three.

I conclude the dissertation by answering objections and clarifying the aim and scope of the account. I extend the scope of my account by offering an account of a general sort of basing that applies to perceptual belief and memory. And I consider the objection that our cognitive capacities do not in fact function to be sensitive to the epistemic reasons. In particular, it might be argued that our supposedly rational capacities do not in fact aim at sensitivity to the normative reasons but instead roughly at survival or procreative success. I argue against this view in part by appealing to the fact that our cognitive capacities can have two kinds of functions: cognitive and biological.