Sometimes what we sincerely say -- aloud or even just silently to ourselves -- doesn't fit with the rest of our cognition, reactions, and behavior. Someone might sincerely say, for example, that women and men are equally intelligent, but be consistently sexist in his assessments of intelligence. (See the literature on implicit bias.) Someone might sincerely say that her dear friend has gone to Heaven, while her emotional reactions don't at all fit with that.
On intellectualist views of belief, what we really believe is the thing we sincerely endorse, despite any other seemingly contrary aspects of our psychology. On the more broad-based view I prefer, what you believe depends, instead, on how you act and react in a broad range of ways, and sincere endorsements are only one small part of the picture.
Intellectualism might be defended on four grounds.
(1.) Intellectualism might be intuitive. Maybe the most natural or intuitive thing to say about the implicit sexism case is that the person really believes that women are just as smart; he just has trouble putting that belief into action. The person really believes that her friend is in Heaven, but it's hard to avoid reacting emotionally as if her friend is ineradicably dead rather than just "departed".
Reply: Sometimes we do seem to want to say that people believe what they intellectually endorse in cases like this, but I don't think our intuitions are univocal. It can also seem natural or intuitive to say that the implicit sexist doesn't really or wholly or deep-down believe that the sexes are equal, and that the mourner maybe has more doubt about Heaven than she is willing to admit to herself. So the intuitive case could go either way.
(2.) Intellectualism might fit well with our theoretical conceptualization of belief. Maybe it's in the nature of belief to be responsive to evidence and deployable in reasoning. And maybe only intellectually endorsed or endorsable states can play that cognitive role. The implicit sexist's bias might be insufficiently responsive to evidence and insufficiently apt to be deployed in reasoning for it to qualify as belief, while her intellectual endorsement is responsive to evidence and deployable in reasoning.
Reply: Zimmerman and Gendler, in influential essays, have nicely articulated versions of this defense of intellectualism [caveat: see Zimmerman's comment below]. I raised some objections here, and Jack Marley-Payne has objected in more explicit detail, so I won't elaborate in this post. Marley-Payne's and my point is that people's implicit reactions are often sensitive to evidence and deployable in what looks like reasoning, while our intellectual endorsements are often resistant to evidence and rationally inert -- so at least it doesn't seem that there's a sharp difference in kind.
(3.) Intellectualism about belief might cohere well with the conception of "belief" generally used in current Anglophone philosophy. Epistemologists commonly regard knowledge as a type of belief. Philosophers of action commonly think of beliefs coupling with desires to form intentions. Philosophers of language discuss the weird semantics of "belief reports" (such as "Lois believes that Superman is strong" and "Lois believes that Clark Kent is not strong"). Possibly, an intellectualist approach to belief fits best with existing work in these other areas of philosophy.
Reply: I concede that something like intellectualism seems to be presupposed in much of the epistemological literature on knowledge and much of the philosophy-of-language literature on belief reports. However, it's not clear that philosophy of action and moral psychology are intellectualistic. Philosophy of action uses belief mainly to explain what people do, not what they say. For example: Why did Ralph, the implicit sexist, reject Linda for the job? Well, maybe because he wants to hire someone smart for the job and he doesn't think women are smart. Why does the mourner feel sorry for the deceased? Maybe because she doesn't completely accept that the deceased is in Heaven.
Furthermore, maybe coherence with intellectualist views of belief in epistemology and philosophy of language is a mistaken ideal and not in the best interest of the discipline as a whole. For example, it could be that a less intellectualist philosophy of mind, imported into philosophy of language, would help us better see our way through some famous puzzles about belief reports.
(4.) Intellectualism might be the best practical choice because of its effects on people's self-understanding. For example, it might be more effective, in reducing unjustified sexism, to say to an implicit sexist, "I know you believe that women are just as smart, but look at all these spontaneous responses you have" than to say "I know you are sincere when you say women are just as smart, but it appears that you don't through-and-through believe it". Tamar Gendler, Aaron Zimmerman, and Karen Jones have all defended attribution of egalitarian beliefs partly on these grounds, in conversation with me.
Reply: I don't doubt that Gendler, Zimmerman, and Jones are right that many people will react negatively to being told they don't entirely or fully possess all the handsome-sounding egalitarian and spiritual beliefs they think they have. (Neither, would I say, do they entirely lack the handsome beliefs; these are "in-between" cases.) They'll react more positively, and be more open to rigorous self-examination perhaps, if you start on a positive note and coddle them a bit. But I don't know if I want to coddle people in this way. I'm not sure it's really the best thing in the long term. There's something painfully salutary in thinking to yourself, "Maybe deep down I don't entirely or thoroughly believe that women (or racial minorities, or...) are very smart. Similarly, maybe my spiritual attitudes are also mixed up and multivocal." This is a more profound kind of self-challenge, a fuller refusal to indulge in self-flattery. It highlights the uncomfortable truth that our self-image is often ill-tuned to reality.
Although all four defenses of intellectualism have some merit, none is decisive. This tangle of reasons leaves us in approximately a tie so far. But we haven't yet come to...
The most important reason to reject intellectualism about belief:
Given the central role of the term "belief" in philosophy of mind, philosophy of action, epistemology, and philosophy of language, we should reserve the term for the most important thing in the vicinity.
Both intellectualism and broad-based views have some grounding in ordinary and philosophical usage. We are at liberty to choose between them. Given that choice, we should prefer the account that picks out the aspect of our psychology that most deserves the central role that "belief" plays in philosophy and folk psychology.
What we sincerely say, what we intellectually endorse, is important. But it is not as important as how we live our way through the world generally. What I say about the intellectual equality of the sexes is important, but not as important as how I actually treat people. My sincere endorsements of religious or atheistic attitudes are important, but they are only a small slice of my overall religiosity or lack of religiosity.
On a broad-based view of belief, to believe that the sexes are equal, or that Heaven exists, or that snow is white, is to steer one's way through the world, in general, as though these propositions are true, not only to be disposed to say they are true. It is this overall pattern of self-steering that we should care most about, and to which we should, if we can do so without violence, attach the philosophically important term "belief".