A. If False, Probably Not

As we saw above, Freud doesn’t really argue that theistic belief has no warrant if taken in the basic way: he seems to assume that such belief is false, and then infers in rather quick and casual fashion that it is produced by wish-fulfillment and hence doesn’t have warrant. Here (despite the appearance of carelessness) perhaps Freud’s instincts are right: I shall argue that if theistic belief is false, but taken in the basic way,²²⁰²²⁰   And, let’s add, not taken on testimony. That is because testimony, like inference, is not an ultimate source of warrant; a belief taken on testimony has warrant for someone only if that belief has warrant for the testifier. See WPF, p. 83. then it probably has no warrant. First, as we saw above, no false belief has warrant sufficient for knowledge; therefore, if theistic belief is false, it doesn’t have that degree of warrant. Still, couldn’t it nonetheless have some warrant? There are at least two reasons for 187thinking not. First, when does a false belief have warrant? Typically, in a case where the faculty that produces the belief is working at the limit of its capability. You see a mountain goat on a distant crag and mistakenly think you see that it has horns; as a matter of fact, it is just too far away for you to see clearly that it doesn’t have horns. You are a particle physicist and mistakenly believe that a certain subatomic model is close to the truth: working as you are at the outer limits of the cognitive domain for which our faculties are designed, your belief is false but not without warrant. If there is no such person as God, of course, then there is no such thing as a sensus divinitatis; and what truth-aimed faculty would be such that it is working at the limit of its ability in producing the belief that there is such a person as God, if that latter belief is false? It is exceedingly hard to think of decent candidates. Further, if your faculties are functioning properly and are unimpeded by desire for fame, ambition, lust, and the like, then if they are working at the limit of their capability, you will not ordinarily believe the proposition in question with great firmness—you will not believe it with anything like the degree of firmness often displayed by theistic belief. Thus you won’t be sure that you see horns on that goat: you will instead think to yourself, “Well, it looks as if it has horns, but it’s too far away to be sure.” You won’t insist that your physical model is correct; if you believe it is, it will be with a certain tentativeness. These considerations suggest that if theistic belief is false, it is not produced by cognitive processes successfully aimed at the truth, and hence does not have warrant.

There is another and more important consideration, which we can approach indirectly as follows. A belief has warrant only if the cognitive process that produces it is successfully aimed at the truth—that is, only if there is a high objective probability that a belief produced by this process is true (given that the process is functioning properly in the sort of epistemic environment for which it is designed). Now from the fact that a belief is false, it doesn’t follow that it is not produced by a process or faculty successfully aimed at truth. It could be that on a given occasion a process issues a false belief, even though there is a substantial objective probability that any belief it produces will be true (given the satisfaction of the other conditions of warrant). For example, a reliable barometer may give a false reading because of an unusual and improbable confluence of circumstances. Physicists tell us that it is possible (though extremely unlikely) that, for just a moment, all the air molecules in the room should congregate in the upper northwest corner of the room. Suppose this happens; at that moment, the air pressure in the vicinity of the barometer in the lower southeast corner of the room is zero; the barometer, however, still registers 29.72, because there hasn’t been a long enough time for it to react to the change. The fact that it issues a false reading under these circumstances doesn’t mean it is not a reliable 188instrument. Similarly for a cognitive process: there might in fact be a high probability that a belief it produces is true, despite the fact that on a given occasion (even though the other conditions of warrant are satisfied) it issues a false belief. Couldn’t something similar hold for the processes that produce belief in God? Might it not be that belief in God is produced by cognitive processes successfully aimed at the truth, even if that belief is, as a matter of fact, false?

I think not. A proposition is objectively probable, with respect to some condition C, only if that proposition is true in most of the nearby possible worlds that display C.²²¹²²¹   For an account of the connection between possible worlds and objective probability, see WPF, p. 162. But now consider the process that produces theistic belief: if it is successfully aimed at truth, then in most of the nearby possible worlds it produces a true belief. Assuming that in those nearby possible worlds it produces the same belief as it does in fact (i.e., belief in God) it follows that in most of the nearby possible worlds that belief is in fact true: in most of the nearby possible worlds there is such a person as God. However, that can’t be, if the fact is there is no such person as God. For if in fact (in the actual world) there is no such person as God, then a world in which there is such a person—an omniscient, omnipotent, wholly good person who has created the world—would be enormously, unimaginably different from the actual world, and enormously dissimilar from it. So if there is no such person as God, it is probably not the case that the process that produces theistic belief produces a true belief in most of the nearby possible worlds. Therefore, it is unlikely that belief in God is produced by a process that is functioning properly in a congenial epistemic environment according to a design plan successfully aimed at the production of true belief. So if theistic belief is false, it probably has no warrant. Freud is right: if theistic belief is false, then it is at the least very likely that it has little or no warrant.