Timothy Keller’s /The Reason for God/ is a step up from the first two books I reviewed in my series on Dawkins’ fleas. It’s not directed solely at Dawkins and his fans. Rather, large chunks are directed at generic, unreflective urban liberals, which Keller no doubt encounters lots of as the pastor of a Manhattan megachurch. Thus, there’s actually a fair amount of material I agree with here, in particular Keller’s criticisms of relativism and so on.

Some of the most interesting material in the book comes in the form of epistemology. Early on, Keller claims that “All doubts, however skeptical and cynical they may seem, are really a set of alternate beliefs. You cannot doubt Belief A except from a position of faith in Belief B” (p. xvii). A footnote to that statement says that “It is often agreed that a ‘fact’ is either a) something that is self-evident to virtually everyone (e.g., “There’s a rock in the road”) or b) something that is not self-evident to the senses but can be proved scientifically. If we hold to something that can’t be demonstrated in either of those ways then it is a ‘belief’ or an act of faith.” Among the consequences of this are that, when skeptics come to understand this fact, “there will be an understanding, sympathy, and respect for the other side that did not exist before.”

Most of this definitional stuff could be accepted–I’m not convinced that there is a one true definition for the word “belief,” for example. I’m comfortable saying that beliefs are, by definition, not facts. Not much follows from that, though. Obviously, something can be true even though no one can demonstrate its truth. Furthermore, something might be irrational to believe, even if in some sense the contrary view cannot be demonstrated. The notion of demonstration used is a bit sketchy, as it doesn’t allow for things to be demonstrated by testimony, and suggests a sharp distinction where none exists. Finally, his claim about needing beliefs to doubt beliefs suggests that beliefs cannot be doubted on the basis of facts. This is a strange assertion, unless, as a definitional matter, Keller wants to say that the denial of a fact is not a belief but, say, a “patent delusion.”

Towards the middle of the book, Keller again returns to epistemology. He first accuses Dawkins et al. of demanding scientific proof for everything, which can’t be gotten, and eventually gets around to saying that Dawkins doesn’t think there’s any absolute proof in science. Why, then, he would attribute to Dawkins a belief in absolute proof is beyond me. Then he tries to advance a positive view, which he calls “critical rationality.” According to Keller, this involves the belief “We come to every individual evaluation with all sorts of experiences and background beliefs that strong influence our thinking and the way our reason works. It is not fair, then, to demand an argument that all rational people would have to bow to,” which Keller goes on to tie to being biased due to self-interest. But clearly these are two different issues: if I’ve had the experience of seeing Smith murder Jones, and you haven’t and don’t know about my experience, then we can rationally believe different things, but that doesn’t validate disagreeing with someone based on self-interest. Part of Keller’s point is that bias is hard to avoid, but that doesn’t mean we shouldn’t try or treat bias as OK.

Finally, as far as Keller’s positive epistemic views, he says “some systems of belief are more reasonable than others, but that all arguments are rationally avoidable in the end. That is, you can always find reason to escape it that is not sheer bias or stubbornness. Nevertheless, this doesn’t mean that we can’t evaluate beliefs, only that we should not expect conclusive proof, and to demand it is unfair. Not even scientists proceed this way.” The talk of avoiding arguments without being bias goes against the grain of his attempts to validate bias. It’s also very odd to realize that Keller seems to come around to the view that we can in fact proceed in religion just as we do in science. But the biggest problem is talk of arguments being rationally avoidable in the end. He treats Dawkins et al. as merely saying that theistic arguments are avoidable, but I think they mean to claim something stronger: that we have no reason at all to accept the assumptions that go in to theistic arguments. Thus, his response to atheists is often to say “oh, sure, you can deny my assumptions, but it’s still some evidence for my views.”

That, though, can’t be taken for granted. As far as I can tell, we have no reason to prefer an uncaused God to an uncaused universe (cosmological argument), we have no reason to think the situation with the laws of physics is similar to the various long-odds scenarios Keller compares it to (design argument), we have no reason to prefer God-based morality to human-based morality (moral argument), we have no reason to think that the stories of Jesus’ life and miracles couldn’t have been corrupted by myth-making (historical arguments for Christianity).