In the next-to-last paragraph of his book, C.S. Lewis’ Dangerous Idea: In Defense of the Argument from Reason, Victor Reppert makes a very interesting statement:

However, I contend that the arguments from reason do provide some substantial reasons for preferring theism to naturalism. The “problem of reason” is a huge problem for reason, as serious or, I would say, more serious, than the problem of evil is for theists. (emphasis mine)

I think this is a very interesting statement for two reasons. First, Reppert acknowledges that the so-called “problem of evil” — which is probably misnamed (see here) — is an evidential problem for theism. All by itself, that is a significant concession that is all too rare among theistic philosophers. But second (and more important), Reppert claims that naturalism’s ‘problem of reason’ is as big of a problem, if not a bigger problem, for naturalism as the ‘problem of evil’ is for theism. I want to focus on this second feature of interest about Reppert’s statement.

I recently asked, “Why Do So Many People Have a “Winner Takes All” Approach to Evidence about Gods?” Suppose you agree with my conclusion that there can be evidence for false propositions, so there can be evidence for atheism if God exists, and so there can be evidence for theism if God does not exist.

As soon as you admit that possibility, you have to be prepared to confront another possibility. How do you weigh competing items of evidence, especially when we don’t have numerical probability values (or likelihoods or Bayes’ factors) to work with? Here are two versions of this problem.

(1) Weighing Two Individual Items of Evidence

Suppose you have two items of evidence, E1 and E2, and two rival hypotheses, H1 and H2. E1 is evidence favoring H1 over H2, i.e., Pr(E1 | H1)  > Pr(E1 | H2). Let B1 the “Bayes’ factor” for E1 , i.e., the ratio of Pr(E1 | H1)  to Pr(E1 | H2). E2 is evidence favoring H2 over H1, i.e. Pr(E2 | H2) > Pr(E2 | H1). Let B2 be the Bayes’ factor for E2, i.e., the ratio of Pr(E2 | H1) to Pr(E2 | H2). If E1 is stronger evidence for H1 than E2 is evidence for H2, then B1 > 1/B2. Likewise, if E2 is stronger evidence for H2 than E1 is evidence for H1, then 1/B2 > B1. But how do you show that?

In some cases, it may be possible to show this is true by definition. For example, in my F-inductive argument from consciousness, I argue that Pr(consciousness | theism) =1 because theism entails the existence of consciousness. Now compare that result to a very weak argument against theism, the argument from scale. I have argued before that, as an argument against mere theism, the evidence of scale provides very weak evidence favoring naturalism over theism. So it seems obvious that if Pr(consciousness | theism) = 1, then consciousness is much stronger evidence for theism than scale is against it.

Or consider Paul Draper’s evidential argument from biological evolution. The key insight to understanding that argument is this. It is really an argument against special creationism, combined with a rigorous argument that special creationism is a viable auxiliary hypothesis to theism. In other words, theism provides a significant antecedent reason to expect that special creationism is true conditional upon the assumption that theism is true, where “antecedent” emphasizes the idea that we are abstracting away all of our evidence from biology. Draper’s evidential argument from biological evolution argues that Pr(special creationism | naturalism) = 0, whereas Pr(special creationism | theism) >= 1/2. Now suppose you have some extremely weak argument for theism, such as the argument from beauty. I don’t think beauty provides any evidence for theism, but for the sake of argument let’s pretend that it does. In that case, it would be obvious that the falsity of creationism is much stronger evidence against theism than beauty is evidence for it.

Not all comparisons of evidence will involve cases where at least one hypothesis entails neither the evidence to be explained nor the denial of the evidence to be explained. In those cases, it seems to me it will be more difficult, possibly impossible, to justify an objective comparison of evidential strength. (Whether it is impossible or merely difficult will have to be determined on a case-by-case basis.)

(2) Weighing Cumulative Cases Against One Another

Suppose now you have two “real” cumulative cases done the right way. In favor of H1, you have items of evidence E1 through E5. In favor of H2, you have items of evidence E6-E10. For example, let H1 be theism and H2 be naturalism. Then let our items of evidence be:

E1: the contingency of the universe

E2: the beginning of the universe

E3: the life-permitting conditions of the universe

E4: consciousness

E5: intentionality

E6: the hostility of the universe to life

E7: biological role of pain and pleasure

E8: falsity of special creationism

E9: mind-brain dependence

E10: psychopathy

You believe that E1-E5 are individually and collectively evidence favoring theism over naturalism. Likewise, you believe that E6-E10 are individually and collectively evidence favoring naturalism over theism.

As before, we’ll use “B” to represent the Bayes’ factor. Let B1-5 represent the ratio of Pr(E1 & E2 & E3 & E4 & E5 |T) to Pr(E1 & E2 & E3 & E4 & E5 |N). Let B6-10 represent the ratio of Pr(E6 & E7 & E8 & E9 & E10 | T) to Pr(E6 & E7 & E8 & E9 & E10 | N).

How in the world are you supposed to show that B1-5 > 1/B6-10?

(3) Is Naturalism’s ‘Problem of Reason’ as Big or Bigger than Theism’s ‘Problem of Evil’?

Let us now return to Reppert’s statement I quoted at the beginning of this post:

However, I contend that the arguments from reason do provide some substantial reasons for preferring theism to naturalism. The “problem of reason” is a huge problem for reason, as serious or, I would say, more serious, than the problem of evil is for theists. (emphasis mine)

Reppert does not attempt to defend this claim in his book, but in fairness we should note the argument from reason is a neglected topic in the philosophy of religion. It seems reasonable to devote an entire book just to (re-?)introducing the argument and defending it. But it would be a major accomplishment in the philosophy of religion, I think, if Reppert were able to successfully defend this claim. Perhaps he can devote his considerable philosophical talents to this task in a future book.