Mike Valle gave a presentation yesterday before the ASU philosophy club on the skeptical theist response to the evidential argument from evil. A good discussion ensued among Guleserian, Nemes, Lupu, Reppert, Valle, Vallicella, et al. Peter Lupu made a comment that stuck in my mind and that I thought about some more this morning. For what puzzles him puzzles me as well. It may be that we are both just confused.
1. Let us assume that our concept of God is the concept of a being that has a certain modal property, the property of being such that, if existent, then necessarily existent, and if nonexistent, then necessarily nonexistent. Call this the Anselmian conception of deity. It follows that God exists, if true, is necessarily true, and if false, necessarily false. Simply put, the proposition in question is either necessary or impossible, and thus necessarily noncontingent.
2. Peter's question, I take it, was: how can such a noncontingent proposition have its probability either raised or lowered by any empirical consideration? In particular, how can considerations about the kinds and amounts of natural and moral evil in the world lower the probability of God exists? If true,then necessarily true; if false, then necessarily false. Peter's sense -- and I share it -- is that evidential considerations are simply irrelevant to the probability of noncontingent propositions.
3. The problem -- if it is one -- arises in other contexts as well. I once argued that conceivability does not entail (broadly logical) possibility. I got the response that, though this is true, conceivability of p raises the probability of p's being possible. That is not clear to me. Assuming the modal system S5, if p is possible then necessarily p is possible, and if p is necessary, then necessarily p is necessary. (The possible and the necessary do not vary from world to world.)
I happen to think that S5 caters quite well to our modal intuitions. Assume it does. Then It is possible that there be a talking donkey is necessarily true, if true. If so, how can the fact that I (or anyone or all of us) can conceive of a talking donkey raise the probability of the proposition in question?
4. Reppert made a comment in response to Lupu about the probability being epistemic in nature. I didn't follow it. If p is noncontingent, and we are concerned with the probability of p's being true, and if truth is not an epistemic property (i.e., a property reducible to some such epistemic property as rational acceptability), then I don't see how evidential considerations are relevant.
The ComBox is open if Victor or Peter want to add to their remarks.