FMER, Leuven, June 10-12, cont'd

This is getting overwhelming. I haven't finished posting about FMER and I'm already at another conference (Non-classical mathematics) that I also would love to blog about. I'll do my best to complete the FMER mini-series as soon as possible.

Day 2, after lunch

Alan Hajek talked about Blaise and Bayes. He first surveyed a few variants of arguments usually given as reconstructions of Pascal's wager in terms of "dominance" and "expected utility". It was fun, especially since he also showed that they're invalid, for some slightly surprising but equally obvious reasons. He discussed certain emendations that can be made to salvage the wager.

Joshua Thurow's talk was titled Does religious disagreement actually aid the case for theism? Disagreement trailblazing for the miraculous. He pointed out that disagreement about an inferentially-based belief may not automatically force one to suspend judgment en block. Divide the evidence for and against religions into two sets: A - the testimony to the occurrence of miracles, B - everything else. Suppose there is enough disagreement about the evidence in B that considered alone B supports suspending judgment in all religious belief. Then, using Bayes's theorem, Joshua argued that if A includes even moderate testimonial evidence for the occurrence of a miracle, then A and B together support whatever theositic religion is most supported by the testimony in A.

Michael Tooley discussed The probability that God exists. He employed Canapian-style structure-description approach to inductive logic to arrive at an upper bound on the probablitiy that God exists given only the information that the world contains n events each of which is such that in the light of the totality of known rightmaking and wronkmaking properties, it would be morally wrong to allow the event in question.

Given that there are n such events and that there are k unknow morally significant properties, the probability that none of those n actions is wrong all things considered, argues Tooley, is less than (k/k+1)(1/n+1). So, he argued, the probability that God exists must be less than 1/n+1.

One idealizing assumption that Tooley seems to take is that the total moral status of an action is assessed in terms of the number of morally relevant properties (rightmaking vs. wrongmaking), known or unknown. I think it's unlikely a theist would buy into this: they might insist that some (especially unknown) properties are more important and mere counting them (especially since it's not really obvious how you individuate properties so that counting makes sense) won't help to assess the moral status of an action.