What possible greater good is there in not preventing the rape of a woman, or the physical, emotional and sexual abuse of a little child
I think atheistgirl’s reaction is exactly right.
Inscrutable evil/suffering = an instance of evil/suffering for which we know of no moral justification for.
Unjustified evil/suffering = an instance of evil/suffering for which no moral justification exists.
The noseeum inference is the inference that some event E is an unjustified evil, from its being an instance of inscrutable evil.
She’s making the noseeum inference, and I think we should follow her in doing so.
The most popular reply is some brand of skeptical theism: Since we don’t have epistemic access to God’s moral justifications, we can’t say he doesn’t have any. For that reason, we’re not able to infer that any instance of evil is unjustified; it’s just as likely as not that God has a moral justification for it.
The problems with skeptical theism are legion. But, here’s one brilliant reply:
The observation of
any event lowers the probability of skeptical theism, so long as its prior probability is less than 1 (a reasonable assumption, 1 being the highest degree of confidence, 0 being the least).
Take for instance…the death of Christopher Columbus. Our background knowledge makes it quite probable that Christopher Columbus died because of things like our knowledge of human physiology, the fact that Columbus lived centuries ago etc. So, on atheism it’s quite probable that Christopher Columbus died (since atheism doesn’t tweak with our background knowledge), I’d say 1.0.
But, on skeptical theism, you must introduce an omni-guy into the picture. The question arises, “Does God have morally good reason to prevent Columbus from dying, or is he morally obliged to preserve his life?” The skeptical theist is in an awkward position now. Who knows? He must say. God’s moral justifications are inscrutable to us. Therefore, it’s just as likely as not that God had morally good reasons for miraculously sustaining Columbus’ life, than that God didn’t. So, we can’t say whether Columbus died, given skeptical theism. That is, where Columbus’ death = C, Skeptical Theism = S, and Atheism = ~S:
P(C|S&K)=0.5, whereas P(C|~S&K)=1.0. It follows by the relevance criterion that P(~S|C&K) > P(~S|K). So, C is evidence for ~S against S. (And so forth for any observation)
Now, after you get P(S|E&K) you take that posterior and make it S’s prior in the next evaluation. So, suppose the following probabilities:
P(S)=0.99
P(~S)=0.01
P(W|S&K)=0.5
P(W|~S&K)=1.0
Plug these into Bayes’ theorem:
0.99 * 0.5 / [0.99 * 0.5] + [0.01 * 1.0] = 0.98
You take this posterior (0.98) and make it S’s prior in the next evaluation. As you can see, S’s posterior will eventually drop to near 0 (it just went from 0.99 to 0.98), considering the
incredible amount of events we could and should evaluate. (Maintaining the same numbers above but adjusting the priors accordingly will give S a posterior of 0.96, this decrease will speed up).
The skeptical theist may reply that the relevant probability isn’t “0.5” but is inscrutable, hence no number can be assigned! (somewhere between 1-0.) But, the skeptical theist must generalize this idea to all observations, therefore the probability of anything at all given skeptical theism is inscrutable. So, there isn’t any evidence for anything. Super.