Standard sceptical theism focuses on our ignorance of the realm of values. I want to suggest a different kind of sceptical response to an evil E. This response identifies a good G such that it is clear that the occurrence of a good relevantly like G logically requires the permission of an evil relevantly like E, but instead the scepticism is in that we have on balance no significant evidence against the conjunction:
- G obtains and
- G outweighs E and
- there is no alternative good G* dissimilar from G that doesn't require anything nearly as bad as E and that would be more or approximately equally worth having.
And it's fairly easy to generate examples of G that do the job for particular E. Take Rowe's case of the horrendous evil inflicted on Sue. Let G be Sue's having forgiven E's perpetrator. We have no significant evidence against the conjunction (1)-(3), then. Granted, we may have significant evidence that G did not obtain in this life, though even that is probably a stretch, but we have no balance no significant evidence that G didn't obtain in an afterlife. My intuitions strongly favor (2)--there is a way in which forgiveness seems to defeat evil--but in any case we have no significant evidence against (2). As for (3), granted there are many great moral goods that don't require anything nearly as bad as E, but I don't think we have on balance significant evidence that these goods are roughly as good as or better than G. Now, of course, it can be the case (whether due to a logical contradiction or dwindling probabilities) that we don't have significant evidence against any conjunct but we do have significant evidence against the conjunction. But I don't think this happens here.
Assign probabilities. Let's assign 1/2 to each of (1)-(3), and suppose they're independent. Then the probability of the conjunction is 1/8, which isn't very significant as compared to the kinds of probabilities one gets in fine tuning arguments. And in a large random sample of justified evils, we'd expect there to be lots of them looking like the probability of their being justified, bracketing the existence of God, is 1/8 or less.
One might think that there are many evils, though, and one could multiply the 1/8s from them to get a really tiny probability that they're all justified. But one can't do that, because the cases aren't independent on the hypotheses under consideration. If God exists, they're all justified.
With a bit of creativity, it's not hard finding G's that logically require E. And for what it's worth I think it's easier for horrendous evils E than for minor evils. Now, it would be hard showing that (1)-(3) in fact hold, but the theist doesn't need to do that. A weaker sceptical move is enough.