[cross-posted at Prosblogion]
I've often heard the charge that theists have a harder time responding to the problem of evil if they hold to a deontological ethical view. Deontology recognizes duties that can't easily be overridden by consequences the way consequentialism allows. Consequentialists say the right thing to do is to do whatever leads to the best consequences. If God does this, then God can do things that lead to bad consequences as long as the good consequences that also happen are better enough to outweigh the bad. So it's easier to deal with the problem of evil if all it takes to justify God's allowance of evil is that it leads to a slightly better outcome overall, even if it's worse with respect to the evil itself. Deontologists, on the other hand, might just say that the duty not to harm or not even to allow harm can't so easily be outweighed by the overall good. Some things are just wrong, and God shouldn't therefore do them. Allowing very great evils seems to be a pretty good candidate for that category of action. It's thus harder to respond to the problem of evil with a deontological view than it is with a consequentialist view.
I used to be a little disturbed at this problem, wondering whether a "higher goods" type of defense that I favor requires a consequentialist view, a view I'm not otherwise attracted to. But it's occurred to me recently that the problem assumes a kind of deontology that I don't agree with. It assumes the absolutism of Immanuel Kant's deontology, not a more moderated kind of deontology such as that of W.D. Ross, which I favor. On Ross's view, we have prima facie duties, none of which are absolute the way duties for Kant are. Duties can often conflict for Ross, and when they conflict only one will turn out to be an actual duty, whichever one is morally more important. In a case of lying to save a life, the life is more important than the normal duty not to lie, but in a case of lying to protect your reputation it's still going to be wrong to lie, even if the consequences are better from lying. So this is not consequentialism, but it's not absolutism either.
Now apply this to the problem of evil. There will be potential cases when God would not do something wrong, because even though the consequences are better it would be wrong to do it. But it leaves open that some goods are so important that God might allow pretty serious harm in order to achieve them. This means that the moderate deontologist can have consequence-based responses to the problem of evil that an absolutist deontologist can't have. This may have been all I was worried about losing by adopting a deontological ethical view, even if consequentialists might have yet more to say to defend a divinity from being immoral for allowing evil.