Peter van Inwagen is a premier freewill theist. I'm going to quote and comment on some of his statements regarding the problem of evil.
But an omnipotent and omniscient creator could be called to moral account for creating a world in which there was even one horror. And the reason is obvious: that horror could have been "left out" of creation without the sacrifice of any great good or the permitting of some even greater horror. And leaving it out is just what a morally perfect being would do: such good things as might depend causally on the horror could–given the being's omnipotence and omniscience–be secured by (if the word is not morally offensive in this context) more "economical" means. The Problem of Evil (Oxford, 2006), 96.
i) One problem is the distinction between a good and a "great" good. What makes something a great good? "Great" for whom?
For instance, suppose a couple have a child who dies of leukemia. As a result, they have another child to compensate for the loss of the child who died. Is the replacement child a great good? Well, it's great for the replacement child. It gives him an opportunity to exist–an opportunity that would not otherwise obtain apart from the tragic death of his older sibling.
ii) When Inwagen says such good things might be secured by more economical means, that fails to distinguish between particular goods and generic goods. Even if certain kinds of goods could be secured by more economical means (which is far from evident), it doesn't follow that particular goods could be secured by more economic means. Take my example of the replacement child. If you leave out the horror of the sibling that died, you do sacrifice the replacement child.
iii) Invoking omnipotence is not a solution, for omnipotence makes feasible the realization of alternate scenarios. But that involves sacrificing one possible outcome for another. God can do either one, but he can't instantiate both alternatives in the same timeline.
A defense cannot simply take the form of a story about how God brings some great good out of the evils of the word, a good that outweighs those evils. At the very least, a defense will have to include the proposition that God was unable to bring about the greater good without allowing the evils we observe (or some other evils as bad or worse). Ibid. 68.
One problem with this statement is framing the justification in terms of bringing about a "greater" good, a good that "outweighs" those evils. But why is that a necessary condition? What about an alternate good? A good that would not obtain apart from attendant evil? Why is that insufficient justification?
To recur to my previous example, suppose the replacement child doesn't "outweigh" the evil of his sibling's premature death. But why does the justification for his existence depend on that condition?
If there were no evil, no one would appreciate–perhaps no one would even be aware of–the goodness of the things that are good. You know the idea: you never really appreciate health till you've been ill, you never really understand how great and beautiful a thing friendship is till you've known adversity and known what it is to have friends who stick by you through thick and thin–and so on. Now the obvious criticism of this defense is so immediately obvious that it tends to mask the point that led me to raise it. The immediately obvious criticism is that this defense may be capable of accounting for a certain amount of, for example, physical pain, but it certainly doesn't account for the degree and duration of the pain that many people are subject to–and it doesn't account for the fact that many of the people who experience horrible physical pain do not seem to be granted any subsequent goods to appreciate. If, for example, the final six months of the life of a man dying of cancer are one continuous chapter of excruciating pain, the "appreciation" defense (so to call it) can hardly be said to provide a plausible account of why God would allow someone's life to end this way. Ibid. 69.
i) Up to a point, that's a valid criticism. But it contains some dubious assumptions:
ii) The "appreciation" defense may fail as a stand-alone theodicy. It may, however, make a distinctive contribution to an overall theodicy. There needn't be one uniform reason for every kind of evil.
iii) What if evil is not intended to benefit everyone who suffers evil?
iv) Apropos (iii), the criticism is shortsighted. What if the beneficiary is not the person who suffers, but someone else? In the chain of events, someone further down the line may be the beneficiary. Changing a variable has a domino effect. Every altered variable has a different domino effect. It's not as if one variable can be changed, while leaving everything else in place. Removing the evil removes good side-effects. It changes how all the dominos fall thereafter–for good and ill alike.
v) His example presumes that a person has to let the cancer take its course. But that's subject to debate.
An omnipotent being would certainly be able to provide the knowledge of evil that human beings in fact acquire by bitter experience of real events in some other way. An omnipotent being could, for example, so arrange matters that at a certain point in each persons's life-for a few years during his adolescence, say–that person have very vivid and absolutely convincing nightmares in which he is a prisoner in a concentration camp or dies of some horrible disease or watches his loved ones being raped and murdered by soldiers bent on ethnic cleansing…It seems clear that a world in which horrible things occurred only in nightmares would be better than a world in which the same horrible things occurred in reality… Ibid. 69.
I'm afraid that really isn't clear.
i) There's a sense in which it would be better for your loved ones if they don't actually suffer. If these are merely dream characters.
ii) But would it be better for the dreamer if he's the one who suffers from a horrible disease? Sure, it's only a dream, it isn't real, but since the experience is phenomenologically indistinguishable from reality ("very vivid and absolutely convincing"), how is that clearly better? If it happens to you, and you can't tell the difference, how is that clearly better? Indeed, that would be a very effective form of torture. In reality, you can only die once, however horribly, but in recurring nightmare, that's indefinitely repeatable.
iii) If, conversely, we know it was just a bad dream once we awaken, then we don't take it all that seriously. It's like a video game about combat. An immersive simulation. Because no one is really harmed, it lacks moral weight. It's safe fun.
iv) Inwagen is sketching a scenario in which horrific evil only happens–or seems to happen–in very vivid nightmares. Our waking state is Edenic.
But wouldn't nightmares like that raise doubts about God's benevolence? If these horrors don't happen in the world I inhabit, then where do they come from? Normally, we dream about things that happen in the world we inhabit. If that's not the case, then what's the source of the nightmares? Do they happen in another world? Am I tapping into another world when I dream? Is the fact that I have these nightmares a premonition of what awaits me in the next world? Inwagen's alternative shifts the problem.
v) It isn't clear, moreover, how he can confine moral evil to nightmares. Doesn't that say something about the imagination of the dreamer? How can his imagination be haunted by moral evil without that spilling over into his waking state?
Repeatedly, Inwagan's analysis of the problem of evil, and his objections to proposed solutions, suffer from compartmentalization. The implications of a hypothetical scenario aren't that self-contained.