The Standard Problem of Evil Reformulated part 1
Although theologians are dead wrong in claiming that it is the only argument in our arsenal, the Problem of Evil is by far the most persuasive and powerful of them all. Christian organizations are always falling all over themselves trying to explain natural disasters in God's designed universe.
Their lame, unscientific and downright offensive excuses - Satan creates disasters, sin creates disasters, disasters are needed to make us good people, we bring disasters on ourselves - don't hold water any more. And so they weave those words like there's no tomorrow, in greater and greater rhetorical circles, until they're exhausted and red in the face, to try to keep their worldview intact.
The nonsense, however, does manage to hide the truth sometimes. Theodicies especially have retained their perceived ability to account for evil. I think this is due primarily to the unfortunate approach that standard PoEs have taken. Instead of concentrating on choice and action, as any argument centering around morality should, it looks at final states. So this leaves a gap between choice and outcome, which leaves the PoE open to pseudo-refutations like the theodicies.
The PoEs have also typically suffered from the use of moral terms, like "evil", which are interpreted in different ways by different Christian sects and atheist groups. The use of morality-neutral terms like "suffering" is helpful but also carries its own disadvantages, such as a discontinuity between the fact of suffering and the conclusion of non-benevolence.
These two problems, however, can be eliminated if we recenter the PoE on the divine choice. This makes an "ironclad PoE", in that there is no more gap for theodicies to slip in. I was inspired to take this approach by Everitt in The Non-Existence of God, pages 243 to 244.
First I need to define one term : "morally righteous". In the context of this PoE, I define moral righteousness as :
Posit a volitional being B. When making a choice where there is at least one perfect alternative and the cost of the implementation of all alternatives are identical (or in the case of a god, where the cost is automatically zero), B will choose a perfect alternative if B is morally righteous.
This is an intuitively obvious principle. Suppose that you go to an ice cream store, determined to get an ice cream cone. I'm hungry, hot, whatever. In the store, I'm faced with all sorts of varieties, all with the same cost. This is, in a trivial way, the same kind of scenario as proposed in the definition. If, in such a scenario, I buy ipecac flavour instead of, say, chocolate, I'm plainly being an idiot (unless, for instance, I am fulfilling a bet, but this would make it inherently more desirable).
In short, in a scenario where all costs are equal, there's no reason for me not to do the right thing. This is an intuitive moral principle that, I think, no one is able to reject coherently.
Perhaps one could argue that the moral problem is only pushed back into the word "perfect", and that perfection is in the eye of the beholder. But this would be no different than saying that the standard PoE fails because good and evil are a part of God's mind. If we are talking about a distinct divine form of morality, then all pretenses that God is benevolent must be dropped, and the notion of worship must crumble, thus collapsing the concept of "God" under its own weight. This also, of course, applies to the Apathetic God Paradox.
Also, we can reformulate the definition of moral righteousness without referring to perfect, but rather to a "best alternative" or "best alternatives". All we would need here is to point out to what the "best" is relative to, and in this case it would be evil, suffering, or whatever other moral term one wishes to apply. The tension between omnibenevolence and neutral terms does not apply here, as we are shedding the notion of benevolence altogether for the notion of moral righteousness, which do not have to be directly related. Someone can accept the notion of moral righteousness while rejecting omnibenevolence, and vice-versa (although, as I said, I don't think anyone can reject the notion of moral righteousness coherently).
The problem of looking at the PoE from the perspective of a specific disaster is that it loses sight of the big picture. A believer can try to explain the existence of a disaster by invoking another feature of the world that it supports or improves, such as our character (character-building theodicy), other goods (second-order goods), free will (free will theodicy), or original sin. But if we look at the universe as a whole, and states of the universe as choices, then this avenue vanishes. For example, it makes no sense any more to state that evil exists because of free will, if we consider both evil and free will as part of God's imperfect choice.
I present my formal argument in the second part.