As additional background, it's worth noting Prof. Scott Carson's most recent contribution to the subject. Unsurprisingly to me, Tony also posted his objections to that in Scott's combox, and in due course I shall return to where he left the matter there. But first I should address myself to what he said directly to me.
From my own paper on the subject, it follows that a sufficient "defense" against the problem of evil (PoE), where PoE is understood as a logical conundrum, is twofold: (a) in general, God can bring a greater good out of evil; and (b) it cannot be established that whatever degree of good God can bring out of the evil in this world could also have been brought about without that evil. If that is correct—and I believe it is—then PoE does not succeed in posing an insuperable logical conundrum for classical theism. For if nobody can establish that God could have brought about at least as much good without the evil in this world as he can with it, then the fact of evil cannot be cited to show that classical theism is committed to an account of God whereby divine omnipotence and perfect goodness are mutually incompatible.
That was the basis of my initial reply to Tony. I claimed that he had merely assumed that rational creatures could have attained "the same" condition of blessedness in a world without evil as the one they can attain in this one. I made that claim because I believed, antecedently, that such a proposition cannot be established by argument and can thus only be assumed. Here is Tony's rebuttal (emphasis added):
This is precisely what I was trying to show. The point is not to "prove" that I can come up with a way to achieve every good, in a way that does not involve evil, that occurs in this world as a follow-on of some evil in this world. Such a task would be gargantuan and speculative at best.
My point takes it from the other end: Each specific good achieved in this world, to be good as such, is not defined in terms of evil. Otherwise one would say that evil caused the good, which is nonsense. Therefore, each specific good could come about in a way not through evil. If such a way is not inherently contradictory to God's nature, then He could do it that way.
The only issue left, then, is whether the sum total of all goods (as distinct from each good separately) in this world could exist without evil. While this might seem virtually impossible to us, the question is not whether it is to us, but to God. Again the answer lies in omnipotence: it IS possible to God, if it is not inherently contradictory to His nature.
Initially, I had not seen Tony's argument as an argument because I had taken for granted that it was question-begging, i.e., that it assumed as a premise precisely what he needed to establish. I still think his argument does that. Now that I have in hand a version of his argument that can be conveniently analyzed, I shall show why I still believe it does that.
I gladly concede that if "each specific good could come about" without evil, then each specific good's coming about without evil would not be "inherently contradictory to God's nature." So if it were also the case that each specific good could come about without evil, then PoE would be an insuperable logical conundrum for classical theism. And that's just what Tony argues, or appears to argue. Yet before we get to what Tony considers "the only issue left," we need to consider his argument that "each specific good" could come about without evil.
His argument is by reductio: to deny that each specific good can come about without evil is to imply that, in some cases at least, "evil caused the good, which is nonsense." But to assert, in general and without qualification, that it's "nonsense" to say that evil "causes" good is ambiguous at best. To be sure, in one sense of 'cause' it would be nonsense to assert, as a general truth, that evil causes goodness. Whatever goodness is understood to be in perfectly general terms, evil is, in equally general terms, a privation of it; and if 'cause' be taken to mean efficient cause—i.e., 'agent of production'—it would indeed be nonsense to say that anything is caused by its privation. That is why, in some concrete instances, it would be nonsense to say that evil causes good. E.g., it would be nonsense to say that the virtue of courage is "caused" by cowardice. Moreover it would be false, if not entirely nonsensical, to say that the evils one develops the virtue of courage by facing courageously are thus causes of one's courage—if by 'cause' one means 'efficient cause'. Evils never produce virtues of themselves. Nonetheless, evils often occasion virtues, in such a way that they are in fact the occasions for a person's development of this-or-that virtue. Evils can thus be among the means God uses to help people become better people; and in that sense, evils can be numbered among the "causes" of goods. Nor would it be nonsensical to say that evils which do not occasion the development of virtue in those who undergo them never occasion "good" anywhere else. Severe birth defects, for example, do not occasion the development of virtue in the infants who inherit them; but they can and do occasion such development in those who care for such infants, and later for the infant if and when they develop the capacity for free choice, and thus reach the level of moral responsibility. So, while it is clear that the very concepts of good and evil rule out asserting that evil, in general, causes good in a certain sense of 'causes', in the case of many specific goods one can and ought to say that evils occasion good, in such wise that God uses them as occasions for helping people become better people. And that matters for Tony's purpose, since it is "specific" goods that he is specifically talking about. Accordingly, what he needs to show is not merely that it would be nonsense to assert, in general and without qualification, that evil causes good; on that much, we can agree. What he needs to show is that, consistently with his nature, God could have so created that (a) few or no specific goods are occasioned by evils, and (b) such a created world could have been as "good" if not better than ours. Unless and until he shows that, his argument does not establish his conclusion, which is that each specific good could come about in a way not through evil.
Now (a) would not be difficult to show. But as Tony implicitly recognizes in the very first paragraph of the comment of his I've quoted, (b) simply cannot be shown. What that fact in turn shows is that this particular debate is really about where the burden of proof lies.
At issue is whether classical theism can overcome the problem of evil as a logical conundrum. Tony seems to believe that the burden is on classical theists to show that such good as God's omnipotence can bring about in the world God has created is greater than such good as God's omnipotence could have brought about by creating a world without evil. In other words, what the classical theist needs is a theodicy, which I have said is neither possible nor desirable. And given how Tony has framed the issue, theodicy is impossible whether desirable or not—from which he concludes that the classical theist lacks a logical defense against the problem of evil. But that is simply to beg the question. For, given how Tony has framed the issue, nobody can show what he claims the classical theist needs to show; but that could just as well be evidence that, by demanding from the classical theist what is clearly impossible, he has misframed the question in such a way as to beg it. And my critique of his argument indicates how he has done exactly that.
As further evidence for that claim, I offer what he said to Scott in the latter's combox: "No, you have to make the case that pain is not evil in the sense that the PoE requires , and this requires looking at just what kind of evil the PoE requires." From the prior discussion in that combox it seems pretty clear that, for Tony, the mere existence of a world in which sentient creatures necessarily undergo pain as well as pleasure is an evil "in the sense that PoE requires." That belief is a large part of what once motivated St. Augustine to be a Manichean before his conversion to Catholic Christianity, and I recommend a reading of his Confessions to see how he overcame it. But my immediate concern is with Tony's assumption about where the burden of proof lies. As is also plain from his criticism of my view, his assumption remains so, and thus his argument begs the question.