There is a recent interesting Prosblogion blogpost on the question of whether theodicy can ever be successful, and if so what success conditions a theodicy must meet. I want consider a related, yet distinct question: can theodicies be convincing in the light of specific instances of evil, and the immediate sense this provokes: "God, if he exists, would not have allowed this"? In the wake of the tragic shooting incident at Newtown, I have been thinking a lot about the problem of evil and classical theodicies and defenses, such as John Hick's soul building theodicy and various forms of free will theodicies/defenses (e.g., Plantinga's; Augustine's).
One way to approach the problem of evil is to look at it as an abstract puzzle to be solved. Wielding modal logic and other tools that analytic philosophy offers, we can argue that evil is unavoidable even for a loving, powerful and omniscient God, if he wishes specific goods like free will to obtain. A different option is to focus on concrete, vivid examples. William Rowe presented the case of a fawn, trapped in a forest fire that was caused by lightning, the fawn suffers horrible burns, and lies in dreadful agony for days until its death. A pointless instance of suffering that, Rowe argues, God could have prevented. Now for cases like Newtown we could invoke the free will defense, since - unlike the forest fire in Rowe's example - the incident was caused by a human agent, exercising his free will, and it was made possible by other instances of free will, such as American policies on gun ownership. But it still seems to me quite a different thing to argue in the face of particular, vivid instances like this that suffering is outweighed by the greater good of the unbridled exercise of free will by moral agents. When confronted with concrete evil like this, theodicy, or indeed any theistic response to the problem of evil, becomes a formidable task indeed.
Gellman (in the paper "A new look at the problem of evil, Faith & Philosophy, 1992) points to this double nature of the problem of evil. Gellman writes "the problem of evil is first and foremost grounded on a type of experience that provides defeasible grounds for believing in the non-existence of God". In this respect, Rowe's fawn in the forest, and Dostojevski's problem of the suffering children (depicted in vivid detail) are not rhetorical tricks, but rather, provide a way to frame the problem of evil in a challenging way. In his book The End of Philosophy of Religion (2008), Nick Trakakis goes one step further. Not only do theodicies fail, he believes that attempts to construct them are immoral. He writes "in the presence of burning children, the declarations of theodicists are shown to be not merely morally confused, but morally scandalous." (Interestingly, when John Hick wrote his influential Evil and the God of Love, he also considered whether theodicies are acceptable, but from a very different angle. Hick thought that justifying God might be considered presumptuous on our part.)
Theodicies should not only offer a solution to the abstract problem, but should withstand scrutiny in the face of concrete, horrible instances of evil. and it seems that in concrete cases, theodicies not fare well. For it is one thing to argue that God did not intend the world as a pleasure-garden, but a challenging place fit for spiritual growth (as Hick proposed), quite another to maintain this in the face of concrete instances of evil.
Trent Dougherty, in a (to my knowledge) unpublished paper manuscript argues in detail that not even skeptical theism fares well. Skeptical theism holds that evil occurs for reasons beyond our ken. This view does not sit well with commonsense philosophy that many theists endorse, viz. that an immediate belief, formed non-inferentially that p provides prima facie a reason to believe that p. In Dougherty's formulation " If it seems to S that p, then S thereby has a pro tanto reason for believing p". So, if I have a religious experience, an immediate sense that God loves me, I am prima facie justified that God loves me. But conversely, if I witness some horrible instance of moral or natural evil and immediately form the belief "No loving all-powerful God would allow this", this provides defeasible grounds for believing that an all-powerful God does not exist.
One way out would be to argue that focusing on concrete cases and the visceral response they evoke is not the best way to approach problems in philosophical theology. One might argue that some philosophical problems are best considered dispassionately, and emotions evoked by concrete cases of evil have little, if any, epistemic weight. But this move does not seem very convincing. Visceral responses drive philosophical thought experiments in ethics and many other areas.
Another way out would be to argue that the evil, especially suffered by children and animals, not only serves some greater good for others, but that the victims additionally get significant benefits in an afterlife. it seems to me that this sort of theodicy + benefits for the victim has a more acceptable feel in concrete situations than theodicies where intense suffering of innocents serves some abstract greater good or contributes to some generally desirable state of affairs.
For example, John Schneider, who was forced to retire from Calvin for rejecting the Augustinian interpretation of the doctrine of original sin, argues that the pain and suffering in animals caused by natural selection predates human sinfulness. Given the magnitude of this suffering, he proposes "We should presume that animals do suffer (contra “Neo-Cartesianism”), and that they suffer in consequence of divine design (contra appeals to a Fall). … Meanwhile, Christian traditions warrant believing that God values all species of animals, that the Atonement (somehow) includes animals, that the Resurrection (somehow) “defeated” evils for animals, and that animals will inhabit heaven." Dostojevski explores this option in the Brothers Karamazov, but concludes that not even a justification of the victim can push away the intuition that concrete instances of evil provide powerful, non-inferential evidence against the existence of God.
All this leads me to think that perhaps Trakakis is not of the mark when he argues that theodicies are questionable, also in a moral sense. Remarkably, the book of Job which deals with the problem of evil in detail comes to a similar conclusion: upon witnessing Job's horrible condition, his friends are startled and shocked, and after the initial shock, they attempt to provide various theodicies (e.g., God is testing you, you must have done something wrong). Job rejects all these accounts, and is, at the end of the book, vindicated by God. God admonishes Job's friends for trying to explain the evil in terms of some retribution or greater good, but he does not blame Job or his wife.