The Evidential Form of the Problem
In light of the failure of atheists to produce such a proposition, the majority consensus among philosophers today is that the logical form of the problem of evil is a failure. For this reason the main focus of the discussion has turned in recent decades to the evidential form of the problem. In some ways this marks a retreat on the atheist’s part to a weaker position, but it is one that is nonetheless potentially damaging to the theist. Proponents of the evidential argument admit that theism is logically consistent and that the existence of evil does not in itself disprove the existence of God. The charge they make is that the existence of evil—and, more specifically, the kinds and quantity of evil that we actually find in the world—constitutes powerful evidence against God’s existence. Put differently, the evil that we find renders it unlikely that God exists, and thus it provides us with good reason for not believing in God.
The evidential form of the problem of evil is best understood as a response to the “greater-good” theodicies sketched in the last section. The atheist here admits that it is possible for a perfectly good, omniscient, omnipotent being to allow evil, if by doing so a greater good is achieved that could not be achieved in any other way. So the mere existence of evil does not contradict God’s existence. However, the atheist will contend that much of the actual evil that we observe in the world is pointless: it does not lead to any greater good, or it is not, at any rate, logically necessary for the achievement of any greater good. A good God—by the very definition of “good”—would not allow pointless evil. The argument can be summarized as follows:
1. If God exists, he does not allow any pointless evil.
2. Probably, there is some pointless evil in the world.
3. Therefore, God probably does not exist. 
The qualifying term “probably” is important. The atheist does not presume to be able to prove that there are genuinely pointless evils in the world; she can admit that it is always possible that God has some justifying reason for allowing the evil that is beyond our ability to comprehend. But she thinks that this is improbable for the simple reason that it clearly appears to us that there are pointless evils. The critic will note that here, just like everywhere else in life, we must make our best judgments based on the way things appear to be. Given that there appear to be pointless evils, the most rational conclusion to draw is that a perfectly good and omnipotent God probably does not exist.
In responding to the evidential form of the problem, the theist can try to rely, once again, on the theodicies we examined in the last section, as well as others that have been proposed.  But suppose one does not find any of these arguments convincing. It must not be forgotten that evil is a problem felt by the believer as well as the nonbeliever. Many believers find the occurrence of many evils in the world—especially those that seem particularly egregious or horrendous—baffling and troubling. They wonder why God allows these things, for they admit that these evils do appear to be pointless. What then can the believer say?
Two things, at least. First, the believer can try to refute the second premise by challenging the reasons used to support it. Recall that the reason given by atheists for thinking that it is likely or probable that there are pointless evils is simply this: it appears that there are pointless evils. The claim that it appears that there are some pointless evils, however, is open to challenge. Stephen Wykstra has argued that this claim violates a basic epistemic principle, called the Condition of Reasonable Epistemic Access (CORNEA, for short).  What CORNEA states, in a nutshell, is that one is justified in making a claim like “It appears that there are no Xs” only if one is justified in believing that, if there were any Xs, one would be in a position to perceive them.
This is best seen by way of an example.  Suppose that someone opened the garage door, turned on the light, took a quick glance around and, on the basis of what he saw, declared, “It appears that there are no dogs in the garage.” He would be justified in making this claim, because—assuming there is nothing strange going on in this scenario—if there were any dogs in the garage, a quick glance would be sufficient for him to see them. But suppose this same person opened the garage door, turned on the light, took a quick glance around and, on the basis of what he saw, declared, “It appears that there are no fleas in the garage.” In this case his claim would not be justified. Taking a quick glance around the garage does not suffice to justify one in making this kind of claim, because it is common knowledge that if there were any fleas in the garage, one likely would not be able to perceive them with just a quick glance.
The general lesson here is that one is not justified in claiming that it appears that there are no Xs if one has reason to believe that, in one’s present epistemic state, one is not in a position to be able to perceive any Xs that might be there. When this is applied to the discussion of the evidential argument, the point is this. Given that God is both omniscient and transcendent, there is every reason to believe that God is privy to a vast amount of knowledge about the relations between good and evil of which we are ignorant. We have reason to believe, then, that for any allegedly pointless evil, if there were some justifying reason that God had for allowing it, we very likely would not be in a position to perceive it. If God exists, it is virtually certain that many of his reasons are inscrutable to us. Consequently, we are in no position to claim, for any actual evil that we observe, that that evil even appears to serve no greater good.  The claim “It appears that there are some pointless evils” is unjustified because it violates CORNEA. But without this claim, premise (2) is unfounded. This response to the evidential argument is sometimes called the cognitive limitation defense.
There is a second response to the evidential argument available to the believer—one that ties into the first response nicely, without being as technical. The theist can state that she believes God has reasons for allowing evil, even if she does not know what those reasons are. The believer’s evidence for thinking that God has justifying reasons for allowing evil will simply be her evidence for God’s existence and goodness. If one has good reasons for believing in a benevolent and loving God, then one is justified in believing that God has good reasons for allowing evil.
To appreciate the force of this kind of response, consider the following argument, which inverts the previous, atheistic argument:
1. If God exists, he does not allow any pointless evil.
2. Probably, God exists.
3. Therefore, probably, there is no pointless evil in the world.
This strategy of turning an argument on its head—sometimes called a G. E. Moore shift, after the early twentieth-century philosopher who made it famous—results, every time, in an argument that is just as valid as the original. Which argument is to be preferred in this particular case? The answer requires that an individual make a judgment about her total evidential situation. Does one have more evidence that God exists, or does one have more evidence that pointless evils exist?
Most would admit that the existence of evil is a problem for the theist; it does “count against” the existence of God, in the sense that it provides prima facie evidence that God does not exist. (Note, however, that Wykstra’s argument could be used to challenge even this assumption.) Even if this is true, however, the important question is whether this negative evidence is sufficient to count decisively against God’s existence (whether it provides what we might call ultima facie evidence). If one has strong reasons for believing in God and for believing him to be good, then evil will be regarded as a difficulty, since one does not understand why God allows evil, but not as a decisive difficulty. After all, as we have just discussed, it seems highly doubtful that finite human beings, with our imperfect and selective understanding of the world around us, could justifiably claim to know that any evil really is genuinely pointless. If one had come to know God as a loving, good being—through religious experience and revelation, perhaps—one would then have powerful evidence that God must have good reasons for allowing evil, even if one had no idea what those reasons might be.
It is, in fact, in just this sort of situation that faith is called for. If the faith is to be reasonable, of course, there must be some basis for belief in God. But adherents of a religion normally claim to have evidence of this type. For example, Christians often cite the incarnation of Jesus as providing them with knowledge of God and God’s character. Jesus’ death and resurrection, while not an explanation of why God allows evil, are a demonstration that he loves his creatures to the point of suffering with them and for them, and that he will eventually triumph over evil by turning it to good. (More on this, shortly.) Evil, then, is a serious problem for the theist, but it is not necessarily an insurmountable one. If the theist has good reasons for believing in God, then he also has good reasons for believing that God is justified in permitting evil. In that case, the occurrence of evil is seen as a test of one’s faith in God.
To the atheist, evil constitutes strong evidence against God’s existence. From a theistic perspective, however, the person who doubts God because of the occurrence of evil needs one of two things. If he does not know God and God’s goodness, he needs to come to know God—through experience or perhaps through special revelation—or he needs to come to know God in a fuller way. If he already knows God and God’s goodness, then he needs pastoral encouragement that will help him persevere in his faith.