On the free will defense...
The "problem of evil" has plagued philosophers and theologians for centuries, and with good reason. The apparent self-contradiction of an omnipotent, omnibenevolent, and omniscient being has spawned the theological branch of theodicy: the attempt to reconcile a good and all powerful god with the existence of evil. The problem of evil also plagues internet discussion forums in debates over religion and philosophy. Typically, free will is suggested as an antidote to the problem, and then the discussion spirals downward:

"But if God is omniscient, he knows the future, so we can't have free will."
"Yes we can. God allows us to choose. If we couldn't choose to do evil, we couldn't be free."
"But how can we be free given God's foreknowledge. If God knows you're going to do evil, how can you be free to do good?"
"It's seems mysterious, but that's how it is. Just because we can't understand it doesn't mean God couldn't."
"It's not possible."
"Yes it is."
"No, it's not!"
"Yes, it is!"

The discussion halts in standoff, with each side simply unable to see what the other side sees. Either free will in the face of divine foreknowledge seems impossible, or it doesn't.
The free will defense suffers from a number of philosophical problems, but I wish to focus on just one: even if it's granted that we have free will, an omnipotent God would nonetheless remain responsible for the existence of evil. (Note, I'm considering only moral evil, not natural evils such as earthquakes and the like.)

A Simplified Moral Universe

First, to set the stage for later discussion, imagine a vastly simplified moral universe consisting of only one being other than God -- let's call him "Luc" -- and in the course of his existence he makes exactly one moral choice, between good and evil. I'll abbreviate those choices "G" and "E," respectively, and assume that E leads to great suffering and G to great happiness.
As a premise of our thought experiment, we assume that Luc has free will, possesses rationality, and is a fully capable moral agent. Now, an omniscient God would know all future outcomes of his actions, and thus an omnipotent God could surely choose to create (in actuality) either one of two possible universes:

1. In universe 1, Luc chooses E.
2. In universe 2, Luc chooses G.

Logically, God is equally capable of creating either universe. Yet one leads to suffering, the other to happiness. Surely, it seems that an omnibenevolent God would choose to create universe #2. Conversely, a God who knowingly created universe #1 would not appear to be omnibenevolent. Ultimately, then, God must be responsible for Luc's suffering. Luc's free will does not let God "off the hook," so to speak.

What's the point of this thought experiment? Simply that if the above logic holds in our simplified case, it holds in our actual universe. To see why, we employ an argument based on mathematical induction. Note first that we can represent our moral universe as a series of moral choices made by free agents. We might simply list these as follows:

1. GGGGGGGGGGGGGG...
2. GGGGEEGEGGEEGGG...
3. EEEEEEEEEEEEEEEEE...
   

Given a series of N moral choices, there are 2^N possible universes. (Ie, if N=3: GGG, GGE, GEG, GEE, EGG, EGE, EEG, EEE.) The actual number of possible moral universes is, of course, quite vast.

In universe 1, every moral agent always chooses the good. In universe 2, which we might imagine to be much like ours, some agent makes an initial evil choice (in the traditional Christian mythos, Satan would presumably be that agent.) In universe 3, moral agents always choose the evil. For the sake of brevity, we'll label universe #1 "GGG..." and universe #2 "GGE..."

To see the implications our initial thought experiment, we use mathematical induction:

1. Having established that in the case of a single moral choice (G or E) God is still "on the hook" so to speak, we make the inductive step:
2. Assume God is responsible for the existence of evil in the case of N moral choices, and imagine that these are represtented as: GGG...G. (A finite series of Gs).
   
3. Now, consider the Nth + 1 moral choice. God can choose to make one of two possible universes: GGG...GG, or GGG...GE. The same reasoning as in our original thought experiment quickly leads to the conclusion that an omnibenevolent God would choose to create universe GGG...GG over GGG...GE.

That concludes the induction. Thus, while it may seem strange and unlikely, it's logically possible for God to have created a universe in which every agent, when confronted with a moral choice, chose the good. Clearly, that is not our universe.

Responses to the Argument

The most common response to this argument is to deny that GGG... genuinely allows its inhabitants free will. Mark Piper, in The Perennial Problem of Evil, offers an elegant response:

The free will defense implies that God could not have created humans who both possess genuine free will and yet always choose the good – but it should be asked: why couldn’t an all-powerful God do such a thing? Why couldn’t God create genuinely free humans that invariably freely choose the good? Detractors might respond that such a suggestion is tantamount to making us robots, in effect taking away our free will (and thus depriving us of the capacity for genuine moral goodness), but there appears to be nothing self-contradictory in the notion of a moral agent possessing genuine free will and perfect moral goodness. After all, isn’t God said to be a being who has genuine free will and yet always chooses the good? And if so, why couldn’t God have created humans precisely in that image? Surely an all-powerful and perfectly good God would have done so.

Another approach is to deny that genuine moral choices can be made in universe GGG... That is, some evil must exist and be known of in order for moral agents to actually conceive of the consequences of their actions. Imagine, for example, that you had never learned that fire is hot, can burn, and cause pain. One could argue that your choice not to put your hand in the fire is fundamentally different than my choice, if I knew the actual consequences. Although there is some intuitive appeal here, there are several problems with this argument:

1. It immediately implies that the very first case in which E was chosen over G was not a true moral choice. But then, what was it? If it wasn't a true moral choice, then E wasn't truly chosen over G. But if that's true, then the next choice, by the same logic, can't be considered a moral choice, and we have an infinite regression in which true morality never gets off the ground.
   
2. Furthermore, even if we were to grant that some evil must exist in the universe for moral agents to make true moral choices, it's far from obvious that so much evil must exist. Which is to say, rather than "GGEEGEEEEGEGEEEGEEGEE..." why not "GGEGGGGEGGGGGGE..."? It's easy, even as a less than omniscient being, to imagine how humanity might have been made just a bit less irrational, violent, and prone to self-destruction. Since there seems to be no logical constraint preventing God from creating this "less evil" universe, one must conclude that such a god could not be omnibenevolent.
3. Finally, this argument essentially boils down to the "greater good" argument, in which some evil is neccessary for the greater good. Mark Piper's article also offers a number of critiques of this argument.
   

- posted by mike @ 7:29 PM