Over at Maverick Philosopher last week, Bill Vallicella produced an argument that, while not purporting to refute atheism, endeavored to show that the atheist has a harder argumentative job than the theist in at least one respect. Though the whole post needs to be read if any part of it is, I quote the heart of his post:
Only if the atheist’s concept of God is adequate, is the noninstantiation of this concept equivalent to the nonexistence of God. Only if the atheist's concept of God completely and with total accurancy captures the essence of God would the noninstantiation of this concept be equivalent to the nonexistence of God. It follows that the atheist must have adequate knowledge of what God is in order to deny that God exists. To have such knowledge the atheist would have to be quite the theologian! Clearly, the atheist lacks such knowledge: all he has is his concept of God, a concept that he cannot be sure is anything more than a concoction of his intellect. The atheist lacks such knowledge because we all do. The theist is in no better epistemic position: he operates with a concept of God that he cannot be sure is adequate to the divine essence.
A sophisticated theist like Aquinas does not claim to know what God is, but only that God — whom he refuses to pin down conceptually — exists. The asymmetry, then, is this. To deny God — God himself and not a mere figment of one’s imagination — one must know God’s essence. But that is what the divine transcendence makes it impossible for us to know. To affirm God, however, does not require knowledge of the divine essence, but only knowledge that there exists (in plenary mode) Something on which everything else depends for its existence (in derivative mode).
The combox on that post was lively and, for a time, of general interest—though it eventually petered out in a debate about the causal requirements for successful reference, which is typical of philosophers and would only interest philosophers. The really striking thing, however, is that Vallicella revised the post with this "update":
The commenters have convinced me that the argument in this post is no good. But I'll let the post stand because some of what I say is right, and because I don't want to deprive my critics of their target. My argument has this most excellent anti-Continental virtue: it is clear enough to be refuted.
Bravo! A philosopher humble enough to admit he's been refuted! That's rare enough in philosophers and even rarer in the blogosphere. The dissonance of it is mellifluous; I'm envious, which is a sign I'm not as humble as he. This man has at least one virtue, the sine qua non of all the others. And you don't have to believe the ancient doctrine of the "unity of the virtues" to know he has others too.
Given that I am woefully deficient in said virtue, however, I want to suggest an improvement in Vallicella's argument. He has now acknowledged that the atheist needn't be sure he's got all and only the concept of God right in order to be sure that he's produced an argument against the existence of God. All the atheist needs to focus on is some-or-other concept C that all parties to the debate would agree God falls under if there is a God at all; then the atheist could argue that nothing falls under C; and since God is C if there is a God at all, there being no C entails that there is no God. So the question now becomes: what would be the atheist's candidate for C?
My sense is that the choice is nowhere near as easy to make as some atheists suspect. Some have suggested, as a candidate for C, the compound: "all-just and all-merciful." That's a dud for reasons that a parable of Jesus' makes clear. And in any case, whichever candidate for C the atheist picks, it wouldn't be enough for the atheist to claim that nothing happens to fall under C; he would have to argue that nothing could fall under C. And that's in virtue of something about what God is if there is a God at all.
In support of that I quote myself, from a paper where I critically evaluate various kinds of argument for God's existence and adumbrate my own:
Modal arguments are sometimes classed together under the heading 'the ontological argument'; but for reasons I cannot discuss here, I shall leave aside ontological arguments that are properly so-called. Modal arguments are called such because they employ the logical modalities of possibility and necessity, and may be generically framed in the following simple argument:
(1) If God possibly exists, then God necessarily exists
(2) God possibly exists
(3) Therefore, God necessarily exists.
The first premise may seem odd, but it has to be true. It means that if anything could be God, then there is a God that exists eternally, uncausably, and unpreventably. That is because God cannot be a contingent being: God is not the sort of thing that happens to exist but might not have, or that happens not to exist but might have. God cannot come to be or cease to be: God cannot be caused to exist, or just pop into existence, or die. However extraordinary, no such being could be God, but would be only one more item of our world. The same goes for any alleged God who undergoes change; such a being would not be God, but another constituent of our world, though perhaps a pretty special one. Whatever would count as God would have to be eternal, uncausable, and unpreventable, so that it always exists, never really changes, and could neither come into existence nor pass out of existence. Hence, if it could exist at all, it must exist.
The simple modal argument I have presented is obviously valid. But since Hume, many philosophers have been inclined to reject its conclusion, and hence its second premise, as false on logical grounds alone. Their argument goes like this: no existential statement (i.e., no statement to the effect that such-and-such exists) could be a necessary truth; for a necessary truth is one whose denial entails a contradiction, and no statement of the form 'x exists' entails a contradiction; hence no statement of the form 'x necessarily exists' is true; hence it cannot be true that God necessarily exists. And given (1) as well, we may also conclude that God does not possibly exist—i.e., that (2) is false.
But that argument is itself invalid. For from the fact—if it is a fact—that no existential statement is such that its denial entails a contradiction, it does not follow that any statement of the form 'x necessarily exists' is false. In particular, to say that God necessarily exists is not to imply that 'God does not exist' entails a contradiction; it is rather to imply that, given the sort of thing God would be, God exists eternally, uncausably, and unpreventably if at all. The word 'necessarily' in this context is not about the modal status of the statement 'God exists', but is rather a shorthand description of how God exists if God exists at all. So if (3) is false, that is not because it is saying something that logic alone can teach us is false.
The real problem with the modal argument is that there seems to be no reliable, publicly available method for verifying premise (2) that does not require prior knowledge that God exists. In general, we find out that a thing possibly exists (i.e., could exist) in one of two ways. The first is to extrapolate from what one knows to exist. That, e.g., is what astronomers do when they hold that there could be life on some planet orbiting some other star; given what we know about the physical universe, that is a reasonable view, even if it turns out to be false. The second way to find out whether something could exist is simply to find out that the thing does exist, from which it follows trivially that the thing could exist. Obviously, to verify (2) by finding out that God exists would render the modal argument useless as an argument. But from what may we extrapolate in order to make it reasonable to think that God could exist? God is not the sort of thing, like life on a planet, that might not have existed if indeed it does exist, or that exists by first coming to be.
In view of this, some theists would argue that God, though not physically possible, is logically possible. Thus, given a complete description of the concept of God, we would find that God is, in a certain sense, like planets but unlike things that are both black-all-over and not-black-all-over. Like the former and unlike the latter, the claim that there is a God would not entail a contradiction; hence God could exist (is possible). But of course, there is no agreement on just what such a description would entail or on whether we could fully understand it even if we could agree on all that it would entail. Short of knowing that there is a God, there seems to be no reliable way to find out just what God is; if so, then one cannot know that God's existence is possible without knowing that God exists. It seems we are at a loss here. By itself, the modal argument does not get us very far—though it does highlight something about what God is that must always be kept in mind.