The cosmological argument for the existence of God starts from the assumption that whatever exists contingently requires an explanation. Given the further assumptions that the world around us exists contingently and that infinite regresses of explanation are ruled out, we get the conclusion that there must exist a necessary being – and “this all men call God,” as Aquinas blithely notes.
Actually Aquinas knows perfectly well that the argument isn’t finished at that point, and goes on to argue that a necessary being would have to have the traditional attributes of God – uniqueness, goodness, omnipotence, omniscience, etc. But those arguments, whatever their merits, aren’t my current concern. Nor shall I consider the assumption that the world around us is contingent, though that assumption is open to challenge as well. My present beef is with the initial assumption that whatever exists contingently requires an explanation.
I have two objections. Here’s the first. Suppose X is a contingent being. In that case, X’s existence is supposed to require an explanation. But why? Presumably because if something is the case which might not have been the case, we need an explanation for why it’s the case rather than otherwise. But then it seems to follow that if X did not exist, its nonexistence, being likewise contingent rather than necessary, would likewise require an explanation.
At this point the demand for an explanation of why X exists begins to look puzzling. Since X is a contingent being, X’s existence is somehow supposed to be metaphysically surprising and to require explanation. But now it turns out that X’s nonexistence would also be metaphysically surprising. But X’s existence and X’s nonexistence are the only logically possible options; how can they both be surprising?
Suppose I reach into an urn containing 10,000 marbles, and I randomly pull one out. Then I reflect: “The odds of my getting this marble were one in 10,000! How amazing! What explains this extraordinary event?” This would be a confusion. By reaching into the urn I guaranteed that I would get one or another of those 10,000 marbles; if there’s nothing special about this marble that the other marbles lack, then there’s nothing to be surprised about – since whatever marble I got was guaranteed to be one in 10,000. By the same logic, if X’s existence and X’s nonexistence are both contingent, and yet those are the only two possible options, then it’s guaranteed that some non-necessary state of affairs will be the case. If that’s so, then there’s nothing metaphysically surprising about it – so why must it require an explanation?
So my first objection is that we don’t need an explanation for every contingent being. My second objection goes farther: that there couldn’t be an explanation for every contingent being.
Why not? Well, granting that every explanatory chain must be finite, consider the causal origin Y in which the explanation of contingent being X terminates. Y is ex hypothesi a necessary being. But what about Y’s causing X? Is that necessary or contingent? If it’s contingent – that is, if Y could have existed without causing X – then we still have an unexplained contingent being (and the fact that Y’s own existence is necessary doesn’t help).
On the other hand, if Y’s causing X is necessary, then since Y’s existence is necessary too, it follows that X’s existence is likewise necessary – in which case we haven’t explained the existence of a contingent being at all, since X turns out to be a necessary being rather than a contingent one.
It follows that if there are any contingent beings at all, then necessarily some of them have no explanation for their existence. In which case the cosmological argument can’t get off the ground.
For example, if Y is supposed to be God, then the question is whether the act of will whereby God creates X is necessary or contingent. If it’s necessary, then so is X, belying the original premise of X’s contingency. And if instead God’s act of will is contingent, then we still have an unexplained contingency – now it’s just the act of will leading to X rather than X itself.
Unless, of course, one wishes to say that something can explain X without being sufficient for X. Now in fact I have no problem with saying that. But once one has said that, then one has granted that explanation can be contingent, in which case the whole rationale for chains of explanation terminating in something necessary has been given up.