Goldschmidt’s volume opens with an essay by Timothy O’Connor who defends the traditional answer to the question of why there is something rather than nothing: God. More specifically, the traditional answer O’Connor defends holds that a necessarily existent immaterial agent chose that contingent beings should exist.
There are several well-known difficulties for this kind of view. The first difficulty is, if there must be an explanation of why there are contingent beings, then mustn’t there be an explanation of why there is a God? This is, of course, a version of the much-ridiculed ‘what caused God?’ retort, and O’Connor’s (implicit) answer to it is that God exists necessarily. (O’Connor implies this response by restricting his ‘principles of explanation’ to contingent beings/events/truths; pp. 35-37.) Now this (standard) answer can be understood in one of two ways: either necessary truths don’t need explanations, or else we claim that any necessary truth p is explained by the fact that necessarily p. That is, on the second option, you explain a necessary truth by asserting that it is necessary. However, the second option by itself doesn’t solve the problem, because we can always ask why it is that God necessarily exists. Based on O’Connor’s discussion of ‘opaque necessities’ I suspect that he endorses the first option, denying that necessary truths need explanations. (To me, brute necessities seem intuitively worse than brute contingencies, but I won’t pursue that point here.) So God’s existence, being necessary, doesn’t need an explanation, but the existence of contingent things does.
However, the opponent of the traditional (theistic) view has an easy retort: “Suppose we grant, for the sake of argument, that God exists necessarily. Surely God’s decision to create this world must be contingent, since the world could have been otherwise. So there must be an explanation of why God chose this world.” We actually still haven’t got much deeper than the ‘what caused God?’ question at this point, for there is quite an obvious answer to this challenge. According to the traditional view, the universe’s existence depends on a free choice, and we know how to explain free choices: we cite the agent’s reasons, desires, character, etc.
In traditional treatments of this issue (e.g., Aquinas, Leibniz), the theist would now go on to give some account of the reason why God created this world. O’Connor makes a different move: he argues that the theist need not do this. According to O’Connor, the superiority of theism over its competitors is shown by the fact that it provides an intelligible explanation schema: that is, we can see how an explanation could go, and what sorts of questions would have to be answered in order to complete the explanation.
O’Connor seems to me to be correct that a hypothesis which implies that something is in principle explicable, and specifies a particular sort of explanation it must have, is ceteris paribus to be preferred over a hypothesis which renders that thing in principle inexplicable. This is so even if the hypothesis doesn’t actually explain the phenomenon in question. Now, it is widely held that the existence of contingent beings is in principle inexplicable unless there is a necessary being. Further, since we have some kind of conception of how agential explanations go, the hypothesis that contingent existence is caused by a necessarily existent agent is ceteris paribus to be preferred to the hypothesis that no necessary beings enter into causal relations.
Two important limitations must be observed here. First, no argument has been presented for the claim that the conception of the necessary being as an agent is superior to alternative necessary being theories. Second, the result is merely a ceteris paribus claim. O’Connor accepts both of these limitations, though he does give some consideration to the question of how an all-things-considered comparison of the two views might go. On this latter point, he is criticized by Oppy in the following chapter, so I will leave off discussion of that until my next post.
I should also briefly mention O’Connor’s response to the modal collapse objection. This objection holds that whatever has a necessary explanation is itself necessary, and so the traditional view, far from explaining contingency, denies the existence of contingency. O’Connor’s response is simple: to cite a cause of something is to give one kind of explanation of it, and that’s the kind of explanation he thinks contingent existence needs. Not all causation involves ‘necessary connection.’ Hence, a necessary thing might contingently cause contingent things, and this would not take away their contingency. (O’Connor does not here discuss the regress worry: not only is the proposition this world exists contingent, so is the proposition God causes this world. What’s the explanation of the second proposition? Since O’Connor has written a lot about agent causation, I’m sure he’s discussed this somewhere.) O’Connor thinks that if you are unsatisfied with this it must be because you are looking, as Leibniz was, for a contrastive explanation, an explanation of why things are so rather than otherwise. O’Connor is happy to deny that such explanations exist.
I’m a little concerned about this response; I tend to think that if one has explicability intuitions strong enough to support the argument from contingency, one is unlikely to be satisfied by weak explanations of this sort.
On the whole, O’Connor’s essay is a competent presentation of the traditional view in the context of contemporary analytic philosophy. He departs from the traditional view mostly in his exhortations to epistemic humility. In a way, this essay was a good choice to begin the volume: it lays out the view that most of the other papers will be, in one way or another, attacking. On the other hand, I found each of the three following essays (by Oppy, Kleinschmidt, and Ross – that’s as far as I’ve read) far more interesting. For the specialist, O’Connor’s essay is rather a slow start to the volume.
(Cross-posted at blog.kennypearce.net.)