Since Quine, the following ontological principle has seen much popularity: we should accept as existing only those entities which are postulated in the best explanation of what we accept to be the case. It follows that we have good reason to believe that x exists if and only if x has some explanatory power: only those entities the postulation of which is necessary for explaining something should be believed to exist. I will argue that appeal to God is always explanatorily impotent.
For us to be justified in accepting the existence of God, the postulation of God must have some explanatory significance. A common argument for the non-existence of God is that there is nothing the (best) explanation of which requires the postulation of God. This was Laplace’s line when he told Napoleon that he did not believe in God because he had ‘no need for that hypothesis.’
It has often been assumed that a canonical explanandum consists of a single event or fact. More recently, however, Peter Lipton has claimed that a canonical explanandum is a ‘contrastive phenomenon’: a pair consisting in a fact and a foil. A fully specified explanatory statement does not take the form ‘p explains why q’, but rather the form ‘p explains why q rather than r’. In its minimal form, an explanatory statement would be ‘p explains why q rather than ~q’.
Let us further assume that one way to explain something (e.g., a contrastive phenomenon) is by citing its cause – what brought it about or made it the case. The assumption is that causal explanation is a genuine and legitimate kind of explanation. This is much weaker than holding a causal account of explanation, according to which all or most explanation is causal. The claim is merely that citing of a cause is explanatory.
Some contrastive explanatory statements are therefore causal statements. They are statements of the form ‘p caused q rather than r’, or at least ‘p caused q rather than ~q’, Such a statement may be ambiguous as between ‘p caused q rather than causing r’ and ‘p caused q-rather-than-r’. That is, there is an ambiguity between a contrast of causation and a causation of contrast. But this ambiguity will not matter for present purposes.
We are now in a position to see why appeal to God is always explanatorily impotent. Because God is omnipotent, there is no possible fact, event, or state of affairs He would be unable to bring about. There is no proposition He could not make true. Moreover, His infinite power implies that differences in the effort required on God’s part to bring about different states of affairs are negligible. It follows that for any proposition p, ‘God caused p’ (Or ‘God caused it to be the case that p’) is just as plausible as ‘God caused ~p’ (Or ‘God caused it to be the case that ~p’). That is, for any p, citing God is just as good a causal explanation of p as of ~p.
This creates a principled problem for employing God in contrastive explanation. Statements of the form ‘God explains why p rather than ~p‘ can never be true because citing God can never account for the contrast between p and ~p. Ironically, then, it is God’s omnipotence that makes Him explanatorily impotent.
I now turn to consider three objections to this argument. First, it might be objected that although God could cause anything, it does not follow that He would. God may have good reasons for preferring the obtaining of p over the obtaining of ~p, and act on those reasons.
Note, however, that in the entertained explanation it is not the appeal to God that does the explanatory work. Rather, it is the appeal to the reasons attributed to God. What explains the fact that p rather than ~p is thus the fact that God has the reasons He does. But that explanation already presupposes that there is a God. What we are still missing is an explanatory context in which God might be introduced into our ontology in the first place.
Second, it might be objected that even though an appeal to God cannot figure in contrastive causal explanation, it can still figure in different forms of explanation.
One reply would be to embrace the thesis that all explanation is causal. But that is unnecessary. Even if p can be explained non-causally whereas ~p cannot, it remains the case that ~p can be explained causally, since everything can be explained causally by appeal to God’s doing. So for any proposition, some explanation of its negation by appeal to God would be available.
Third, it might be claimed that the argument commits a verificationist fallacy. It may show that we cannot establish whether God actually brought it about that p or that ~p. From this, it does not follow, however, that there is no fact of the matter as to whether God brought it about that p or that ~p.
This objection is an ignoratio elenchi. The above argument is not supposed to show that God does not exist. Rather, it attempts to show that there is not — or more accurately, could not be –reason to believe that He does. That is, the claim is that belief in the existence of God is epistemically unjustified. The argument does not impose a verificationist constraint on truth or reality, though it may impose a verificationist constraint on justification.
I conclude that appeal to God is explanatorily impotent, and necessarily so, given the contrastive structure of explanation. This means that there could be no grounds on which to introduce God into our ontology.
This is also the vein of Russell’s ‘teapot atheism’: there is no more evidence for the existence of God – no more explananda calling out for the postulation of God – then there is for the existence of a miniature China teapot orbiting the sun. See Bertrand Russell, ‘Is There a God?’ in The Collected Papers of Bertrand Russell, Volume 11: Last Philosophical Testament, ed. J. G. Slater and P. Köllner, (London: Routledge, 1997).
Peter Lipton, ‘Contrastive Explanation’ in D. Knowles (ed.), Explanation and its Limits, (Cambridge: Cambridge University Press, 1990).
Lipton’s favorite example is the following. Lipton’s three-year-old son throws the food on the floor. When asked to explain his misdeed, the child says that he was not hungry anymore. This explains why the child threw the food on the floor rather than eat it, but it does not explain why he threw the food on the floor rather than leave it in his plate.