A Version of the Contingency Argument I Don't Hate

It might work

Ever since I first heard the contingency argument I thought it obviously failed. I thought that it was among the least convincing theistic arguments, with both premises being clearly unsuccessful, and the conclusion absurd. I’ve since moved away from a belief in the absurdity of the conclusion, and over time, I’ve become significantly more sympathetic to each of the premises. It’s not one of the arguments that move me the most, or that moves me very much at all, but I don’t think it’s obviously wrong, the way I used to. Here, I’ll explain the version that I like.

It begins with the principle of sufficient reason (PSR) which says that everything has a sufficient explanation for why it is. Some things, like the basic mathematical, moral, modal, and logical axioms don’t need some external explanation; if you ask why there can’t be contradictions, that doesn’t require anything else to explain it—it seems like the very nature of what a contradiction is provides a full account of why it can’t exist. So therefore this view doesn’t create any infinite regress, for some things can have a sufficient explanation for why they are without having some outside, external explanation.

Why accept the PSR? There’s something intuitively plausible about it. It seems like things do need some explanation. A theory that explains everything is better, ceteris paribus, than one that says of some fundamental thing “it just is—that’s the end of the story—now stop asking questions.” So it seems that there’s a potent intuitive case in favor of the PSR. If there’s a good reason to give up the PSR, that could outweigh the reason in favor of it, but the fact that it’s intuitively plausible at least gives some reason to accept it.

The argument that really sold me, however, came from Alexander Pruss in The Blackwell Companion to Natural Theology. He notes first that we all—or nearly all—accept a more limited causal principle. We accept that the high-level objects in our experience—signs, trains, buses, grains of sand, and people—need to have causes. Pruss, in a blog post, tells a story to illustrate this:

A couple of days ago, my wife found a significant pool of water in the morning on the top surface of our clothes dryer. When I looked at it, it was like 250ml or more. If it were on the floor or on the washer, I would expect it was from a washer-related leak. If our clothes dryer had a water connection for steaming clothes, a leak would make sense (ChatGPT 3.5 suggested this hypothesis). If the quantity were lower, it could easily be from wet clothes put carelessly on top of the dryer or condensation. If there was wetness in the cabinets above the dryer, it would likely be a leak in one of the many containers of cleaning, photo-developing and other chemicals stored there. If the ceiling showed a discoloration above the dryer, it would be a leak from upstairs. If the liquid smelled, it might be urine from the cat sneaking in.

He notes that even when confronted with such a bizarre appearance, no one would reasonably think that this pool of water just came from nowhere, for no reason. We’d always try to look for an explanation. Yet, Pruss claims, restricting the principle that things need explanations to only being about things of the type that we experience in the world would be arbitrary.

Imagine that you lived your entire life in California. You observed that things in your experience need explanations. The correct inference would not be “I guess things in California need explanations.” That would be an arbitrary restriction. Yet what non-arbitrary restriction on the types of things that need explanations can be given? What rule, other than the PSR, can explain why things in our experience need explanations without being arbitrary?

Here’s a first thought: maybe only things within time and space need explanations. But this is unsatisfactory for three reasons. First, it seems unfortunately arbitrary. Why just things within spacetime? Why not other things? This seems intuitively like arbitrarily restricting the principle to things in California.

Second, this is insufficient. We generally accept that non-natural facts need explanations too. Suppose one asserted that having more than 500,000,000,000 grains of rice was impossible, but had no deeper account of that. This would be a terrible theory. It would be terrible precisely because it posits a modal fact with no explanation either in its own nature or in something else. The same is true with various ad hoc mathematical facts, moral facts, and metaphysical facts.

Yet these facts are not temporal. Even if the universe had never existed, 1+1 would still equal 2, married bachelors would be impossible, things would need shapes to have colors, and pain would still be bad. Thus, the principle must be extended to cover these facts which are non-temporal. But then it becomes really arbitrary.

Third, even this causal principle seems insufficient. Suppose that time only holds within the universe. It seems utterly bizarre that the if the universe ends and time stops, a bicycle could just appear from nowhere. On what grounds could we even declare that improbable?

For a while, I worried that the PSR conflicted with modal rationalism, which says that the modal facts all could be figured out by one sufficiently intelligent without having to observe the world. Just as a mathematician in a dark room could prove any particular mathematical truth, so too could they deduce all the modal facts—facts about what things are possible and necessary. So one could figure out that married bachelors are impossible, that things can’t have colors without shapes, and much more.

However, I think not only does the PSR not conflict with modal rationalism, it actually requires it. If all facts need explanations, then so do the modal facts. However, if this is right then, it won’t do to suppose, as some defenders of the contingency argument do, that God just is necessary for no reason. Instead, we must suppose that there is some foundational causal entity, whose existence is fully explainable, either through its own nature or that of another, that brings about all things.

It’s not clear exactly why God would be necessary. Various arguments have been proposed—my argument from modal limits, Godel’s ontological argument, and various others. Perhaps it’s some reason that we can’t understand, or perhaps the reason is to be found beneath the dusty cover of some book about Thomistic metaphysics. Yet I think it’s clear, from these arguments, that there’s a much more plausible story to be told about why God might be necessary than about other beings, like the strings that string theory takes to be fundamental.

Furthermore, I think there’s a convincing argument for why if there is something necessary, it must be God. If we accept the PSR, then for some state of affairs to obtain, it must be explained. But none of the other necessary facts are causal. So it must be explained by the foundational first thing. Thus, for any world to be metaphysically possible, it must be creatable by the first thing. But that means that the first thing can create any metaphysically possible world. Any way the world can be, the first being can bring about.

But if it can do that, then it is omnipotent, it can bring anything about. If it’s omnipotent, however, then it would also be omniscient, for it can bring to its mind any piece of information. If it’s omniscient, it can understand at a deep level the folly and irrationality of immorality, so therefore it would be omnibenevolent. Just as clearly as we grasp the folly of sticking our fingers in the oven, it can see the folly of acting immorally.

To illustrate this argument, suppose, like Graham Oppy, that you think that the initial singularity is necessary. Well then a world with different laws of nature would be impossible. But clearly, such a world is possible. There could have been Newtonian laws! Such laws aren’t able to be ruled out a priori, so modal rationalism entails that they’re possible. Therefore, the first necessary thing must be able to bring about any state of affairs.

Now, maybe you think that the first thing might have infinite power but not have a mind or goodness. Perhaps it’s just some indeterministic system that can bring about any state of affairs. This seems to be able to explain the robustness of the modal domain.

However, this has a few problems. First, once we’ve granted that the first thing must be able to bring about any possible state of affairs, that’s ridiculously strong evidence for God. Ruling out all the non-God options except one makes God way more likely to exist, just as proving that no one in any state other than Florida can outrun you would give strong evidence that no one can outrun you at all. This is especially so because this alternative candidate for the first thing is quite hard to believe in—a theory that there’s some process that randomly generates a universe, out of the possible universes, is not a good theory. Second, it’s hard to see in what sense other states of affairs are really possible. If the initial singularity will bring about some random state of affairs, in what sense are the others possible? Probability is in the mind. In what sense are the other worlds things that it actually could generate? Only a being freely choosing to bring about one particular state of affairs can explain the sense in which the other options are really open.

Third, given that this thing is necessary and fully self-explanatory, it can’t have any chanciness or contingency in it. On the alternative view, this initial singularity would bring about a universe by chance, but then it would have some contingency in it. Only a God who, through constitutive features of his psychology would only bring about one world, can explain why one world emerges over another.

Even if you’re not sure about either step, you shouldn’t be totally sure that they fail. This means theism will start with a pretty high prior probability, for it takes up a significant share of the probability space where there exists something necessary, and that takes up a reasonable share of the initial probability space. Thus, even if you think that each step only has a 10% chance of succeeding, the prior in a necessary God will still be 1%—which isn’t too low, especially given its breathtaking explanatory power. No other model of fundamental reality has a prior probability anywhere near that, especially given that there are infinite possible ways fundamental reality could be. Theism starts off, in terms of priors, infinitely better than average!

One who believes in an infinite regress might accept the PSR but think that they are in a position to explain the regress. Perhaps each thing does have an explanation going infinitely far back. Thus, everything is explained by some previous event. Yet this is not a full explanation for it doesn’t explain why there’s an infinite chain at all.

Suppose that you find a universe in your drawer and wonder why it exists. It would not be helpful to say “well, it’s always been there, so each moment of its existence is explained by a previous moment.” Yes, okay, maybe it’s always been there, but why, pray tell, has it always ben there? Each part of the chain may be explained by a previous part, but why the heck does the chain exist in the first place. If there are infinite train cars propelled by the car in front of them, even if each is explained by the one before them, it’s reasonable to ask: why is there an infinite sequence of train-cars at all? Or, to use an example Pruss has given, if you find a cannonball hurtling towards the earth, that appeared from nowhere, this would seem bizarre—you should look for an explanation. Yet suppose that the cannonball appeared 1 second ago. It’s movement half a second ago is explained by its movement a quarter of a second ago, which is explained by its movement an eighth of a second ago, going infinitely far back. Yet it’s still quite reasonable to ask: why is it there at all?