A New Cosmological Argument

Towards an uncausable cause

0

I’ve always found the abductive arguments of the form “God best explains feature X of the world,” much better than deductive arguments which argue that some element of the world—existence, change, or causation—necessitates God. I still feel this way and, in my most detailed case for God, stuck almost exclusively to these abductive arguments. Still, however, over the years I’ve grown more sympathetic to a few deductive arguments. I’ve recently found a new deductive argument for theism that I think might work, though I’m not sure. Certainly, if I were trying to convince people of God’s existence, I’d stick with something like fine-tuning, but as part of a cumulative case, I think this is potentially successful.

The argument is broadly inspired by one from Rob Koons and has just a few premises:

1. Everything that exists and is possibly caused is actually caused.

2. There cannot be an infinite chain of possibly caused things without some deeper cause of the chain.

3. Therefore, there exists at least one uncausable cause.

4. If there exists at least one uncausable cause, then God exists.

5. Therefore, God exists.

Unlike, say, the Kalam, the premises have the advantage of being probably true :P. Let’s take them in order.

1

The first premise is fairly standard cosmological reasoning. In explaining the world, as Pat Flynn notes in his book The Best Argument For God, there are three broad sorts of options. First, one can posit an infinite regress of causes, in which case each thing is explained by an earlier thing. Second, one can posit something that could be caused, but just exists for no reason—say, positing that the universe just exists. This is unsatisfactory—we want a theory that explains as much as possible. Positing something simply exists for no reason is a last-ditch effort. A worldview where things don’t just randomly exist for no reason is better than one that does. The final option is to posit something that exists but isn’t the sort of thing that needs an explanation—if, for example, the number 2 could cause things, it might be a good explanation of the universe, for it doesn’t seem like the sort of thing that needs an explanation.

This first premise is intended to rule out the first option—that things just exist for no reason, without any deeper account of why they’re the sorts of things that might exist for no reason. I think there are several things that can be said in support of this premise.

First, it just seems intuitively plausible. It would be very strange if there was something that simply existed for no reason, that wasn’t explained by its nature, but that just existed. One shouldn’t deviate from the commonsense notion that things need explanations unless they have a really compelling reason to do so.

Second, we normally take the things in our experience—tables, chairs, collections of Jellybeans arranged to spell out the phrase “the anthropic argument is clearly successful, and anyone who denies is is a cursed self-sampling assumption proponent,” and the like—to be the sorts of things that need explanation. But what could, other than their causability, explain why they need an explanation? If causable things have causes, it explains why we look for causes of the things in our experience, but it’s hard to see what else could explain why we should look for explanations of those things but not of the cause of the universe.

Some (e.g. me like 2 months ago) suggested that the explanation is that they’re in time, but this is a poor explanation. If time terminates at some point, one shouldn’t expect a bartender or golf ball to subsequently pop into existence for no reason.

Third, simplicity is clearly a virtue. If we denied that simplicity was a virtue, we’d have no grounds on which to trust induction, and doubt that there aren’t more complicated laws that would interfere with the laws we’ve come to observe up until this point. Even a bit of extra complexity massively lowers probability—the theory that God knows every fact except the Capital of Florida has an either zero or infinitesimal prior, even though it’s just an extra arbitrary limit, and the theory that the laws of nature break down in a random patch of space has a similarly diminutive prior.

So the theory that there’s an extra unexplained bit of a theory massively reduces its prior probability, potentially by orders of magnitude, as in the case of the theory that there’s a patch of space where the laws of physics break down. But this means that a theory with zero unexplained elements has a prior probability orders of magnitude higher than a theory with a few unexplained elements. Therefore, one should reject brute facts, and expect all of reality to be fundamentally explicable.

Fourth, this causal principle best explains the data of our experience. It explains, for example, why we never observe tables or chairs without a cause, why things don’t simply pop into being for no reason. The things that seem to need a cause—tables, people, and chairs—are all possibly caused, while the things that don’t, like numbers and moral facts, don’t seem to need a cause.

I’m not super confident of this premise, but I think there’s a lot to be said in favor of it. It establishes at least a strong consideration favoring the premise. It could, of course, be overturned if sufficient reasons were presented, but absent that, I think the first premise is a safe bet.

2

Premise 2 says that you can’t have an infinite causal chain without a deeper cause of the chain. There are three reasons to think this.

First, infinite causal chains are probably impossible. They lead to all sorts of horrifying paradoxes. For example, say there are an infinite number of reapers, one of whom will kill Bob if he’s alive after 30 seconds, one of whom will kill Bob if he’s alive after 15 seconds, one after 7.5 seconds, and so on forever. Does Bob survive? No particular reaper can kill him (because he’d always be killed by an earlier reaper), but he can’t survive, because if he survives any amount of time, a reaper would kill him. Paradox!

I’ll just address one response to this which is the common unsatisfiable pairs diagnosis. On this picture, this scenario is impossible simply because it leads to paradox! You can have infinite reapers, as long as they don’t do anything impossible, like this, just like you can have a red bucket, even though you couldn’t simultaneously have it be blue. Infinite reapers are possible, and these instructions are possible, but not in combination.

Alexander Pruss’s wonderful book Infinity, Causation, and Paradox is, I believe, decisive in refuting this objection. First, he imagines a reversed reaper scenario—one reaper will kill the guy after 30 seconds, one after 45, one after 52.5, and so on. This doesn’t lead to contradiction—he’s simply killed by the first one. But this implies that once this scenario is set up, it would be impossible for each of the reapers to change their instructions—the one currently slated to kill him after 45 seconds would instead do it after 15 seconds, the one who will kill him after 52.5 would switch to killing him after 7.5 seconds. But what magic force could possibly stop these reapers?

More worryingly, imagine that one reaper will kill the person at 12:59, another at 1:30, another at 1:15, another at 1:7.5, another at 1:3.725, and so on. There’s no paradox—the man would simply be executed at 12:59. But now, on such a picture, the first reaper couldn’t switch from killing the person at 12:59, because then the traditional reaper paradox would arise.

The second point to be made in defense of premise 2 is that even if infinite causal chains are possible, they seem to need some deeper explanation. Imagine you see a tiger in the park and wonder why it’s there. Suppose a person pops out from a bush (uncaused, mind you, but don’t worry, because they’re infinitely old) and explains that it doesn’t need an explanation. After all, it’s an infinitely old tiger, so its existence at each moment is explained by its existence at an earlier moment. Nothing’s unexplained—every moment of its existence is explained.

This would be unsatisfying. Sure, it makes sense that each moment of the tiger exists given the one before it, but why the heck is there are infinitely old tiger at all? Similarly, we can ask “even if there’s an infinite causal chain, what explains why it’s there?” An infinite regress doesn’t remove the need for explanation.

Or here’s another case. Imagine we have an infinite chain of boxcars that are stationary. Suddently, they all start moving. You can’t figure out why they all started moving. However, someone helpfully explains that each boxcar’s movement is explained by the one before it. This wouldn’t be an explanation—even if each member in the sequence is explained by something before it, what explains the sequence as a whole??!

The third point that supports premise 3 is that even if an infinite chain can exist without a cause, such a thing is unlikely. If you doubt that it’s unlikely, you have no reason to guess that, say, a tiger won’t appear in your room—as long as it’s explained by an infinite succession of past tigers. But if infinite causal chains are unlikely to exist absent an explanation, then we should have a strong presumption in favor of explaining everything, rather than invoking infinite causal chains.

I’m pretty confident of 2, for this reason. It just seems impossible for there to be an infinite chain, especially absent an explanation.

3-5

Premise 3 says “Therefore, there exists at least one uncausable cause.” It follows from premises 1-2.

Premise 4 says “If there exists at least one uncausable cause, then God exists.” The thought behind this is simple: God’s fundamental nature is something unlimited—unlimited mind, being, actuality, goodness, perfection, or something like that. But all those things seem like they couldn’t be unlimited if the thing was caused. If, for example, a mind was caused, then it’s limited—it’s limited in when, in the causal series, it began, and it’s limited in that it depends on something else.

So God is an uncausable cause. But it’s not clear what else could be. The initial singularity isn’t uncausable—the physical world could be caused by something else. Maybe if you think the physical world is necessary—as Graham Oppy does—then you’ll think it’s uncausable, but that view is so implausible! Surely something else could have caused the physical world? There’s a possible world with, for example, a being that caused the physical world. Oppy’s view also implies different laws of nature would be impossible, for the initial laws are necessary.

Other than God, I can’t think of any things that are essentially uncausable. So then God seems, tentatively, to be the only remaining option that could be an uncausable cause. But if there’s an uncausable cause, then it must be God. One shouldn’t believe in a bunch of different uncasuable causes because that would be less parsimonious.

Premise 5 says that God exists. It follows from the earlier premises.

This is my favorite cosmological argument, but it might go awry—I’m less sold on these arguments than the sort of abductive arguments I laid out here. But if you disagree with the argument, I’d be curious: where does it go wrong?