Is This The Best Argument for God?

Matthew Adelstein presents what he believes to be the best argument for the existence of God.

From the editor of JAT: This guest post comes from Matthew Adelstein (

Bentham's Bulldog

), who presents what he believes to be the best argument for the existence of God. Is it? That, of course, is for you, dear reader, to decide.

Interested in contributing to the Journal of Absolute Truth? Click here for details.

I was an atheist for the vast majority of my life. In this essay, I am going to tell you about the primary argument that convinced me that there was a God. It was not the only argument—fine-tuning also played a major role, as did psychophysical harmony, and a variety of other cumulative considerations. But it was one of the major ones that convinced me that theism makes more sense of the world than atheism.

The details of the argument get slightly technical (I’ve discussed it in great detail here, if you want a deeper overview)—various technical premises need to be defended. But let me first state the brief summary of it. It proceeds from two basic premises. First, it claims that you’re likelier to come to exist if more people exist. If a coin gets flipped that creates ten people if it comes up heads and one person if it comes up tails, and you get created from this process, you should think it’s ten times as likely that the coin came up heads as that it came up tails. The basic idea: if ten times as many people come to exist, it’s ten times likelier that you would come to exist.

The second basic premise is that far more people are likely to exist if God exists than if there’s no God. God has no limits on how much he can create, and he is perfect. It’s good to create a person and give them a good life, so God would want to create. A being of unbounded perfection and creative potential would be at least reasonably likely to engage in unbounded creation. Such a being would be likely, therefore, to create a stupendously large multiverse. It turns out that, as the mathematicians tell us, there are bigger and smaller infinities—if there’s a God, therefore, it’s decently likely that the number of people that get created would be the biggest infinity worth of people that could be created.

In contrast, how many people would exist if there is no God? Probably many fewer. On atheism, maybe you can cook up a multiverse or a very large universe that contains infinite people, but that’s probably the limit. But here’s the thing: as we saw, your existence is ten times likelier if ten people exist than if one person exists, and it is a thousand times likelier if a thousand people exist than if one person exists. But then if atheism predicts a small infinity worth of people existing, and theism predicts a big infinity worth of people existing, because the big infinity is infinitely bigger than the small infinity, you’re infinitely likelier to exist given standard theism than standard atheism.

Or to put it in one sentence: you’re likelier to exist if there are more people than fewer, but because theism predicts that many more people exist on average than atheism does, it makes your existence likelier.

Just to make things clearer, let’s consider a few cases to see how this could work. Suppose first that a coin is flipped: if it comes up heads, ten people gets created. If it comes up tails, one person gets created. After being created you should think that heads is ten times likelier than heads (and I’ll defend this in more detail later).

But now suppose that instead of ten people being created if the coin came up heads, now infinite people will be created. Well now, by the same logic, you should think that it’s infinitely likelier that the coin came up heads.

But now consider a final case. The coin is flipped. If it comes up heads, a really large infinity of people get created. If it comes up tails, the smallest infinity of people get created. Upon being created, by the same logic, you should think heads is infinitely likelier than tails. The argument in a nutshell: theism is like the coin coming up heads in this scenario, and atheism is like the coin coming up tails. Theism predicts a bigger infinity of people than atheism under even the most optimistic scenario, and so it makes your existence likelier.

A clarificatory note: really your existence is only likelier if there are more people of the class of people you might be. If a coin gets flipped that creates ten shrimp if heads and one person if tails, you should think it probably came up tails, because that results in more people that you might presently be. If a coin gets flipped that creates ten men if heads and one man and nine women tails, upon being created, if you’re a man, you should think heads is ten times likelier than tails. You might be any of the ten created men if the coin comes up heads, but can only be the single man created if the coin comes up tails. What matters, according to this principle, called the self-indication assumption (SIA), is that more people exist that you might presently be. The SIA holds: a theory that predicts X times as many people exist that you might presently be makes your present existence X times likelier.

Put in syllogism form:

1. Your existence gives you infinitely strong evidence that the biggest number of people that could exist does exist.

2. That’s way likelier given that God exists than given that he doesn’t.

3. Therefore, the fact that the biggest number of people exists is strong evidence for theism.

Why think 2? Simple—atheism has no plausible story of how you get the biggest number of people that could exist. The smallest infinity is called aleph null, and it’s equal to the number of natural numbers—1, 2, 3, 4, etc. Aleph null is the most people you could fit in a single universe, so if you want more than aleph null people, you need a multiverse. Bigger than aleph null is Beth 1—it’s equal to the number of numbers, which includes integers as well as all the infinitely non-repeating decimals. Even bigger than Beth 1 is Beth 2—in fact, Beth 1 is literally 0% of Beth 2, just as aleph null is 0% of Beth 1.

A plausible lower bound on the number of possible people is Beth 2.

In fact, I think it’s way more than that. As Alexander Pruss has shown, so long as there’s no limit to how much a thing can be duplicated, there can’t be any set of all people. You can duplicate people Beth 1 or Beth 2 or Beth 3 times—so the number of possible people isn’t even a set (the Beth numbers are all sets), and is infinitely far above Beth 2 in the hierarchy of infinities. Also, if you buy that your existence is likelier if more people exist, rather than less, then you should think that the number of people is too big to be a set, because that makes it so that more people exist than alternative theories.

But if the number of possible people is at least more that the number of numbers and at most so incomprehensibly large that it’s beyond the range at which mathematicians talk about it—literally an infinite number of successions of infinite growth over any infinite that mathematicians talk about—how the heck does atheism predict that many people exist? If there’s a God, this many people being created is pretty likely—God has no limit to how much he could create and would want to create because of his perfection. But on atheism, this large a number of people is ridiculously unlikely. None of the multiverse models discussed by physicists get anywhere close to the number of people needed to make your existence anything but infinitely unlikely.

Now, there is one way for atheism to explain this. The atheist could be a modal realist. Modal realists think that every possible world is real. Every single way reality could be exists somewhere in the multiverse. Thus, imagine a world made of just eggs: modal realism says that that world exists somewhere! So does a world where Richard Dawkins became the pope and where unicorns exist.

But even though modal realism can explain your existence—if everyone possible exists, you have to exist—if it’s the only way to explain your existence consistent with atheism, then the anthropic argument massively lowers the probability of atheism. After all, if you eliminate 99.9% of the ways atheism could be true, leaving only some highly contentious premise that implies that unicorns are real and that gods exist somewhere, that is significant progress. A doesn’t have to be the only way to explain B for B to make A more likely—the hypothesis that a fairy rigged a deck by magic and that Jon is cheating both explain Jon getting ten royal flushes, but the second is far more likely.

But modal realism has a big problem—and this problem will likely be shared by every atheistic view that can explain the data (see section 2 here)—it undermines induction. If modal realism is true, there are infinite worlds exactly like ours that spontaneously collapse one second from now. In some of those worlds, the laws are replaced with new chaotic laws, in others the entire world is replaced by a sea slug and thing of spaghetti. For every world where things continue chugging along according to plan, there are infinite worlds where the world collapses. But if we know there are infinite worlds that collapse one second from now, we should have no confidence that, say, the sun will rise tomorrow. So modal realism leads to complete radical skepticism, where you can’t make any predictions about anything!

Let’s turn now to the first premise that “your existence gives you infinitely strong evidence that the biggest number of people that could exist does exist.” Before I defend this, I want to head off a misconception: this doesn’t require every possible person exist. God does not need to create Pennywise the clown, for instance.¹ All this requires is that the number of people that exists is equal in number to the possible people.

An analogy: the number of prime numbers is equal to the number of natural numbers. Two infinite sets are the same size if you can put them into one to one correspondence with each other—pairing all the members of the first with members of the second. You can do that with the primes and naturals—pair one with the first prime number, two with the second, and so on. This is a weird property of infinity that mathematicians have discovered: two infinite sets may be the same size even if one intuitively seems bigger, because it has all the members of the first set, plus extra.

So then why believed this principle, called the self-indication assumption (SIA), that says your existence is likelier if there are more people? I find it has great intuitive pull—it just seems like your existence is likelier if there are more people. It’s also supported by about 27 arguments—in my view, it’s favored by probably the most powerful cumulative case in all of philosophy. I’ve discussed at length the four best arguments for SIA here. Obviously I’m not going to discuss all 27 arguments, but let me discuss two of them.

First, if you reject SIA, you inevitably get the result that the odds of fair coins that haven’t been flipped yet coming up heads is more than 50%.

To see this, imagine one person is created and then a coin is flipped. If it comes up heads nine other people get created. Suppose that you learn you get created but you don’t know your birth rank (that is, you don’t know if you’re the first person or a later person). Then you’re told that you’re the first person.

Because the odds you’d be the first person are only one in ten if there are ten people, while they’re 100% if there’s only one person, unless you start out thinking ten people existing makes your existence ten times likelier, after learning you’re the first person, it inevitably follows that the odds of the coin coming up heads is more than 50%. But fair coins only come up heads half the time! Worse, we can get even more absurd results by changing around the numbers so that you end up arbitrarily certain in improbable chancy events, like that you’ll get many royal flushes in poker, so long as if you don’t get many royal flushes many people will be created.

We can model this mathematically. Consider two hypotheses:

1. The coin comes up heads.

2. The coin comes up tails.

Upon being created, SIA holds that you should think 1 is ten times likelier than 2, and those who reject SIA reject that. But now consider what happens when you learn that you’re the first person. Well, the odds of that are one in ten conditional on 1, while they’re 100% conditional on 2—if the coin comes up heads, you might have been any of the ten people, while if it comes up tails, you have to be the first person. So now you learned something that’s ten times likelier given 2 than 1. If you learn something that’s ten times likelier given a theory than given a different theory, your credence in the theory that makes it ten times likelier should go up by a factor of ten. Therefore, if you don’t start with the credences of SIA—thinking that the ten people existing on 1 makes your existence ten times likelier—you end up with the crazy result that the odds of a fair coin coming up heads are more than .5.

(You might think this only applies to finite cases, but if you just switch around the numbers, so that a small infinity worth of people get created, and then a coin is flipped that creates a big infinity worth of people if it comes up heads, this can be shown to be wrong. Any view that denies that your existence is infinitely likelier if a bigger infinity is created rather than a smaller infinity implies absurdly that you should be certain that the coin will come up tails, after learning that you’re one of the small infinity people created before the coinflip).

Thus, so long as you think that the odds of a fair coin that hasn’t been flipped yet coming up heads is 50% rather than less than 10% and that learning something that’s likelier given one theory than given another makes the first theory likelier, you must accept SIA.

One last worry—particularly common among theists who like the Kalam—is that the argument assumes that God would create an infinite world, but infinites are impossible. However, I think that the argument gives you a reason to accept that infinites are possible. If infinites are impossible, then as we’ve seen, it’s infinitely unlikely that you’d be created—only 0% of the infinite possible people could be created.

Additionally, I think rejecting the possibility of an actual infinite is wildly untenable. I share this judgment with the substantial majority of theistic philosophers, excepting William Lane Craig, including Richard Swinburne, Alexander Pruss, and Peter Van Inwagen. There just are too many things that seem either to be infinite or conceivably infinite: the extent of space, the extent of time, the divisibility of space, the number of numbers, the number of thoughts in the divine mind, and so on. While infinity may have some counterintuitive properties, it’s not contradictory, nor is there any particularly good reason to reject it.

Best of all, one can still make the Kalam work! While infinites existing is, in my view, obviously possible, infinite causal chains are far more dubious. The Kalam can be run based on the impossibility of infinite causal chains, and this makes it, in my view, far more persuasive. You don’t have to twist yourself into knots trying to deny that there are infinite divine beliefs and numbers in order to run the Kalam. Alternatively, you could jettison the Kalam for better cosmological arguments like the contingency argument.

You might still have some worries about the argument (you shouldn’t :) but you might)! If so, I’d encourage you to read my more comprehensive post on the subject, where I provide many more arguments for SIA, defended in quite a bit more detail, and address all the obvious objections. If you’re thinking of some objection right now, I probably addressed it in that post.

But for now, hopefully I’ve given at least a basic sketch of the argument and made clear why it’s so powerful. While atheism might predict that someone would exist—a few isolated people—only theism predicts enough people to make your existence anything other than ridiculously unlikely. Anthropic reasoning—probabilistic reasoning about your own existence—testifies to infinite creation, which testifies to an infinite creator.

About Matthew Adelstein

Matthew Adelstein is an undergraduate studying philosophy at the University of Michigan with publications in ethics and anthropics https://philpeople.org/profiles/matthew-adelstein as well as an amazing blog that you should immediately check out

1. God still might create every possible person—I think he likely does because people are essentially souls and no soul is essentially wicked—but the argument doesn’t require that.