A New Argument For God: The Argument From Modal Limits

I'm still an atheist, but I think this is a fun argument nonetheless

I wish politicians, would stand up for Anselm, not just Gaunilo on an island somewhere.

—Amos Wollen, paraphrasing Oliver Anthony.

The original ontological argument is the philosophical equivalent of a parlor trick. It begins by defining God as a maximally great being, before noting that a being is greater if it exists than if it doesn’t, therefore God must exist. Almost everyone who hears this argument is aware that something has gone awry, and those who aren’t aware of this frequently fall for scams involving Nigerian princes, but it takes a few minutes to figure out what exactly the error in reasoning is. For the simple version, the answer is that the fact that a being exists by definition doesn’t mean it exists—a necessary Martian exists by definition but doesn’t actually exist.

To make a successful ontological argument, we’ll have to do better. First, unlike Anselm’s argument, it will have to not rest on a basic logical error. Second, it will have to be invulnerable to parodies. Parody arguments rely on the same format to prove other conclusions that are clearly false. Anselm’s original reasoning could also be used to prove that there is a maximally perfect cheesecake because such a cheesecake is better if it exists than if it doesn’t.

Can we do better? I think we can, with the following argument:

1. Limits on possibility require explanations.

2. There is no explanation for a limit on God’s possibility.

3. So God is possible.

4. If God is possible, he exists.

5. So God exists.

The first premise is quite plausible. Married bachelors are impossible, but there’s an explanation of that—they’re impossible because they entail a contradiction and contradictions are impossible. Furthermore, this principle follows from the principle that limits require an explanation which is also plausible—if something is restricted in some way, it seems there must be an explanation of why it is so. The speed of light couldn’t just be an arbitrary uncaused speed limit—it requires an explanation, as Josh Rasmussen has argued. The third premise follows from the first two.

The fourth premise sounds controversial but it is actually rather trivial. God is defined as being a necessary being. As a consequence, if God exists in some possible worlds then he exists in all of them. Compare: if Fermat’s last theorem is true in some possible worlds then it’s true in all possible worlds. So if God possibly exists then he actually exists. This follows from a formal logical system called S5. For more on this, see the modal ontological argument which employs a similar premise, though is flagrantly question-begging.

This argument is also invulnerable to parody arguments. It couldn’t be used to prove a maximally perfect island, because such an island would be limited in some way. If limits require explanation, then there can’t be a necessary limited thing, because its limits would need something else to explain it.

It also couldn’t be used to prove a maximally evil being. I accept a view of rationality, argued for in Reasons and Persons, according to which immorality is necessarily irrational. As such, a maximally evil being would have to be limited in terms of its rationality.

You might object that a perfect being is also limited—it is limited in the number of persons it is, for example, and how horrible it is is severely limited. Here, however, we have to distinguish fundamental limits from higher-order limits. A limit is fundamental if one of the most basic features of reality is limited. A limit is higher-order if it derives in some way from the most fundamental features of reality.

Something can have limits as long as they are not fundamental. Remember, the claim is that limits need explanations, but they can have explanations in terms of more fundamental and unlimited features. If God is perfection itself, that fundamental property is unlimited, but that results in other higher-order limits. Fortunately, because they’re explained, a necessary God is still possible.

Another way you could go to escape these arguments is to be a mereological nihilist. Thus, you’d claim that only what is fundamental exists. But minds are fundamental, so only an unlimited mind is truly unlimited. But an unlimited mind would possess the constitutive features of a mind in infinite quantities—knowledge, freedom, and ability to will things, meaning it would be omniscient and omnipotent. It would thus be omnibenevolent if we accept that wrong action is a form of irrationality.

Finally, if you endorse divine simplicity, you can think that a being like God is truly unlimited in all respects—limits only exist analogically.

Where do I think this argument goes wrong? I reject premise 2. I think that if something is defined as necessary then it isn’t possible absent some explanation of its possibility. Contingent things are possible unless something restricts their possibility, but necessary things need an explanation of why they’re possible. This seems quite intuitive—the reason that a necessary leprechaun is impossible is that there’s nothing that would explain why it is possible. The claim that a necessary leprechaun is possible is the same as the claim that there is a necessary leprechaun—but such a thing would require an explanation.

Still, the endorser of this argument can reject this principle. They can claim that a leprechaun can’t be necessary because it’s limited—an that needs an explanation. Furthermore, they could note that, as long as their metaphysical principle about limits on possibility requiring explanations is somewhat plausible, it raises the prior probability of God. Finally, they could claim that it’s a better explanation of existence—if fundamental reality is contingent, then the odds of stuff existing is less than 1, while if it’s necessary then something will necessarily exist. Thus, their principle that limits require explanations and this is why necessary leprechauns are impossible, best explains why something exists at all.

I don’t think this is a great argument. But it’s probably the best version of the ontological argument and perhaps raises the prior probability of theism a bit.