Arguments For God Tier-list

#4238497328974 will SHOCK you

Tier lists on arguments for God are all the rage! Tier lists rank arguments for God on a scale from F to S, where F is the worst, S is the best, and the rest follow a traditional letter grade—A better than B, B better than C, and so on. If you want a good heuristic for an ideal tier list, just take the opposite of Redeemed Zoomer’s (his channel is quite good though)!

I thought I’d join the fun and make a tier list myself (I already did this in video form with my friend Amos (blog here), so watch that if you’re interested). You’ll see that when it comes to arguments for God, I am, like Sarah Palin, a maverick, with mavericky choices of the best arguments for God. My views on which arguments for God are good diverges sharply from the standard views.

Note, in this piece I will only basically describe my thoughts on each of the arguments. I will not address them in much depth—there are too many to do adequate justice to—but I’ll generally link to related literature.

F (Arguments with zero or negative force)

The moral argument/other arguments from necessary truths

The moral argument is the most overrated argument for God in the history of the human species! It goes:

1. If there is no God, then there are no objective moral values and duties.

2. But there are objective moral values and duties.

3. So God exists.

The problem is with the first premise. There’s simply never been any convincing reason to think that, say, a Platonist account of morality is defective. Given the many different accounts of morality, it’s hard to see on what grounds one would declare them all defective.

The argument can only work if God is the sort of thing capable of grounding the moral facts. But how can an agent ground a moral fact? It seems that, say, the wrongness of torture depends on facts about torture, not facts about some extra agent. God can’t ground the moral facts any more than physical things could ground the logical facts. And there isn’t even a parsimony advantage to using God to explain the moral facts, because for every fact the atheist posits as brute, the theist brutely grounds it in God.

I’ve laid out all these charges in more detail here. Ironically, if I became convinced that the first premise was true, then I’d stop being a theist. If morality can’t be grounded in God, and wouldn’t exist absent God, then objective morality doesn’t exist. But theism can only be true if objective morality exists (how can God be perfect if there’s no such thing as objective perfection). Thus, I think the argument is a massive own goal—the stronger the arguments for premise 1, the lower your credence in theism should be.

I’ve only talked about morality, but the same points apply, mutatis mutandis (wow, aren’t I smart, using SAT words), to arguments from math, logic, and modality.

Nomological argument

I feel bad including this argument because I think that it’s a clever argument and it’s not obviously wrong. Nonetheless, I felt I had to include it because I think that it has negative force—it raises the likelihood of atheism.

The nomological argument claims that God is the best explanation of causal regularities. All of the particles in the universe obey the same laws. Hildebrand and Metcalf—two quite good philosophers—give the following analogy: suppose that you’re playing poker and you see a person keeps getting royal flushes. It would be reasonable to conclude they were cheating. This is because:

1. The regularity is very surprising by chance. It needs an explanation. Why is it that they keep getting the same good hand, out of all the hands they might get.

2. The regularity is plausibly explained by the actions of an intentional agent.

Then it’s argued that the regularities in physics are unlikely to be explained by chance. After all, there are many more ways the particles could behave that aren’t regular than that are. Thus, it’s surprising that they behave regularly—and we should infer an agent, who might have a desire for regularity, is responsible.

I think both premises are suspect.

First of all, it’s very mysterious why God follows these universal regularities. Why doesn’t God suspend the natural regularities to prevent various bad things from happening? Unless one simply has an intrinsic desire for atoms to dance in accordance with predictable laws, it’s quite surprising that they’d direct atoms to follow predictable laws.

Like, imagine that you could control atoms to bring about various goals of yours. It’s highly unlikely the atoms would follow universal physical laws. Instead, you’d constantly have atoms zigging and zagging—so long as such a thing was conducive to your aims. Thus, I think the regularity of the world is surprising on theism.

In contrast, I think it’s positively expected on atheism. Fundamental reality is likely to be simple. A world where every particle behaves the same laws—where they’re all the same kind of thing—is simpler to specify than another world. It takes a shorter time to describe uniform physics than non-uniform physics. Thus, for simplicity reasons, on atheism we’d expect uniform physical laws!

I think, therefore, that the presence of uniform regularities ends up favoring atheism, for this reason.

The Anselmian ontological argument (F)

This argument goes roughly as follows:

1. God is the greatest conceivable being, possessing everything that makes one great.

2. Existence makes one great.

3. Therefore, God exists.

The culprit lies in premise 1. Suppose we say “unicorns are horses with horns.” This can mean two things. The first thing it might mean is that it’s part of the definition of being a unicorn that they have horns, and the second is that there actually are horses with horns. Translated into logic, the first is “if something is a unicorn, it is a horse with a horn,” and the second is “there exists some unicorn that is a horse with a horn.”

Given the first reading, the argument just shows either that God exists by definition or that if God exists he exists. Neither is of much significance. The fact that it follows from the definition of something that it exists, doesn’t tell us whether it actually exists: the smallest actually existing Martian exists by definition, but this doesn’t tell us that it actually exists.

Additionally, the argument is vulnerable to all sorts of parodies. Consider the worst possible being. Well, it’s worse if it exists than if it doesn’t. Thus, by the same reasoning, it would exist. Or consider a being possessing every great-making feature except for knowing the number of atoms on mars. Well, it’s better if this being exists—thus, there would have to be some other, limited God that knows every fact except the number of atoms on mars.

The normal Anselmian argument thus rests on a basic logical error. It has exactly zero force.

The modal ontological argument (F)

This argument goes roughly as follows:

1. It’s possible that God exists.

2. If it’s possible that God exists, then he actually exists.

3. Therefore, God exists.

While it may seem like 2. is suspicious, that premise turns out to be fine. Here, talk of possibility is in terms of possible worlds—ways for reality to be. Something is possible if there’s a way for reality to be that would have it. For instance, though there are no unicorns, they’re possible, because reality could have had them.

Now, the argument goes, God is the sort of thing who is necessary if he exists. Because it’s good to exist in all possible worlds, by definition a perfect being would exist in all possible worlds. Thus, God if he’s possible at all would be necessary.

But if something is possibly necessary then it is necessary. If, for instance, Fermat’s last theorem is true in one possible world then it’s true in all of them. Thus, if God—a necessary being—exists in any possible world, he must exist in every possible world, including the actual world.

The problem is, of course, the first. If a premise asserts the possible truth of a claim that is necessarily true if true at all, then it’s straightforwardly question-begging. For example, if I said “possibly a necessary unicorn exists,” this would be just the same as asserting that a necessary unicorn exists. If I said “possibly Fermat’s last theorem is true,” this would be the same as asserting that Fermat’s last theorem is actually—and necessarily—true.

So the first premise is obviously question-begging—it amounts to asserting that a necessary God exists. The argument is just a trick. And there’s no reason to grant the first premise instead of the inverse premise: possible God doesn’t exist. One could similarly argue:

1. Possibly God doesn’t exist.

2. If God possibly doesn’t exist then he doesn’t exist.

3. Therefore, God doesn’t exist.

Proponents of the modal ontological argument would agree with premise two. But there’s no reason to prefer premise one in the original modal ontological argument to premise one in the modified modal ontological argument.

Now, there are ways that people try to support the possibility premise—none of these succeed, unfortunately. But those at least have the possibility of having force. So if you include those, the modal ontological argument could work. However, the standard version as presented here, without one of those fancy symmetry breakers, is just a complete and utter failure. It’s simply question-begging.

Additionally, the argument is also vulnerable to the parodies I listed earlier. We could similarly argue:

1. Possibly god- (a being like god except he doesn’t know the population of Paris) exists.

2. If possibly god-exists then he exists.

3. Therefore, god- exists.

Similarly, we could argue:

1. Possibly a maximally horrible being exists.

2. If a maximally horrible being exists possibly, then it actually exists.

3. Therefore, a maximally horrible being exists.

People usually respond to this by suggesting that existence is a great making property, so a maximally horrible being wouldn’t have it. But surely whether existence is good or bad depends on the sort of thing possessing it. It’s bad for the maximally horrible being to exist, while it’s good for God to exist. Thus, a being with all the worst properties—for it to have—would have necessary existence.

D

Biological design (D-)

The argument from biological design claims that certain biological features that we possess exhibit irreducible complexity. This means that they have parts that fit together but couldn’t arise by chance. For example, it’s claimed, a mousetrap couldn’t evolve gradually, because no part of the mousetrap has survival advantage absent the presence of the others. It’s argued similarly that, say, the bacterial flagella could not have evolved.

Now, I’m not a biologist, so it’s hard to know quite what to make of this argument. But the biological claims don’t seem justified, and evolution has tremendous explanatory power.

The argument from motion (presentation from Feser) (D-)

There are different formulations of the argument from motion, but it starts from the premise that change is the actualization of potential. When something changes, such people claim, it goes from being a way it can be to a way that it is. In order to actualize the potential, something else needs to actualize it. For instance, my coffee has the potential to be hot, but for it to become hot, something (E.g. a microwave) must actualize that potential.

Now, if each thing to be actual at any moment depends on some earlier entity to actualize it, then this pushes things back to a purely actual actualizer (proponents claim). Then they argue that this purely actual actualizer must be God.

I’ve never understand why Thomists seem to love arguments of the form:

1. Controversial premise.

2. Controversial premise.

3. Controversial premise.

4. Controversial premise.

5. Poem.

6. Controversial premise.

7. Gibberish.

8. Controversial premise.

9. Therefore there must be esse in some subject only insofar as the esse’s esse depends for it on the workings of some being not containing, through itself independently of another the beginning of esse qua being.

10. Controversial premise.

11. Controversial premise.

12. Therefore, God exists.

I mean, first of all, the argument from motion relies on a particular account of change. It says change is actualization of potential. Almost no-one else thinks this is what change is. This view:

1. Doesn’t really seem to explain change. What is it for potential to be actual other than for it to change to being actual. Thus, the explanation is parasitic on the thing being explained.

2. Requires that things can exist in different ways—as actual or potential—which is probably false.

Second, the argument requires that for anything to be actual at some time, another thing must be actualizing it at that time. Thus, the view assumes that existential inertia is false, but it’s almost definitely true.

Third, the view only gets to a being that actualizes other things, not a purely actual being. There’s no way the argument rules out the first being having potentials.

Fourth, the arguments that attempt to bridge the gap from a purely actual being to God are just hopeless. None of them come close to succeeding.

The argument has many more controversial steps. But, as I think Schmid and Linford argued convincingly, there are many different steps of the argument that are almost definitely false, or that don’t follow from previous steps.

I’ve made this charge specifically against Feser’s presentation. I think this might also apply to Aquinas’s, but I don’t know how to interpret Aquinas. There are, I think, some versions of the argument with better prospects. For instance, my friend James has a version that is maybe C- tier. This argument makes it as essentially an improved version of the Kalam. It says that changes need cause, and there can’t be infinite causal chains, so by regress, there has to be at least one thing that causes others to change without being previously changed. Now, I don’t think that this must be God, and I’m not sure whether you can have an infinite causal chain, but it’s at least some evidence for God. This version is way better than Feser’s!

The neo-platonic proof (D-)

The neo-platonic proof says that if something is made of parts, there must be something responsible for the parts being combined. By regress, the first thing can’t be made of parts, because it would need something else to bind it together. Thus, there must be something without parts, which is argued to be God. In syllogistic form, the neo-platonic proof (scooped once again from Schmid and Linford’s book)

1. The things of our experience are composite.

2. A composite exists at any moment only insofar as its parts are combined at that moment.

3. This composition of parts requires a concurrent cause.

4. So, any composite has a cause of its existence at any moment at which it exists.

5. So, each of the things of our experience has a cause at any moment at which it exists.

6. If the cause of a composite thing’s existence at any moment is itself composite, then it will in turn require a cause of its own existence at that moment.

7. The regress of causes this entails is hierarchical in nature, and such a regress must have a first member.

8. Only something absolutely simple or noncomposite could be the frst member of such a series.

9. So, the existence of each of the things of our experience presupposes an absolutely simple or noncomposite cause.

The problem is that the argument fails on every plausible view of composition. Mereological nihilism says that there aren’t composite objects—if this is right then 1 is false. Mereological universalism says that every collection of simples forms its own composite object; for example, there is a composite object composed of my left ear, the pope, and my computer. On this view, the needed ingredients for composition are: nothing. All you need for composition is for things to exist. Lastly, there are restrictivist views that say that composition only sometimes occurs—but these views specify the conditions needed for composition. Thus, either 1. or 4. is false.

The premise is also just wildly unintuitive. If there are some atoms that make up a chair, it seems you don’t need some extra sustaining cause for them to make up a chair. All you need is the atoms arranged properly for them to compose a chair. The parts properties alone is enough to explain why they form a composite object.

Additionally, even if you can get to an object that is not composed, there isn’t a plausible route from this to God. You could have a simple physical object, for instance. Thus, the argument is a total failure.

The Kalam (D+)

I hesitated putting this in D tier as I think it’s a lot better than the other arguments I’ve listed so far. The problem is that the Kalam ought to have very little persuasive power given the large number of questionable steps. For the Kalam to be right, as I’ve argued recently, it must be that:

1. B-theory is false.

2. The past is of finite age.

3. Things that start existing need causes.

4. If the universe had a cause it was God.

The problem is that there aren’t knockdown arguments for any of them. I suspect 1 and 3 are false, and only believe 4 based on other arguments—I don’t think there are reasons given by the Kalam to accept 4. I don’t think you should be very confident in any of the premises, given their extreme degree of philosophical controversy, and even if you give them each 70% odds, there’s only around a 24% chance that they’re all right. Thus, I think the argument wholly lacks significant argumentative power.

In contrast, the other arguments on the list appeal largely to uncontroversial premises, or premises supported by knockdown arguments. They’re thus much more successful.

C

Arguments from miracles (C)

Normally, when people argue to God from miracles, they’ll argue that whichever miracles are of their chosen tradition are historically well-evidenced. Jews will argue that the Exodus is well-attested historically; Muslims argue that Muhammad splitting the moon is a historical fact; and most famously and persuasively, Christians will argue that Jesus rising from the dead best explains the facts.

While I think the historical evidence for Christianity is okay, it’s not enough to make me a Christian, and the historical evidence for other religions is even worse. Given this, why does this argument score so highly in C tier?

First of all, I think the best argument from miracles won’t appeal to the central miracle claims made by religions. Instead, it will appeal to the miracles that are remarkably well-attested, like the levitation of Joseph of Cupertino or our lady of Zeitoun. These fit better in a theistic worldview, where a God performs miracles.

Second, while I am not moved by the resurrection evidence, I can see how a person might be. While I’m pretty confident that people shouldn’t be very moved by the Kalam, for instance, it’s much less obvious how the historical evidence should be evaluated. It might be good for quite a high Bayes factor, as e.g. Wall has argued. I’d also be worried to have a debate with someone about the argument from miracles, in a way I wouldn’t about the Kalam, so I’d feel bad giving it below C. It’s a small chip in favor of theism, but not at all decisive.

The argument from physical reality (C-)

The physical world exists. This argument claims that the presence of a physical world favors theism. While God could have created a non-physical world full of only spirits, it’s at least reasonably likely that he’d create a physical world. In contrast, on atheism, it’s likelier that there’d be nothing concrete than a physical world. I think this is, like most of the considerations C and above, a small chip in favor of theism—just one more thing that theism explains a bit better—but not at all decisive. It’s not that unlikely there’d be a physical world on atheism, and not guaranteed on theism.

Argument from laws (C+)

The argument from laws is quite similar to the argument from physical reality. It argues:

1. There are laws.

2. The fact that there are laws is likelier on theism than on naturalism.

3. Therefore, laws provide evidence for theism.

Given that for the argument from laws, we’ll have the existence of a physical universe in the background, I think this is better than the argument from the mere existence of physical reality. Given that physical reality exists, it’s pretty surprising on atheism that it’s governed by laws. It would surely be more natural and parsimonious for physical reality to simply do nothing, rather than evolve in accordance with laws. In contrast, on theism, it’s pretty expected that once God bothers to make a physical universe, he’d have it do something.

Haecceities (C+)

This is a weird one. I give it a low ranking because, while it might have immense force, it’s really hard to figure out if it works, so it can’t get too high of a ranking. The argument is original to me.

Lots of philosophers believe in haecceities. These are non-qualitative markers of identity. For example, imagine God creates two identical universes that are spatiotemporally disconnected from each other. Clearly, they’re not the same thing, because they’re different universes. But what could distinguish them from each other? They have all the same qualitative properties—the same number of people, materials, distance across, and so on.

Or to give another example, imagine that God makes two iron balls of identical size. They sit dormant, except every year, they swap locations. This seems possible! But given that nothing qualitative about the world is changing, for this to be possible, there must be irreducible haecceities. It has to be that, while a the world remains qualitatively similar, now this ball there and that ball is here.

Haecceities don’t reduce to any more fundamental properties by definition. They’re just a basic, irreducible thisness. So if haecceities are real and fundamental, then it seems that for every fundamental entity, there’s an extra irreducible property. So long as you think lots of different particles are fundamental, your ontology will be bloated by huge numbers of different fundamental entities. Theism avoids this by explaining all the fundamental properties by just one: God! At most theism has one irreducible haecceity, while naturalism has many.

Still, I’m not sure if this argument works. I have two main worries:

1. Though I can’t figure out exactly where it goes wrong, there just seems something sus about treating haecceities this way.

2. Perhaps haecceities serve to demarcate things that actually exist, but there isn’t a menu of distinct haecceities prior to creation. For instance, it might be that once a particle exists, its haecceities makes it different from other particles—so that it could be swapped with another particle—but prior to creating any particle, there aren’t multiple different haecceities that could be instantiated. Now, I think this is probably wrong for at least haecceities concerning people—it seems God could have made a qualitative duplicate of this world, without ever making me—but it weakens the argument for haecceities not involving people.

3. The argument is new, tricky, and confusing, so hard to put that much stock into it.

B

B Common consent

Here is a fact: almost everyone for almost all of human history has believed in gods of some sort or another. Nearly every view with this kind of systematic, widespread support, is true. Other examples include:

• Trees exist.

• It’s brighter during day than night.

• If you drop heavy things on your foot, that is typically painful.

• It’s better to have ice cream than be poked in the eye.

Thus, it seems there should be a presumption in favor of believing things so long as they’re widespread, in the absence of strong counterargument. This isn’t completely decisive—most people probably haven’t believed in an infinite universe—but it’s at least some evidence.

More decisive, in my view, is the coincidental convergence between what is philosophically defensible and what is actually true. Here’s an analogy: suppose you’re not sure if there is a king. However, you know if there is a king, he has all sorts of highly specific properties: maybe he’s 5’11, named Gerard, and the seventh son of Philip.

Crucially, he has the ability to provide messages to the masses in his country. You don’t know if the messages are purveyed. But when you go to the country that he might be the king of, you learn that most everyone believes there is a 5’11 King named Gerard who is the seventh son of Philip. This would be great evidence that there is a King and that the King is giving messages.

This is rather like the situation regarding theism.

Suppose you join me in thinking that there are strong philosophical arguments for the existence of a perfect God. If you think that it’s a delusion on the order of believing in Santa Clause, this argument won’t convince you, but that cannot be the standard an argument must meet to demonstrate success. Most modern people believe in such a God.

Yet strangely, these people were not convinced by the good reasons. They were mostly convinced by some combination of personal experience, testimony from others, and being brought up in a tradition from a young age. Even those convinced by argument were mostly convinced by bad, F-tier arguments like the moral argument. The historically most influential Catholic arguments, for instance, were the five ways, which do not work.

Thus, it seems like the situation is similar: on supposition that there is a being with various properties, it’s no surprise that people believe there is such a being. On the supposition there isn’t, it’s a bizarre and mysterious coincidence that they have such a belief.

Philip Swenson had an interesting reply to this argument. He argued that perhaps people believe mostly for a good reason: the world is awesome in lots of surprising respects. This is a good reason to believe and explains why many people think there’s a God.

But crucially, that won’t distinguish a limitless God from a limited one. You need philosophical argument to do that. Therefore, even granting that this is why most people believe, it’s still a coincidence that the most philosophically plausible model of God won out.

B Contingency argument

The best of the cosmological arguments (except maybe this one). The contingency argument holds that things that exist need explanations for their existence. Mathematical facts are explained by the more basic mathematical facts, or are self-explanatory, elements of the physical world are explained by other more basic elements of the physical world, and so on. It would be very weird if something was unexplained. Thus, by explanatory regress, there must be something self-explanatory (it’s not adequate to simply hold that there’s an infinite regress, because then the chain is unexplained. It would be like if an infinite line of boxcars started moving for no reason—sure, you can explain each boxcar’s movement by the one in front of it moving, but why did the chain start moving?)

Now, you might doubt that things strictly need explanations. But at the very least there ought to be a strong presumption in favor of thinking things have explanations. This presumption is the very basis of science. If you see that, say, your floor is wet, you’d never entertain the hypothesis that it is simply brutely unexplained. But so long as there is a very strong presumption in favor of holding that things are explained, a worldview on which everything is explained—brought about by some self-explanatory being—is vastly superior to alternative worldviews.

If there’s something self-explanatory that brings about everything else, there’s a decent case to be made for it being God for a few reasons. First, if we grant that everything needs an explanation, then for something to be possible, it must be able to be produced by the unexplained thing. After all, nothing else could explain it, for everything traces back to the unexplained thing. If there’s one necessary thing, and everything needs an explanation, then the explanation for everything that can be ultimately traces back to that single necessary thing.

Now, God can do anything, so if you’re a theist, this isn’t so bad. But if you deny theism, then the realm of possible world’s is bizarrely shrunken. For example, suppose that, following renowned atheist philosopher Graham Oppy, you think that the initial singularity is necessary. For something to be possible, it must follow from the initial singularity. But this is very counterintuitive! It implies that Newtonian laws are metaphysically impossible; that the world genuinely couldn’t have been Newtonian.

A second reason the necessary thing must be God is that God is the only thing that has a plausible case to being self-explanatory. Merely positing brutely that something is necessary—for no reason—is not an adequate explanation of its necessity. This seems to be the situation that those who posit non-God necessities are in, as there’s never any explanation of their necessity. One plausible view of necessity—indeed, a view I’m very confident is correct—is called modal rationalism. Modal rationalists hold that all the modal facts—facts about what is possible and necessary—can be figured out a priori. In theory, a sufficiently perceptive reasoner in a dark room with a TREMENDOUS BRAIN would be able to figure out all the modal facts.

If modal rationalism is right, then the necessary entity could be discovered a priori. There’s never been any plausible story about a non-God concrete entity whose concrete existence can be deduced a priori. Modal rationalism would rule out Oppy’s view, as the complete laws of physics could obviously not be deduced from the armchair, wholly in the absence of experimental confirmation.

Now, modal rationalism does pose problems for thinking that God is necessary. If God is necessary and modal rationalism is right, there must be some successful ontological argument. It must be that his existence is the sort of thing you could figure out from the armchair. But there are at least three ways this could be true:

1. It might be that one of the existing ontological arguments works. I haven’t explored them all, they have some smart proponents, and they just need to be sound, not persuasive. This doesn’t seem so improbable.

2. It might be that there is some unknown ontological argument, that has yet to be formulated by humans, that establishes God’s existence. Presumably the best ontological argument has never been formulated.

3. It might be that, as Aquinas famously thought, God’s essence is the sort of thing that we cannot grasp, but if we could, we could see that it must exist. God is an infinite mind, further above us than we are above ants. Just as an ant cannot see why some mathematical theorem must be true, or why truth must be, perhaps we cannot see the necessity of the divine reality.

Now, I’m a bit doubtful of the success of this argument. I find it hard to see how God could be necessary, and have considerable sympathy to Swinburne’s claim that God is contingent. But it has at the very least considerable force; if you think all of reality is ultimately explained, then in my view, it is hard to escape the pull of theism.

A

Consciousness (A-)

How strange it is to be anything at all

—In the aeroplane over the sea

Assume that you are not Dan Dennett; you are conscious. This is a very surprising fact. Somehow, when you put bits of matter together into brains, an associated private, mental inner life appears. In perhaps the most shocking event in the history of the created order, there’s something it’s like to be certain bits of arranged matter. Were (per impossible) a non-conscious observer to look on from the sidelines, they would not expect this qualitative shift, wherein properties exhaustively described by mathematics somehow give rise to an inner life.

Now, suppose that like me you are a dualist. You hold that mind does not reduce to matter, no matter how it is arranged. This is supported by an array of very powerful arguments, most convincingly laid forth in David Chalmers’s magnum opus, The Conscious Mind.

In this case, mind is utterly unlike everything else in the created order. It is a prerequisite for at the very least most of the world’s value. Additionally, if dualism is right, it is not explained by other bits of the created order, but instead is built in at the ground level. Dualists hold that there are basic laws of the form “when physical arrangements like this exist, mental states like that exist.” These are called psychophysical laws—they’re about the relationship between the material world and the mind.

To a theist, the presence of psychophysical laws is not at all surprising. We would expect the world to contain laws giving rise to the source of all—or at the very least most—of that which is good. God’s plan in creation makes no sense in a barren world devoid of any conscious beings.

In contrast, on atheism, the existence of psychophysical laws is quite a great mystery. Any time an extra fundamental law is posited, the probability of such a thing will be low. It’s quite surprising that there are laws operating at the level of human brains that give rise to the primary source of the universe’s value. Were an atheist and a theist who hadn’t yet been told that there were psychophysical laws to bet on the odds of there being such things, the theist would predict psychophysical laws with higher credence.

Now, suppose that you accept that the mind is fundamentally physical, that all that’s needed for consciousness is the right arrangement of atoms. Should you still find this argument convincing? Yes for two reasons.

First, even if consciousness is physical, and the physical statements that give rise to mental states do so necessarily, it’s still surprising that there are such things. Imagine that there was a chemical compound which gave rise to crystals spelling out made by God. Even though they’d do so necessarily, it’s surprising that there is something that does that necessarily. Consciousness is surprising on atheism, even if it’s necessary.

Second, in order to get consciousness, one needs various physical states. Only some kinds of physical states can give rise to consciousness. But theism makes it likelier that those physical states would exist than atheism does. Thus, even a physicalist should be moved by the argument from consciousness.

I think this argument is quite decisive on the best view of consciousness—non-physicalism. On physicalism, it has some force, but not as much. This is why the argument gets ranked a low A-, just above B-tier.

Nomological Harmony (A)

The stuff in the world interacts. The world does not lie in a dormant, stillborn state, never moving or changing. This is surprisingly surprising on every view of laws (see the brilliant paper of Cutter and Saad for more on this). There are three main views of laws:

First, governing accounts. On these views, laws are what give rise to the behavior of physical entities. The law of gravity, for instance, governs the universe, and this is what makes particles move in accordance with the law of gravity. If this view is true, then laws are equivalent to a set of rules that exists over and above the physical entities and tells the physical things what to do.

But on this view, it’s quite a coincidence that the laws are applicable. Suppose that the most basic physical components of our world are As. Well, insofar as our world does anything, our laws are of the form “all As X,” where X is some action. But crucially, the laws could have been different: they could have been “all Bs X,” or “all Cs X,” or “all Bs Y.”

Cutter and Saad, in their paper, give the analogy of a rulebook. Imagine that you come across a series of pieces that look like part of a game. Then you discover a rulebook that describes how the pieces are supposed to move around—for instance, you discover a chess board and the rules of chess. This would require design; after all, it’s quite unlikely that by chance there would be a harmonious pairing between the content of the rule book and the game pieces.

Similarly, it’s a big coincidence that our laws apply to the things that exist on the governing account. The overwhelming majority of conceivable laws would not be this way. The overwhelming majority of imaginable laws wouldn’t apply to the things that exist; it’s thus a major coincidence that the actual laws are applicable.

The second major account of laws is the powers based view of laws. On this view, laws represent the powers that things have. The law of gravity is really just a summary of the fact that particles have the power to attract other particles. Entities have powers—stuff that they do—and laws are summaries of their powers, not what makes them have the power.

Now, the power based view of laws avoids the earlier problem, but falls into a problem of its own. Things have powers to produce various effects under certain conditions. For instance, suppose the fundamental things are Xs. Well, in order for the Xs to do anything, they must have the power to interact with other Xs. But crucially, most of the imaginable powers Xs could have wouldn’t involve them interacting with other Xs. There could have been Xs that lie dormant, in a stillborn state, because they only interact with Ys.

Thus, on the powers view, it’s overwhelmingly implausible that the world would ever do anything. But it does: theism makes sense of that, while atheism leaves it mysterious.

The third view is called the Humean view of laws. Humean views of laws hold that laws don’t explain why entities behave as they do or summarize the powers of entities, but merely provide a general description of how things behave. When we say that there’s a law of gravity, all that means is that the observed behavior of particles is such that they attract other particles. There’s no deeper explanation of why they behave this way in terms of powers of governing laws; for a Humean, such an explanation is not needed.

Now, Cutter and Saad in the paper think that Humeanism solves the problem but argue that Humeanism is independently implausible. Why, if Humeanism is true, are there universal regularities? Why do all things behave in the same way? On Humeanism, insofar as there’s no deeper explanation of behavior, universal behavior seems like a huge coincidence.

I don’t find this very convincing. Just as the non-Humean should hold that simpler laws are inherently likelier, the Humean should hold that it’s simply intrinsically likelier that reality would behave in accordance with simple and general regularities. Explaining why we should prefer simpler laws seems to be everyone’s problem.

But Humeanism has other problems. One problem for Humeanism is that it just seems like stuff doing stuff needs an explanation. Humeanism takes the fact that things behave in regular ways as brute, but this doesn’t seem like it could be brute. Another problem for Humeanism is that insofar as laws are just regularities, then insofar as two theories both predict all the observed data, and differ only in imagined cases, there’s no fact of the matter about which is correct.

For instance, suppose there are two theories of physics that both predict all the observed data. However, they differ about what would happen if someone built a particle accelerator. Suppose that, in fact, no one builds a particle accelerator. Humeanism holds that there isn’t a fact of the matter about which theory is correct. Insofar as two theories correctly describe all regularities, there’s no fact about which is right.

However, most importantly, I don’t think Humeanism actually solves the problem. Sure, on Humeanism, laws are just summaries of how physical entities behave. But importantly, entities behave in various ways under various conditions. It’s a coincidence that the conditions under which the entities of the actual world do anything, happen to arise.

Thus, even if you’re a Humean, I don’t think there’s any way out of this argument. It thus has lots of force. So why don’t I give it above A tier? A few reasons:

1. It’s a very new argument. And while there are some new arguments like the anthropic argument that you can just see work—their internal kinematics are simple once you grasp them—this one is weird and hard to think about. It’s hard to be that confident it works.

2. It’s unclear exactly how it interacts with Humeanism. While I think it still works, it’s hard to think through, and thus it can’t have the kind of overwhelming force that the S-tier arguments have.

Thus, nomological harmony argument, I give you A-tier—for now!

Moral knowledge (A+)

The moral knowledge argument for God is misnamed. It’s not just about moral knowledge; the same points apply to mathematical knowledge, modal knowledge, and knowledge of which world is likelier than which other world.

Here’s a plausible constraint on knowledge: suppose that you believe A based on B. If there’s no connection between A and B, such that A doesn’t explain the features of B, and you know that, then you shouldn’t believe A based on B. For instance, if I think there’s a table because I see it, but I know I’m hallucinating the table, so that I’d see it even if it weren’t there, then I shouldn’t really believe there’s a table. More precisely:

If some agent infers some proposition A based on some experience B, but there’s no connection between A and B, such that a full explanation of B could be given without making any reference to the truth of A, and the agent knows this, the agent is not justified in their inference.

But now suppose that naturalism is true. The explanation of our beliefs is to be given fully in terms of the states of our brains. However, the fact that torture is wrong or simpler worlds are inherently likelier to be isn’t the sort of thing that can move around atoms. But we only believe those based on intuition. Thus, on naturalism, we have no justification for our beliefs about math, logic, or the relative likelihood of worlds, meaning we can’t even justify induction. We have no way to rule out the world having been created five minutes ago with the appearance of age.

The default naturalistic view seems to make our knowledge of morality, modality, and reasonable priors rather like if you had an intuition that some mathematical judgment was correct—but only because aliens hypnotized you to think it was correct. That would undercut your basis for trusting your judgment. But on atheism, it seems, this is what all our judgments are like.

Now, the atheist can hold that we have a non-physical faculty that allows us to grasp non-natural truths directly. This is what I held toward the end of my atheism. But crucially, the existence of such a faculty is unlikely and ad hoc on atheism. Only theism provides an account of why there would be such a thing. In addition, theists don’t even need to hold that there is such a faculty. All they need to hold is that God sets up the evolutionary process so that our beliefs about math, logic, morality, and the likelihood of different states of affairs is accurate.

I think this argument is easily A-tier. I remember when I was an atheist it being one of the major things that gnawed at me and kept me awake at night. It’s just one more element of the ideal atheist worldview that simply does not hang together.

Skepticism (A+)

A skeptical scenario is a scenario in which you’re massively deluded in all of your core beliefs. For example, if you’re a brain in a vat and your friends and family aren’t real, then you’re in a skeptical scenario. If you’re a theist, there’s a unified explanation of why none of the skeptical scenarios arise: God doesn’t want us to be massively deceived.

In contrast, if you’re a naturalist, you have no core, unified explanation of why every skeptical scenario is false. This leads to two separate but related problems. First, given that naturalism has no plausible way of ruling out skeptical scenarios, it seems if you’re a naturalist you should simply be a skeptic—about induction, and various other obvious things. If this is the cost of naturalism, it’s too high a cost. Second, even if the naturalist can adopt some epistemological contortion to explain why they shouldn’t think they’re in a skeptical scenario, theism provides a better explanation of why all the skeptical scenarios are false. In other words, even if you know that you’re not in a skeptical scenario, theism best explains why you’re not in a skeptical scenario.

Now, of course, for this argument to succeed it must be that the skeptical scenarios are likely conditional on naturalism. But, in fact, they are. In fact, on naturalism, I think there are three separate skeptical scenarios each of which are pretty likely.

First, you might be a Boltzmann brain. A Boltzmann brain is a brain that briefly fizzes into existence in the recesses of outer space—with beliefs, desires, and so on. Boltzmann brains are generally short lived, as if you materialize into being in outer space, your prospects are unlikely to be very good. But generally, it seems the default theory of physics produces the result that you’re probably a Boltzmann brain. This isn’t surprising—it’s easier to get a small patch of order than a whole universe worth of order. Non-finely-tuned universes can produce Boltzmann brains, but they cannot produce a life-permitting universe.

Additionally, default physics—without addendums—seems to result in entropy increasing over time. A high entropy universe is too chaotic to produce complex stable structures, but it can produce a randomly fluctuating Boltzmann brain. Given that the high entropy state lasts forever, it seems on naturalism, you should expect to be a Boltzmann brain.

Second, every theory of anthropics seems to imply skepticism. There are three main theory of anthropics—and indeed, as I argue in a paper I’m currently working on, every view of anthropics will have the problems of one of the existing ones—called the self-indication assumption, the self-sampling assumption, and compartmentalized conditionalization. On each one of them, if atheism is true, skepticism would seem to follow.

The least popular theory of anthropics is called compartmentalized conditionalization (CC) (boo!) (you should be so drunk on Purim that you cannot tell SIA from CC) (only real Jews will get that joke). CC holds that upon having some experience, you should favor theories proportional to the probability that someone would have that experience if the theory were true. For instance, suppose that your experience in life consists of every day waking up and going to the bagel shop. Well, CC holds that theories are likelier proportional to the odds that if they were true someone would have the experience of waking up and going to the bagel shop.

CC implies skepticism for a slightly tricky reason that’s hard to wrap your mind around. Suppose I google “laws of physics,” and find that this is what the standard model is:

The prior probability that this would be right description of the laws of physics is very low—it’s quite specific and random. In contrast, suppose that the world is very large and has a bunch of Boltzmann brains. Well, in that world, it’s guaranteed that someone would have my experience of believing there to be a past and believing the above to be the right view of physics. Thus, the theory that this is the real law of physics starts out with a super low prior, and it gets no update to overcome the low prior. My credence in that being the right description of physics should, therefore, be very low. The same point applies to any specific fact about math.

Note: this is quite specific to CC. On SIA or SSA, a theory that predicts a larger share of people having some experience is thereby rendered more likely. Thus, if one theory predicts a large share of observers with beliefs about physics thinking that the above is the standard model, it gets a massive probabilistic boost after you update on the fact that you think the above is the standard model. Only CC, neglecting the number of people having an experience, only gets a minimal update from your observing specific facts, and thus implies skepticism. On CC, there’s nothing to overcome the very low prior in specific facts.

Here’s another way to think of it: any time there’s a highly specific fact that you learn, it will have a low prior. For example, suppose that someone comes up to you and tells you “my name is Thomas Philipson,” or suppose someone gives you their phone number as 14149321135. The prior probability of that being their phone number or name is very low—on the order of 1/10^9. But nonetheless, you should think that is their phone number.

The reason is that if that’s their phone number, it’s highly likely that that’s the number they’d tell you. So on the supposition that their phone number is 14149321135, it’s highly probable that they’d tell you their phone number is 14149321135. In contrast, if they are lying about their phone number, they could lie about any sequence of numbers—thus, on the supposition that they’re lying about their phone number, the odds they’d tell you that sequence of digits is on the order of 1/10^9. The probabilities cancel out, and thus it makes sense to believe them (see also Aron Wall’s discussion of the prosecutor’s fallacy and the LessWrong post Strong Evidence Is Common.

But now suppose that you adopt compartmentalized conditionalization. When you learn some specific fact about physics, the prior of that being the true theory of physics is very low. Thus, in order to overcome the low prior, you need very strong evidence. We normally have this; while the prior in the standard model being what I pictured above, if that’s not the standard model, then the standard model might be anything. Thus, while it’s super unlikely the above thing would be the standard model, it’s also super unlikely that I’d mistakenly believe the above thing was the standard model—as many other things as could be the standard model are things I could be mistaken about as the standard model.

But on CC, you don’t have very strong evidence. All you need to explain the relevant data, on CC, is that someone has the experience of thinking that’s the standard model (along with the rest of your life experiences). But in a big universe, it’s almost guaranteed that someone would have that experience. Therefore, the odds of someone having the experience are whatever the odds of a big universe—maybe 10%. The big universe theory, on CC, explains the relevant data—someone having your experience. Therefore, nothing on CC overcomes the low prior, and total skepticism reigns!

The second major view of anthropics is the self-sampling assumption. SSA holds that you should treat your present existence as randomly selected from among those in your reference class. Your reference class is the set of people that are enough like you that it is in some way surprising, according to SSAers, that you’re not them. For instance, suppose a coin gets flipped: if it comes up heads, one person gets created with a red shirt, and if it comes up tails, one person gets created with a red shirt and one person with a blue shirt. SSA holds that after being created with a red shirt, you should think heads is twice as likely as tails: on tails you might have been either person, while heads guarantees that you’d be the red-shirted person.

So to recap, SSA favors theories according to which a larger percentage of people in general are those you might presently be. Heads is favored in the above case, because on heads everyone is someone you might be, while on tails, only half of people are those you might be.

Now consider two theories:

1. The world is much as we think it is. There are billions of currently existing people, and about 110 billion people who have ever lived.

2. There’s just a single brain in a vat—and it’s you!

The second theory predicts that everyone in the world is someone you might be. In contrast, the first theory predicts that only 1/110 billion people are those you might be. Thus, on SSA, you get a 110 billion to one update in favor of the theory that you’re a lone brain in a vat. Few views survive a 110 billion to one update against them. Thus, SSA makes skepticism almost guaranteed.

The last view is called the self-indication assumption. On SIA, you should think that theories are more likely if they predict more people that you might presently be. For instance, suppose a coin is flipped: if it comes up heads one person gets created, while if it comes up tails, ten people get created. Suppose that you get created from this process. SIA holds that tails is ten times likelier than heads, because there are ten people that you might be rather than just one.

If SIA is right, you should think a giant infinity worth of people exists. Whatever the most people is that there could be, that’s how many people you should think there are. Now, if you adopt theism, this isn’t so bad—God creates every possible person and places them in a flourishing world. But if you reject theism, then this probably implies skepticism.

There are two main naturalistic theories that predict a number of people existing that is the maximum number possible. The first is modal realism, which says that every possible world exists. But because there are infinite worlds where you’re deceived for every world where you’re not, on modal realism, you should think you’re deceived (well, actually probably you should have no credences on modal realism, but that’s worse!)

The second is a view endorsed by Max Tegmark which says that every mathematical structure is instantiated. Thus, everything that can be mathematically modeled is its own separate world. This view also implies skepticism, because there are the same cardinality of worlds where people are deceived as of worlds where people have veridical perception (I had a long back-and-forth with Scott Alexander about that, if you’re interested).

Now you might worry: if God makes infinite worlds, won’t there still be infinite massively deceived people? So then shouldn’t this undermine induction? No, because for every particular person, they’re placed in a world optimal for their flourishing, which is unlikely to be a counterinductive world.

Here’s an analogy: suppose you’re in Hilbert’s hotel (that’s an infinitely big hotel). There are infinite copies of you in the hotel. You roll a six-sided die. What’s your credence in the die coming up 1-5? The answer is, of course, 5/6. But note: the cardinality of the people who get 1-5 is the same as the cardinality of people who get 6. In fact, by moving people around, you could make it so that every room has a hundred people who get 6 and only one who gets 1-5—or the opposite. This is a weird property of infinity—if you have two infinites of the same cardinality, you can match them up any which way. You can match ten members of one set to each member of the other set or do the opposite.

On theism, the situation is rather like that in Hilbert’s hotel when you roll dice. Every particular person is in the scenario ideal for their flourishing: that probably doesn’t involve being deceived. However, on non-theistic views, there’s no analogous process: there are simply infinite deceived and non-deceived people, and the infinites are the same cardinality. Thus, non-theistic views consistent with SIA probably result in skepticism.

A third way that atheism probably implies skepticism is that if the universe is infinite in size, which seems to be favored by our best scientific theories, then skepticism follows, on non-theistic views, for the same reason it does if SIA is true. There are infinite people, both deceived and non-deceived. I won’t go into much detail about this as I’ve laid it out here and this article is already quite literally over 9,000 words.

S-tier

Psychophysical harmony (S)

This is a hard one to get your head around. However, once you do, you come to see that it’s absolutely devastating. I’ll probably have an article laying this argument out in more detail sometime soon. There are plenty of helpful resources for getting the argument. There’s a law of nature that everyone who finds the argument persuasive gets so irritated by people misunderstanding it that they write a brief explainer on it. My friend Amos has a several-part series on it; my other friend Apologetics Squared (not the name his parents gave him) has an incredible series on YouTube of 38 very short videos, roughly 1 minute in length, explaining and defending the argument, as well as a pretty short and comprehensive video on it; Dustin Crummett and Brian Cutter have a paper on the argument (I’m starting to think I’m friends with everyone in the world who likes the argument); and I have an article on it.

In short, the argument proceeds by noting that there’s a kind of harmony between the mental and the physical. Your internal map of reality roughly matches reality as it is—when you see a four-foot table in front of you, there really is a four-foot table. Similarly, there’s a kind of harmony between our desires and the actions we take. When you want to move your limbs, your limbs move.

Now, while this seems ordinary enough, in reality, it is very surprising. Think about all the ways the mental and the physical could pair. Most of them wouldn’t produce any interesting harmony—most of them would produce random mental static, rather than anything interesting. It would be much simpler for every brain to have the experience of seeing a red wall, or a yellow one, or a green one, than for them to have the complex experiences that they actually have.

Most ways the mental and the physical could pair wouldn’t produce anything interesting. No desires, no coherent mental life, just chaos. Even if you do have a mental life complicated enough to be called an agent—vanishingly unlikely on atheism—it’s unlikely your desires would pair with the world. It might be that, for instance, your desires produce no physical effects. In such a world, we’d be like a person with locked-in syndrome—able to want things but never able to act.

Now, the most common reply is that evolution solves the problem. It really, really doesn’t, and this is something that everyone who understands the argument agrees on. Evolution cares about how you behave, not whether you’re psychophysically harmonious. Insofar as we got another set of psychophysical laws that caused the same behavior as the actual laws, then we’d behave the same way, but be radically disharmonious.

In short, because evolution only cares about how we behave, and psychophysical harmony is about the pairing between our behavior and mental life, rather than how we behave, it doesn’t guarantee harmony. In fact, the simplest psychophysical laws preclude harmony—if every experience was just of seeing a red wall, then there couldn’t be a rich mental life. Even if the psychophysical laws are complicated enough to give rise to agents, given that you could have disharmonious agents act any which way, evolution’s fixing of behavior doesn’t guarantee harmony. Evolution guarantees creatures will act to survive, but the whole puzzle is why a being’s conscious states match up with its actions to survive.

I can’t hope to go into depth on every objection—or even many of them—as this article is already 10,000-words, and I doubt anyone wants a 20,000-word article. But let me just head off one potential worry. You might think that physicalism avoids the problem. On physicalism, after all, the pairing between the mental and the physical is necessary—how can it be surprising?

Dustin and Brian address this in the paper, as do various other people in various other places. In short, psychophysical harmony can be surprising even if it’s stipulated to be necessary. If a person gets ten royal flushes in a row in poker, even if they’re a necessitarian—holding everything is necessary—you should still suspect that they’re cheating. Merely positing that a fact is necessary isn’t an explanation; thus, theism has a better explanation of psychophysical harmony.

Analogy: suppose that every conscious creature had the appearance of the words “God loves you,” constantly in its visual field. This would be strong evidence for the existence of God. Sure, physicalists could posit that this fact is necessary, but it’s so unlikely that would be necessary rather than something else. Psychophysical harmony is the same.

When I became a theist, this was one of the main arguments that convinced me.

Fine-tuning (S)

The fine-tuning argument is widely seen as the best argument for God. But crucially, there are actually three kinds of fine-tuning, and two of them dodge most of the standard objections (for more on the argument, see my very long post about it):

1. A priori fine-tuning: this kind isn’t about the specific laws. Instead, it’s about the more general point that most ways the world could be wouldn’t produce anything interesting. If the laws are very simple, then probably they’d just result in a basic pattern—too basic to produce anything. For instance, the ultimate laws could have just involved particles aimlessly bouncing around, or moving in a circle, or disappearing after a second, or moving in a line. If the laws aren’t simple, then they’d be almost guaranteed to produce random chaos. This kind of fine-tuning is probably the most convincing, and isn’t threatened by findings from physics.

2. Anthropic fine-tuning from physics: this kind proceeds from the striking observations that the constants of physics—the values that are plugged into the laws—fall into an incredibly narrow range needed to give rise to life. For example, if the cosmological constant weren’t in a tiny range, on the order of one part in 10^120 of its possible values, no life or complex structures of any sort would arise.

3. Fine-tuning for discoverability: this builds on work mostly from Robin Collins. What Collins argues is that some of the constants are precisely set in a way ideal for scientific discovery. For instance, he claims that the masses of many of the particles in particle physics happen to be an ideal quantity for us to measure them. This is expected if God set the constants in ways ideal for us to do science, but unexpected if they took their values by chance.

Taking into account all three kinds of fine-tuning, this argument is utterly devastating. Maybe an atheist can explain the second kind by a multiverse—though, as I’ve explored, a multiverse has various problems—but certainly they can’t use a multiverse to explain the first and third kind.

The problem for the atheist is that the improbability is so vast—it’s so wildly unlikely that we’d get a fine-tuned universe by chance—that they must, for their view to be plausible, have some explanation of fine-tuning. But the explanations of the first kind of fine-tuning just transfer the fine-tuning back a level—if the atheist invokes a multiverse, for instance, the multiverse itself is just a physical system that generates universes. But to generate the right kinds of universes—and universes at all, rather than producing nothing—it needs fine-tuning. Thus, even the multiverse just transfers things back a level.

While there are some mild ways to criticize the likelihood of fine-tuning on theism, the atheistic replies are not enough to overcome the vanishingly low probability of fine-tuning on theism. Atheism predicts a barren wasteland—the fact that isn’t what we observe strongly undermines it.

The anthropic argument (S❤️❤️❤️)

(The hearts aren’t a higher rank, they just represent how I feel about it).

I’ve defended the anthropic argument in various places—if you want to read about it, I’d recommend reading this, this, this, this, and this, but especially the first one. In short, the argument runs as follows:

1. You exist.

2. You’re likelier to exist if there are more total people that exist. Suppose that a coin gets flipped which creates one person if heads and ten people if tails. You should, after being created by the coinflip, think tails is ten times likelier than heads.

3. If 10 people existing makes your existence ten times likelier and 100 people existing makes your existence 100 times likelier, infinity people existing makes your existence infinitely likelier.

4. Thus, you should think there are infinite people. This doesn’t stop at the smallest infinity—you should think the number of people that exist is the most that there could be.

5. That’s a really huge number. Theism can nicely explain why that number of people exists, but atheism has no comparable explanation. In fact, because it’s good to create, theism actively predicts that number of people existing, while atheism does not.

The main controversial premise is the second one. But the second one is, in my view, as well attested as anything in philosophy—having far more to be said in its favor even than other very plausible views. In fact, I’ve compiled a list of 27 arguments for SIA, and many of them are good.

If SIA is right, then it’s trivial that you should think a huge infinity worth of people exists. Theism naturally predicts this, from the goodness of God’s creation, while atheist’s have no plausible explanation. In fact, as I discussed in the section on the argument from skepticism, I think any naturalistic explanation will inevitably imply skepticism—that you shouldn’t be confident that the sun will rise tomorrow.

I find this argument compelling because I do not know what a naturalist can plausibly say about it. Denying the self-indication assumption is, in my view, off the table—it just has too much in its favor. But I do not know of any remotely tenable naturalistic view that predicts the existence of enough people to be plausible without undermining induction. While for the other arguments, you can posit that we got lucky, with the anthropic argument, I don’t even know if it’s possible to do that. I do not think there is a story of what ultimate reality works that doesn’t undermine induction but that explains the anthropic data. I do not know what I’d say about it if I were still an atheist.

Conclusion

Hopefully you’ve enjoyed my ranking of the various arguments. Sorry that this article was so long—there were lots of arguments, so it was hard to be concise. Do you think I got anything wrong? Let me know!