Against The Kalam Cosmological Argument For God

A deep dive

Introduction

One of the most popular arguments for the existence of God is the Kalam cosmological argument. It says (roughly):

1. Everything which begins to exist has a cause.

2. The universe began to exist.

3. Therefore the universe had a cause.

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

5. Therefore, God exists.

William Lane Craig—one of the most influential Christian apologists—is responsible for resurrecting this argument, and it’s become a staple in the apologists’ argumentative quiver. Most apologists endorse some form of the argument and are prepared to defend it. Despite this, I do not think it is a very good argument. I think it lacks any significant persuasive power, and should be abandoned. While it’s possible that every step is right, there are enough steps in the argument that are quite non-obvious, that it doesn’t give an atheist substantial reason to abandon their atheism.

In this article, I thought I’d explain why I think these things. I’ll go through each premise and argue that they range from plausible to not obvious to probably false. An argument built on a tower of controversial assumptions without substantial evidence in their favor is largely ineffective.

I say this not as an opponent of theism. I think that God exists, and that this is rendered likely by many powerful arguments including the ones here, here, here, here, here, here, and here. But even though I broadly agree with the conclusion of Kalam proponents, one should criticize bad arguments, even in support of true conclusions.

1 “Everything which begins to exist has a cause”

The first premise of the Kalam states “everything which begins to exist has a cause.” Generally two things are said in favor of this premise:

1. The premise is intuitive. It just seems like a thing can’t start existing uncaused. As Craig is fond of saying, it seems like something can’t simply pop into being wholly uncaused, out of nothing.

2. This best explains that in our experience things don’t begin existing uncaused. If universes could pop into being wholly uncaused, why couldn’t, say, an elephant simply appear in my room?

I don’t find either of these convincing. As for the first, it only seems intuitive when phrased in a misleading way. Yes, it’s weird for things to just pop into being at some time. But the universe’s beginning isn’t like that. It’s not that the universe, on the supposition that it lacks a cause, popped into being out of nothing. Rather, there simply wasn’t anything before the universe.

The distinction between nothing and not anything sounds innocuous but is crucially important. If there was ever a state of nothing, then it’s plausible that it couldn’t cause anything. But those who reject premise one don’t think that at some point there was nothing—they simply deny that there is such a thing as before the universe. There isn’t anything hotter than the hottest object, but nothing—pure emptiness—isn’t hotter than the hottest object.

When one is clear about this, not only do I find this justification not obvious, I think it’s actively wrong. It seems obviously possible for the universe to simply be a finite age and have not been preceded by anything. Note: we’re not talking about it being preceded by a state of nothingness, but about there not being anything preceding it, just as we might say no one is older than a very old man, not in the sense that the absence of people is older than the man, but in the sense that there isn’t anyone older than the man.

In fact, it even seems possible—to me at least—that something begins in time without a cause. It seems possible that, say, a giraffe would simply appear uncaused. Sure, it’s very unlikely, but nothing about it seems absurd or contradictory. I can imagine it distinctly—it strikes me as clearly ideally conceivable.

Thus, I think this first justification backfires. Not only do I think intuition can’t rule out things existing without cause, intuition makes clear that things can exist without a cause.

What about the second justification, that if things can exist wholly without a cause then this would lead to total skepticism, because bicycles and root beer might pop into being uncaused? I don’t think this justification works either:

1. This alone doesn’t seem enough to explain why there isn’t total skepticism. If, after all, you think that things that begin to exist need causes, but things that are eternal don’t need causes, then you get a similar puzzle: why not think that there’s an infinitely old unicorn that will teleport into my room one second from now? This scenario can’t be ruled out by the causal principle, because the unicorn didn’t begin to exist. Similarly, why not think that there’s an atemporal ghost that will haunt your room one second from now, or simply a series of distinct atemporal ghosts.

2. The solution to these puzzles, in my view, isn’t a causal principle that rules them out, but a preference for simplicity. It’s unlikely that a unicorn will pop into your room, because that would require a new unexplained thing utterly unlike all the other fundamental things. The probability of a new fundamental thing is inevitably quite low, because it drastically raises the number and kinds of fundamental things. This solution can explain why a unicorn existing uncaused is metaphysically possible, but can maintain the intuition that it’s unlikely—in fact, it’s unlikely for exactly the reason that a frozen timeless unicorn is unlikely.

3. Worst case scenario, you could adopt another version of the causal principle. For instance, you would hold that things that begin within spacetime need a cause. While I don’t find this plausible, it strikes me as roughly as plausible as the Kalam’s causal principle.

Some have proposed the principle of Principle of Material Causality (PMC), according to which “All concrete objects that have an originating or sustaining efficient cause have an originating or sustaining material cause, respectively.” In short, this means that for some material thing to be caused, it must be caused by existing material things. Or, as Leon says it, “In simple terms, PMC says that all made things are made from other things.” This would explain why unicorns don’t pop into being out of nowhere—to be made, they’d have to be made from existing material.

I don’t think this response succeeds. It just seems false that things are necessarily made from preexisting material. In Harry Potter, for instance, when a spell is cast that created a new object, this seems metaphysically possible. Now, if the principle isn’t modal, but is just a generalization about the actual world, then we would only have reason to accept the principle if we don’t think God exists, in which case it’s toothless as an argument against God. It would be like arguing against God by adopting:

PNG (principle of non-God): things that are caused are not caused by God.

Obviously, this principle is only worth accepting if you already disbelieve in God. But one could support it with a similar inductive argument—there are loads of things not caused by God and none uncontroversially caused by God. I regard the PMC to be the same, so long as we think it isn’t necessarily true.

People also support the first premise of the Kalam by saying “out of nothing, nothing comes.” But again, those who reject this premise can agree with this claim. Nothing comes out of nothing. There simply never was nothing.

Now, you might not share my intuitions here. Perhaps—and this is, I suspect, the main reason that people adopt this principle—you just find the notion absurd that something could begin existing uncaused. But it’s hard to be super confident in this premise, if it just comes down to conflicting intuitions, particularly if there are no arguments that can be given in defense of the intuition, and plausible debunking accounts on the table (from the fact that nothing in our experience ever does start existing uncaused, we overgeneralize and assume that nothing can).

Given that the argument has several steps, so long as we can show that none of the steps are supremely obvious, we’ll have successfully defanged the arguments—even if each of the three contentious steps has a 2/3 chance of being right, the odds they’re all right is just above 30%. And an argument isn’t very successful if it has only a 30% chance of succeeding. Even if you’re 80% sure of each premise, the odds all three are right is just above 50%.

Now, lastly you might argue that there should be at least a presumption in favor of thinking that things that begin to exist have causes. Perhaps it’s possible for a thing to exist without a cause, but most likely that if something exists it has a cause. But I think this same presumption applies generally—uncaused things are unlikely, no matter what they are (with the exception of abstract objects, but that’s a story for another day). If this is right, then positing an uncaused God is no better than positing an uncaused universe that began.

For these reasons, I think the first premise fails. By having a preference for simplicity—which we need to avoid other skeptical scenarios—we can avoid the intolerable conclusion that elephants are likely to appear in your room, and explain why in our experience nothing appears uncaused. Contrary to its proponents, the premise is not only not intuitive but actively unintuitive, and supported by nothing.

(By the way, if you want a much deeper analysis of this premise, from one of the best philosophers I know, check out Joe Schmid’s deep dive into the subject. Dude’s a beast.)

2 “The universe began to exist”

2.1 B theory

The second premise says that the universe began to exist. I also find this premise dubious—it requires rejecting two separate controversial theses.

The first view this requires rejecting is known as the B-theory of time. The B-theory denies that there’s an objective present—it says that the universe eternally exists as a static block. Just as there are different points in space, but no objective privileged “here,” there are different points in time, but no objective privileged now. Each time is equally real, and there is no objective present.

Put precisely, the B-theory says that every time exists equally. In contrast, the A-theory says that only now exists—the past used to exist and the future will exist, but only now currently exists. A-theory holds that there is no time like the present—in that it’s real. It’s most helpful to think about the B-theory with the analogy to space: just as all points in space are equally real, all points in time are equally real.

If the B-theory is true then the universe didn’t begin to exist. There was no first moment when it started existing. Instead, it exists “eternally” as a 4-D block. Time exists within the universe, and is relative to an observer at a moment, but the universe itself doesn’t have a first moment of existence, just as there’s no space outside the universe that the universe dwells in.

If the B-theory is true, there’s nothing wrong with the universe simply existing. No popping into being ever occurred—there simply is a 4-dimensional block that contains events at different times. On the B-theory, the first moment of time isn’t special—it’s like the right-most point in space of an object—and thus not needing a special cause.

Some people argue for the Kalam consistent with B-theory. Notably Craig, it’s most famous proponent, thinks that the A-theory needs to be right for the Kalam to be true [Edit: he seems to have changed his mind about this]. However, others, like I believe Pruss, think that the Kalam works even if the B-theory is right. A proponent of the B-theory Kalam could say that on B-theory, the universe still began to exist in some sense. But I’m doubtful the argument can get off the ground. One such attempt to reconcile this has involved defining “beginning to exist as,”

1. x begins to exist at t if (i) x exists at t, and (ii) there is no moment earlier-than t at which x exists.

But this would imply that God needs a cause. After all, at the first moment of creation, God exists and there isn’t an earlier moment in which he existed. Now, you could get around this by saying:

x begins to exist at t if (i) x exists at t, (ii) there is no moment earlier-than t at which x exists and (iii) x didn’t exist timelessly.

In this sense, the block universe at the first moment of creation began to exist, because it did not exist timelessly, existed at the first moment of creation, and there’s no earlier moment in which it existed. But then the argument loses its force.

First of all, once we accept B-theory, there’s nothing strange about the universe beginning to exist. While a thing popping into being—starting at some time in the normal sense—might be suspicious, a thing starting to exist in the above highly technical sense doesn’t seem at all suspicious. It’s no stranger for the universe to have a first moment in time than a rightmost edge in space. Thus, if this is how causation is defined, we can simply reject the causal principle.

Second, if we accept B-theory, we can give an alternative more plausible causal principle. We can say that everything which begins to exist has a cause—where “beginning to exist,” is defined as “having not existed at any earlier time.” This can explain why Beethoven doesn’t pop into being, but doesn’t rule out the timeless block.

Third, this causal principle is unmotivated. While it might be intuitive that a thing can’t pop into being at some moment uncaused, it’s not intuitive that this timeless block can’t be uncaused. The only reason to adopt it is abductive—that it explains why things don’t pop into being—but (1) this doesn’t explain why it’s better than other causal principles and (2) this doesn’t give a reason to favor the causal principle over a mere preference for simplicity.

The question is, then, is the B-theory true. There’s a lot to like about it (most of this comes from Ted Siders’ excellent book on the subject):

1. Sometimes—and intuitions conflict about this—there seems to be something incoherent about the notion of now. It’s often hard to place one’s finger on what’s supposed to be incoherent—Richard has an attempt here, and Siders discusses it more here—but the notion of a present, such that only the current notion exists, sometimes seems incoherent.

2. Physics seems to suggest that time passes at different speeds for different observers. This gives reason to think that there’s no one objective present. Now, there might be ways around this, but they struggle.

3. One very attractive view holds that there are temporal parts. This allows us to resolve certain semantic paradoxes. For instance, the ship of Theseus paradox can be resolved by holding that there are two distinct temporal “worms,” one of which contains the old ship and new ship, the other of which just contains the old ship. But something can’t have a non-existent part. For this to be viable, therefore, it must be that things have temporal parts.

4. B-theory is more parsimonious. Rather than positing that the present is special and privileged, it posits each time as equally real.

5. One very plausible metaphysical view holds that truth supervenes on being. In plain English, what this means is that the things that are true are a function solely of the things that exist. Nonexistent entities don’t exist, so they can’t affect what’s true. But if this is right, then because presentism implies that only the present exist, there can’t be truths about the past.

There are many more arguments for B-theory, but this is not an area that I’m very informed about, so I won’t try to discuss them all. They are enough, however, to make it so that an argument that depends crucially on rejecting the B-theory is not dialectically effective. An argument can’t be very successful if it requires holding multiple distinct views that are very controversial.

However, there are some things to be said in favor of A-theory. I once again can’t discuss them all, so let me list just two—the first is the most common argument, and the second is one that is unique to me.

First, the B-theory is counterintuitive, conflicting with our ordinary judgments about the kinds of things that we are. It sure seems like now is all that currently exists. Surely there’s some sense in which the present is more real than the past. Non-philosophers almost always assume that the A-theory is right.

Second, infinite ethics seems to show that if there are two collections of infinite people, each containing infinite well-off and infinite poorly off people, and neither contains the same people, then neither is better than the other. For instance, while we might ordinarily assume that the world would be better if every galaxy had a hundred happy people and two miserable people than if it contained a hundred miserable people and two happy people, this can’t be right, because the first world could be transformed into the second world just by moving people around.

But now consider two states of affairs. In state of affairs one, the person has ten days of being happy and then one of being unhappy, before having another ten days of happiness and a day of happiness, repeated forever. In state of affairs two, the person has ten days of being unhappy, then one day of happiness, then ten days of happiness, then one day of unhappiness, and so on, repeated forever.

If, as B-theorists claim, time and space are analogous, then each of these days is equally real. If each of these days were to be equally real, then it would seem that the states of affairs were equally good, just like they’d be equally good if they were simply being experienced at different points in space. But intuitively, it seems obvious that they’re not equally good! The A-theorist has a better explanation: they can say the world is good because the days aren’t all actual, and at any given time, there are more happy actual days than sad days in state of affairs one. Now, this is still a bit murky for the A-theorist, but probably they have a better account than the B-theorist does.

Ultimately I lean towards B-theory, but am largely uncertain about which view is better. Thus, given that the Kalam, I’ve argued, can only succeed if B-theory is false, it’s on quite shaky ground.

2.2 The universe is finitely old

Even if the B-theory is false, the Kalam might fail if the universe is infinitely old. After all, the second premise says that the universe began to exist, and this will generally be false if the universe is infinitely old. Generally, four things are said in defense of this premise.

First, people say that scientific evidence provides a strong case for the beginning of the universe. I don’t know much about the scientific evidence, so I’ll just point you in the direction of two people I respect discussing the topic. The second link is especially good for a complete explanation of the topic. My sense is that physics imposes constraints on models according to which the past is infinite, but doesn’t come close to ruling them out. But once again, I don’t know what I’m talking about and you shouldn’t listen to me.

Second, people say that you can’t have actual infinites. Craig argues that while you can have things that approach infinity as a limit, you can’t have anything that’s actually infinite. I think his argument here is quite lame—first of all, there aren’t compelling reasons to rule out an actual infinite. While the infinite has some counterintuitive implications—e.g. that a thing can be placed into 1:1 correspondence with some subset of itself—the mere existence of an infinite doesn’t entail a contradiction. While there are some paradoxes that arise involving infinites, those can be avoided in other ways (we’ll talk about that later).

For example, Craig argues that if there was an actual infinite, you could construct an infinite hotel in which every room is full. But then you get the strange result that you could open up new rooms simply by having people move a room over. If there are people A, B, C, D…through infinity occupying rooms 1, 2, 3…through infinity, and every room is full, you could keep everyone housed and open up a new room by having A move to 2, B move to 3, and so on. But surely moving people over, Craig claims, couldn’t generate new rooms.

But while this is a bit strange, it’s not paradoxical. There’s no contradiction that it entails. Wouldn’t we expect the infinite to be strange? If you want a more comprehensive reply, see here. There are some genuinely trippy infinite paradoxes, but many of those still arise for Craig’s view. In addition, there are lots of things that we have strong evidence are infinite, including:

1. The number of actual people (this is certain if SIA is true, which it is).

2. The number of future events (Craig always responds to this objection by changing the subject. He notes that the number of events that will have been past is a potential infinite that approaches infinity, but that’s different from the number of future events).

3. The number of possible worlds.

4. The number of things God knows (he knows that he exists, he knows that he knows that he exists, etc).

5. The number of apples he could create (there’s no finite number N such that God could only make N apples and no more).

6. The number of numbers.

7. Space, as is attested to by our best theory of physics.

8. The divisibility of space (it’s true that the size of the units in which space can be divided into are a potential infinite, but the number of times space could be divided is an actual infinite).

Craig’s view is quite widely rejected among philosophers and mathematicians generally—and for good reason. As Peter Van Inwagen—sassy as always—says of Craig’s arguments “those arguments are really awful…it’s just depressing.”

The third argument for the beginning of the universe being some finite time ago is that just as you can never count to infinity, the past can’t be infinite because it would never have gotten to the present. It’s called the successive addition argument. The standard reply is that while you can’t get from a finite number to an infinite by successive addition—counting one at a time—since there was no beginning, every moment in the past the universe was infinitely old. Thus, you never went from a finite to an infinite, but instead from an infinite to an infinite.

I think this is right if B-theory is true. If B-theory is true, just like there’s nothing wrong with space being infinite, there’d also be nothing wrong with time being infinite.

If A-theory is true, however, I find there to be something fishy about this reply, but it’s hard to articulate why. In order to get to anything that’s infinite years old, you’d have to have each of the previous years pass—but they could never pass! It seems just as impossible to reach a past infinite without a beginning as to get to a past infinite with a beginning. I confess I don’t have much to say that would convince someone who doesn’t share this intuition, beyond simply stomping my foot and repeating it. So that’s what I will do! My sense is most other smart people do not share this judgment.

The last argument for why the universe can’t be old appeals to causal finitism. Causal finitism says that infinite causal chains are impossible—more precisely, that you can’t have one effect determined by infinitely many causes. Alexander Pruss, for instance, defends it in his excellent book Infinity, Causation, and Paradox. It’s often defended with paradoxes like the following:

1. The grim reaper paradox: there are an infinite number of reapers, one set to kill you at 12:30, one at 12:15, one at 12:7.5, etc. Each will only kill you if you’re still alive by the time that you reach them. This scenario leads to paradox: each reaper will only kill you if you’re not killed by an earlier reaper, but no reaper can kill you because there’s always an earlier reaper. But if no reaper earlier than any particular reaper kills you, then that particular reaper would kill you. So you must be killed by a reaper, but not particular reaper can kill you.

2. The grim messenger paradox: there are an infinite succession of messengers, each with their own special symbol. Each writes down their symbol on a slip of paper, but only if the paper doesn’t have a message from an earlier messenger. Here, the same paradox arises: no particular messenger can write down their number, but if none of them do, then inevitably every particular one would. Contradiction: a message must be written down, but no message can be written down!

Causal finitism has been extensively criticized. Many have argued for what’s known as the unsatisfiable pairs diagnosis. They suggest that just as red cars are possible and blue cars are possible, but some car can’t be both red and blue, infinite causal chains are possible, and so are grim reapers, but you can’t combine them. On this view, each individual one is possible, but they can’t be combined.

This is a monstrously complicated topic. I very weakly lean towards causal finitism, but I’m very uncertain about it and have no idea what to say about certain objections, so it doesn’t seem very worth opining on it publicly. Instead, let me just recommend Pruss’s book on the subject, this excellent paper, and this very good debate between two experts (I’ve been watching and reading a lot of Schmid’s stuff on the subject, hence the high number of Schmid recommendations).

All in all, I’d guess that probably the universe began some finite number of years ago. For there to be a past infinite, it must be that infinite causal chains are possible, aren’t ruled out by mere addition, and that actual infinites are possible. I think all of these are likely right, but enough together, when combined with the evidence from physics, to make this premise somewhat plausible. Thus, I think of the four controversial steps in the Kalam—those being 1) everything which begins to exist had a cause, 2) B-theory is false, 3) the universe is at most finitely old, and 4) if the universe had a cause, that cause is God, I think that 3) is probably the most defensible. But even then, I’d hesitate to give it much more than 60% odds—and if all four steps had 60% odds, the odds they’re all right are only 12.96%.

3 “If the universe had a cause, it was God.”

This last premise states “if the universe had a cause, that cause was God.” Generally, three things are said in favor of this premise.

First, people argue that if something caused the universe—all of time, space, and matter—it must be timeless, spaceless, immaterial, outside of the universe, and very powerful. But this sounds a lot like God. However, this relies crucially on equivocation.

When people say the universe had a cause, they might mean “the collection all three: time, space, and matter.” In this case, the universe might be caused by something material—so long as the thing is material but not temporal. This is proposed by some physicists—that some basic physical thing gives rise to the physical and temporal manifold that we know about.

People might instead mean “the collection containing at least one of the following: time, space, and matter.” In this case, while this premise is more defensible, the second premise is doubtful. Insofar as time and our local universe emerge from some simple kind of matter, that kind of matter—which counts as the universe—wouldn’t have begun to exist.

Additionally, even if the universe was brought about by something immaterial, my sense is that some physicists think even physical stuff emerged from more fundamental things. If this is right, then the universe might have been caused to exist by something else more fundamental.

A second thing said in support of this premise is that there are only three kinds of things: God, abstract objects, and material stuff. But material stuff can’t cause the universe, because the universe includes all the material stuff. Similarly, an abstract object—like the number 4 or the fact that contradictions are impossible—can’t cause the universe because abstract objects don’t cause anything.

This, however, runs into the earlier crucial ambiguity. If the universe refers to all three things: time, space, and matter, then it could have a material cause. If the universe refers to time, space, or matter, then the universe didn’t necessarily begin to exist.

Additionally, even if the universe was caused by a mind of some sort, this doesn’t tell us that it was caused by God. Perhaps it was caused by some deistic creator, for instance, or some other kind of agent. Now, I don’t think this is likely, because I think we have independent arguments for the high prior (and posterior) probability of a perfect being. But insofar as the Kalam requires this, it cannot serve as a standalone deductive argument. It could still perhaps raise the likelihood of theism by disbarring certain kinds of naturalism, but that’s all it can do.

Additionally, there might be all sorts of other causes of the universe. While minds, abstract objects, and physical things are the only kinds of things that we know about, nothing rules out some other exotic cause of the universe.

Lastly, people argue that the cause of time must be a mind because if it wasn’t then the conditions by which it creates the time would always have been timelessly. But then it would always have been making the universe! Only a mind can decide to go from a state of not making the universe to making the universe. This argument is dubious:

1. It seems about as mysterious for a mind to be thinking and choosing without the existence of time as for a non-mind thing to do so. Thus, while I can’t really imagine the universe being caused by atemporal matter, I also can’t really imagine it being caused by an atemporal mind.

2. More broadly, I have no intuitions about how causation is supposed to work timelessly. This seems like a particularly brain-breaking area, and we shouldn’t place much stock in it.

3. Perhaps the thing that existed sans time simply was always in the process of generating the universe. It atemporally generated the universe. The necessary and sufficient conditions for the universe were atemporally present, and consequently, a universe began in time.

Thus, I think nothing remotely persuasive can be said in defense of this premise. The arguments for it are quite poor, and nowhere near the kind of significant persuasive power needed for a successful argument.

4 Conclusion

Here I’ve criticized the Kalam cosmological argument. I’ve argued it has four main controversial steps—that things which begin needs causes, that the universe is finitely old, that the B-theory is false, and that the cause of the universe must be God. Contrast it with, for instance, the anthropic argument, which has just one philosophically controversial premise, and that premise shouldn’t be controversial and only is because anthropics is treated as a no-go-zone by philosophers, where they simply declare that the puzzles are difficult and mostly don’t investigate the relative plausibility of the opposing views.

For every step in the argument, I suspect it is false. If asked to bet on each premise, I would bet against any particular one—though probably some are true. An argument that relies on a mound of controversial assumptions, none of which have much to be said in their favor, is bound to be unsuccessful. It would be like if a person had an argument of the form:

1. Moral nihilism is false—some things we might do really are wrong.

2. If you take an act, and you have twins that deterministically do the same things as you, this counts as a consequence of your action.

3. The universe is infinite.

4. If atheism is true determinism is true, while if theism is true, we have libertarian free will.

5. Actions with infinite consequences, where neither their positive nor negative consequences is of greater cardinality than the other, are neither right nor wrong.

6. Therefore, God exists.

Even if this argument were valid—there actually is an argument along these lines, and I’ll leave working out the details as an exercise for the reader—it wouldn’t be convincing, because it has too many non-obvious steps. The Kalam is similar, though its steps are even less likely: some are, I think, almost definitely false. Thus, proponents of theism should instead make other, much better arguments like the anthropic argument!