Fixing William Lane Craig’s Biggest Mistake

Welcome to the Counsel of Trent podcast, a production of Catholic Answers.

Hey, everyone. Welcome to the Counsel of Trent podcast. I’m your host, Catholic Answers apologist, Trent Horn. And today I want to talk about one of my apologetic role models and why he’s wrong. But before I do that, I hope that you will be a role model to other viewers of this channel and leave a like on this video if you like it and subscribe to help us reach more people and equip them to defend the truths of the faith. And if you really want to help us out, please consider becoming a premium subscriber at trenthornpodcast.com where you get access to weekly live streams and my private study series. All right, so one of the people who helped me convert from deism, belief in a generic creator God, to Christianity, was the Protestant philosopher and theologian William Lane Craig. He’s also been a role model to me for how to do apologetics. Because Craig embodies the rare combination of someone who is well-read in Philosophy, Science, History, Theology. He can communicate those ideas in a winsome way and defend them in a debate by being quick on his feet.

Although I feel like Catholics get into the following cycle with William Lane Craig, stage one, unrestrained hero worship. “Oh my gosh, guys, have you heard of William Lane Craig? He’s the greatest Christian thinker of all time.” Stage two, unrestrained contempt. “Really? You listen to William Lang Craig? Why would you care what a heretic has to say about anything? Along with being a Protestant, which is bad enough, he thinks Christ had only one will and that God is in time. Get rid of your WLC, replace it with some TST. Thomas’s Summa Theologiae in Latin.”

Stage three, balanced depreciation. William Lane Craig is a thoughtful, intelligent Protestant who has made significant contributions to the philosophy of religion, and he’s modeled how to be a winsome communicator. While his theology has errors and some of his philosophical arguments are problematic, Christians would do well to critically engage his arguments and his methods. And I’m basically at stage three. So today I want to talk about Craig’s most famous argument, the Kalam cosmological argument. Of course, it’s not Craig’s argument. Versions of it have gone back to John Philoponus in the fifth century and medieval Muslim scholars of the Kalam School of Theology.

Ghazali formulates his argument very simply. I quote, “Every being which begins has a cause for its beginning. Now, the world is a being which begins, therefore it possesses a cause for its beginning.” Ghazali’s reasoning involves three simple steps. Premise one, whatever begins to exist has a cause of its beginning. Premise two, the universe began to exist, from which it follows three, therefore, the universe has a cause of its beginning.

Craig defends premise two with scientific arguments based on contemporary cosmology and philosophical arguments based on the impossibility of actual infinites. The first approach uses modern science, and that can be helpful if someone believes that things like the Big Bang Theory show the universe had an absolute beginning. The problem is that cosmological models about the origins of the universe are very speculative. I remember when I was 19 years old and Craig started including the 2003 Borde-Guth-Vilenkin Theorem, BVG theorem in his arguments for the absolute beginning of time. He said the theorem showed that even in a multiverse, the universe would have to have an absolute beginning of time. Craig brought up the theorem in a 2015 debate with the atheistic cosmologist Sean Carroll, and Carroll had a clever response.

Now, you might think that, there’s a theorem by Alan Guth and Arvind Borde and Alex Vilenkin that says the universe had a beginning. I’ve explained to you why that’s not true, but in case you don’t trust me, I happen to have Alan Guth right here. One of the authors of the Borde-Guth-Vilenkin Theorem. Alan, what do you say? He says, “I don’t know whether the universe had a beginning. I suspect the universe didn’t have a beginning. It’s very likely eternal, but nobody knows.” Now, how in the world can the author of the Borde-Guth-Vilenkin Theorem say the universe is probably eternal? For the reasons I’ve already told you, the theorem is only about classical descriptions of the universe, not about the universe itself.

Now, you can argue that Guth is mistaken or that the other models he proposes aren’t plausible. But all of this shows that using theoretical physics and big bang cosmology to prove the universe had a beginning, it can become unwieldy. Now, I’ve defended the scientific evidence for the beginning of the universe, including the BVG Theorem. I did that in my first book Answering Atheism published in 2013. Although I did add this caveat in the book, “The tentative nature of science is why I consider the philosophical argument for the finite past to be the main part of the Kalam cosmological argument. The scientific evidence simply serves as a confirmation of the conclusion reached in the philosophical arguments. It’s the scientific icing on the philosophical cake.” End quote. I basically hold the same view today, but I would say that science at best has not proven the universe is eternal. That means if you want to defend the Kalam argument, you’ll need a good philosophical argument for the beginning of the past. Craig offers two.

The first argument is based on the impossibility of an actually infinite number of things. It goes like this, an actually infinite number of things cannot exist. A beginning less universe involves an actually infinite number of past things, therefore a beginning less universe cannot exist. Now in this argument, premise two is obvious. If the universe never began to exist, then prior to today, there have been an infinite number of past causes or events in the history of the universe. The second argument that I’ll give on behalf of premise two is based on the impossibility of forming an actually infinite number of things by adding one member after another. It goes like this, a collection formed by adding one member after another cannot be actually infinite. The series of past events is a collection formed by adding one member after another. Therefore, the series of past events cannot be actually infinite.

So the first argument is that actual infinites cannot exist because they create contradictions, and since an infinite past is an actual infinite, it can exist. The second argument says that even if an actual infinite could exist, you could never traverse it. Today could never have happened if time had to traverse an infinite number of days before today. When I defended the Kalam argument in the past, I focused on the impossibility of traversing the infinite, the second argument. But one of the weaknesses of this argument is that it relies on a theory of time that is not popular among physicists and philosophers. It’s called presentism or the view that only the present moment is real. Past events no longer exist and future events don’t exist yet. Most laypeople hold this view, but many experts defend eternalism the view that past, present and future events are all equally real and the passage of time is somewhat illusory.

Now, there are other views than these two, and people use different labels to describe them. For simplicity, I’m just going to talk about eternalism and presentism, but if you want to go deeper in the subject, I’ll link in the description below to an episode I did just on the philosophy of time. The traversal argument for a finite past relies on the idea that the past is formed one event at a time through a successive series of events. But of all events are equally real, they aren’t formed successively one event at a time. Everything would just exist in a single block universe. That means the traversal argument wouldn’t work because nothing is being traversed.

All right, a question from Frank, “Should the B or any other A temporal theory of time proven to be correct, how would you modify the Kalam if you wouldn’t discard it completely?

Well, I would give up the second philosophical argument based upon the impossibility of forming an actual infinite by successive edition.

But it gets worse because if Craig’s model of presentism were true, that would actually undermine his first and primary argument for a finite past. That’s the claim that since actual infinites cannot exist, that means a beginning less past can’t exist because a beginning less past would’ve an actual infinite number of past events. To show actual infinites cannot exist, Craig uses thought experiments like Hilbert’s Hotel or a hotel with an actually infinite number of rooms. He says actual infinites lead to contradictions and that’s why they can’t exist.

The mathematician David Hilbert illustrates the problem by imagining a hotel with an infinite number of rooms, all of which are occupied, there’s not a single vacancy. Every room in the infinite hotel is full. It gets even crazier when the guests start to check out. Suppose all the guests in the odd numbered rooms check out. In that case, an infinite number of people have left the hotel and yet there are no fewer people in the hotel. But suppose instead, all the guests in rooms numbered four and above checkout. In that case, only three people are left, and yet exactly the same number of people left the hotel this time as when all the odd numbered guests checked out. Thus, we have a contradiction. We subtract identical quantities from identical quantities and get different answers. These absurdities show that an actually infinite number of things cannot exist in the real world.

Craig also uses infinite libraries and infinite collections of baseball cards, but they all make the same point. Actual infinites are contradictory so they can’t exist in the real world. One problem though is that even if these arguments work, they only show an actual infinite number of objects like a collection of hotel rooms can’t exist at the same time. Not that actual infinites can’t exist at all. For example, imagine a Hilbert room, that the room is always existed and at the start of every minute a small iron pellet pops into existence and vanishes one second later. By the present moment, this room would’ve held an actually infinite number of pellets. Each pellet would correspond to a negative number in the past, but nothing seems absurd about the Hilbert room because the pellets never existed together at the same time. Past events aren’t like rooms at Hilbert’s Hotel that all stick around to contradict each other.

They seem more like appearing and disappearing iron pellets in a Hilbert room that don’t cause any contradictions because they’re never together at the same time. So even if Hilbert’s Hotel can’t exist, that doesn’t prove an infinite past can’t exist because an infinite past is not an actual infinite collection of objects at a certain moment of time like a hotel. Another problem with Craig’s argument is the symmetry objection. Basically, if a beginning less past can’t exist because it contains an actually infinite number of past events, then an endless future can’t exist either because it also contains an actually infinite number of future events. This objection has been made by philosophers like Wes Morriston and Alex Malpass. They show there is a symmetry between the arguments for a finite past and the arguments for a finite future. Even if you can prove the past must have a beginning, that same argument could be used to show the future must have an end, which isn’t good for Christians who believe that heaven will be an endless existence with God.

Craig’s standard reply is that an endless future only has a potentially infinite number of events while beginning less past has an actual infinite number of past events. The past will always have had an infinite number of past days, but the future approaches infinity as a limit and never reaches it. But Malpass has pointed out that Craig makes a mistake here in tenses. In an endless future, there will never be a day number infinity. That’s true, but the total number of days going forever forward on a timeline is the same as the total number of days going forever backward on a timeline infinite. This is true even if there is no single moment in the future when someone notice, an infinite number of days has elapsed. So Craig’s argument faces a calamity. A beginning less past may not be an impossible actual infinite because past days don’t all exist together at the same time like at Hilbert’s Hotel.

A beginning less past is infinite, but it’s not an actual infinite collection of things. So Craig’s argument faces a calamity. First, a beginning less past may not be an impossible actual infinite because past events don’t all exist together at the same time. A beginning less past is infinite, but it’s not an actual infinite collection of things. Second, any argument to show the past must have a beginning because if it didn’t, it would be an impossible actual infinite. It could be used to show the future must have an end so that it’s not an actual infinite either. So how do we escape these objections? I believe that Craig was on the right track in arguing that the past cannot be infinite or beginning less because this leads to contradictions. Where he errs is that it is not the infinite past itself that is the contradiction or the thing that can’t exist.

An infinite past is not impossible in the same way an infinite hotel is impossible. Instead, an infinite past is impossible because it allows things like infinite hotels to exist that are impossible. The difference is subtle but important. To make an analogy, imagine you want to show time travel into the past is impossible. You don’t have to show the act of time travel as a square circle. I mean, you can imagine going into the past, being there for one nanosecond, then returning to the present without causing any paradoxes, but you can’t imagine a square circle. Likewise, a critic of the Kalam argument might say, “You can imagine every previous day having a previous day before that,” and there’s no contradiction that arises from that simple thought. In fact, a philosopher Arif Ahmed said something like that in his 2005 debate with Craig.

That doesn’t mean the notion of infinity is contradictory. It certainly doesn’t mean the universe had to have a beginning. It’s very easy to refute this part of Dr. Craig’s position. It could be true that before every event that took place, there was another one. That makes perfectly good sense and if it were true, there’s no first event. There’s nothing for God to cause. Let me repeat that. Every event follows another one. I hope Dr. Craig can tell us later on tonight which of those five words he doesn’t understand.

But something can be impossible even if it seems possible at first. Time travel movies make it seem like you could go into the past and remain on the same timeline and return to the present, but if you stop to think about it, this can’t happen. If you could be in the past for a nanosecond, could you be there for one second? How about 60 seconds? If you can displace the air in the past, could you displace other molecules? Could a gun that you brought displace air and then human tissue when you use it to kill your grandfather? If you could travel into the past and remain on the same timeline, what would stop you from doing something impossible like killing your grandfather before he met your grandmother? If you did that, you would never be born and then you would never go back in time to kill your grandfather, thus causing the famous grandfather paradox.

That’s why some philosophers say the best way to avoid these kinds of paradoxes is just to say that time travel into the past is impossible. We can make a similar move to strengthen the Kalam argument. An infinite past is impossible, not because the infinite past itself is impossible, but because an infinite past would allow impossible things to happen. The argument could be formulated like this, one, if the past were infinite, then contradictions would be possible. Two, contradictions are not possible. Three, therefore, the past is not infinite. In order to bring out the impossible effects of an infinite past, the philosopher Andrew Loke asks us to imagine a Hilbert’s Hotel that is built at the rate of one room every year, and that each room is equipped with a device that creates an immortal guest who then checks into the hotel.

And if you think the whole Hilbert’s Hotel thing is silly, it doesn’t say anything about whether the past can be infinite, we can change the example to make it more realistic and boring. You could do something like imagining an indestructible iron pellet pops into existence once every 10 to the 500 years, that’s a one followed by 500 zeros. In that amount of time you’d think anything could happen. And if the past were infinite, then in the present moment there would’ve been an infinite number of one in 10 to the 500 segments of time. And so in the present there would be an actual infinite number of indestructible iron pellets. But back to hotel rooms, they’re more interesting to talk about. Under this view, the objection that past events don’t exist at the same time evaporates. We can point out that it’s possible for the effects of all these past events to exist at the same time in the present, even though the past events themselves never existed together in this way. By focusing on the effects of an infinite past, this answers the objection that an infinite past is not an actual infinite collection.

The past itself may not be, but it can result in actual infinite collection just as time travel can result in things like the grandfather paradox. To prevent both paradoxes, it’s best to conclude that the past cannot be infinite and you can’t travel to the past from the present. Next, by comparing the effects of an infinite past to the effects of an infinite future, we can avoid the symmetry objection even if the past, present and future are all equally real, there is still a difference between the past and the future. Present effects exist because of past causes, not future causes. For example, if you started today and built one room of Hilbert’s Hotel every year into an endless future, there would never be a time when an impossible Hilbert’s Hotel exists. The hotel would always have a finite number of rooms. This captures the essence of Craig’s objection, that an endless future is potentially infinite.

It is potentially infinite if each moment involves actualizing something as part of a collection like incrementally building a hotel. But paradoxically, if Hilbert’s Hotel had been incrementally constructed in the same manner every year from an infinite past, then at every point in the past there would be an infinite number of rooms that cannot exist just as at every moment in an infinite past there are an infinite number of moments that proceed every past moment. The symmetry is broken when we see that in classical causation, causation flows in one direction from the past into the future. In my dialogue with Alex Malpass, I brought up Loke’s example of creating a Hilbert Hotel incrementally as a way to avoid the symmetry objection. Malpass countered with an attempt at a symmetry argument by imagining a hotel destroyer that destroys one room of an infinite hotel every year into the future. Here’s the exchange in my response to him.

If for him to be able to, obviously he could destroy one hotel room tomorrow, it doesn’t mean Hilbert’s hotel exists today. It could just be one hotel room and that’s it, but he could destroy that. But for him to be able to keep destroying hotel rooms once per day forever, they’d have to be a Hilbert’s Hotel in the present. But aha, second premise, Hilbert’s Hotel is impossible. Therefore, God can’t destroy hotel rooms once a day for the rest of the endless future. It might be he lacks that capacity because there couldn’t be a Hilbert’s Hotel. That seems to be the sensible thing to say there. I mean, otherwise you could say maybe the future has an end, but that doesn’t… You probably don’t want to admit that either. The easiest thing to, suppose is just because Hilbert’s Hotel is impossible, God couldn’t in fact destroy a unique hotel room every day for the rest of the future.

Mike, this is the first time I’ve heard this example, so let me stumble my way through it. It seems like to me it is not the destruction that creates the problem in the future example, it’s just the existence of the hotel, but it is the construction in the beginning less past that gives us the problematic element and why it exists in the first place. It seems like the existence of the hotel is quite separate from the act of destruction versus the paradoxical element that’s involved. I think that’s like a first.

Yeah, I think some people have that same thought as well. I mean, that’s fair enough.

Loke’s modified version of Hilbert’s Hotel preserves the essence of Craig’s claims about the impossibility of an infinite past, but it escapes common objections to it by focusing on the impossible effects of the past rather than the past itself being impossible. Another way to improve Craig’s argument would be to say that if the past were infinite, then it could be possible for there to be an infinitely long causal chain. But philosophers like Rob Koons and Alex Press have shown that causal chains cannot be infinite. If they were infinite, they too would yield contradictions. For example, in my dialogue with Malpass, he said there was nothing impossible about someone existing for all eternity.

There are other beginning the sequence of events, so they’re innocent of that logical contradiction. Ones where everybody passes everyone else a note that says, “I love you,” or whatever. Nothing wrong there. There’s just a nice, cozy, perfectly logically consistent world even if it’s got no beginning.

But I think the way that it would work here, it goes back to what I was saying about, could there be a beginning less past and there’s a causal thread running through it? Because even if you have the examples like passing a note, “I love you,” and it arrives at the present, it doesn’t seem to be problematic. But then-

… [inaudible 00:23:07] sequence.

… yeah. But then it’s like, well, let’s say we have an infinite past and there is always a thread at negative four… Sorry, negative 1, 2, 3, 4, and they count or recite the number or they write down something. It still seems to me it’s getting us closer more to the number line I was thinking of before, that there’s nothing… I guess, maybe it’s lacking things or transformative. To draw my point, imagine if instead of just reciting numbers for all eternity, our eternal counter wrote each negative number down on a separate sheet of paper God made for him in that moment.

In the present moment when this countdown is finished, there will be an actual infinite number of papers each with a different negative number on it. This would be similar to Loke’s example involving Hilbert’s Hotel and how an infinite past creates an actual infinite collection. Or to use a modified example from Koons, imagine God creates an infinite number of immortal people standing in a line. The line terminates in a person whose name is zero. Every person in the line besides zero, has a name that corresponds to one of the negative integer and no integer is repeated. Now, imagine two scenarios. One, each person in the line says their name after the person to their left says their name. Each person knows what their name is by subtracting negative one from the name of the person to their left who just said their name. Eventually the line terminates in a person named zero.

So far so good, but let’s change the example slightly. Two, each person in the line already knows their name that corresponds to a negative integer, but they receive a piece of paper from the person on their left. If the paper is blank, they write their name on it. If the paper already has a name on it, they then pass the paper on to the next person. Eventually the paper reaches Mr Zero. Now, here’s the question. What number is written on the paper Mr. Zero received? There has to be some number written on the paper because if it were blank, then Mr. Negative one would’ve written his name on the paper, but the paper could not have been blank when Mr. Negative One got it because Mr. Negative Two would’ve written his name on the paper and so on. We end up with a piece of paper that arrives in the present that must have some number on it, but it cannot have any particular number written on it, which of course, leads to a contradiction. And like Hilbert’s Hotel, you can modify the example to make it more mundane.

The point is just to show that a certain kind of causal sequence that requires every member to act only when a previous member acts, but also has no first member always causes a contradiction. But it seems that you can have causal chains where each member pauses until another one acts. The impossible element we should reject then is the claim that there could ever be no first member, but then we have to reject an infinite past because it would contain causal chains without first members or infinite chains. So once again, we’re back to the modified Kalam argument. It would go like this. One, if the past were infinite, then contradictions like infinite causal chains would be possible. Two, contradictions like infinite causal chains are not possible. Three, therefore, the past is not infinite. Now, one might say that time travel into the past and infinite pass are both possible, but some force will keep paper passers and grandpa killers from causing paradoxes.

You’ll slip on a banana peel when you try to shoot your grandfather, or the pencils of the paper passers will break and prevent the chain from reaching Mr. Zero. However, this solution results in what Alex Press calls a strange metaphysical force that acts in the universe to stop paradoxes. That sounds a lot like God. So it’s a pretty high price to pay to avoid something that provides evidence for an uncaused cause of the universe. Finally, I want to make clear that this is not the end of the conversation on the Kalam argument, far from it. There are many other objections that have been made to Loke’s version of the argument and objections to both premises of the causal finitism argument that I’ve laid out, and I’ll be addressing those objections in a new book I’m working on, on the case for theism. I’ll be focusing especially on arguments for the existence of God that have been made in the last 20 years and the best route forward for those arguments.

But I do think that these arguments I’ve shared with you represent probably the best way forward with something like the Kalam argument. So just to summarize, this newer version of the Kalam argument would go like this. One, if the past were infinite, then contradictions would be possible. Two, contradictions are not possible. Three, therefore, the past is not infinite. In conclusion, there’s a lot of work to be done to defend these premises and answer objections to the newer Kalam argument, but this formulation helps avoid the problems I discussed earlier in Craig’s classic approach to the Kalam argument.

And while I think his case can be improved, I do want to thank William Lane Craig for all the work that he has done in this area, how he’s modeled to present this argument with those who disagree with him and personally, all the work that he has done for me to lead me to faith in Christ and model for me how to be a good Christian apologist. So I’m thankful to him for that. I’m thankful for you all for checking out today’s episode. If you want more resources on these newer versions of the Kalam argument, check out the links in the description below, and I just hope that you all have a very blessed day.

If you like today’s episode, become a premium subscriber at our Patreon page and get access to member only content. For more information, visit trenthornpodcast.com.