William Lane Craig CV


Existence of God (part 12)

Transcript of William Lane Craig's Defenders 2 class.

Excursus: Natural Theology
§ II. Kalam Cosmological Argument
Lecture 7

To set the context, we have been talking about the Kalam Cosmological Argument for the existence of God. That argument, you will remember, goes like this:

1. Whatever begins to exist has a cause.

2. The universe began to exist.

3. Therefore, the universe has a cause.

We have been looking at evidence for that controversial second premise that the universe began to exist. We looked at philosophical arguments, and then we looked at scientific confirmation of this premise. We most recently have been looking at the confirmation of this premise from the second law of thermodynamics. You’ll remember we saw that the application of the second law of thermodynamics to the universe as a whole implies that given a sufficient finite amount of time, the universe will reach a state in which it becomes cold, dark, dilute, and dead. The question then arises: if, given a finite amount of time, the universe will arrive at such a state, then why is it not now in such a state, if it has already existed for infinite time? If the past is infinite, then the universe should have already reached a condition of being lifeless, dark, dilute, and dead. And yet it has not.

All of this seems to raise the implication that the assumption behind the problem is wrong. Namely, this assumes that the universe has existed forever. If the universe began to exist, then it has simply been a finite amount of time since the initial energy was put in at the beginning of the universe, and it is now winding down toward some sort of thermodynamic heat death, which it will arrive at in the future. The application of the second law to the universe as a whole implies that the universe began to exist, which is the second premise of the Kalam Cosmological Argument.

Attempts to Avoid the Beginning of the Universe

Of course, attempts have been made to try to avoid the beginning of the universe which is implied by the second law. But so far, none of these has been successful in avoiding the beginning of the universe at some time in the finite past. For example, during the 1960s, a number of scientists suggested that perhaps the universe is in an eternal process of oscillating, that is to say, it expands and then it contracts and then it expands again, and then re-contracts, over and over again in a sort of concertina-like fashion. This process of oscillating has been going on from eternity past, so if you were to trace the history of the universe back in time it would look like a series of humps, which represents the cycles of expansion and contraction going back in time (where the vertical axis is space and horizontal axis is time). In this way, the beginning of the universe would be avoided. [see Figure 1]

Figure 1 - Oscillating Model. Each expansion phase is preceded and succeeded by a contraction phase, so that the universe in concertina-like fashion exists beginninglessly and endlessly.

Ironically, however, the thermodynamic properties of the universe imply the very beginning of the universe that these theorists sought to avoid. For it has been shown that entropy, or thermodynamic disorder, is conserved from cycle to cycle. That is to say, the energy is not completely recycled each time it contracts and expands again. Rather any thermodynamic disorder that accumulated in one cycle is pulled through to the second, where it accumulates even further and then is pulled through to the one following that. So over time the thermodynamic disorder will continue to accumulate from cycle to cycle. This has a very interesting effect upon the behavior of an oscillating model. This entropy accumulation causes each cycle to be larger than the cycle before it and to have a longer duration [see Figure 2].

Figure 2 - Oscillating Model with Entropy Increase. Due to the conservation of entropy each successive oscillation has a larger radius and longer expansion time.

So as you trace the expansions back in time, they would get smaller and smaller until one comes to a first oscillation and an absolute beginning of the universe.1 So, ironically, the thermodynamic properties of the oscillating model implied the very beginning that its proponents sought to avoid.

In fact, since entropy is accumulating from cycle to cycle, if the past were infinite, you would have an infinite amount of entropy in the universe. But astronomers have estimated on the basis of current entropy levels in the universe that even if the universe were oscillating, it could not have gone through more than 100 previous oscillations before you reached the first oscillation and the absolute beginning of the universe.


Question: The last few Scientific American magazines had articles on a variant on this where they talk about some kind of convergence and also on a Big Bounce where you accumulate density and then it repulses much like when you put two poles of magnets together. But it seems very (inaudible)

Answer: It sounds like it. I have not read those specific articles. But one of the other problems with the oscillating models that I didn’t even mention is that up to now there hasn’t been any known physics that would cause a re-contracting universe to bounce back to an expansion. The physics seem to predict it would just collapse into a black hole from which it would never emerge. So physicists are looking for some sort of mechanism that would cause this kind of oscillating behavior. But what this argument shows is that even if this were possible for the universe to be oscillating, the thermodynamic properties imply that it still has to have a beginning of the process.

Question: You mention that the energy is not always reserved going from one cycle to another, correct?

Answer: What I said was that the energy isn’t completely recycled in the sense that you begin with a new low entropy level. It is not as though entropy builds up to a certain quantity here, and then in the next cycle you start again with zero entropy all over again. What I am saying is that whatever accumulates in the first cycle gets passed through to the second. This is called the conservation of entropy from cycle to cycle.

Followup: Why would that cause it to be . . .?

Answer: The effect of this increased entropy is that each cycle will be longer than the previous one and the radius will be greater of the expansion. So that causes this behavior of diminishing cycles as you go back in time.

Question: Wouldn’t this make it for an atheist less probable that if there had been many previous cycles that the universe could rise again each time from purely natural processes?

Answer: I think what you are raising here is the question of the fine-tuning of the universe for life. One of the problems with this oscillating model is that in order to get a universe that would oscillate from eternity like this, so that each cycle would be exactly similar to its predecessor, it involves a kind of infinitely precise fine-tuning of initial conditions in order for this to take place. Moreover, this fine-tuning would have to be of an extraordinary sort because it would have to be set at infinity past, which is kind of crazy – how can you have initial conditions be set at past infinity? The model, even on its face, involves a really bizarre form of fine-tuning in order to have a cyclical behavior like this that would allow universes to exist each time that would be characterized by observers like us.2

Question: How popular is this model among cosmologists?

Answer: According to Stephen Hawking, this was popular back in the 1960s, especially among Russian physicists. But, with the enunciation of the Hawking-Penrose singularity theorems, it fell into disfavor because what those theorems showed is that any universe collapsing toward a singularity simply ends at that point – there is no way to bounce back. In recent days, there have been some attempts to resuscitate oscillating models by figuring out ways to avoid collapsing down to a singularity and to try to have the universe bounce back before it reaches a singular state. The problem is that those models, even if they succeed in avoiding the singularity, have not been able to be extended into infinity past. You still cannot have an infinite past with such models, even if you can avoid collapsing down to a singular point at the end of a cycle.

Question: What could I point someone to who is questioning whether or not entropy is conserved between cycles?

Answer: If you look at the article in the Blackwell Companion to Natural Theology, it is extensively footnoted with the scientific literature, so I would recommend looking at that. Also, this is really a very well-known feature of these models. This isn’t some esoteric fact; this has been known for decades that entropy is conserved. But you can look at the literature that is footnoted there. I may have it in Reasonable Faith as well.

Question: When you first started talking here, I envisioned just pulsing like a heartbeat – expanding and contracting but not to the extreme of a single point.

Answer: That gets into what I was just talking about. The Hawking-Penrose singularity theorems showed that a universe which is under gravitational self-collapse just goes right down to a singularity, like a black hole. It just collapses down to a boundary point. By exploiting quantum theories of gravity and trying to marry those with general relativity, scientists are trying to show how you can maybe have a pulsating model where the universe wouldn’t collapse all the way down to a singularity, but you can get down to some prior state. But, as I say, these models still can’t be extended to infinity past. So even if the universe didn’t begin with a singular state, as in the standard model, the problem of extrapolating it to infinity remains. That is the real issue. That is important to understand. The real issue is not whether the universe had a beginning in a singular point. The issue is whether the universe began to exist – whether its beginning was singular, as in the point of a cone, where you have a singular point, or whether it is non-singular, as in the Hartle-Hawking model, where it is rounded off at the beginning. In either case, the past is still finite not infinite. And that is what the second premise of the argument is. It is not that the universe began as a singular point, but that the universe began to exist. So don’t be misled by folks who say, “Well, the singularity may not have been real; the singularity is just an artifact of the standard model, but we can adopt theories of the universe that avoid the singularity.” That is not really the issue. The issue is: Did the universe begin to exist? Is the past finite or infinite? The question is not, “Was the beginning state a singular state or not?” By a singular state, one means a state at which space-time curvature, density, and temperature become infinite. That is what one means by that – it would be like a point of a cone.3

Let me talk about another more recent attempt to avoid the beginning of the universe. That would be by saying that our universe is not, in fact, the entire universe. Our universe is just a pocket universe, which is part of a much wider reality, sometimes called the multiverse. So our universe is just a bubble in a sea of similar bubbles, and each of these is expanding, as well as the sea of energy in which these bubbles exist. Our universe has a beginning, but that doesn’t mean the multiverse as a whole has a beginning. The multiverse as a whole can still be eternal and infinite in the past. The second law of thermodynamics, it would be claimed, only applies to our bubble universe, not to the multiverse as a whole.

Whether or not the second law of thermodynamics applies to the multiverse as a whole is a moot point. That is a controversial point. If the second law does apply to the whole multiverse, then it implies that the multiverse itself must have a beginning and cannot have existed for an infinite time. But in any case, even that aside, we’ve already seen that the theorem developed by Borde, Guth, and Vilenkin implies that even the multiverse itself cannot be extended into the infinite past but must have a beginning point. So even this foam of bubbles that is forming is something that must have begun to exist, which is again the second premise of the Kalam argument.


Question: What if someone were to claim that the multiverse itself, as a whole, is a static entity that is not expanding or contracting and that there is just this sort of existent sea of energy?

Answer: This is very similar to a kind of model that developed during the 1970s that I have called Vacuum Fluctuation Models. This says that there is a mother vacuum, which is a sort of womb in which these baby universes are formed, and it is a static and eternal entity. These baby universes are expanding into this great quantum vacuum in which they exist. This model ran into a very serious problem, namely, at any point in the quantum vacuum, there is a non-zero probability that a universe would form at that point by a quantum fluctuation. But given infinite past time, universes will have come into being at every point in the quantum vacuum because given any non-zero probability and enough time, eventually that probability will be actualized. But then those universes will have by now so expanded as to fill the entire quantum vacuum and so will run into each other, coalesce, and form one infinitely large, infinitely old universe – which contradicts observations that we exist in a relatively young universe. So it is not enough to have the bubble universes expanding in this static mother universe; the mother universe has to be expanding as well. But once you do that, then it falls under the Borde-Guth-Vilenkin theorem which says any universe which has, on average, been in a state of expansion throughout its history cannot be infinite in the past but must have a beginning. So the static model was tried, and more recently these multiverse inflationary models have been tried, but what scientists have discovered is once again they cannot be extrapolated to infinity past.

Question: Are you saying there are scientists that believe these things always were?

Answer: There is a good number of people who believe that the universe is eternal in the past and is uncreated and has always existed.4

Followup: And those people struggle with the thought that there could be a deity who always was?

Answer: Let’s put it this way: not necessarily! Thomas Aquinas, for example, the great medieval theologian, thought the universe could be eternal in the past but still be created by God. It just wouldn’t be created at some beginning point. It would just be eternally dependent on God in the way that, say, a heavy weight that is dependent on a chain hanging from the ceiling is always dependent on the chain, even if they have been hanging there from eternity past. So it is not an argument for atheism or against the existence of God, but it would be simply a way of undercutting the Kalam Cosmological Argument, which has a premise in it that says the universe began to exist. If you could undercut that premise, then that would make that argument unsound.

Followup: It amazes me that people can think that something can exist without having a starting point. At one time an oak tree was not an oak tree.

Answer: That gets into those philosophical arguments we talked about – whether or not the series of causes can regress infinitely into the past. I tried to show that that is a very, very difficult notion and very problematic. What we are finding now is that we have not simply philosophical reasons that the past is finite, but actual scientific reasons as well.

Question: It seems that many of the people that I talk to that use this are actually trying to avoid the fine-tuning of the universe. Should we point out that that is their real motive?

Answer: Your point is well-taken in that today there is another argument for the existence of God, the design argument, that is based on the fine-tuning of the universe. And the odds of this fine-tuning occurring by chance are so astronomically low (that would be an understatement), they are so incomprehensibly improbable, that the only way to save the chance hypothesis is through the multiverse. If you can’t get the roulette wheel to land on a certain number in one spin, then you posit an infinite number of spins and an infinite number of roulette wheels, and then that way, by chance alone, the improbable will happen. So if our universe is improbably fine-tuned for the existence of intelligent life, the way you get it to happen by chance is you have an infinite number of other universes, most of which are dead and unobserved. So this same multiverse hypothesis is the principal means by which the design argument is refuted today. The whole debate between practitioners or proponents of design versus skeptics of design is whether or not we live in a multiverse like this and whether or not that will eliminate the fine-tuning. But as I tried to explain, it has some relevance as well to the Kalam argument in that the proponent of the multiverse would agree that our bubble had a beginning but not the whole thing.

Let me go into one final attempt to avoid the implications of the thermodynamic properties of the universe. This is the speculation that has been floated in some quarters that perhaps our universe is the baby of some prior mother universe which has spawned it. And the idea here is that perhaps black holes are really portals through which energy can tunnel to some other unobservable universe. As the energy goes into the black hole, it goes through the worm hole and then is ejected into this other space-time region.5 The speculation is that, with time, the worm hole gets thinner and thinner until finally it pinches off and the baby universe becomes a separate entity in and of itself [see Figure 3].

Figure 3 - Baby universe spawned from a mother universe via worm hole

The idea here might be that perhaps this process has been going on from eternity past – that our universe is simply the product of some prior universe which was itself the product of some prior universe, and so on ad infinitum, so that the universe which began to exist is merely the product of an infinite series of prior universes, each spawning baby universes through black hole production.

Could this scenario be extended into the infinite past to avoid an absolute beginning? Well, sorry – it won’t work. It has been shown to contradict the laws of subatomic physics, or quantum physics. What physicists have discovered is that the information that goes into a black hole remains in our universe. It cannot escape our universe and go to another universe. So this scenario postulating that this baby universe could pinch off and thereby isolate the information that went into the black hole into another world is physically impossible.

This scenario was the subject of a bet between Stephen Hawking and an American physicist named John Preskill. Preskill held that this scenario is impossible and that it contradicts the laws of quantum physics, whereas Hawking was espousing this idea. Hawking, who was one of the last holdouts, admitted in 2004 that he had lost the bet. Offering his apologies to science fiction fans everywhere, Hawking admitted there is no baby universe branching off and the information remains solidly in our own universe. So once again this attempt to avoid the beginning of the universe through very speculative cosmological conjectures was shown to be a failure.


Question: Given all these failed attempts to explain our universe and its beginning, where do cosmologists today hold their faith? What model or what idea? Are they still searching or is there some credible model that is being considered?

Answer: It is wide open. There are all sorts of competing research programs to try to develop cosmological models of the beginning of the universe. Some of these will involve a beginning in the finite past. Others will be attempts to avoid the singular beginning and extrapolate back to a pre-Big Bang condition, for example. But none of them has succeeded in extrapolating to infinity past, so as to restore the eternal universe and avoid the prediction of the standard model that the universe began to exist.

Followup: So is the Big Bang the current one that has yet to be disproved and the one being taught in school?

Answer: I think what one would say would be that the standard model is taught as describing accurately the history of the universe right back to very close to the beginning. But then before you get to that point, prior to that Planck time, then they would say we don’t have a physics to describe that early region of the universe, and so we don’t really know. We don’t have a theory that will allow us to discover what it was like. That is where it is wide open, and there are all sorts of speculations and different models. But remember, as we shared earlier, the Borde-Guth-Vilenkin theorem applies regardless of your physical description of that era. And the second law of thermodynamics is a scientific theory that is in a field of science that is so well understood that it is almost a closed field of science. So the prospects of avoiding or revising that second law in some way as to avoid the beginning are pretty remote.6

Question: What were you trying to explain in (unintelligible)?

Answer: If I understand your question, the second theory that I was talking about was the multiverse model that says our universe is just a bubble in an expanding, wider universe: while our bubble may have begun to exist, the whole universe didn’t begin to exist. The third one here is based upon black hold production of baby universes.

Followup: I was talking about [Vacuum Fluctuation models].

Answer: That was a response to another question about how we could perhaps adjust the model so that the wider mother universe in which these bubbles are formed is not itself expanding, but it is just static. And I said that ran into the problem that then the bubbles would run into each other and coalesce. But if you make it expanding, so that it expands more rapidly than the bubbles do, so the bubbles can’t run into each other, which is the multiverse model – the vacuum is expanding so quickly that it outpaces even the expansion of the bubbles, so they can’t coalesce – but if you say that, then the Borde-Guth-Vilenkin theorem applies to the multiverse, and you have to have a beginning.

In conclusion to this section of the argument, the scientific evidence of thermodynamics confirms the conclusion that we already reached based on the expansion of the universe, that we have good grounds that the universe began to exist. And this evidence is especially powerful because the field of thermodynamics is so well understood that it is virtually a closed field. That makes it highly unlikely that these findings are going to be eventually reversed.

So then, on the basis on both philosophical argument and scientific evidence, we have good reason for thinking the second premise of the Kalam Cosmological Argument is true: the universe began to exist. In conjunction with the first premise, that whatever begins to exist has a cause, the conclusion therefore follows with logical necessity: therefore, the universe has a cause.

What we will now do is explore the theological implications of there existing a cause of the universe, and we will see that a striking number of divine attributes can be deduced from a conceptual analysis of what it is to be a cause of the universe.7


1 5:12

2 10:18

3 15:00

4 20:12

5 25:02

6 30:06

7 Total Running Time: 33:49