This is the weekly Q & A blog post by our Research Professor in Philosophy, Dr. William Lane Craig.


Dr Craig,

My question is based on your formulation of the argument from contingency, specifically, your restricted version of the PSR.

Restricted PSR: everything that exists has an explanation for its existence, whether in the nature of its own necessity or an external cause.

There are good reasons to prefer a restricted PSR over the strong version - it avoids the famous objection by Peter Van Inwagen, which argues that the PSR is false because it has the absurd consequence on making all facts necessary. I am aware that you have of Alexander Pruss's work on defending the strong version and am on the fence at the moment as to whether Inwagen's objection succeeds.

Back to your formation, it seems that there exists an argument from Van Inwagem against your version too! The argument can be stated:

1 Restricted PSR: For every object, there is a sufficient reason for why that object exists.

2 Restricted PSR is plausible, but it won't help in demonstrating that there is a necessary being.

2.1 We can only use Restricted PSR to demonstrate that a necessary being exists only if the universe as a whole is an individual object.

2.2 But the universe can't plausibly be taken to be an individual object

2.3 Rather, the only plausible way to think of the universe is that of an enormous collection of objects (See Peter Van Inwagen's book, Material Beings, for the details of the argument here).

2.4 But if so, then perhaps there are infinitely many objects in the universe stretching through a beginningless past, and the existence of each member of the collection can be explained in terms of one or more other members of the collection.

2.5 And if so, then we need not posit a necessary being to explain the existence of the universe.

3 Therefore, on the restricted version of PSR, the Leibnizian cosmological argument fails.

One of the reasons I find this quite a troubling argument, is that I've always seemed to agree with 2.3, that the universe should be seen as a collection of objects rather than a single objection (and why I've never been persuaded by your example of the ball in the forest, which is a single object).

First, I'd like to ask the general question: do you think the argument is valid, if not, why?

I think the leg work needs to be done in premises 2.1 and 2.4. Regarding 2.4, am I correct in thinking this is similar to what Pruss has called the Hume-Edwards principle? The principle that a whole set can be explained by explaining each individual constituent? If so, that can (I think has) been addressed and offered a powerful defeated. By favourite example is found in Anthony Flew's book 'There is a God'. Where he briefly addresses the cosmological argument, he notes that Swinburne and Conway had given persuasive arguments against Hume's objection. He quotes Conway: " the causal explanations of the parts of any such whole in terms of other parts cannot add up to a causal explanation of the whole, if the items mentioned as causes are items whose own existence stands in need of a causal explanation'. Flew demonstrates this with the example of a self replicating computer virus that has infected computers connected by a network. We could have the fact a 100 computers are infected with the virus, but it wouldn't explain the existence of the virus. However, this would only seem to hold if I have understood the premise correctly.

However, I have absolutely no idea how to address 2.1?!

To summarise my question as best I can: do you think the argument given succeeds in making your restricted version of the PSR more implausible than plausible? Do you have any ideas on how to address premise 2.1? Have I adequately addressed 2.4 or have I mistaken it for something it isn't?

I really hope you can reply to me Dr. Craig, I've always thought that this was the strongest of the arguments from natural theology, but that is in serious doubt now.



United Kingdom

Dr. William Lane Craig’s Response

Dr. William Lane Craig

This is a really good question, Callum, that impinges, not just on the Leibnizian version of the cosmological argument but also on the kalam cosmological argument. It raises profound metaphysical issues, which the natural theologian would prefer to leave open, so as to broaden the appeal of his theistic argument as much as possible.

I’m inclined, therefore, just to stick with my own restricted version of the Principle of Sufficient Reason (PSR), namely,

Everything that exists has an explanation of its existence, either in the necessity of its own nature or in an external cause.

and just leave the PSR formulated in your (1) to others. For (1) states merely a sufficient condition, not a necessary condition, for something’s having an explanation. It tells us that if something is an object, then it has an explanation; it doesn’t require that something be an object in order for it to have an explanation. That’s good. There are plenty of things that may not count as objects, e.g., a school of fish, that plausibly have explanations of their existence. So (1) is an even more restricted version of PSR than mine and its truth is in no way incompatible with the truth of my version. So I agree that (1) is true and forget about it.

So, since I’m not using your Restricted PSR, (2) is a matter of indifference to me. It’s a problem only for those who are using your version, along with its very peculiar notion of what counts as an object. If (3) is true, then all that follows is that proponents of the Leibnizian cosmological argument ought not to abandon my version of the PSR for yours!

But what might be said of the argument in support of (2)? You’re not sure how to address (2.1). Well, what sort of argument for a necessary being is supposed to be in view here? Presumably, one which infers from the universe’s being a contingent object that it has an explanation. Since your restricted PSR applies only to objects, one needs the premiss that the universe is an object in order to get to the conclusion that it has an explanation. There might be other sorts of arguments for a necessary being that employ your restricted PSR, but if we’re talking about an argument that infers from the universe’s being a contingent object to the existence of a necessary being, (2.1) seems unobjectionable: it just underlines how severely restricted your PSR is.

The real heart of the argument is (2.3) and its implication (2.2). In affirming (2.3), I don’t think that you realize how radical van Inwagen’s view of objects is. For you say, “I've always seemed to agree with 2.3, that the universe should be seen as a collection of objects rather than a single objection [sic] (and why I've never been persuaded by your example of the ball in the forest, which is a single object).” For on van Inwagen’s view the ball in the forest is not a single object. Indeed, on his view there is no such object as the ball in the forest; it is no more an object than the universe is an object. On van Inwagen’s view, the ball in the forest literally does not exist. So you cannot consistently say that the ball in the forest must have an explanation but that the universe need not, for the ball and the universe are ontologically on a par. No such objects exist. So they do not exist without an explanation, since they do not exist at all.

Now maybe you’ll want to retreat from van Inwagen’s radical metaphysics and affirm that the ball has sufficient unity for the particles of which it is composed to constitute an object, whereas the universe does not. I think that claim could plausibly be disputed: the universe has all sorts of properties unique to it, like a certain spacetime curvature, a certain density, a certain temperature, a certain expansion rate, and so on, enough for it to count as an object, just like the ball. But never mind: even if the universe is not now an object, it certainly was in the past, when it was so dense that it was shrunk down to subatomic proportions. Maybe it was once an object which has now blown to bits. In that case (2.1) requires that it have an explanation.

I prefer to avoid this whole conundrum by using “existence” in the light sense employed in ordinary language. Even if the universe is not an object and so does not exist in a metaphysically heavy sense, still the universe exists in the same sense that the Eiffel Tower exists. So if you think that the Eiffel Tower needs an explanation, so does the universe. That is the point of my version of the PSR. There’s no reason to restrict our attention to “objects” in the metaphysically heavy sense that van Inwagen is talking about.

So, in sum, “do you think the argument given succeeds in making your restricted version of the PSR more implausible than plausible?” Just the opposite! The lesson of your objection is not to base a Leibnizian contingency argument on a bad PSR! Mine comes out smelling like a rose by comparison. “Do you have any ideas on how to address premise 2.1?” See my comment above. “Have I adequately addressed 2.4 or have I mistaken it for something it isn't?” Your refutation of (2.4) works nicely on my light conception of existence, for we still want to know why an eternal universe exists. Whether it works given your metaphysically heavy sense is another question.

This post and other resources are available on Dr. William Lane Craig's website:

Learn more about Dr. Craig’s book, A Reasonable Response.