Austin Cline of Atheism.About has offered an interesting argument as to ‘why God can’t be used as a cause of the universe’ based on ‘causes and physical laws.’ Cline invites us to consider his case:

Theists often insist that the universe needs an explanation and science can’t provide one; therefore, their god is necessary and must be believed in. Such theists miss something, however: causal explanations are functions of natural laws which are themselves products of the universe. Therefore, they can’t be used to explain the universe. Perhaps there is no ’cause’ of the universe.

Cline then quotes two atheists whom have advanced similar lines of reasoning, Robin Le Poidevin and B.C. Johnson. The former writes:

A world in which there can be causal explanation is not a chaotic world; it is a world tightly constrained by the laws of nature. Causal generalisations are simply reflections of these laws: that is, they are true because of the existence of fundamental laws. Causal explanation, then, takes place against a background of laws.

But when we come to the explanation of the universe as a whole, part of what we are required to explain is the existence of the laws themselves. We cannot therefore help ourselves to any laws in order to explain the existence of the universe. Consequently, the explanation of the universe cannot take place against a background of laws. But, since causal explanation requires such a background, there can be no causal explanation of the universe.¹

The only way any sense can be made of this argument would be for Le Poidevin to spell out exactly what he means by “laws of nature” and “fundamental laws.” Does he mean those physical laws which contingently exist as unique expressions of our spatiotemporal universe, such as gravity, electromagnetism, and the nuclear strong and weak forces? More generally, is he suggesting a Humean regularity view (e.g. “copper expands when heated”)? Or does he mean those deeper, perhaps metaphysical laws, which necessarily exist as universal expressions of objective reality, such as the law of the excluded middle, noncontradiction and identity? The distinction is crucial, for the former kind of laws operate strictly according to the latter and, moreover, the former are contingent upon the existence of our universe whereas the latter are not. The distinction is more crucial still, for it is needed in order to proceed with an accurate understanding of what he thinks constrains causal explanation and what said causal explanation is a reflection of. Despite his ambiguity, it seems that Le Poidevin is referring the physical laws which contingently exist as unique expressions of our spatiotemporal universe, for only by assuming these types of laws can his conclusion, that they cannot be used to explain the universe’s origin, follow logically.² Accordingly, the argument can be summarized:

1. It is impossible to explain the origin of a cause by its effect.
2. Natural laws are an effect of the universe.
3. Causal explanations are constrained by and expressions of natural laws.
4. Therefore, it is impossible for the origin of the universe to have a causal explanation.

If it is impossible to explain the origin of the universe by natural laws, and causal explanations are expressions of natural laws, then it is impossible to have a causal explanation for the origin of the universe. So what is to be said of this argument?

It is telling that we never get a clear definition of the type of laws Le Poidevin is referring to. One reason for this could be that once it is clear that the laws to which he refers are physical ones, his argument blatantly begs the question by assuming that only the physical order of things merit causal explanation. Technically, his argument still begs the question even when we substitute “physical laws” with “laws of nature,” only this time we meet the fallacy of ambiguity because, as we’ve seen, there are several ways one can take “laws of nature” and “fundamental laws.” But lets give Le Poidevin a little more credit and proceed with a substantive evaluation of his argument anyway (hereafter “laws of nature” therefore will be understood as those physical laws which serve as unique descriptions of our spatiotemporal universe unless otherwise specified).

Premise (1) is reasonable enough, for in it lies the metaphysical insight that something cannot be caused by nothing. For example, if some entity x causes or brings about some effect y, then y cannot be used to explain the origin of x seeing that x is ontologically prior to y. Though Le Poidevin doesn’t explicitly state this, it is upon the truth of this supposition his entire argument rests. Otherwise, we’d have the premise

1′. It is impossible to explain a cause by its effect

which, though it has been defended, is generally unacceptable. We utilize what we know to be effects to explain causes all the time—in principle and in scientific inquiry. For example, suppose you fly over a forest and observe that all of the trees and foliage below are burnt. It is most reasonable to infer that the cause of such was fire, even though the only part of the causal sequence available to you is the effect. Or suppose you observe a sheet of ice on your front step. From this you know that the cause was the there being water molecules present in conjunction with the temperature’s being below 32° F. Again, if the universe of space and time had a cause, then we know that the cause must be spaceless and timeless, et alii. And so there are numerous cases where we justifiably explain causes by way of their effects, namely by deducing features we know must be ontologically anterior the effect.

Premise (2) also should be uncontroversial. But to clarify, when I say that natural laws are an ‘effect’ of the universe, I do not mean to imply that they came into existence in that there was a time before which they did not exist. All I mean to say is that said laws are caused by and therefore contingent upon the spatiotemporal universe for their existence. This can be seen by noting the difficulty in conceiving of a possible world wherein said laws existed independently from a spatiotemporal universe; conversely, the quite easily conceivable possible world in which there existed a spatiotemporal universe such that there were no laws. This is true whether or not the universe or the laws therein had an absolute origin. So we still have a genuine cause-effect relationship between the universe and its natural laws, even though there is no temporal gap which separates them as an effect of the universe as their cause. In this case, the effect could occur simultaneously with the cause, each being eternally present with one another. So said laws are the effect of the universe and are contingent (or dependent) upon the spatiotemporal universe as their cause.³ The key premise, then, is obviously

3. Causal explanations are constrained by and expressions of natural laws

To establish this, a bit more clarification is in order. First, a word should be said on what constitutes the kind of “causal explanation” Le Poidevin is referring to. There is what Richard Swinburne calls the ‘what’ and ‘why’ of an exhaustive causal explanation with respect to natural phenomena. The ‘what’ refers to the necessary and sufficient initial conditions C for a cause to produce an effect E. The ‘why’ refers to the reasons why the cause, under the specified initial conditions, produced the effect it did. The why is a set of natural laws L. C and L then fully explain E if E follows deductively from them.⁴

So what exactly would such a causal explanation look like? It can be captured as follows: an event of kind K in initial conditions of kind C (the ‘what’) occurring to an entity of kind E causes an event of kind Q for reasons L (the natural laws which guide sequence, the ‘why’). For example, suppose a brick breaks a glass window. The moving of the brick and its touching of the surface of the glass is the event K, the glass being in a solid and not liquid state is an initial condition C, the glass window itself would represent the entity E, and the causing of the glass to break is the event of kind Q. Additionally, the brick’s course of action would be an expression of and constrained by gravitational laws, L. This type of causation is generally known as efficient causation (where causes are means by which effects are produced), a notable feature of which is that causal explanations can be sufficiently (but not necessarily) exhausted by way of an event-event or state-state causal thesis alone. And so all causes and effects are reduced to spatiotemporal events or states that constitute causal sequences construed either deterministically or probabilistically, the former designated by causal conditions sufficient for an effect to occur and the latter by causal conditions sufficient to fix the chances for an effect to occur. In sum, the conjunction of the ‘what’ and the ‘why’ represent a full causal explanation of some effect E. For simplicity’s sake, hereafter we can refer to this type of explanation as Hempelian causation, named after C. G Hempel, to whom its articulation is primarily credited.⁵

At this point we start to run into difficulties with Le Poidevin’s argument. First, note that with only the type of causation explained above, for every cause, we can meaningfully inquire of a cause for that cause, and a cause for that one, ad infinitum. Every cause is itself always the effect of a prior cause. So it seems we are quickly caught facing the dilemma of whether or not there can be an infinite regress of causes. So either we deny the paradoxical nature of the actual infinite and defend its reality or cut the chain of causes off at some point in the finite past. Assuming that the idea of an infinite regress of causes and/or the nature of an actual infinite presents intuitive absurdities, or rather that the universe is not infinite on account of some other reason, we can safely continue without following this side of the dilemma. Therefore, that leaves us with cutting off the chain of causes at some point in the finite past; obviously, in this case the proper stopping point would be the origin of the universe—the moment at which the sequence of natural laws and, hence, causal explanations begin by coming into existence. But this side of the dilemma entails the conclusion that while the universe did begin to exist, and perhaps even that it had a cause, the universe nonetheless lacks the kind of causal explanation explicated above.

But surely it doesn’t follow that the universe does not have a causal explanation. All that follows from Le Poidevin’s argument, as we’ve seen, is that the universe cannot utilize Hempelian causation for its explanans. In fact, it is from this very observation that we gather evidence to support a Leibnizian cosmological argument, which can be summarized thus:

1. Everything that exists has an explanation of its existence, either in the necessity of its own nature or in some external cause.
2. The universe (including any singular state) exists.

It follows from (1) and (2) that the universe has an explanation of its existence. Premise (3) states:

3. The universe (including any singular state) does not exist by a necessity of its own nature.
4. Therefore, the universe has an external cause.

If the universe can’t be explained by the necessity of its own nature (e.g. by the natural laws thereof) then it must have an external explanation. What would such an external explanation look like? Remembering back to our initial support for (1), it must transcend space and time, be metaphysically necessary, and changeless and immaterial. The only things we know of that can exist in that way are either abstract objects (numbers, laws, sets, values, propositions, properties, et cetera), or minds. But abstract entities do not have causal efficacy (e.g. the number two cannot cause anything) and so cannot be the explanation of the universe. So the explanation of the universe is most plausibly taken to be a mind, which is minimally what everybody means by “God”.⁶

Premise (3) in this Leibnizian argument is precisely what Le Poidevin’s argument supports. But something tells me this is not exactly what he intended to show with his argument. We are left, then, vis a vis with his central assumption: it is from only Hempelian causation we get genuine explanation. But I see no reason to accept this assumption as true.

Indeed, Le Poidevin’s argument is self-refuting, for by assuming only Hempelian causation in (3), we are led to reduce the causal principle represented in (1) to a mere physical principle, contingent upon the natural laws of our universe. But this can’t be right, for (1) seems rightly categorized as a metaphysical principle, explanatorily relevant not only to our universe. (1), like other metaphysical principles, seems necessarily true in all possible worlds and so also explanatorily relevant in all possible worlds, not just those with the kind of natural laws Poidevin thinks constrains them. When I say a principle is “explanatorily relevant,” I simply mean a principle whose description can be applied and accurately correspond to the activity of any and all causal phenomena whatever. In short, since we can easily conceive of a world where (3) is false but not (1), (1) therefore proves to be an exception to (3).

In other words, it is more than obvious that all causal explanations are not constrained by and expressions of the Hempelian kind. While it can be admitted that such is an accurate description of causal phenomena in the natural world, it would be utterly false to say that “causality, as a concept, is based entirely upon the context of our universe,” as Cline does, for there are causal principles which are not contingent upon the existence of the natural laws of our universe, causal principles such as “whatever begins to exist has a cause” and “no effect precedes its cause” and ironically, "it is impossible to explain the origin of a cause by its effect".

If these considerations weren’t enough, we can show the argument’s key premise, (3), false on other grounds. Suppose someone accuses me of building a straw man and contends that (1) wasn’t really part of Le Poidevin’s argument and so cannot really be guilty of being self-refuting. Even though I see no way Le Poidevin’s argument can escape assuming (1) as its foundation (other than (1′)), all that would be necessary to then show his argument false is citing examples of causal explanations which might exist apart from those which can be exhausted by or reduced to an event-event or state-state causal thesis alone. Examples would include the entire world of final causes—causes for the sake of which effects are produced.⁷ Final causes always involve teleological goals, ends or purposes expressed by an agent and its intentions. This is the kind of causation often used to describe libertarian agency; a necessary condition of which is for the agent to be an immaterial substance, for neither a physical state or event within the agent could account for a totally free act. Moreover, an overarching physical law cannot capture such actions, for they can neither be determined nor can they be probabilified.⁸ A very different picture of causation is painted here—one that falls outside the kind of natural laws Le Poidevin assumes causal explanation is constrained by and an expression of. In fact, it is this very type of causation which is understood to be exercised in God’s primary act of creation. William Lane Craig nicely summarizes the situation:

The event of the universe’s coming into being cannot be an instance of state–state causation or event–event causation, since the origination of the universe is not a state and the condition of the timeless cause not an event. But neither can it be an instance of state–event causation, for this seems clearly impossible: If the unchanging cause is sufficient for the production of the effect, then the cause should not exist without the effect, that is to say, we should have state–state causation. If the cause is not sufficient for the production of the effect, then some change must take place in the cause to produce the effect, in which we have event–event causation and we must inquire all over again for the cause of the first event. The best way out of this dilemma is agent causation, whereby the agent freely brings about some event in the absence of prior determining conditions.⁹

Causality, then, is not as easily defined as Le Poidevin had hoped. Now what about B.C. Johnson, the second atheist Cline quotes to warrant the conclusion that the God cannot be the cause of the universe? Johnson writes:

The theist innocently demands a cause for orderliness, forgetting, of course, that “cause” presupposes “orderliness.” Without the laws of causality, no causes would be operative. The laws of causality must therefore exist before any cause can operate. Therefore the laws of causality cannot be the result of any cause. These are laws which cannot be caused even by God.¹⁰

Attentive readers will notice that the arguments of Le Poidevin and Johnson are quite different. In fact, I’ll argue that Johnson’s argument implicitly contradicts Le Poidevin’s. Despite the alleged connection between causation and orderliness, such a parallel is irrelevant to what Johnson is trying to argue. Johnson’s argument can be stated as follows:

1. The laws of causality existed before any cause could operate.
2. Therefore, the laws of causality cannot be the result of any cause.
3. Therefore, God cannot be the cause of the laws of causality.

The hidden assumption this argument makes is that the laws of causality were never themselves caused and so must exist necessarily. This is exactly the opposite assumption Le Poidevin made in his argument. Le Poidevin’s assumption was that causal explanation was constrained by and an expression of natural laws which were themselves contingent features of the universe. And so the arguments are quite different. As for a critique of Johnson’s argument, well, I don’t find myself in disagreement with it. As a result, I’m forced to ask myself where Johnson’s argument against theism is. Actually, I think it implicitly counts as evidence for theism! The assumption that the laws of causality are themselves uncaused and necessarily existent is only at home within a theistic worldview, for laws as such are what philosophers call abstract objects, the existence and grounding of which is most plausibly an ultimate (omniscient), divine mind. Hence, it is a hopeless non sequitur to infer that Johnson’s argument shows that “it’s not even logically possible for ‘god’ to ‘cause’ the existence of. . . causality. . . If there is any sort of connection between ‘god’ and ‘the universe,’ it’s not a causal relationship,” as Cline suggests. Rather, Johnson’s argument only shows that God cannot be a causal explanation of the laws of causality, which is radically different than concluding God cannot be a causal explanation of the universe.

Ergo, the arguments of Robin Le Poidevin and B.C. Johnson to which Cline alludes provide no good grounds for believing God cannot be used as a causal explanation of the universe. Cline mistakenly thought Le Poidevin’s and Johnson’s arguments were similar, but instead they implicitly contradict one another. Le Poidevin’s argument, properly understood, should be rejected on several accounts. First, it begs the question by assuming that only the physical order of things merit causal explanation. Second, to avoid this first problem, Le Poidevin uses the phrase ‘laws of nature’ instead of ‘physical laws,’ and therefore commits the fallacy of ambiguity in that such a move equivocates the phrase ‘laws of nature.’ Third, the argument is greeted with hardship by assuming the universe is either infinite or finite: if the former, Le Poidevin is faced with affirming the existence of an actual infinite and the various absurdities that follow. If the latter, we are given a Leibnizian cosmological argument for the existence of God. The argument is also self-refuting, for premise (1) proves to be an exception to the argument’s key premise, (3). Lastly, the argument’s main assumption, that only Hempelian causation yields explanation, is false. Johnson’s argument, on the other hand, while logically valid and in my opinion sound, proves itself difficult not for the theist, but rather the atheist; for its main assumption, that there are laws which are uncaused and necessarily existent, can be used as a key premise in an argument to establish the existence of God.
__________________________________________________________________________________________

Endnotes

1. Robin Le Poidevin, Arguing for Atheism: An Introduction to the Philosophy of Religion (Routledge: New York, 2000).
2. There are currently three views regarding natural law: the regularity view, the nomic necessity view, and the causal dispositions view. The lattermost cannot be what Le Poidevin has in mind, for on this view natural laws are metaphysically necessary. We’ve already seen that the natural laws Le Poidevin assumes cannot be metaphysically necessary ones, for that which is metaphysically necessary isn’t constrained by this universe. But Le Poidevin thinks said natural laws are constrained by this universe. So he must be referring to either Hume’s regularity view or the nomic necessity view (or some variant thereof), both of which take natural law to be inductive generalizations—which is precisely how Le Poidevin reads. On natural law, see Alfred J. Freddoso, “The Necessity of Nature,” Midwest Studies in Philosophy 11 (1986): 215-42.
3. This idea isn’t so new. David Lewis argues that the laws of nature are supervenient features of the universe, being the product of a certain spatiotemporal arrangement of intrinsic properties thereof. See David Lewis, ‘A Subjectivist’s Guide to Objective Chance,’ Postscript, in Philosophical Papers (Oxford University Press, 1986), ii. 122.
4. Richard Swinburne, The Existence of God (Oxford University Press: New York, 2004), pp. 24-25.
5. For a general overview, see C. G. Hempel, Philosophy of Natural Science (Prentice-Hall, 1966), ch. 5.
6. William Lane Craig presented the argument this way in his debate with Quentin Smith. See the Craig-Smith Debate: Does God Exist? (Harvard University April, 2003). See also Craig, The Cosmological argument from Plato to Leibniz (Wipf and Stock: Oregon, 2001) pp.257-281.
7. Swinburne calls this personal explanation. See Swinburne, The Existence of God (Oxford University Press: New York, 2004), chs. 2-4, particularly ch. 2 on the nature of this kind of causation and ch. 3 for its justification.
8. I do not mean to say that acts of free agents can’t be inductively probabilified, but that they cannot be physically probabilified. The distinction here is between inductive and physical probability. Richard Swinburne explains the latter: “The physical (or natural) probability of an event (and so of the proposition that records it) is a matter of the extent to which at some earlier time the event is predetermined by its causes.” Swinburne, The Existence of God (Oxford University Press: New York, 2004), p. 15. Inductive probability, on the other hand, is concerned with the epistemic likelihood of q given p where p is some evidence and q is some hypothesis. To further contrast, p could itself represent some physical probability which constitutes evidence that makes q a more inductively probable hypothesis.
9. William Lane Craig, “Must the Beginning of the Universe Have a Personal Cause?: A Rejoinder.” Faith and Philosophy 19 (2002): 94-105.
10. B.C. Johnson, The Atheist Debater’s Handbook (Prometheus Books; New Ed., 1983).