Home » Library » Modern Library » Contra Carrier: Why Theism is Needed to Make Sense of Everything

Contra Carrier: Why Theism is Needed to Make Sense of Everything


No necessary being can explain existence; contingency is not an illusion, an appearance which can be dissipated; it is absolute; and consequently perfectly gratuitous. Everything is gratuitous, that park, this town, and myself. When you realize that, it turns your stomach over and everything starts floating about.

     — John-Paul Sartre, Nausea (1938), p. 188

That there is a contingent being actually existing has to be discovered by experience, and the proposition that there is a contingent being is certainly not an analytic proposition, though once you know, I should maintain, that there is a contingent being, it follows of necessity that there is a necessary being.

     — Father Joseph Copleston, debate with Bertrand Russell, BBC Radio (1948)

Arguments for God’s existence often take the form of an inference to the best explanation, the pattern of reasoning Charles Sanders Pierce called “abduction.” An inference to the best explanation (IBE) is an inductive argument having the following logical form:

  1. The argument begins with facts that need explaining.
  2. Ideally, all plausible explanations of the facts are examined.
  3. One explanation is selected as the “best” explanation, based on the standard criteria used in science, history, and academic fields generally (internal consistency, external consistency, explanatory power, scope, and so on).
  4. It is concluded that this is probably the correct explanation of the cited facts. In other words, the selected explanation is probably true.[1]

A “theistic” IBE claims that, of all available explanations, a theistic one is the best or most reasonable explanation of the data, based on the standard criteria. But if an explanation is correct, then the explanatory entities invoked by the explanation exist. Thus, a theistic IBE is an inductive argument for the existence of God insofar as it includes at least one reference to God. More generally, I shall call an argument a “theistic explanatory argument” if it posits the existence of God as part of an explanation of observed phenomena. (Thus, for example, ontological arguments are not theistic explanatory arguments because they do not aim to explain any observable phenomena.)

Richard Carrier’s General Critique of Theism

In “Ten Things Wrong with Cosmological Creationism” (2000), Richard Carrier offers a general argument against any theistic explanatory argument:

If we posit a god, we would still be in the same position of trying to explain his existence. Anything sufficient to explain a god’s existence will also apply to the universe, or at least the physical principles of the universe which give it form and motion, and thus there is no reason for the added element of a god.

Fleshed out, I take Carrier’s argument to consist in the following steps:

  1. Any theistic explanation of the universe invites the question: “So, the creative activity of a god is proposed as an explanation for the existence of the universe, but what accounts for, or explains, this god’s existence?”
  2. Either the theist has an answer or he does not.
  3. Suppose the theist has no answer. If the theist has no answer, then theists and atheists are in the same position vis-à-vis explaining the whole of reality: The theist has no explanation for his rock-bottom reality, God’s existence, while the atheist has no explanation for his ultimate reality, namely, the existence of the material universe. In which case theism has no explanatory advantage over scientific materialism, since both views ultimately invoke brute fact in the end. This outcome leaves theism in the awkward position of asking us to believe in additional entities while offering no explanatory advantage in return. Thus, if the theist has no answer to Carrier’s question–“What explains or accounts for God’s existence?”–then there’s no rational motivation to accept the theistic explanation of the universe (along with its extra ontological commitments).
  4. Suppose the theist does have an answer. If the theist does have an answer, i.e., an account or explanation for God’s existence, then this explanation can be logically detached from its subject and applied directly to the material universe alone, without mentioning God, thus eliminating any rational need to refer to God in our explanation of the universe.
  5. Thus, either way, a theistic explanation is an unnecessary explanatory epicycle–a spinning wheel doing no useful explanatory work. When it comes to explaining the universe, theism gets us nowhere. In short, any theistic explanation will inevitably be superfluous and therefore rationally unacceptable.

Carrier’s objection, in one form or another, has become almost an obligatory reply when a theistic explanatory argument is presented. Nevertheless, I am convinced that it is fundamentally unsound. My completely unrealistic aim in this essay is to put this little reply or objection to rest once and for all.

Against Carrier, I shall argue for three points. (1) Theists have available a reasonable reply to the question, “If the creative activity of God is proposed as an explanation for the existence of the universe, what accounts for, or explains, God’s existence?” (2) Moreover, it would be a category mistake to take the theist’s reply out of its theistic context and apply it directly to the material universe, with no mention of God. That is, the theist’s reply cannot provide a nontheistic explanation for the existence of the universe because it logically requires a being possessing at least one essential characteristic that necessarily cannot be possessed by any material object or collection of material objects. (3) Since the theist’s ultimate explanation necessarily cannot be applied to the material universe without mention of God, theism offers an explanatory advantage over its chief rival, scientific naturalism, collapsing a crucial premise of Carrier’s argument.

Theistic Explanation

In the history of Western philosophy, theistic explanatory arguments appear in great variety and with regularity, going back at least to Anaxagoras, Plato, and Aristotle, and continuing to the present in the works of such widely respected philosophers as Richard Swinburne and Robert M. Adams. My focus in this paper will be the modal cosmological argument, perhaps the most prominent philosophical case for traditional theism, although I could have chosen at least ten other theistic arguments. Let us keep in mind that this venerable argument for God’s existence stems from two lines of thought, both of which began within classical Greek philosophy before the Christian era. It is also worth remarking that both lines of thought originated apart from any specifically theistic context, and both can be motivated apart from theistic considerations.

The first is the modal distinction between necessary and contingent. Essentially, necessary means “cannot be otherwise,” while contingent means “can be otherwise.” This distinction has been applied to both truth and being, although I shall not rehearse the matter here. The pre-Socratics were the first to give philosophical expression to this modal distinction, but Aristotle was the first to develop it in a theoretical and systematic way.

With the work of modern theorists such as Saul Kripke, David Lewis, Alvin Plantinga, and Robert Stalnaker, the logic of necessity and contingency has come of age, possessing, I believe, an eminently respectable semantical foundation. Given the success of contemporary modal logic, I take the distinction between the necessary and the contingent seriously, as applied to both truth and falsity (in classical modal logic) and existence and nonexistence (in quantified modal logic).

The second line of thought, which goes back at least to Parmenides and Anaxagoras, was expressed nicely by Sir Isaiah Berlin in a famous essay in which the Oxford don described the phenomenology of coming to know and accept necessary truth:

If I am a schoolboy, all but the simplest truths of mathematics obtrude themselves as obstacles to the free functioning of my mind, as theorems whose necessity I do not understand; they are pronounced to be true by some external authority, and present themselves to me as foreign bodies which I am expected to mechanically absorb into my system. But when I understand the functions of the symbols, the axioms…. [and] the logic whereby the conclusions are obtained–and grasp that these things cannot be otherwise, because they appear to follow from the laws that govern the processes of my own reason, then mathematical truths no longer obtrude themselves as external entities forced upon me which I now must receive whether I want to or not, but as something which I now freely will in the course of the natural functioning of my own rational activity…. To want necessary laws to be other than they are is to be prey to an irrational desire–a desire that what must be X should also be not X. To go further, and believe these laws to be other than what they necessarily are is to be insane. That is the metaphysical heart of rationalism…. I am a rational being; whatever I can demonstrate to myself as being necessary, as incapable of being otherwise in a rational society–that is, in a society directed by rational minds, towards goals such as a rational being would have–I cannot, being rational, wish to sweep out of the way. I assimilate it into my substance as I do the laws of logic, of mathematics…. and therefore will the rational purpose, by which I can never be thwarted, since I cannot want it to be other than it is.[2]

Gloss: In philosophy, we explain one thing in terms of another, or one level of reality in terms of a more fundamental level, which in turn is explained in terms of something more fundamental, and so on. In other words, explanation usually takes the form of an explanatory regress. Normally, a regress starts with a contingent object or state of affairs and then proceeds back through various stages that may or may not also be contingent. The ultimate philosophical goal is explanatory closure: a final explanation that ends the questioning in a rationally acceptable way. In the passage above, Berlin describes what explanatory closure must be like for a rational being: Only a rational grasp and acceptance of necessity can finally end an explanatory regress. Once we grasp the nature of necessity as rational beings, we give up as irrational the desire to “sweep it away” or have it be otherwise, realizing that certain aspects of reality “cannot be otherwise than they are.” Along with the laws of logic and mathematics, we seek to “assimilate” the necessary into our substance. In so doing, we freely accept and “will” (presumably in a Kantian sense) the necessary. After experiencing and willing necessity, no explanatory regress ending in a purely contingent state of affairs will ever be rationally satisfying. Only when a regress “bottoms out” in that which is metaphysically necessary will it reach a rationally acceptable conclusion. I shall call this the “modal grounding principle.”

Berlin’s reflection reminds us that every philosophical explanatory regress has a gas pedal and needs a brake. The gas pedal is the initial question that starts the regress. (Regarding the existence of the universe, the question is: “Why does it exist?”) The brake, if there is one, must be a real “stopper,” that is, a final explanation that stops the regress by ending the questioning. According to the modal grounding principle, only a necessary ground of being can finally serve as a rationally acceptable brake in a philosophical regress.

The Modal Version of the Cosmological Argument

Let us now see how these two ideas, the distinction between the necessary and the contingent, and the modal grounding principle, support traditional theism and undermine Carrier’s argument. I begin with a statement of the modal cosmological argument in three steps.

1. The Contingent Universe

By “the universe” I mean “the material universe,” the complete collection of all existing things composed of matter. By “matter” I mean “that which physics studies.” Thus, the material universe is a large collection of particles and fields composed of, or reducible to, quanta of mass-energy existing within space and time. I also take it that each of the particles and fields composing the material universe is contingent, since current physical theory holds that each of these items came into existence at a moment in time in the past, and anything that came into existence is certainly contingent.

Incidentally, whether or not the latest inflationary Big Bang model is the correct account of the birth of the universe, the mere fact that this model appears to be logically coherent is sufficient to ground the claim that matter is contingent. Since (a) the current model is a coherent scientific theory which posits that matter had a beginning in time, and thus (b) it is reasonable to suppose that it is at least possible that matter had a beginning in time, (c) it is therefore reasonable to suppose that matter is contingent–for anything that could have come into existence in time is contingent.

Of course, there are other reasons to suppose that the physical universe is contingent. As Richard Taylor observes in his classic text Metaphysics, there is absolutely nothing about the physical universe that, upon critical examination, seems to be in the least bit necessary.[3] Furthermore, according to the latest theories in scientific cosmology, each thing within the universe either has a cause or at least originates within a quantum field; and if something has a cause or a scientifically explicable origination, it is certainly contingent. Thus, viewed through a scientific lense, the universe certainly appears to be contingent through and through. And it is reasonable to suppose that things are as they appear unless there are good reasons to suppose that they are not.

2. Wolfe’s “Question of Questions”

One of the oldest philosophical questions of all time is surely “Why is there something rather than nothing?” In other words, what accounts for the fact that a material universe exists? The historian Bertram Wolfe (1896-1977) wonderfully expressed this question in his autobiography, A Life in Two Centuries:

There was one large problem that I tackled that I could not solve at 11 or 12 or 13, and have not solved yet, nor do I expect to: Why does the universe exist at all, and why is there life on the earth and perhaps elsewhere? I tried to imagine what would be if the universe did not exist…. The more I wrestled with such problems the less I could explain to myself, until at last I was driven to the question of questions: Why is there something? Why is there anything? Why is there not nothing? How did it all come to be…? And there my mind stopped as a boy, and there my mind stops now in the closing years of my life.[4]

3. Grounding the Regress

By the modal grounding principle, it just won’t do to explain one part of the material universe in terms of some more fundamental material part. For that would only explain the contingent in terms of the contingent, leaving the explanatory regress at rock bottom poised precariously upon a mere contingency–and thus ultimately ungrounded and rationally incomplete. To stop with the contingent is to stop thinking too soon. Only if the regress terminates in a metaphysically necessary ground of being can it reach a rationally acceptable end.

Incidentally, since a purely scientific explanation must always refer to or invoke only entities and laws that are part of the material universe, i.e., objects that exist within the whole, it follows that any scientific explanation merely explains one contingent part of the material universe in terms of another contingent part. Thus, it follows (by the modal grounding principle) that no purely scientific explanation can ever end the cosmological explanatory regress in a rationally acceptable way. In other words, from a philosophical view, and supposing the modal grounding principle, a purely scientific explanation of the universe “stops too soon.”

Given this, it looks like the only rationally acceptable end to the cosmological regress, and thus the only rationally acceptable explanation for the existence of the contingent universe, would be a metaphysically necessary ground of being. Thus a theistic IBE:

  1. A vast contingent universe exists.
  2. The only rationally acceptable explanation for the existence of this enormous entity or collection of entities is theism, i.e., the claim that a metaphysically necessary ground of being exists and is responsible for the existence of the universe. (Following Ockham’s razor, we postulate the bare minimum needed to explain the phenomenon, which in this case means positing just one necessary being; to posit any larger number would be to invoke unnecessary explanatory entities.)
  3. Therefore, it is reasonable to suppose that a necessary being exists and is responsible for the existence of the universe.

I interpret this as a modal cosmological argument. Of course, strictly speaking, a further argument is needed if one is constructing a philosophical case for theism, namely, an argument attributing characteristics of divinity to the necessary being and thus showing why this is a theistic argument. For example, after he concludes his famous Five Ways, Aquinas goes on to give a separate argument for the conclusion that the First Mover must have the characteristics traditionally ascribed to God.

Thus, after presenting a modal cosmological argument, a philosophical theist needs to argue further that (a) the necessary being must, insofar as it is necessary, have an existence apart from the material universe since (b) the universe is contingent, not necessary. And an argument would also be wanted for the claim that the necessary being is personal in the sense of possessing knowledge, reason, and a will. With such secondary arguments added, the theist, it seems to me, has a strong philosophical case.

Of course, for every philosophical argument, questions arise, along with objections and counterarguments. (This statement may be the only exceptionless yet true and nontrivial generalization concerning the philosophical enterprise!) At any rate, upon hearing the above argument, the first question likely to arise is this: Why accept the modal grounding principle?

In Defense of the Modal Grounding Principle

Suppose a philosophical regress ends in something contingent, call it C. By its very nature, C has the possibility of nonexistence; that is, among the set of all logically possible states of affairs, there is the possibility that C does not exist. (Following Plantinga, I mean possibility in the broadest sense.) For if there was no possibility of C’s nonexistence, then C would be necessary, not contingent.

But the possibility of C’s nonexistence logically grounds the question “Why does C exist?” And if C is contingent, the explanation of C’s existence will refer to external circumstances or entities that caused C to exist. (Thus, if someone asks why that large tree over there exists, the explanation will refer to a seed that took root some 50 years ago, and if someone asks why that seed existed, the explanation will refer to a tree that existed before it, and so on.) The explanation of C cannot be completely internal to C; otherwise, C would be ontologically independent and thus necessary, not contingent.

If a philosophical regress bottoms out in something purely contingent, then live questions pertaining to existence at the bottom level of being will remain unanswered (since the final step in the regress will, by its intrinsic nature, raise questions that could only be answered by appeal to a deeper level). Thus, no purely contingent entity can ever serve as an adequate “regress brake” or rationally acceptable final explanation.

Peter van Inwagen has identified a general problem with any regress consisting of nothing but contingent entities. Imagine an explanatory regress composed of contingent beings alone, with no reference whatsoever to a necessary being. Any statement that would be true in such a possible world would be true in virtue of relations obtaining among various contingent beings. A statement’s truth would therefore depend on the existence of contingent beings, logically presupposing their existence. But, as van Inwagen observes, surely no statement presupposing the existence of contingent beings, and depending on their existence for its truth-value, could itself explain or account for the fact that contingent beings exist.[5]

Van Inwagen’s point is supported by the practice of philosophical explanation in general, for when we explain the existence of the members of one class, we normally must appeal to something completely outside that class. For instance, to explain the existence of cars, we must appeal to something beyond cars themselves (i.e., to engineers); to explain the existence of human beings, we must appeal to something nonhuman; and so on. To continue the pattern, it would seem that if we are to explain the general existence of contingent being, an appeal to the existence of noncontingent being would be the natural move.

On the other hand, suppose a philosophical regress ends in a metaphysically necessary ground of being. Does this endpoint, by its intrinsic nature, also raise its own unanswerable questions–for instance about its own existence–thus also precluding a rationally acceptable closure? If so, wouldn’t it follow that we have no hope of ever attaining a rationally acceptable explanatory endpoint?

To answer this, recall that a necessary being, by hypothesis, is a being whose intrinsic nature is such that its nonexistence is impossible. This means that in the case of (and only in the case of) a necessary being, there is necessarily a logically sufficient explanation for its existence. As van Inwagen has so succinctly put it: “There could hardly be a more satisfying explanation for the existence of a thing than that its non-existence was impossible.”[6]

The following thought experiment motivates the point. Suppose you go to a friend’s house to watch TV, and all evening the only channel you see is ABC. Finally, around 11 PM, you ask, “Why must we watch ABC all night?” Your friend answers, “This is a screwy television–no matter what channel you turn the dial to, it only shows ABC, it can’t show any other channel.” Hasn’t your friend provided a perfectly acceptable answer to the question? Now you understand why the TV shows ABC: It shows ABC because it cannot possibly show any other channel. In other words, it actually shows ABC because it necessarily must show it. (This is a case of physical necessity, but that is no matter.)

I think the general principle is the following: If someone is wondering why something is actually so and not otherwise, and they learn that it is necessarily so and cannot possibly be otherwise, then they have a rationally acceptable explanation for why it is actually so.

Notice that the explanation for the existence of a necessary being, as expressed by van Inwagen, does not refer to any other being, circumstances external to the being, or another ontological level of being. If this is right, then an intrinsically necessary ground of being at the end of an explanatory regress, unlike a contingent one, raises no unanswerable questions of existence and therefore constitutes a rationally acceptable regress brake–a rationally satisfying closure for an explanatory regression.

Richard Carrier’s Objections

Carrier writes that “If we posit a god, we would still be in the same position of trying to explain his existence.” Perhaps so, but the theist has a rational reply available–namely, the reply articulated and argued for by van Inwagen. And this reply is an eminently reasonable one–or so it seems to me. Carrier then argues:

Anything sufficient to explain a god’s existence will also apply to the universe, or at least the physical principles of the universe which give it form and motion, and thus there is no reason for the added element of a god.

But this is false. The theist’s reply cannot be applied directly to the material universe, for the universe is contingent. Any attempt to do so would commit a category mistake, as the reply only works in reference to that which is metaphysically necessary.

How Would Carrier Reply?

I think we can see how Carrier would reply to my argument. Later in his piece, in response to an argument from design, Carrier writes:

But there is a third option that Mr. Walker omits, for intelligent creation is not the only possible cause for order. The universe may be a necessary thing–that is, it was not produced by chance or design, but could not have not existed, and could not have been any different than it is.

In other words, in place of the theist’s necessary being (God), Carrier would simply posit that the material universe is itself the necessary being.

Unfortunately, Carrier’s alternative hypothesis simply won’t do, for a number of reasons. First, it is completely ad hoc: There is absolutely nothing about the material universe that seems in the least bit metaphysically necessary, and Carrier offers no independent, positive, theoretical reason to support positing the necessity of the universe. Absent an independent reason for attributing necessity to the cosmos, the hypothesis of a necessary universe is for Carrier nothing more than a way to avoid an unwanted (theistic) conclusion.

Second, there are strong theoretical reasons to suppose that the necessary being at the end of the cosmological explanatory regress, if such a being were to exist, would have to be metaphysically simple, i.e., noncomposite. For instance, it would seem that any composite being depends for its existence on its parts, or on the condition that its parts stay together, which would make it contingent rather than necessary.

Moreover, any composite being, by its intrinsic nature, inevitably raises further questions. How did its parts come together? Why those parts and not others? Why that many parts and not some other number? What holds the parts together? That the above questions logically arise in the case of any composite entity suggests that nothing composite can serve as a rationally satisfying regress stopper. Therefore, since the physical universe is composed of something like 10^56 protons and like numbers of electrons, photons, and other subatomic particles, and since the whole system is governed by an enormously complex set of mathematical laws of nature, the material universe hardly has the look of a metaphysical simple. Thus there are strong philosophical reasons, independent of theism, to suppose that the material universe is not itself a necessary being.

Furthermore, there are scientific reasons to suppose that the physical universe is contingent. As I have already observed, current cosmology tells us that the physical universe began to exist a finite time ago. If contemporary science is right, the physical universe is not a necessary being, since a necessary being cannot possibly be something that came into existence. Notice that Stephen Hawking and Steven Weinberg, two important authorities on the matter, strongly affirm the contingency of the universe in their discussions of final theories in physics.[7]

I conclude that Carrier has no reason to posit the necessity of the material universe other than to avoid a theistic explanation.

Further Objections

I will now turn to an extremely interesting objection raised by an anonymous reviewer during this paper’s review process. The reviewer’s objection is not only very powerful, but also very penetrating, cutting to the heart of the matter while opening up new philosophical issues that are interesting and illuminating in their own right. I shall slightly paraphrase the reviewer’s argument as follows.

The modal grounding principle is essential to this paper’s argument: “No explanatory regress ending in a purely contingent state of affairs is rationally satisfying.” However, it is a small step from acceptance of this principle to the conclusion that there are no contingent states of affairs. If all contingent states of affairs are “grounded” in a necessary being or a necessary state of affairs, such that the “grounding” is itself necessary (e.g., if a necessary being’s act of creation is itself always necessary), then there are no contingent states of affairs. Thus, if the universe is explanatorily grounded in a necessary being, then it follows that–if there are to be any contingent states of affairs at all–the act of creation itself can only be contingent. But if the act of creation is itself contingent, then there exists at least one brute (unexplained) contingency. Therefore, the modal grounding principle is false: unless we suppose that everything is necessary, we cannot reasonably deny the existence of at least some brute contingency. Carrier’s line of thought therefore survives unscathed: Given that we must have brute contingency, why not rest with the (brutely contingent) origin of the physical universe? In other words, why take the unnecessary additional step of hypothesizing brutely contingent creative preferences in some postulated necessary being?

Essentially, the reviewer poses a dilemma for the theist: Either the act of creation is necessary, in which case the theist must deny the contingency of the universe (since what follows from the necessary is itself necessary), or else the act of creation is contingent, in which case the theist is ultimately left with brute contingency, and any appeal to a necessary being accomplishes nothing. Either way, the modal cosmological argument fails by losing an essential premise–namely, the modal grounding principle.

Although this is a very strong objection that raises interesting and important issues, I believe that the theist has a philosophically satisfactory reply. First, while it may be a “very small step” from the modal grounding principle to the claim that there are no contingent states of affairs, it is a step that the theist need not (and does not) take. As the reviewer observes, if a necessary being’s act of creation is itself necessary, then there are no contingent states of affairs. Classical theism recognizes and accepts this point, which is partly why the tradition has held that the act of creation was not itself necessary. However, the theist’s admission that the act of creation was itself non-necessary does not entail that the theist (like the atheist) is saddled with brute or unexplained contingency. More specifically, it does not entail that the theist is stuck “hypothesizing brutely contingent creative preferences in some postulated necessary being.” Here is why.

Classical theism hypothesizes that God’s decision to create a (material) universe was free in the sense that it was both uncaused and not logically necessary. It follows from this that the existence of the universe is metaphysically contingent: the universe would not have existed had things (in the broadest possible sense) been sufficiently different. However, it does not follow from this that the existence of the universe is, for the theist, a brute or unexplained fact. The theist can explain the existence of the universe by relating it, via explanatory relations we already understand, to a necessary feature of the divine nature–namely, that God is an essentially loving being. This explanatory relationship, which grounds the contingent existence of the universe in a necessary feature of a necessary being (thus satisfying the modal grounding principle), can be modeled in the following terms.

It is the inherent nature of love to freely share itself, to freely go out to another, in other words, to give of itself in a supererogatory act–one that is not compelled or motivated by self-interest, but performed freely and for the sake of the intrinsic good of the other. This is a necessary truth about the intrinsic nature of love that each of us comes to understand through our own direct experience of love. An act is not a purely loving act if it is done out of selfishness, compulsion, or duty; an act of love is by its nature freely undertaken for the sake of the intrinsic good of the other. The theist thus hypothesizes that God freely creates separate beings, out of love, in order to freely share what God intrinsically is–i.e., to share being with separately existing (finite) beings.

For explanatory purposes, theism need not propose God’s creative love as the cause of the universe; it need only hypothesize love as the motive explaining why God creates. But it does not follow from this that the act of creation is a brute, unintelligible act; for freely given acts of love, though neither caused nor necessitated, are inherently intelligible. Indeed, explanations by reference to acts of freely given love are among the most intelligible and intellectually satisfying accounts of things that we possess. That Ann and Bob married out of love, rather than out of self-interest or compulsion or because their marriage was arranged by others, explains why they married and renders their marriage intelligible. Their free choice in this case is a perfectly good “explanation stopper.” Similarly, the theist can argue, the contingent existence of the universe is “grounded” in a necessary being in virtue of being explained and made intelligible, via explanatory relationships we already understand, by reference to a necessary feature of the divine nature, namely, supreme, creative love. In this way, the theist can avoid both horns of the reviewer’s dilemma: the theist avoids being saddled with rock-bottom brute contingent facts while holding fast to the contingency of the universe.

Thus, it is not the case that the existence of the material universe can only be explained or rendered intelligible by showing that it follows necessarily from a necessary act of a necessary being (which would make the universe necessary). The theist has available an alternative explanatory model, rooted in our direct experience of love, that explanatorily grounds the contingent existence of the universe in a metaphysically necessary being without destroying the contingency of the universe. This explanation in turn satisfies the intuitively very appealing modal grounding principle. This model or explanatory structure is not available to the atheist, of course, since atheism rejects both a metaphysically necessary ground of being and (more importantly) a personal source of the existence of the material universe. The theist is therefore not saddled with the brute unexplained fact that a contingent universe exists, although, by the reviewer’s own admission, the atheist is. It is therefore a mistake, I think, to claim that the theist must hypothesize “brutely contingent creative preferences in some postulated necessary being.” The objection thus fails to show that the theist’s appeal to a necessary being is, in the end, superfluous.

A Reply to This

The wonderful thing about the reviewer’s argument is the way it opens up deeper philosophical issues regarding the differences between theistic and nontheistic views of the universe. In response to the above reply, the reviewer posed a very interesting and penetrating counterargument, which I here quote:

On Herrick’s own admission, there are possible worlds in which God makes a different universe from the one that God makes in our world. So, on his own principles, there must be something to which one can appeal that explains why God makes our universe in our world, and that other universe in the other possible world. Given that, again on Herrick’s own admission, it is God’s creative intentions, etc., that determine which world is made, there must be a difference in God’s creative intentions between the two worlds. Either this difference in creative intentions is a brute contingent difference, or there is something that explains why God has one set of creative intentions in one world, and a different set of creative intentions in the other. Moreover, if there is something that explains why God has one set of creative intentions in one world, and a different set of creative intentions in the other, then we’re back on the same merry-go-round: either this in turn is a brute contingent difference, or [it is necessary and there really are no contingent facts].

So, we have a choice. Given that we hold onto the assumption that the existence of the universe is contingent, either we suppose that the coming into existence of the universe is a brute, contingent fact, or we suppose that there are brute contingent facts–concerning God’s creative intentions, or something of that kind–that explain why God made this world, rather than some other world that God might have made. Hence, in point of avoiding brute contingencies, there is no advantage in postulating God as the explainer of the existence of the universe. Moreover, when it comes to a choice between supposing that there is a necessarily existing being with brutely contingent creative intentions or supposing that the coming into existence of the universe is a brute contingent fact, there is no semblance of an argument in Herrick’s letter (or his paper) that favors the latter choice over the former.

The reviewer is raising here, in regard to God’s creative intentions, essentially the same questions that critics of agent causation raise concerning the adequacy of that particular conception of free agency. A satisfactory reply would therefore require entering that enormously complex issue. But the reviewer’s new objection again raises another important issue: At bottom, when the replies and counter-replies are all in, the modal cosmological argument, as I have defended it, depends in an important way on the conceptual adequacy of a libertarian account of free will. Thus, the modal cosmological argument, as I have defended it, is connected at a deep level with the issue of free will, which is surprising and illuminating.

However, to enter that well-known debate here would be to go beyond the scope of this paper; rather, I will press a simple point. In everyday life, one “why” question sometimes leads to another, which leads to another, and so on, generating an explanatory regress. It is perfectly reasonable, and it can be intellectually satisfying, to end such an explanatory regress with reference to the free choice of an agent, provided that the choice is informed and morally autonomous; and if that condition is met, then the choice, as an act of origination, is not brute (although it is contingent). In the end, when all is said and done, that Ann and Bob ultimately chose to marry, that their choice was fully informed and morally autonomous, and that they married out of love, ends the regress of questions by grounding it in an intellectually satisfying and reasonable stopping point. Granted, it is a contingent fact that they chose to marry, but it is not, on that account, a brute fact. A brute contingent fact has no explanation, whereas a free, morally autonomous act of the type under consideration does have an explanation, albeit one that is not deterministic. Love, as a motive, can make an action intelligible without making it metaphysically necessary. The more general point is that reason-explanations that cite motives are sometimes perfectly acceptable stoppers of explanatory regresses. I am appealing here to our everyday explanatory practices, which, I would argue, should guide us unless and until they are shown to be based in error or confusion.

Thus, I concede that “there are possible worlds in which God makes a different universe from the one that God makes in our world.” However, I deny that “there must be something to which one can appeal that explains why God makes our universe in our world, and that other universe in the other possible world.” This does not follow, at least if we accept the legitimacy of our ordinary explanatory practices with respect to informed and morally autonomous choices of agents. I thus resist the contention that the theist is saddled with “brute contingent facts concerning God’s creative intentions.”

I am sure that this reply will not be satisfying to everyone. The following concern, expressed by the reviewer in the final round of comments, will probably continue to trouble many who resist the conclusion of the cosmological argument. As the reviewer observes, “on a libertarian account of freedom, if X chooses A rather than B, then there are two possible worlds exactly alike up until the moment of choice, but in one of which X chooses A, and in the other of which X chooses B…. [in which case] therefore, there is no explanation of why X chooses A rather than B.” And this in turn, the reviewer concludes, ultimately leaves the theist saddled with brute contrastive contingent facts.

And so we stop at what may be very near the heart of the matter: That which the theist accepts as a rationally intelligible and satisfactory ender of an explanatory regress, both in daily life and in metaphysics–namely, the free but purposive choice of a rational agent. In other words, what some see as a reason-explanation citing a motive, others see as simply another brute contingent fact. As the reviewer observes, continuing this argument would require entering the free will issue and the debate between compatibilist and libertarian analyses of agency. (Isn’t it always the case that if one philosophical issue is pressed far enough, it eventually leads into every other philosophical issue?) And so the discussion continues.

Steven Weinberg on the Contingency of the Universe

Before drawing this discussion to a close, if I may be permitted an argument from authority, it appears that Steven Weinberg, one of the world’s greatest living physicists, agrees with at least a major strand of the argument I am presenting. In “A Designer Universe?” Weinberg writes:

I have to admit that, even when physicists will have gone as far as they can go, when we have a final theory, we will not have a completely satisfying picture of the world, because we will still be left with the question “Why?” Why this theory, rather than some other theory? For example, why is the world described by quantum mechanics? Quantum mechanics is the one part of our present physics that is likely to survive intact in any future theory, but there is nothing logically inevitable about quantum mechanics; I can imagine a universe governed by Newtonian mechanics instead. So there seems to be an irreducible mystery that science will not eliminate.[8] [emphasis mine]

Thus, after a lifetime of work in theoretical physics, and after a Nobel Prize for unifying the electromagnetic and the weak nuclear forces, Weinberg concludes that the universe appears to be, from a purely scientific standpoint, irreducibly contingent. A final, completed scientific theory of everything, in Weinberg’s judgment, would still leave the universe partly unexplained, for it would leave unexplained why there is a contingent universe governed by this complete scientific theory rather than by another.

I think Weinberg is giving expression here to a sense of incompleteness or intellectual dissatisfaction felt by many people when they consider the prospect of a completed scientific account of the existence of the universe. It is this same sense of incompleteness, I believe, that drives theistic philosophers beyond the scientific account of the universe to a philosophical one, and finally to the conclusion that there must exist a metaphysically necessary ground of being, which some may call God.


Carrier’s argument against theism fails. At least one rational case for theism, the modal cosmological argument, has its roots in the general philosophical tradition, specifically in two major philosophical ideas that originated within Greek philosophy, before the Christian era, apart from theistic considerations, both of which can be motivated apart from theistic considerations. Carrier’s argument does not place a dent in either.

Ever since the first philosophical theory was put forward by Thales of Miletus in the early sixth century BC, philosophers have sought a rational explanation of the world. The modal cosmological argument, with its roots in pre-Socratic thought, gives expression to this explanatory demand and provides rational support for the theistic claim that the universe ultimately makes sense only if it is rooted in a metaphysically necessary ground of being. Thus L. P. Gerson, a scholar of classical Greek thought, has noted:

That which explains the existence of things can’t be something whose existence also needs explaining. So, all the major Greek philosophers conceived of God as a being in some sense necessary–not just another thing that might not have existed–hence not in itself in need of further explanation.[9]

Recall the question from his childhood that haunted the historian Bertram Wolfe: “Why is there something rather than nothing?” While it is remarkable that a 12-year-old boy would feel the force of, and wrestle with, one of the greatest questions in philosophy, it is also remarkable that there is an answer staring us in the face, for as van Inwagen has noted:

Why should there be anything at all?… If we could show that there was a necessary being, a necessarily existing individual thing, we should have an answer to our question. For if there were a necessary being, then it would be impossible for there to be nothing. And if we could show that it was impossible for there to be nothing, that, surely, would count as an answer to our question.[10]

In other words, whether or not one believes that a necessary being exists, a necessary being hypothesis does indeed account for the extremely basic fact that there is something rather than nothing. Which is to say that theism at least offers a solution to one of the oldest problems in philosophy.

To anyone trained in the Western philosophical tradition, classical theism should be attractive for the same reasons that any philosophical theory is attractive: It answers questions, accounts for phenomena, solves philosophical problems, and explains things about our world that would otherwise remain unexplained. Why else do we hold theories in philosophy? Why else do we find philosophical theories reasonable?[11]


[1] For more on inference to the best explanation, see Paul Herrick, The Many Worlds of Logic, 2nd edition (Oxford University Press, 1999), chapter 25.

[2] Isaiah Berlin, Four Essays on Liberty (Oxford: Oxford University Press, 1969), p. 144.

[3] Richard Taylor, Metaphysics (Englewood Cliffs, New Jersey: Prentice Hall, 1974), p. 110.

[4] Bertram Wolfe, A Life in Two Centuries (New York: Stein and Day Publishers, 1981), p. 82.

[5] Peter van Inwagen, Metaphysics (San Francisco: Westview Press, 1993), p. 103.

[6] Peter van Inwagen, Metaphysics (San Francisco: Westview Press, 1993), p. 102.

[7] See Stephen W. Hawking, A Brief History of Time (New York: Bantam Books, 1988), p. 174, and Steven Weinberg, Dreams of a Final Theory (New York: Vintage Books, 1994), p. 6.

[8] See Steven Weinberg, “A Designer Universe?” in Paul Kurtz, ed. Science and Religion: Are They Compatible? (Amherst, NY: Prometheus Books, 2003), p. 33.

[9] L. P. Gerson, God and Greek Philosophy: Studies in the Early History of Natural Theology (London: Routledge, 1990), p. 14.

[10] Peter van Inwagen, Metaphysics (San Francisco: Westview Press, 1993), p. 98.

[11] I wish to thank my colleague, Dr. Steven Duncan, for helpful comments on parts of the paper. I also wish to extend my appreciation to the anonymous reviewer who posed very sharp and insightful comments that caused me to take the argument to a deeper level.

Copyright ©2006 by Paul Herrick. This electronic version is copyright ©2006 by Internet Infidels, Inc. with the written permission of Paul Herrick. All rights reserved.

all rights reserved