The universe (or the totality of contingent things) can be an unexplained brute fact, so cosmological and contingency arguments give no reason to posit God

The universe can be an unexplained brute fact, so...

The objection is not that cosmological arguments are silly. In its strongest form, it is a serious challenge to a family of arguments that move from the contingency of the world to a necessary explanation of it. The challenge says: those arguments require a strong Principle of Sufficient Reason, and that principle is more controversial than the conclusion it is meant to support.

Steelman: the brute-fact objection

A typical Leibnizian contingency argument - defended in different forms by thinkers such as Leibniz, Samuel Clarke, Alexander Pruss, Robert Koons, and Joshua Rasmussen - says something like this:

• Contingent things exist.
• Every contingent thing, or every contingent fact, has an explanation.
• The totality of contingent things cannot be explained by another contingent thing within that totality.
• So the explanation must be non-contingent, or necessary.

The crucial premise is the Principle of Sufficient Reason, or PSR. In one strong form, it says that every contingent fact has a sufficient explanation. The atheist objection is that this principle is not self-evident, not scientifically established, and not forced on us by reason.

Bertrand Russell gave the popular version in his 1948 radio debate with Frederick Copleston:

"I should say that the universe is just there, and that's all."

The more careful philosophical version appears in different forms across the literature. Hume questions whether causal and explanatory principles learned within experience can be extended to the universe as a whole; J. L. Mackie criticizes cosmological arguments for depending on an overstrong PSR; Peter van Inwagen presses the modal-collapse problem; and Graham Oppy argues that naturalism may reasonably take the fundamental natural order as ultimate. These are not all the same argument, but they share a common point: not everything inside the universe lacks explanation. Local explanations are everywhere. Stars form from gas clouds; organisms arise through evolutionary processes; particular events have causes or probabilistic histories. The disputed claim is that the whole explanatory system must itself have an explanation outside itself.

Several considerations motivate this resistance.

First, quantum theory is often taken to challenge deterministic versions of causation. On some interpretations, radioactive decay or particular measurement outcomes are not fixed in advance by prior physical states, even though they are described by a quantum state, probabilities, and laws. Other interpretations handle indeterminacy differently, so this should not be presented as a settled proof that uncaused events occur. The more modest point is enough for the objection: if fundamental physics is not uniformly deterministic, then a PSR requiring sufficient determining reasons for every event is harder to defend.

Second, van Inwagen's modal-collapse objection targets strong versions of the PSR. Suppose every contingent truth has an explanation, and consider the total conjunction of all contingent truths: the complete contingent fact about reality. An explanation of that whole conjunction cannot be just one more contingent truth of the same order, because that truth would itself be part of what is to be explained and would not explain why the complete conjunction obtains. So the explanation seems to have to be necessary. But if a necessary truth sufficiently explains the complete contingent truth in a way that entails it, then the complete contingent truth could not have been otherwise. Contingency collapses into necessity. That is absurd, so the strong entailment-style PSR should be rejected.

Third, Hume warned against extending causal principles beyond experience. We observe relations among events within the world. We do not observe worlds being caused. To infer from "things in the universe have causes" to "the universe has a cause" risks treating the universe as one more object inside the universe.

Fourth, the move may commit the fallacy of composition. Even if every part of a collection has some explanation, it does not automatically follow that the collection as a whole has an explanation of the same kind. Every human has a mother; humanity as a whole does not have a mother.

So the atheist can say: cosmological and contingency arguments depend on a disputed metaphysical principle. Once that principle is denied, the universe, or the totality of contingent reality, may be taken as a brute fact. This is not a contradiction. It is a stopping point.

What should be conceded

A fair critique must concede that this objection has real force.

It is not enough to say, "Everything needs a cause," because many philosophers, including many theists, do not formulate the argument that crudely. Nor is it enough to insist that explanation must continue forever until God appears. That simply assumes the PSR rather than defending it.

The modal-collapse objection is especially important. A very strong PSR - every contingent truth has a sufficient explanation that entails, or necessitates, that truth - faces deep problems. If the explanation of the complete contingent fact is necessary and necessitates that fact, then it becomes difficult to preserve the thought that reality could have been otherwise. Defenders of contingency arguments often respond by weakening or restricting the PSR, or by denying that explanation must entail what it explains. Those moves may help, but they also show the burden: the argument needs a carefully defended PSR, not a slogan. This is not an internet objection. It is a serious challenge in analytic philosophy.

It should also be conceded that brute facts are not obviously incoherent. A person can consistently say: explanations eventually stop. The dispute is not over whether the phrase "brute fact" is meaningless. It is over whether stopping at the universe is intellectually satisfying, and whether rejecting the PSR is less costly than accepting it.

Where the brute-fact objection overreaches

The first overreach concerns quantum indeterminacy. Quantum theory may undermine deterministic causation, but that is not the same as showing that events have no explanation at all. A radioactive decay, for example, may not have a deterministic prior cause fixing its exact time, but it still occurs within a physical system governed by a quantum state, decay probabilities, and relevant laws. That is an explanation of a kind, even if not a necessitating one.

This matters because contingency arguments need not rely on the claim that every event is deterministically caused. Many defenders of cosmological reasoning distinguish causes from explanations, and deterministic causes from probabilistic or grounding explanations. Quantum mechanics may support the existence of irreducibly chancy outcomes. It does not by itself establish that the existence of the whole physical order is the same kind of brute fact as a particular quantum outcome.

The second overreach concerns the fallacy of composition. The warning is valid, but it is not decisive by itself. Some arguments really do commit the fallacy: "Every part has a cause, so the whole has a cause" is too quick. But the better contingency arguments are not merely adding up causes of parts. They ask why there is any contingent reality at all, or why this entire dependent order obtains.

Composition fallacies are fallacies only when the inference is invalid in the relevant case. Some properties do transfer from parts to wholes, and some do not. If every tile in a mosaic is blue, the mosaic is blue. If every member of a committee is small, the committee need not be small. So saying "fallacy of composition" names a possible error; it does not yet show that the contingency argument has made it.

The third overreach concerns Hume. Hume's critique of causation does make it difficult to claim that observation alone gives us a necessary causal principle applying to the cosmos as a whole. That is a real limitation. But the conclusion should be stated modestly. A Humean or empiricist need not abandon science; they can treat scientific explanation in terms of laws, models, probabilities, counterfactual dependence, or inference to the best explanation without accepting a metaphysically necessary PSR. What Hume blocks is the quick move from familiar causal explanation within the world to a compulsory cause of the world as a whole. He does not, by that alone, prove that an ultimate explanation is misguided.

So Hume supports modesty. He does not automatically vindicate the stronger claim that no ultimate explanation should be sought.

The strongest atheist reply

The atheist can respond that these concessions do not rescue the cosmological argument. Weaker versions of the PSR may avoid modal collapse, but they can look ad hoc. If the principle is weakened enough to escape counterexamples, it may no longer be strong enough to generate a necessary being. If it is strong enough to generate the conclusion, it may again face modal-collapse worries.

The atheist can also argue that everyone must stop somewhere. If the theist says that God exists necessarily, the atheist may ask whether this necessity is genuinely explanatory, independently intelligible, and theoretically better than taking the natural order as ultimate. The theist will reply that a necessary reality is not a brute contingent stopping point; the atheist will answer that this reply only helps if the alleged necessity is well motivated rather than stipulated. This is the line developed in different ways by philosophers such as Graham Oppy: naturalism and theism should be compared as total theories, using theoretical virtues such as explanatory scope, simplicity, coherence, and fit with our best background knowledge. On that comparison, Oppy argues, naturalism need not be inferior merely because it has an ultimate natural stopping point.

This reply succeeds in showing that contingency arguments are not coercive. A person who rejects the PSR, or accepts only a modest methodological version of it, is not thereby contradicting himself. The brute-universe position is a coherent position.

But coherence is not the same as explanatory superiority. To say "the totality of contingent things is just there" is not an explanation of that totality in the sense sought by the contingency argument. It is a denial that such an explanation is required, or perhaps a claim that explanation legitimately terminates in fundamental reality. That may be legitimate, but it has a cost: it limits the scope of rational explanation at precisely the largest question.

What the critique establishes

The brute-fact objection establishes that cosmological and contingency arguments depend on substantive metaphysical assumptions. In particular, they require some defensible version of the PSR. If that principle is rejected, the arguments lose much of their force.

It does not establish that the PSR is false. It does not establish that the universe is a brute fact. It does not show that naturalism explains the existence of contingent reality better than theism. It shows that the atheist has a coherent stopping point available.

The deeper issue is comparative. Which is more reasonable: accepting a carefully limited principle that contingent reality calls for explanation, while avoiding modal collapse and other excesses, or accepting that the entire contingent order simply exists without further explanation? The brute-fact answer is possible, and it is not irrational merely because it stops. The theistic answer is also not refuted merely because it uses a substantive metaphysical principle. The live question is which stopping point, with its costs included, gives the better overall account.