The Humean challenge
The argument under examination is not the crude claim that "science has disproved causation." It is a more serious empiricist challenge, associated above all with David Hume in A Treatise of Human Nature and An Enquiry concerning Human Understanding. Hume's point is that when we attend carefully to experience, we never perceive causal necessity itself. We see one kind of event regularly followed by another: flame and heat, impact and motion, striking a match and ignition. But the alleged "must" connecting cause and effect is not itself an object of sight, touch, or introspection. What we call necessary connection is, on Hume's account, the mind's expectation formed by habit after repeated conjunctions.
Applied to cosmological arguments, the challenge is sharp. Many such arguments rely on a causal or explanatory premise: "whatever begins to exist has a cause," "every contingent fact has an explanation," or "there cannot be an uncaused beginning of existence." The Humean asks: what licenses this principle? It is not directly observed. It is not a truth of reason in the way that "all bachelors are unmarried" is. Its denial does not appear formally contradictory. Hume's separability point is that, since the idea of an object's beginning to exist and the idea of a cause are distinct, we can at least conceive the former without conjoining the latter. The conclusion is not that we never expect causes; plainly we do. It is that the causal premise, if used as a universal and modal principle, needs more defense than bare experience or mere obviousness supplies.
This is a serious objection. It targets not a detail of physics but the epistemic status of a premise. It says: before arguing from the world's existence to a cause, explain why the causal rule is entitled to govern reality as a whole, especially at the boundary case of the origin of the universe.
What Hume is right about
Hume's critique succeeds against a simplistic form of the causal premise. If someone says, "We literally see necessary causal power passing from one object to another," Hume's response is powerful. Ordinary perception gives us succession, regularity, and correlation; it does not present necessity as an additional visible ingredient. Nor is the universal causal principle straightforwardly derived by induction. We have observed many things beginning within the universe that have causes, but we have not observed universes beginning to exist, nor have we surveyed all possible beginnings.
Hume also exposes a weakness in some popular presentations of cosmological arguments. "Everything has a cause" is often asserted as though it were simply undeniable. But that premise is not even the premise used by more careful defenders, since if everything had a cause, the proposed first cause would require one too. More refined arguments use claims about things that begin to exist, or contingent realities, or dependent beings. Still, even those more careful claims require defense. They cannot be treated as observationally obvious in the same way that "this glass fell when released" is observationally obvious.
There is another important concession. Hume's point about conceivability has some force. We can imagine, at least in a thin sense, a flash of light appearing without a prior producing event. We do not immediately find ourselves affirming a logical contradiction. That matters because some defenders of cosmological arguments speak as if "uncaused beginning" is as incoherent as "square circle." It is not. If the causal premise is true, it is not true as a trivial analytic truth.
Where the Humean argument overreaches
The strongest reply is that Hume's argument proves less than it is often taken to prove. From "we do not observe necessary connection as a sensory item," it does not follow that we have no rational entitlement to any causal principles. A careful Humean can allow that causal reasoning is psychologically unavoidable and practically indispensable while denying that it reveals objective necessity. Even so, if the skeptical standard becomes "a principle is warranted only if it is directly sensed or analytically true," then the standard threatens more than cosmological arguments. Mathematical structure, logical validity, personal identity over time, the existence of other minds, and the uniformity assumptions involved in empirical inquiry are not simply read off as sensory items. The challenge is therefore to explain why causal and explanatory principles are uniquely disqualified.
This is not a cheap tu quoque. It identifies a pressure within the empiricist criterion. Scientific practice does not merely record constant conjunctions; it looks for mechanisms, dependencies, counterfactual structure, interventions, and explanatory relations. Contemporary philosophers of causation disagree deeply about how to analyze these things. David Lewis developed an influential counterfactual account; J. L. Mackie analyzed causes as INUS conditions, insufficient but necessary parts of unnecessary but sufficient conditions; Elizabeth Anscombe criticized the assumption that causation must be reduced to exceptionless regular succession; and Nancy Cartwright emphasized causal capacities and the plural, model-relative character of many scientific laws. Humean and neo-Humean accounts remain serious options. The point is only that Hume's regularity picture is not the uncontested final word, and that rejecting directly perceived necessity is not the same as refuting causal realism.
More importantly, the move from "not observed" to "unwarranted" is too quick. Necessary connection might not be an object of observation because it is not that kind of thing. A relation of dependence need not be perceptible in the way color or motion is perceptible. We often infer causal dependence from patterns, interventions, background theory, and explanatory success. If I infer that a virus caused an illness, I do not perceive necessity itself. Yet the inference may still be rational. The defender of a causal premise can say that causal principles are not read off a single experience; they are part of the framework by which experience is made intelligible.
That reply must be handled carefully. It would be overreach to say that because science uses causal reasoning, therefore a cosmological causal principle is established. Scientific causation normally concerns transformations within an existing order: particles, fields, organisms, events, and systems. The beginning of the universe, if there is such a beginning, is not simply another laboratory event. Still, Hume's point does not by itself show that causal reasoning is invalid at such a boundary. It shows that extending it there requires argument.
Conceivability and possibility
Hume also relies on a conceivability claim: we can conceive of something beginning to exist without a cause, so such a beginning is possible. This is plausible if "conceive" means "form a loose mental picture without noticing contradiction." But that is only a defeasible guide to metaphysical possibility. Saul Kripke's discussion of necessary a posteriori truths is relevant here: before chemistry, one might coherently imagine watery stuff that is not H2O, but if "water" rigidly names the actual substance H2O, then water's being H2O is not a merely optional fact. Likewise, some apparent mathematical or modal possibilities disappear once the concepts are examined more carefully. Conceivability can provide evidence of possibility, but it is not automatically decisive.
The defender of the causal premise can therefore respond: perhaps an uncaused beginning is imaginable in a superficial way, but not genuinely possible once the metaphysics of existence, dependence, and actuality are properly understood. That response is not a proof. It is a challenge to Hume's inference from conceivability to real possibility.
At the same time, the Humean can reply that the burden of proof remains on the person asserting necessity. If the causal principle is not analytic, not directly observed, and not easily derived from finite experience, then calling its denial "metaphysically impossible" demands an account of why. This is fair. A serious cosmological argument must do more than insist that uncaused beginnings feel absurd.
Kalam and contingency arguments
The Humean critique lands differently on different cosmological arguments.
- Kalam-style arguments depend on a premise such as "whatever begins to exist has a cause." Hume's critique presses hard here because the premise is modal and universal. Defenders such as William Lane Craig often support it by appeal to metaphysical intuition, the idea that being cannot arise from absolute non-being, and the claim that alleged counterexamples in quantum theory are not examples of things emerging from literal nothingness. Critics such as Graham Oppy and Wes Morriston reply that this still may not establish a universal principle, especially when the origin of spacetime may not resemble ordinary beginnings within spacetime. The fair result is not that the kalam premise is obviously false, but that it needs a substantive metaphysical defense rather than a quick appeal to everyday causation.
- Leibnizian contingency arguments usually rely not on a simple causal rule but on a principle of sufficient reason: contingent reality requires an explanation. Hume's general skepticism about necessary connections is relevant, but not decisive, because the PSR is broader than event causation. Its defenders, such as Leibniz and in contemporary forms Alexander Pruss and Robert Koons, argue that rejecting sufficient reason undermines explanation too deeply and leaves brute contingency where reason should still ask "why?" Its critics, including many influenced by Hume and later Bertrand Russell's "brute fact" response, argue that explanation may legitimately terminate in the universe itself, and that strong versions of the PSR may carry controversial modal costs. So the dispute is not simply "cause or no cause"; it is also about how strong a demand for explanation reason is entitled to impose.
- Thomistic or Aristotelian arguments often concern present dependence, actuality, or sustaining causes rather than a first event in time. Hume's critique of observed necessary connection still matters, but these arguments are not always vulnerable to the same "we can imagine an uncaused beginning" objection, because their target is not merely temporal beginning.
This distinction is important. There is no single "causal premise of cosmological arguments." Hume creates trouble for several families of arguments, but the amount of trouble varies.
The strongest Humean rejoinder
The best Humean reply is modest and powerful: even if causation is indispensable within experience, we are not entitled to project it beyond the conditions under which it was learned. Our causal concepts arise within the universe, among events ordered by time and governed by regularities. A cosmological argument often applies them to the universe as a whole, or to the beginning of time itself. That extension is not automatically licensed.
This rejoinder succeeds against any argument that moves casually from everyday examples to ultimate reality. "Houses have builders, paintings have painters, therefore the universe has a cause" is philosophically weak. The universe is not one more artifact inside the universe. Likewise, appeals to quantum physics must be made cautiously. It is misleading to say that quantum events simply prove things can come from nothing; quantum events occur within a physical framework with fields, states, and laws. But it is also misleading to pretend that modern physics straightforwardly confirms classical causal intuitions. The honest conclusion is that physics complicates the discussion rather than settling it.
What the critique establishes
Hume's argument establishes that the causal premise of a cosmological argument cannot be treated as a deliverance of bare observation. We do not perceive necessary connection as a sensory datum. Nor is "whatever begins to exist has a cause" a trivial truth whose denial is formally contradictory. A defender of a cosmological argument owes an account of the causal or explanatory principle being used, why it is rationally warranted, and why it applies to the relevant case.
But Hume's argument does not establish that causal principles are irrational, that uncaused beginnings are genuinely possible in the full metaphysical sense, or that every serious cosmological argument fails. It does show that one popular route to a causal premise is inadequate: simply pointing to ordinary cases and announcing a universal necessity. It does not show that no other route is available. Causal realism, metaphysical principles of dependence, inference to the best explanation, and versions of the principle of sufficient reason remain live options in contemporary philosophy, though each is contested and none may be invoked as if it were a settled result of science or common sense alone.
The balanced verdict is this: Hume gives atheists and skeptics a strong tool against overconfident cosmological arguments. He forces the theist to stop saying "obviously, everything needs a cause" and to begin doing metaphysics. But the Humean critique also has its own burden. If it reduces causation to habit or regularity too aggressively, it risks weakening the very explanatory practices on which rational inquiry depends. The result is not a refutation of theism, nor a proof of theism, but a clarification of the real issue: whether causal and explanatory principles are merely psychological expectations formed inside the world, or rational principles that can legitimately guide thought about why there is a world at all.