The cosmological argument

The cosmological argument is one of the oldest and most debated arguments for the existence of God. In its broadest form, it reasons from some general feature of the observable world — the existence of change, the presence of causal chains, or the contingency of the things that happen to exist — to a being that exists necessarily and serves as the ultimate explanation for why anything exists at all. Unlike the ontological argument, which proceeds from the concept of God alone, the cosmological argument begins with an empirical observation about the world and argues that the world as we find it requires an explanation that transcends the natural order. Versions of the argument appear in Plato, Aristotle, Islamic and Jewish medieval philosophy, Thomas Aquinas, Gottfried Wilhelm Leibniz, and, in the twentieth and twenty-first centuries, in the work of William Lane Craig and other analytic philosophers of religion.^{2, 17}

The argument takes several distinct forms, but all share a common structure: they identify a feature of the world that appears to demand explanation, argue that an infinite regress of explanations is either impossible or insufficient, and conclude that a first or necessary being must exist to terminate the regress. The principal variants are the Aristotelian–Thomistic argument from motion and causation, the Leibnizian argument from contingency and the principle of sufficient reason, and the kalam cosmological argument from the finitude of the past. Each has generated its own tradition of defense and critique, and together they constitute the most sustained line of natural-theological reasoning in Western philosophy.^{2, 11}

Aristotelian foundations

The philosophical lineage of the cosmological argument begins with Aristotle, who argued in the Physics and the Metaphysics that the existence of motion (kinesis) in the natural world requires a first unmoved mover. Aristotle understood motion broadly as any transition from potentiality to actuality — not merely locomotion, but qualitative change, growth, generation, and corruption. He observed that whatever is moved is moved by another, since nothing can actualize its own potential (a thing cannot be simultaneously in potency and in act with respect to the same property). If every mover is itself moved by a prior mover, one faces a regress of movers. Aristotle argued that such a regress cannot proceed to infinity in the case of essentially ordered (per se) causal series, because in such a series each intermediate mover transmits causal power it receives from the prior mover, and without a first mover the entire series would lack any motive power at all. He concluded that there must exist an unmoved mover — a being of pure actuality with no unrealized potentiality — which initiates motion without itself being moved.^{13, 14}

Aristotle’s unmoved mover is not a creator God in the Judeo-Christian sense. It is an eternal, immaterial, self-thinking intellect that moves the outermost celestial sphere as a final cause — the sphere moves out of desire or love for the perfection of the unmoved mover, not because the mover exerts efficient causation. Aristotle did not argue that the universe had a temporal beginning; he believed the cosmos was eternal. His argument is therefore not about what initiated the universe in time but about what sustains the ongoing process of change at every moment. This distinction between temporal and sustaining causation would become crucial in the later development of the cosmological argument, particularly in the hands of Aquinas.^{13, 14, 2}

Aquinas and the Five Ways

Thomas Aquinas adapted Aristotle’s reasoning in his Summa Theologiae (1265–1274), where the first three of his Five Ways constitute versions of the cosmological argument. The First Way (the argument from motion) follows Aristotle closely, arguing that the existence of change requires an unmoved mover. The Second Way (the argument from efficient causation) extends the reasoning to efficient causes more broadly: in the observable world, nothing is the efficient cause of itself, every effect depends on a prior cause, and an infinite regress of per se ordered efficient causes is impossible, so there must be a first efficient cause. The Third Way (the argument from contingency) argues that the existence of beings that are possible — capable of existing or not existing — requires a being that exists necessarily, since if everything were merely possible, there would be nothing to bring any of them into existence, and at some point nothing would have existed at all.^{1, 12}

P1. We observe that things in the world are in motion (i.e., undergoing change from potentiality to actuality).

P2. Whatever is in motion is moved by another, since nothing can actualize its own potentiality.

P3. An infinite regress of essentially ordered movers is impossible.

C. Therefore, there must exist a first unmoved mover, which all understand to be God.

A critical feature of Aquinas’s arguments, often misunderstood by later commentators, is that they do not require the universe to have a temporal beginning. Aquinas explicitly held that philosophy cannot demonstrate whether the world had a beginning in time; he accepted on faith that it did, but he regarded the question as irrelevant to his proofs. His arguments concern essentially ordered (per se) causal series operating in the present moment, not accidentally ordered (per accidens) series extending back through time. In an essentially ordered series, each member derives its causal power from the member prior to it simultaneously: a hand moves a stick, which moves a stone. Remove the hand and the stick cannot move the stone, regardless of how many intermediate sticks one interposes. Aquinas argues that such a series requires a first member that has causal power underivatively, and this is what he means by a first cause or necessary being.^{1, 12, 21}

The Third Way introduces a modal element absent from the first two. Aquinas distinguishes between beings whose existence is contingent (they might not have existed) and a being whose existence is necessary (it cannot not exist). He argues that if all beings were merely contingent, then given infinite past time, every possible state of affairs would have been realized, including the state in which nothing exists. But if at some point nothing existed, nothing could have come into existence (since nothing comes from nothing), and therefore nothing would exist now — which contradicts observation. Aquinas concludes that there must be at least one being that exists necessarily, and he traces the chain of necessary beings (some of which may owe their necessity to another) back to a being that is necessary of itself. This argument has been criticized on multiple grounds, including its assumption that infinite time guarantees the realization of every possibility, but it introduced the concept of necessary existence that would become central to all subsequent cosmological arguments.^{1, 17, 21}

The Leibnizian cosmological argument

Gottfried Wilhelm Leibniz developed a distinct form of the cosmological argument that shifts the focus from causation to explanation. In his On the Ultimate Origination of Things (1697) and the Monadology (1714), Leibniz argued that the existence of the universe requires not merely a cause but a sufficient reason — an explanation that renders it intelligible why this universe exists rather than some other, or rather than nothing at all. The foundation of his argument is the principle of sufficient reason (PSR): for every fact, there is a reason why it is so and not otherwise. Leibniz applied the PSR to the totality of contingent things and argued that the sufficient reason for the entire series of contingent beings cannot itself be a contingent being (since that would merely push the question back a step) but must be a necessary being whose existence is explained by its own nature.^{4, 7}

P1. Every contingent fact has an explanation (the principle of sufficient reason).

P2. The universe (the totality of contingent beings and states) is a contingent fact.

P3. The explanation for a contingent fact cannot lie wholly within the series of contingent facts.

C. Therefore, there exists a necessary being that provides the sufficient reason for the existence of the universe.

A distinctive strength of the Leibnizian argument is that it does not depend on whether the universe had a temporal beginning. Even if the universe has existed eternally, the question remains why this particular eternal series of states obtains rather than some other. An infinite regress of contingent explanations does not provide a sufficient reason for the whole series, just as an infinite chain of photocopies does not explain the content of the original document. Leibniz was explicit on this point: even if one imagines an eternal series of geometry textbooks, each copied from the prior one, one must still ask why geometry textbooks exist at all rather than natural history books or nothing. The sufficient reason must be found in a being whose existence is not contingent on anything else — a being that exists necessarily.^{4, 2, 16}

Contemporary defenders of the Leibnizian argument, most notably Alexander Pruss, have developed increasingly sophisticated formulations of the PSR designed to avoid objections. Pruss has argued for a restricted version of the PSR that applies to contingent propositions without entailing the kind of global necessitarianism that some critics have claimed follows from the unrestricted principle. Richard Swinburne has offered a related argument that treats the existence of God as the simplest explanation for the existence of the universe, invoking the PSR not as a strict logical principle but as a presupposition of rational inquiry. The Leibnizian tradition thus continues as an active research program within analytic philosophy of religion.^{7, 16, 11}

The kalam cosmological argument

The kalam cosmological argument takes a different approach from the Thomistic and Leibnizian versions by insisting on the temporal beginning of the universe as an essential premise. The argument has roots in medieval Islamic theology, particularly in the work of al-Kindi (c. 801–873) and al-Ghazali (1058–1111), who argued against the Aristotelian doctrine of an eternal cosmos. The term “kalam” refers to Islamic scholastic theology (ilm al-kalam). The argument was revived and reformulated in analytic terms by William Lane Craig beginning in 1979, and it has since become one of the most widely discussed arguments in contemporary philosophy of religion.^{3, 2}

P1. Everything that begins to exist has a cause.

P2. The universe began to exist.

C. Therefore, the universe has a cause.

Craig defends the first premise on the grounds that the alternative — that something could come into being from nothing, without any cause whatsoever — is metaphysically absurd. If things could spring into existence uncaused, there would be no reason why anything and everything does not appear spontaneously. The second premise receives both philosophical and scientific support. Philosophically, Craig argues that an actually infinite number of past events is impossible (drawing on paradoxes of the actual infinite, such as Hilbert’s Hotel) and that even if an actual infinite could exist, it could not be formed by successive addition — one cannot count to infinity. Scientifically, Craig appeals to standard Big Bang cosmology, which describes the universe as expanding from an initial singularity approximately 13.8 billion years ago, and to the Borde–Guth–Vilenkin (BGV) theorem, which demonstrates that any universe that has, on average, been expanding throughout its history must have a past boundary — a beginning — and cannot be past-eternal.^{3, 8, 11}

The BGV theorem, published in 2003 by Arvind Borde, Alan Guth, and Alexander Vilenkin, is particularly significant because it applies to a wide class of cosmological models, including many inflationary scenarios. Vilenkin himself has stated that “all the evidence we have says that the universe had a beginning,” though he has also explored speculative models (such as quantum tunneling from nothing) that attempt to describe the origin event in purely physical terms. Craig argues that even if such models succeed on their own terms, they still require a cause, since the quantum vacuum from which the universe allegedly tunnels is not “nothing” in the metaphysically relevant sense but is itself a physical state with structure and causal powers.^{8, 9, 11}

Scientific premises and cosmological evidence

The relationship between the cosmological argument and modern physics is complex and contested. Standard Big Bang cosmology describes an expanding universe that, when extrapolated backward, converges on an initial state of extreme density and temperature approximately 13.8 billion years ago. For proponents of the kalam argument, this provides empirical confirmation that the universe began to exist. The discovery of the cosmic microwave background radiation by Arno Penzias and Robert Wilson in 1965, the observed abundances of light elements predicted by Big Bang nucleosynthesis, and the detailed measurements of the cosmic microwave background by the COBE, WMAP, and Planck satellites all support the standard model of an expanding universe with a finite past.^{9, 15}

Key cosmological evidence relevant to the cosmological argument^{8, 10, 15}

Critics have pointed out that the initial singularity in classical Big Bang cosmology does not straightforwardly support the philosophical claim that the universe “began to exist.” The singularity represents a breakdown of general relativity, not a point at which the universe popped into being from nothing. A complete theory of quantum gravity — which does not yet exist — may describe the very early universe in terms that avoid a singular beginning altogether. Stephen Hawking and James Hartle proposed a “no-boundary” model in which the universe is finite in the past but has no initial boundary or edge, much as the surface of a sphere is finite but has no starting point. Hawking suggested that in this model “the universe would be completely self-contained and not affected by anything outside itself. It would neither be created nor destroyed. It would just BE.” If correct, this would undermine the kalam argument’s second premise by providing a physical description in which the universe is finite but does not “begin to exist” in the sense the argument requires.^{10, 15}

Hume’s and Kant’s critiques

The cosmological argument has faced philosophical criticism since at least the eighteenth century. David Hume, in the Dialogues Concerning Natural Religion (1779) and the Enquiry Concerning Human Understanding (1748), challenged several of the argument’s key assumptions. First, Hume questioned the causal principle itself. He argued that our concept of causation is derived entirely from experience — from the observed constant conjunction of events — and that we have no a priori warrant for asserting that every event must have a cause. The claim that something cannot come into being uncaused is a deeply ingrained habit of thought, Hume contended, but not a necessary truth of reason. If the causal principle is empirical rather than necessary, it cannot ground a demonstrative proof of God’s existence.^{5, 18}

Second, Hume raised what has become known as the fallacy of composition objection. Even if every individual member of a collection has a cause, it does not follow that the collection as a whole has a cause. Hume put the point memorably in the Dialogues: if one explains each member of a collection of twenty particles by reference to the member that caused it, one has explained all the members, and there is no further fact — the existence of the collection itself — left over to explain. Applied to the cosmological argument, this suggests that even if every contingent being has a cause, the totality of contingent beings may not require an external cause. Defenders of the argument have responded that the fallacy of composition charge is itself fallacious in this case: while it is sometimes illegitimate to infer properties of a whole from properties of its parts (each brick is small; therefore the wall is small), it is legitimate in many other cases (each brick has mass; therefore the wall has mass). Whether contingency is the kind of property that transfers from parts to wholes is itself a contested philosophical question.^{5, 17, 11}

Third, Hume questioned whether the notion of a necessary being is coherent. He argued that whatever we can conceive as existing, we can also conceive as not existing, and therefore no being’s existence is logically necessary. This objection targets the conclusion of the cosmological argument directly: if necessary existence is incoherent, the argument cannot establish a necessary being even if its premises are granted.^{5, 18}

Immanuel Kant, in the Critique of Pure Reason (1781), offered a different set of objections. Kant argued that the cosmological argument covertly depends on the ontological argument, because the step from “there must be a necessary being” to “that being is God” requires identifying necessary existence with the most perfect being, and this identification presupposes that existence is a real predicate — the very assumption that Kant believed he had refuted in his critique of the ontological argument. If existence is not a genuine property that can be predicated of a concept, then the notion of a being whose essence entails its existence is incoherent, and the cosmological argument collapses at the crucial step of identifying its first cause with God. Kant also argued that the principle of sufficient reason, which grounds the Leibnizian version, is a regulative principle of inquiry (useful for guiding investigation) rather than a constitutive principle of reality (an established truth about the way things are). Used regulatively, the PSR tells us to keep looking for explanations; used constitutively, it makes a metaphysical claim that Kant believed could not be justified by pure reason.^{6, 17}

Modern objections and responses

Twentieth- and twenty-first-century philosophy has added several new objections and refinements to the classical critiques. The most prominent modern objection to the kalam argument draws on quantum mechanics. In standard quantum theory, certain events — the radioactive decay of an atom, the spontaneous emission of a photon by an excited electron — are genuinely indeterminate: they occur without sufficient antecedent conditions that determine when or whether they will happen. If quantum events can occur without deterministic causes, this seems to undermine the kalam argument’s first premise that everything which begins to exist has a cause. Craig and other defenders have responded by distinguishing between deterministic and probabilistic causation: quantum events may lack deterministic causes but still have probabilistic causes (the quantum vacuum, the laws of quantum mechanics, the prior state of the system), and the kalam premise requires only that things do not come into being from absolute nothing, not that every event has a sufficient deterministic cause.^{3, 11, 17}

A related challenge comes from quantum cosmology. Lawrence Krauss, in A Universe from Nothing (2012), argued that modern physics shows how a universe can arise from “nothing” via quantum fluctuations in a vacuum state, without requiring a divine creator. Krauss’s argument generated extensive philosophical criticism, including from sympathizers with his atheistic conclusion. The quantum vacuum from which the universe supposedly emerges is not “nothing” in the philosopher’s sense — it is a quantum field governed by physical laws, possessing energy, structure, and causal dispositions. The philosophical question at the heart of the cosmological argument — why is there something rather than nothing? — is not answered by pointing to a physical state, however minimal, from which other physical states emerge, since the existence of that minimal state itself demands explanation. Critics such as David Albert and Edward Feser have argued that Krauss equivocates between the physicist’s “nothing” (the quantum vacuum) and the metaphysician’s “nothing” (the complete absence of any entity, property, law, or state whatsoever).^{20, 12, 17}

Another important objection is the question of what caused God. If everything requires a cause or explanation, then God too requires one, and the argument generates an infinite regress rather than terminating it. Defenders respond that the argument’s premises do not assert that everything requires a cause, but only that everything that begins to exist (kalam), or everything that is contingent (Leibniz), or everything that is in motion (Aquinas) requires a cause. God, as a necessary being that does not begin to exist and is not contingent, is exempt from the premise. The accusation of special pleading can be met only if the proponent can give a principled reason for the exemption — and proponents argue that the concept of a necessary being, one whose nature entails its existence, provides exactly such a reason. Whether this is ultimately coherent remains one of the deepest questions in the debate.^{3, 7, 16}

The “brute fact” response offers a different strategy for resisting the argument. Rather than challenging a specific premise, this response denies that the universe requires any external explanation at all. The universe, on this view, simply exists as a brute fact — an unexplained given that serves as the terminus of inquiry. Bertrand Russell famously expressed this position in his 1948 BBC debate with Frederick Copleston: “I should say that the universe is just there, and that’s all.” Defenders of the cosmological argument object that accepting brute facts represents an abandonment of the rational demand for explanation — a demand that is presupposed by all scientific and philosophical inquiry. If we accept that some facts are simply brute, we undermine the very principle that motivates us to seek explanations for anything. Pruss has argued that the rejection of the PSR has unacceptable modal consequences: if brute contingent facts are possible, then it becomes possible for any contingent proposition to be true for no reason, which threatens to undermine our confidence in the regularity of nature.^{7, 16, 17}

Current state of the debate

The cosmological argument remains one of the most actively debated topics in analytic philosophy of religion. Surveys of professional philosophers suggest that a majority do not find the argument persuasive, consistent with the broader trend of atheism or agnosticism among academic philosophers, but a significant minority regard it as providing substantial evidence for theism. The argument has benefited from the broader revival of natural theology in analytic philosophy since the late twentieth century, led by figures such as Alvin Plantinga, Richard Swinburne, and William Lane Craig, who have brought new standards of rigor and precision to arguments that were widely regarded as defunct after Kant.^{11, 16, 17}

Several developments characterize the current state of play. First, the different versions of the argument are now typically treated as independent lines of reasoning rather than as variations of a single argument, since they rest on different premises and face different objections. The kalam argument depends on the finitude of the past and is sensitive to empirical discoveries in cosmology; the Leibnizian argument depends on the PSR and is primarily a debate in metaphysics; the Thomistic argument depends on the distinction between per se and per accidens causal series and draws on Aristotelian metaphysics that many analytic philosophers reject. The strength or weakness of one version does not straightforwardly transfer to the others.^{2, 17}

Second, the debate over the principle of sufficient reason has become increasingly technical. Pruss and other defenders have formulated restricted versions of the PSR designed to avoid the charge of necessitarianism (the objection, raised by Peter van Inwagen, that the unrestricted PSR entails that everything that exists exists necessarily, eliminating all contingency). Whether these restricted principles are strong enough to ground a cosmological argument while avoiding modal collapse remains an open question in modal metaphysics.^{7, 11}

Third, the interface between cosmology and philosophy of religion continues to generate productive exchange. The BGV theorem has given kalam proponents a powerful scientific premise, but cosmologists have also proposed models — including cyclic cosmologies, emergent spacetime models, and string-theoretic landscapes — that may avoid the theorem’s conditions. The ultimate resolution of the question of whether the universe had a beginning may depend on a future theory of quantum gravity, which does not yet exist. In the meantime, the question of whether a physical beginning would constitute a “beginning of existence” in the metaphysically loaded sense required by the kalam argument remains a live philosophical issue, with some commentators arguing that the kalam premise trades on an equivocation between the physical and metaphysical senses of “begins to exist.”^{8, 9, 19}

Fourth, the Thomistic version has seen a revival among analytic philosophers and theologians who emphasize that its core reasoning does not depend on modern physics at all. Edward Feser and other contemporary Thomists have argued that the argument from motion, properly understood as an argument about the metaphysical dependence of changing things on a purely actual being in the present moment, is immune to the objections that target the kalam argument’s dependence on cosmology. Whether the Aristotelian–Thomistic metaphysical framework on which these arguments depend is defensible is itself a major question, since it requires commitments to act–potency composition, real essences, and the principle that nothing can actualize its own potentiality — commitments that many philosophers working in the empiricist and naturalist traditions reject.^{12, 21, 1}

The cosmological argument, in all its forms, thus remains what it has been for over two millennia: a serious philosophical argument that intelligent people reasonably disagree about. It has not been definitively refuted, nor has it achieved the status of a proof that commands universal assent. Its continued vitality is a testament both to the depth of the question it addresses — why does anything exist at all? — and to the difficulty of answering that question without invoking something that transcends the natural world.^{2, 17}