Argument from reason

The argument from reason contends that naturalism — the view that the physical world is all that exists — is self-defeating because it cannot account for the rationality of human thought. If every mental event is entirely determined by prior physical causes (brain chemistry, neural firings, evolutionary conditioning), then no belief is held because it is logically warranted; every belief is held because physical processes caused it. But if beliefs are never held for rational reasons, then the belief that naturalism is true is itself not held for rational reasons, and naturalism undercuts the very reasoning by which anyone might come to accept it. The argument targets the foundations of naturalism by turning its own commitments against itself: if naturalism is true, we have no reason to trust the reasoning that led us to conclude that naturalism is true.^{1, 2}

The argument has roots in the nineteenth century but received its most influential modern formulation from C. S. Lewis in chapter three of Miracles (1947, revised 1960). It has since been developed and defended by Victor Reppert, and it shares structural affinities with Alvin Plantinga’s evolutionary argument against naturalism and Thomas Nagel’s critique of reductive materialism. The argument does not directly prove God’s existence but establishes a negative conclusion — that naturalism is rationally untenable — which its proponents take as indirect evidence for a non-naturalistic worldview in which reason has a secure foundation.^{1, 4, 8}

Historical background

Precursors to the argument appear in the nineteenth century. Arthur Balfour, in The Foundations of Belief (1895), argued that naturalistic evolution undermines confidence in the reliability of human cognition, since natural selection shapes the mind for survival rather than truth. G. K. Chesterton made a similar point in Orthodoxy (1908), observing that the materialist who explains thought as a mere product of matter has sawn off the branch on which he sits. But the argument assumed its recognisable modern form with C. S. Lewis, whose third chapter of Miracles — titled “The Cardinal Difficulty of Naturalism” — presented the problem with a philosophical precision that invited serious academic engagement.^{1, 2}

Lewis’s original 1947 formulation drew a sharp objection from the philosopher Elizabeth Anscombe at the Oxford Socratic Club. Anscombe argued that Lewis conflated two senses of “explanation”: the causal sense (what physical events brought about a belief) and the logical sense (what reasons justify the belief). Lewis took Anscombe’s critique seriously and substantially revised the chapter for the 1960 edition, reformulating his argument to distinguish more carefully between the causes and the grounds of belief. The revised version remains the standard text. The Anscombe episode is itself philosophically significant, as it illustrates the central dialectical tension in the argument: whether physical causation and rational justification can coexist.^{1, 2, 3}

Lewis’s argument

Lewis begins by distinguishing two kinds of connection between events. A physical or “cause-and-effect” connection holds when one event brings about another through natural law: a stone falls because gravity acts on it. A logical or “ground-and-consequent” connection holds when one proposition entails or supports another: the conclusion of a valid argument follows from its premises. Lewis contends that a genuine act of reasoning requires the thinker’s inference to be determined by the logical connection between the propositions — the thinker must accept the conclusion because it follows from the premises, not merely because a chain of physical causes happens to produce the neural state corresponding to that conclusion.¹

On naturalism, Lewis argues, every mental event is entirely the product of prior physical causes. A belief is a brain state, and that brain state was caused by a preceding brain state, which was caused by a preceding brain state, and so on back through an unbroken chain of physical causation. At no point in this chain does a logical relationship between propositions play any causal role. The thinker does not believe the conclusion because it follows from the premises; the thinker believes the conclusion because certain neurons fired in a certain pattern, and those neurons fired because of prior electrochemical events, all the way back to the Big Bang. The logical connection between premises and conclusion is, on the naturalist account, causally inert — an epiphenomenal shadow that accompanies the physical process but does not guide it.^{1, 2}

Lewis concludes that if naturalism is true, then no inference is rationally justified, because no inference is determined by its logical grounds. Every apparent act of reasoning is merely a sequence of physical events that happens, by coincidence, to track a logical pattern. But if no inference is rationally justified, then the inference to naturalism is not rationally justified, and we have no reason to believe naturalism. Naturalism therefore destroys the credentials of the very reasoning on which any argument for naturalism must depend.¹

The formal argument

Victor Reppert has formulated the argument with greater precision, identifying what he calls “the Lewis-Anscombe controversy” as the key philosophical issue. Reppert’s formulation can be summarised as follows:

P1. No belief is rationally justified unless it is held at least in part because of the logical relationship between it and other beliefs that serve as its grounds.

P2. If naturalism is true, then every mental event (including every belief) is entirely determined by prior physical causes operating according to natural law.

P3. If every belief is entirely determined by prior physical causes, then no belief is held even partly because of the logical relationship between it and its grounds.

C1. If naturalism is true, then no belief is rationally justified. (From P2 and P3)

C2. If no belief is rationally justified, then the belief that naturalism is true is not rationally justified.

C3. Therefore, naturalism is self-defeating. (From C1 and C2)

The argument’s force depends on premise P3 — the claim that complete physical determination of belief is incompatible with rational justification. This is the point at which defenders and critics diverge. Reppert identifies four distinct sub-arguments in Lewis’s text that jointly support this premise: an argument from the nature of intentionality, an argument from the nature of truth, an argument from the nature of mental causation, and an argument from psychological unity. Each addresses a different feature of rational thought that Reppert claims naturalism cannot accommodate.²

Reppert’s development

In C. S. Lewis’s Dangerous Idea (2003), Reppert developed and systematised the argument far beyond Lewis’s original formulation. Reppert argues that rational inference requires several features that are difficult to accommodate within a naturalistic framework: intentionality (the “aboutness” of mental states), propositional content (the fact that beliefs have truth conditions), logical relations between propositions (entailment, consistency, evidential support), and the capacity for mental states to be caused by their propositional content rather than merely by their physical substrate.²

Reppert examines and rejects several naturalistic strategies for accommodating reason. Functionalism — the view that mental states are defined by their causal roles rather than their physical composition — does not help, because causal roles are still physical relations, and the question is whether logical relations can be reduced to causal ones. Eliminative materialism — the denial that beliefs, desires, and other propositional attitudes exist — concedes the argument by eliminating the rational agent altogether. Non-reductive physicalism — the view that mental properties supervene on but are not identical to physical properties — faces the problem of mental causation: if mental properties are epiphenomenal (causally inert), they cannot guide reasoning, and if they are causally efficacious, their causal power cannot be explained by the underlying physics without overdetermination.^{2, 12}

Reppert argues that theism avoids these problems because it provides a metaphysical framework in which reason is fundamental rather than derivative. If God is a rational being who created human beings in the divine image — endowed with rational faculties designed to track truth — then the reliability of human reason has a straightforward explanation: reason is reliable because it was designed to be reliable by a being who is himself the paradigm of rationality. This does not prove theism, but it shows that theism can accommodate reason in a way that naturalism cannot, which is itself a significant philosophical advantage.^{2, 13}

Relation to Plantinga’s EAAN

The argument from reason shares significant structural affinities with Alvin Plantinga’s evolutionary argument against naturalism (EAAN), though the two arguments differ in their premises and dialectical strategy. Plantinga argues that if naturalism and evolution are both true, then the probability that human cognitive faculties are reliable is low or inscrutable, because natural selection shapes cognition for survival rather than truth. Since the naturalist must assume the reliability of cognition to assess the probability of naturalism, naturalism is self-defeating in the same way Lewis identified: it undercuts its own rational credentials.^{4, 5}

The key difference is that Lewis’s argument focuses on the logical structure of inference — the claim that physical causation cannot ground logical entailment — while Plantinga’s argument focuses on the evolutionary production of belief-forming mechanisms — the claim that natural selection does not select for truth-tracking. Lewis’s argument is more fundamental: it applies to any form of naturalism, whether evolutionary or not, because its target is the causal closure of the physical world, not the specific mechanism of evolution. Plantinga’s argument is more formally developed and has generated a larger philosophical literature, but it addresses a narrower range of naturalistic positions.^{2, 6}

Thomas Nagel has independently developed a related critique of naturalism in Mind and Cosmos (2012), arguing that the materialist neo-Darwinian conception of nature is “almost certainly false” because it cannot account for consciousness, cognition, and value. Nagel, himself an atheist, does not endorse theism as the alternative but argues that some form of teleological naturalism — a view in which the universe has an inherent tendency to produce rational minds — is needed to accommodate the features of mind that reductive naturalism leaves unexplained. Nagel’s critique thus supports the negative conclusion of the argument from reason (that reductive naturalism is inadequate) without endorsing its positive implication (that theism is required).^{8, 15}

Major objections

The argument from reason has attracted several lines of criticism. The most influential is the compatibilist objection, which challenges premise P3 — the claim that complete physical causation of belief is incompatible with rational justification. Compatibilists argue that a belief can be both physically caused and rationally justified, just as an action can be both physically caused and morally responsible. The physical causal chain that produces a belief may itself instantiate the logical relationships that justify the belief: the neural processes that constitute an inference may be the very processes by which the logical relationship between premises and conclusion exerts its rational force. On this view, there is no conflict between physical causation and rational justification because they operate at different levels of description of the same process.^{3, 10}

Daniel Dennett has argued that the argument from reason rests on a false dichotomy between “mere physical causation” and “rational guidance.” Cognitive science reveals that reasoning is implemented in physical systems (brains, computers) that process information according to rules that track logical relations. A computer following a proof-checking algorithm is entirely determined by its physical states, yet it reliably tracks logical validity. If physical systems can implement logical operations, then the fact that the brain is a physical system does not prevent it from reasoning validly. Lewis’s argument succeeds only if one assumes that physical processes cannot implement rational ones — an assumption that begs the question against the naturalist.¹⁰

Graham Oppy has pressed a tu quoque objection: if the argument from reason shows that naturalism is self-defeating because physical causation cannot ground rational justification, then the same argument threatens theism. If God causes human beliefs (through creation, providence, or direct intervention), then human beliefs are caused by something other than the logical grounds that justify them. The theist faces the same structural problem: explaining how beliefs can be both divinely caused and rationally justified. If the theist can solve this problem (by arguing that God designs the causal process to track logical relations), then the naturalist can make the same move (by arguing that evolution designs the causal process to track logical relations).¹¹

Jaegwon Kim’s work on mental causation has complicated the dialectic. Kim argues that non-reductive physicalism — the most popular form of naturalism in the philosophy of mind — does face a genuine problem of mental causation: if mental properties supervene on physical properties but are not identical to them, it is difficult to see how they can be causally efficacious without violating the causal closure of the physical. Kim’s analysis lends partial support to the argument from reason by identifying a real tension within naturalism, though Kim himself does not draw theistic conclusions from it.^{12, 16}

Responses to objections

Reppert has responded to the compatibilist objection by arguing that it misidentifies the problem. The question is not whether physical systems can track logical relations (a computer can correlate symbols according to syntactic rules) but whether physical causation can constitute genuine rational insight — the awareness that a conclusion follows from its premises. John Searle’s Chinese Room argument is relevant here: a system that manipulates symbols according to rules may produce outputs that match what a rational agent would produce, but it does not thereby understand the logical relations between the symbols. Similarly, a brain that produces belief-states matching valid inferences may do so through physical causation without the rational agent genuinely grasping the logical connections. Reppert contends that the naturalist conflates computational reliability with rational understanding.²

Against the tu quoque objection, defenders argue that theism and naturalism are not symmetrically positioned. On theism, God is himself a rational agent — indeed, the paradigm of rationality — and creates human beings with rational faculties that are designed to track truth. The causal chain from God through creation to human cognition is not a chain of blind physical events but a chain initiated and sustained by a rational agent with the intention that human reasoning be reliable. The theist can therefore explain the coincidence between physical causation and rational justification: God designed the physical causal processes to implement rational operations. The naturalist cannot make an analogous move, because on naturalism the physical causal processes are not designed by anyone and have no teleological orientation toward truth.^{2, 13}

Edward Feser has developed a related response from within the Aristotelian-Thomistic tradition, arguing that the argument from reason reveals the inadequacy of the mechanistic conception of nature that underlies modern naturalism. If nature is conceived as devoid of inherent teleology — as modern physics since Galileo and Descartes has characteristically conceived it — then rational thought, which is inherently teleological (directed toward truth), cannot be a natural phenomenon. But if nature is conceived in Aristotelian terms as inherently teleological, then rational thought can be a natural phenomenon without reducing to blind physical causation. Feser contends that the argument from reason is best understood not as a proof of theism per se but as a demonstration that the mechanistic naturalism of modern science is philosophically untenable and that some form of teleological metaphysics is required.⁹

The argument and the cumulative case

The argument from reason is typically presented not as a standalone proof of God’s existence but as one element in a cumulative case against naturalism and for theism. Richard Swinburne incorporates a version of the argument into his broader Bayesian case, treating the existence of rational agents as one of several features of the universe that theism explains better than naturalism. William Lane Craig similarly regards the argument as supporting the conclusion that naturalism is inadequate, which in turn strengthens the case for the theistic alternative established by the cosmological, teleological, and moral arguments.^{13, 14}

The argument intersects with several other debates in the philosophy of religion. It connects to the argument from consciousness, which reasons from the existence of subjective experience to the inadequacy of physicalism. It connects to the evolutionary argument against naturalism, which targets the specifically Darwinian form of naturalism. And it connects to the transcendental argument for God, which contends that the preconditions of rational thought presuppose a theistic framework. Together, these arguments constitute a family of anti-naturalist considerations that target the adequacy of materialism as a comprehensive worldview.^{2, 7}

Family of anti-naturalist arguments from mind^{2, 7}

Contemporary assessment

The argument from reason occupies a distinctive position in contemporary philosophy of religion. It is less formally developed than Plantinga’s EAAN and less widely discussed in the professional literature, but it addresses a more fundamental issue: whereas Plantinga targets the evolutionary production of cognitive faculties, Lewis targets the very possibility of rational inference within a causally closed physical world. If Lewis’s argument succeeds, it undermines not only evolutionary naturalism but any form of physicalism that maintains the causal closure of the physical domain.^{2, 4}

The argument’s assessment turns on deep and unresolved questions in the philosophy of mind: the nature of mental causation, the relationship between reasons and causes, the possibility of reducing intentionality to physical properties, and the adequacy of functionalist accounts of cognition. Jaegwon Kim’s influential work has shown that the problem of mental causation is genuine and not easily solved within a physicalist framework, lending indirect support to the argument’s central claim. But whether the problem is insoluble — whether physicalism in principle cannot accommodate rational inference — remains disputed.^{12, 16}

The argument’s enduring significance may lie less in its capacity to prove theism than in its capacity to expose a tension within naturalism that naturalists themselves acknowledge. The problem of mental causation, the hard problem of consciousness, and the question of how meaning arises in a physical world are among the deepest unsolved problems in philosophy of mind. The argument from reason draws attention to these problems and contends that they are not merely unsolved but unsolvable within a naturalistic framework — a claim that, if correct, would have profound implications for the viability of naturalism as a comprehensive philosophical worldview.^{8, 7}