Plantinga's Evolutionary Argument Against Naturalism

Plantinga's EAAN

Abstract

Plantinga's argument is a probabilistic reductio aimed at the naturalist who accepts evolution. Its engine is the gap between adaptiveness and truth. Natural selection can only "reward" behavior, and a given adaptive behavior is compatible with indefinitely many assignments of belief-content, most of which are false. So if our minds are the products of unguided evolution and there is nothing more to them than what naturalism allows, the probability that they are reliable — that they mostly yield true beliefs — is low, or at least inscrutable. Grasping this gives the naturalist a defeater for trusting her faculties, and since she has no faculties other than these to vindicate them, the defeater spreads to all her beliefs, naturalism included. Theism, by contrast, gives an antecedent reason to expect reliable faculties, because it holds that they were designed by a truthful God. The argument is, in the SEP's own placement, the reason Plantinga relocates the science–religion conflict to a science–naturalism conflict.

Formal statement

1. If naturalism and evolution are true, then our cognitive faculties arose through unguided processes selecting for adaptive behavior, not for true belief.
2. For any adaptive behavior, many belief–desire combinations (most involving false beliefs) would produce it equally well; content is invisible to selection except through its behavioral output.
3. Hence P(our faculties are reliable | naturalism & evolution) is low or inscrutable.
4. One who accepts naturalism & evolution and sees (3) has an undefeated defeater for "my faculties are reliable."
5. That defeater defeats every deliverance of those faculties, including belief in naturalism & evolution.
6. Therefore naturalism & evolution is epistemically self-defeating and cannot be rationally held; theism, predicting reliable faculties, does not fall to the parallel argument.

Key evidence / textual basis

The primary texts — Plantinga's Warrant and Proper Function (1993), his exchanges in Beilby's Naturalism Defeated? (2002), and the fullest statement in Where the Conflict Really Lies (2011) — are copyright-locked and not in this corpus; the argument is reconstructed here from open-access encyclopedia summaries, and the reconstruction awaits verification against those primaries. {{UNSOURCED: Plantinga's Warrant and Proper Function (1993) and Where the Conflict Really Lies (2011) are not in corpus; the EAAN's premise-by-premise statement (esp. the P(R|N&E) calculation and the "defeater-deflector" reply to the generalization worry) is here reconstructed only via SEP glosses. Primary texts needed to verify the formal statement and Plantinga's own handling of the self-referential step.}} What the corpus does contain is the SEP's crisp statement of the thesis: naturalism "is epistemically self-defeating: if both naturalism and evolution are true, then it's unlikely we would have reliable cognitive faculties" (SEP 'Religion and Science' §1.3), together with the claim that for Plantinga "the conflict is not between science and religion, but between science and naturalism" (ibid.).

The corpus also grounds the positive half of Plantinga's program — his theistic account of evolution. On divine action, "Plantinga suggests that randomness is a physicalist interpretation of the evidence. God may have guided every mutation along the evolutionary process," so that God could "guide the course of evolutionary history by causing the right mutations to arise at the right time and preserving the forms of life that lead to the results he intends" (Plantinga 2011: 121, quoted at SEP 'Religion and Science' §3.1). The point for the EAAN is structural: on theism the process that produced our minds was aimed, so reliable faculties are expected; on naturalism it was blind, so reliability is a lucky accident whose antecedent probability is the very thing in question.

Scripturally, the theistic side of the contrast draws on the imago Dei: humanity is created "in the image of God" (Gen 1:27 (bib)), which the tradition reads as including a rational nature apt for knowing truth; on "the hearing ear and the seeing eye, the LORD has made them both" (Prov 20:12 (bib)), the faculties are a designed gift; and Rom 1:20 (bib) — God's attributes "clearly perceived... in the things that have been made" — presupposes that human perception can reliably read the natural order. Col 2:3 (bib), locating "all the treasures of wisdom and knowledge" in Christ, supplies the deeper claim that reason itself is grounded in a rational source rather than in blind matter.

Leading proponents

• Alvin Plantinga — originator of the EAAN (1993) and its fullest statement in Where the Conflict Really Lies (2011). Primary anti-naturalism works not in corpus; the argument is cited via SEP (see gap flag). Papineau's entry independently notes Plantinga's dissent from the "no relevance" version of methodological naturalism: "Some think that religious doctrines do make a difference to scientific practice, yet are defensible for all that (Plantinga 1996)" (SEP 'Naturalism' §2.1).
• C.S. Lewis — C. S. Lewis — a historical antecedent via the argument from reason (treated as the third view below); Plantinga's EAAN is the probabilistic descendant of Lewis's causal worry.

Strongest counter-arguments

The most damaging objection is internal to the argument's own logic: the generalization worry. If the defeater in premise (4) really defeats every deliverance of our faculties, it is hard to quarantine — it appears to give anyone who entertains it a reason to doubt all their beliefs, which looks like a recipe for global skepticism rather than a targeted refutation of naturalism. Critics press that a "defeater" that defeats too much defeats nothing, since a self-consuming skeptical argument cannot rationally compel any particular conclusion (including "therefore theism"). This is developed under the naturalist view below.

Second, the probability of (3) is contested at its root: whether P(R | N&E) is genuinely low, rather than high or simply unassignable, depends on substantive claims about the relation between belief-content and behavior that the naturalist rejects (see the content objection below). If true beliefs are, as a matter of the best cognitive science, the normal and expected products of adaptive cognition, premise (2) fails and with it the whole argument.

Third, the theistic escape in premise (6) is not obviously clean: the theist must show that P(R | theism) is high, which requires assuming that a designing God would want and succeed at making reliable knowers — an assumption a critic can press (a deceiving or indifferent designer, or one with non-epistemic priorities, is not ruled out by bare theism).

Responses

Plantinga's camp replies, first, that the generalization worry misreads the dialectic: the argument is a conditional — it shows that if you accept N&E you are in trouble — and the rational response is not global skepticism but abandoning one conjunct (naturalism), since theism is precisely the position on which the defeater never arises. Second, to the content objection they reply that its assumption — that true beliefs are what drive adaptive behavior — is exactly what naturalism is not entitled to help itself to, since on naturalism it is the neurophysiology that does the causal work and the semantic content rides along; the burden is on the naturalist to show that selection tracks content-truth rather than behavioral output. Third, on the theistic-escape worry, Plantinga's move is comparative and modest: he needs only that P(R | theism) is not low while P(R | N&E) is, and the doctrine of the imago Dei (Gen 1:27 (bib)) supplies a principled reason for the former that naturalism lacks for its analogue. Whether these replies succeed is a genuinely live question in the literature.

Assessment

Assessment: Live — the EAAN is not the institutionally defeated intelligent-design program (see Methodological Naturalism — Scope and Limits for that contrast); it is a formal epistemological argument that continues to generate serious response and reply in analytic philosophy of religion. Its weakest joint is premise (2)/(3) — the claim that reliability is improbable rather than merely un-provable — and its most-debated joint is the generalization worry. Neither has produced a consensus verdict, which is why the article rates it live rather than strong.

Naturalist Rebuttals

Abstract

The naturalist replies come in two main families, and the SEP's framing of evolution as "the non-random retention of random variations" (SEP 'Religion and Science' §3.1) supplies the wedge for both. The content objection holds that Plantinga's premise (2) is false: true beliefs are adaptive, because an organism that mostly misreads its environment — mistaking predators for prey, cliffs for meadows — is selected against, so selection does after all favor reliability. The proves-too-much objection holds that even if the argument's logic were granted, its defeater cannot be contained: it would undercut all belief whatever, theistic belief included, and a weapon that destroys every position refutes none. A distinct, more sympathetic naturalist strand — associated with Thomas Nagel — concedes that reductive materialism struggles to account for the reliability of reason, but takes this as a reason to expand naturalism (toward a non-theistic teleology) rather than to convert to theism.

Formal statement

1. Adaptive behavior in a complex, dangerous environment is reliably produced only by an organism whose beliefs largely track that environment's real features; systematically false beliefs would, over evolutionary time, be maladaptive.
2. Hence P(R | naturalism & evolution) is not low — selection is an error-correcting filter that favors truth-apt cognition. Premise (2) of the EAAN fails.
3. Independently: if the EAAN's defeater defeated every deliverance of our faculties, it would defeat theistic beliefs and the argument's own premises too; a globally self-consuming defeater licenses no conclusion.
4. Therefore naturalism-and-evolution is not shown to be self-defeating; at most, reductive materialism about mind is under pressure, which motivates a richer naturalism rather than theism.

Key evidence / textual basis

The naturalist has the letter of evolutionary theory on the point that selection retains variations non-randomly — i.e., by their fitness consequences — which is the corpus-grounded basis for insisting that catastrophically unreliable cognition would be filtered out (SEP 'Religion and Science' §3.1). The proves-too-much objection connects to a classical epistemological point that the corpus states directly through Bertrand Russell: all reasoning terminates in "instinctive beliefs" that cannot themselves be non-circularly vindicated. Russell observes that "we come by our belief in an independent external world" not by argument but by finding it "ready in ourselves as soon as we begin to reflect: it is what may be called an instinctive belief" (Russell 1912, ch. II), and that "all knowledge, we find, must be built up upon our instinctive beliefs, and if these are rejected, nothing is left" (Russell 1912, ch. XI). The naturalist deploys this against Plantinga: every epistemology, theistic ones included, must at some point trust faculties it cannot certify from outside, so a demand for external certification of reliability, if it defeats naturalism, defeats knowledge as such — Russell's point that "it is of course possible that all or any of our beliefs may be mistaken... but we cannot have reason to reject a belief except on the ground of some other belief" (Russell 1912, ch. XI) is a naturalist-friendly reply that the defeater is idle because it is universal.

For the more concessive strand, Thomas Nagel — himself an atheist — grants in Mind and Cosmos (2012) that the reliability of reason is exactly what a purely materialist evolutionary story has trouble explaining, and treats this as grounds for a non-theistic, teleological expansion of naturalism rather than a reason to accept design. {{UNSOURCED: Nagel's Mind and Cosmos (2012) is copyright-locked and not in corpus; his concession that reductive materialism cannot account for the reliability of reason, and his non-theistic teleological alternative, are reported here only from secondary framing. Primary text needed to verify that Nagel grants the EAAN's diagnosis while rejecting its theistic conclusion.}} This matters dialectically: Nagel is the strongest possible witness that the EAAN's problem is real, while being the strongest possible witness that its conclusion (theism) does not follow.

Leading proponents

• Thomas Nagel — concedes the difficulty reductive materialism faces with the reliability of reason (Mind and Cosmos, 2012), but infers an expanded naturalist teleology, not theism; primary work not in corpus (gap-flagged).
• Graham Oppy — the wiki's systematic contemporary naturalist; on his general program "no extant theistic argument succeeds, and naturalism beats theism on theoretical simplicity at equal explanatory power" (Graham Oppy profile, from SEP CA §1) — the natural home for the proves-too-much and simplicity replies to the EAAN. Oppy's EAAN-specific writings are not in corpus; the position is represented via his general naturalist stance.
• Ernest Nagel / mid-century naturalists — the SEP notes the naturalist tradition holds "reality is exhausted by nature, containing nothing 'supernatural'" (SEP 'Naturalism', preamble); this is the metaphysical commitment the EAAN targets and these authors defend.

Strongest counter-arguments

Plantinga's rejoinder, restated as an objection to the rebuttal: the content objection assumes that it is the truth of the belief, rather than the underlying neurophysiology, that does the adaptive work — but on naturalism the causal story runs through physical states, and semantic content is either identical to those states (in which case its truth is still not what selection sees) or supervenient on them (in which case truth is causally idle). So the naturalist's "true beliefs are adaptive" begs precisely the question the EAAN raises. Against the proves-too-much reply, Plantinga insists the defeater is not universal: it arises only for one who holds the specific conjunction N&E, so theism is not symmetrically threatened, and the Russellian "everyone trusts instinctive beliefs" point misses that theism warrants that trust while naturalism cannot.

Responses

Naturalists respond that the neurophysiology/content dilemma is a general problem of mental causation afflicting everyone, theists included, and cannot be weaponized selectively; and that Nagel's route shows the honest reading is "materialist reductionism is inadequate," which motivates better naturalism, not theism. On the asymmetry claim, they press that theism's warrant for reliability itself rests on faculties whose reliability is in question — the theist reasons her way to a reliable-faculty-making God using the very faculties — so Plantinga's circle reappears for the theist. Whether the circle is vicious for one side and virtuous for the other is the unresolved crux.

Assessment

Assessment: Live — the content objection is the majority naturalist reply and is taken seriously by both sides; the proves-too-much objection is powerful but meets Plantinga's asymmetry defense head-on without a knockout. Nagel's concession is dialectically important precisely because it grants the premise while denying the conclusion, showing the argument's terminus (theism) is not forced even if its diagnosis bites.

The Argument-from-Reason Cousin

Abstract

The EAAN has an older and distinct relative: the argument from reason, given its best-known form by C. S. Lewis in Miracles (1947) and revived by Victor Reppert. Where the EAAN is probabilistic and Bayesian — it computes a low conditional probability of reliability and derives a defeater — the argument from reason is about the causal and logical status of rational inference itself: it contends that if every mental event, including an act of reasoning, is fully determined by non-rational (physical, ultimately chance-and-necessity) causes, then no belief is ever held because it is warranted by the grounds for it, and the very notion of a valid inference dissolves. The two arguments are cousins, not twins, and the article keeps them distinct: the EAAN can grant that we do reason and ask whether that reasoning is reliable; the argument from reason asks whether, on naturalism, "reasoning" in the truth-tracking sense so much as occurs. A dedicated The Argument from Reason (Lewis / Reppert) article is forthcoming (currently queued); this section marks the relation and the distinction.

Formal statement

1. Rational inference requires that one belief be held because it is a ground for another — a relation of rational, not merely causal, dependence.
2. On naturalism, every mental event is produced by a chain of non-rational causes (physical states governed by law and chance).
3. A belief produced solely by non-rational causes is not held because of its logical grounds, whatever its content.
4. Therefore, if naturalism is true, no belief is held on rational grounds — including belief in naturalism — so naturalism undermines the possibility of reasoned belief in it.

Key evidence / textual basis

The argument from reason is a causal-status argument turning on the ground/cause distinction in premises (1)–(3), not on a probability calculation — which is what separates it from the EAAN. The corpus does not contain Lewis's Miracles, so this section states the argument's structure rather than quoting it. {{UNSOURCED: Lewis's Miracles (1947, rev. ch. 3) and Reppert's C.S. Lewis's Dangerous Idea (2003) are not in corpus; the argument-from-reason's premises are stated here from standard secondary framing. Primary texts needed to verify Lewis's exact formulation and to build out the queued The Argument from Reason (Lewis / Reppert) article.}} The EAAN can be read as Lewis's intuition made probabilistic: instead of the strong claim that naturalism makes reasoning impossible, Plantinga makes the weaker, defeater-generating claim that naturalism makes reliable reasoning improbable. They share a diagnosis — unguided nature does not aim at truth — and diverge on the modal force of the conclusion.

The theistic alternative both arguments presuppose is scriptural in its roots: reason as a participation in a rational source, "in [Christ] are hidden all the treasures of wisdom and knowledge" (Col 2:3 (bib)), and the mind as designed equipment, "the hearing ear and the seeing eye, the LORD has made them both" (Prov 20:12 (bib)).

Leading proponents

• C. S. Lewis — the argument from reason in Miracles (1947), esp. the revised ch. 3; primary work not in corpus (gap-flagged). Lewis's target is naturalism's claim to be a conclusion of reasoning while denying reasoning any non-natural status.
• Victor Reppert — C.S. Lewis's Dangerous Idea (2003), the leading contemporary defense and systematization; not in corpus.

Strongest counter-arguments

The decisive naturalist reply is compatibilist: premises (1) and (3) present a false dilemma between "held because of grounds" and "produced by causes," whereas a naturalist can hold that a belief is caused by physical states that themselves instantiate or track the logical relations — the causal process can be the ground-tracking process, as it is in a well-designed computer or, the naturalist adds, a well-evolved brain. On this reply, rational and causal explanation are not rivals but the same process under two descriptions, and premise (3) simply assumes what it needs to prove. This is a stronger reply against the argument from reason than against the EAAN, because the EAAN can concede that grounds are causally efficacious and still ask whether the process reliably tracks true grounds.

Responses

Defenders answer that the compatibilist "same process, two descriptions" move must show that the physical process tracks logical validity rather than mere behavioral success — and that this is exactly the point at which the argument from reason hands off to the EAAN, since selection rewards success, not validity. The debate thus loops back to the content objection of the second view. Where it remains genuinely open is whether a naturalized account of following a rule or grasping an entailment can be given without smuggling in normativity that naturalism cannot fund — an unresolved question in the philosophy of mind.

Assessment

Assessment: Live — the argument from reason is older and less formally tractable than the EAAN, and the compatibilist reply is widely thought to blunt its strongest form; but the residual question (whether truth-tracking, as opposed to success-tracking, can be naturalized) is exactly the EAAN's live question, which is why the two arguments are best studied together and kept conceptually distinct.