+
Evolution selects for reliable cognitive faculties because having true beliefs usually helps an organism survive and reproduce.
1. Behavior, not belief content, is what evolution observes.
Natural selection operates on behavior and physiology, not on the truth of beliefs. What matters for survival is how an organism acts, not whether the internal representations driving that action accurately describe reality. Two organisms that behave identically in the face of a predator are equally fit, regardless of whether one has true beliefs and the other has a bizarre but coincidentally adaptive set of false beliefs.
2. Many false-belief systems produce identical adaptive behavior.
Plantinga's central illustration makes this concrete. A prehistoric human who believes he must run toward large orange-striped cats to honor the sun god, but also believes those cats will flee if he charges them, will behave in exactly the same way as one who simply believes tigers are dangerous: he runs away and survives. Natural selection rewards the outcome, not the accuracy of the beliefs behind it. Evolution is blind to which belief system was responsible for the behavior.
3. Adaptive success in survival tasks does not extend to abstract domains.
Even granting that evolution might favor accurate perception of immediate physical threats, food sources, and social relationships, this tells us nothing about the reliability of our faculties in domains with no bearing on survival: the philosophy of mathematics, the foundations of logic, cosmological theory, ethical reasoning, or the question of whether naturalism itself is a coherent worldview. These are not survival tasks. Ancestral humans who reasoned accurately about quantum mechanics or modal logic had no reproductive advantage over those who did not. Evolutionary pressure simply does not reach these domains, which are precisely the domains we rely on when evaluating the EAAN argument itself.
4. For the more sophisticated philosophical versions of this objection, see Defeaters 4 and 5.
Philosophers Ruth Millikan and Fred Dretske have developed a detailed theory called teleosemantics, arguing that natural selection establishes a biological function for our beliefs to track environmental features, making truth-tracking a principled part of our evolutionary inheritance rather than a lucky byproduct. Daniel Dennett presses a related but distinct point: that getting things right about the environment is not merely correlated with fitness but is the very mechanism by which fitness is achieved in cognitively complex organisms. Both are serious positions that deserve careful treatment. See Defeater 4 for teleosemantics and Defeater 5 for Dennett's specific argument.
+
Even if evolution doesn't guarantee perfect reliability, it could still make it likely that our faculties are mostly reliable, and that's enough.
1. The EAAN targets justified trust in reliability, not perfect reliability.
Plantinga's argument does not require that evolution produce perfectly reliable faculties. The question is whether, given naturalism and unguided evolution, we have any principled basis for trusting our faculties in the sense required for science, philosophy, and careful reasoning. Even "mostly reliable" is a claim that needs grounding. Where does that grounding come from if evolution cares only about survival?
2. The crucial distinction is between survival-relevant domains and abstract domains.
The objection may seem plausible for narrow, concrete cognitive tasks: detecting nearby predators, identifying edible plants, recognizing familiar faces. In these domains, there is at least a rough argument that accurate perception pays off in fitness.
But our sophisticated theoretical reasoning stretches far beyond these narrow tasks. Consider what is required to evaluate the foundations of mathematics, assess the logic of philosophical arguments, reason about cosmological fine-tuning, or determine whether naturalism as a worldview is coherent. None of these had any bearing on whether our ancestors survived on the African savanna. Evolution had no stake in whether they got these right.
Think of a car navigation system that was calibrated and tested exclusively for navigating a small town. Within that zone it may be perfectly reliable. But hand it an intercontinental route and you have no basis for trusting it, because the system was never designed or tested for that purpose. Our evolved perceptual systems may be the navigation system; abstract theoretical reasoning is the intercontinental route.
3. "Mostly reliable" in survival-relevant domains does not rescue naturalism.
Even granting that evolution produced faculties reliable for immediate, survival-relevant tasks, this does not justify trusting our faculties when they venture into metaphysics, abstract logic, or the assessment of worldviews. The very reasoning the naturalist uses to evaluate the EAAN argument is exactly this kind of extended abstract reasoning. If its reliability in those domains is unjustified, then the naturalist's own evaluation of the argument is called into question.
4. A low or inscrutable probability still generates a defeater.
If, given naturalism and evolution, you should regard the reliability of your cognitive faculties across abstract domains as low or simply impossible to assess, then you have a reason to withhold trust from those faculties when they operate in precisely those domains. That undercutting condition is all the EAAN needs to reach its self-defeat conclusion.
+
If evolution undercuts trust in our cognitive faculties, then theists who accept evolution are also undermining their own beliefs.
1. The EAAN targets a specific combination: naturalism and unguided evolution together.
The argument's target is not evolution as a scientific description of biological history. It is the conjunction of evolutionary biology with naturalism as a complete metaphysical worldview. The problem is not that evolution occurred. The problem is that on naturalism, evolution is the only story there is: a blind, purposeless process with no intention of any kind behind it, producing minds that happened to develop and with no reason to trust that those minds track truth. Remove the metaphysical commitment to naturalism and the problem dissolves.
2. On theism, evolution can be a guided process aimed at truth-tracking creatures.
A theist who accepts evolutionary biology is not committed to unguided evolution. God could have worked through evolutionary processes to bring about beings with cognitive faculties calibrated for truth, not merely for survival. The mechanism is evolutionary, but the intention behind it is divine. This is not a claim that God intervened in each individual mutation. It is the claim that the entire process was grounded in and directed by a rational mind who intended its outcome. When the intention behind a process is truth-tracking, the products of that process can be trusted in a way that the products of a purely blind process cannot.
3. Theism gives a principled, non-circular basis for trusting truth-tracking faculties.
The key question is not merely whether our faculties happen to be reliable, but whether we have principled grounds to expect them to be. On theism, those grounds are substantial and non-circular.
The classical theological tradition identifies God with Logos, the rational ground of all reality. Augustine held that the human mind participates in divine illumination, the light by which truth is known. Aquinas identified God with Veritas, Truth itself, and held that the natural light of human reason participates in God's eternal reason. Calvin spoke of a cognitive sense implanted by God specifically oriented toward true knowledge of God and the world.
On this picture, a God who is himself Truth would not create minds aimed at falsehood or mere survival-fitness. That would contradict His own nature as the rational ground of being. The expectation of truth-tracking minds flows from who God is, not from an arbitrary divine choice.
4. Naturalism has no analogous grounding resource.
The theist's response does not merely assert that God made reliable minds. It grounds that expectation in the nature of God himself. On naturalism, there is no rational personal intention behind the cosmos at all. No designer valued truth, no teleological aim was built into the evolutionary process. The evolutionary story is simply what happened: mutations, selection pressure, reproductive success, with no mind behind any of it.
This asymmetry is decisive. The theist and the theistic evolutionist have a story about why truth-tracking is to be expected, rooted in the character of the Creator. The naturalist has no such story. That is precisely why the EAAN poses a problem for naturalism but not for theism.
+
At best the EAAN pushes us toward global skepticism, not toward God. Why think theism is the solution?
1. The EAAN is primarily an internal critique of naturalism.
The main conclusion of the argument's first stage (C1) is not that theism is true, but that naturalism combined with unguided evolution is self-defeating: it undercuts the rational basis for accepting itself. That is a significant result entirely on its own, before any alternative worldview enters the picture. The naturalist cannot sidestep this by asking where the argument goes next.
2. We cannot actually live as though our cognitive faculties are radically unreliable.
There is a powerful practical dimension here. Wholesale skepticism about our cognitive faculties is not a live option for any human being. We cannot function without trusting perception, memory, inference, and rational judgment. We rely on these at every moment, in every area of life, including the moment of reading this argument.
This inescapable trust is itself a significant fact. If naturalism provides no principled basis for this trust, and yet the trust is inescapable, that is a strong reason to look for a worldview that does provide such a basis. We should be suspicious of any worldview that, when taken seriously, requires us to abandon something we genuinely cannot abandon.
3. Theism provides what naturalism lacks, and does so in a principled way.
On theism, a rational God creates a rationally ordered world and rational minds fitted to comprehend it. The fit between mind and world is not a cosmic accident. It is built into the structure of creation from the beginning. The ancient philosophical concept of Logos, which the opening of John's Gospel deliberately appropriates, names exactly this: the rational principle by which all things exist and through which they can be known. Mind and world share a common rational origin, which is why the one can understand the other.
This is not a vague gesture toward "something beyond nature." It is a specific, positive claim: a personal, rational God who is himself the ground of truth and intelligibility provides a principled basis for cognitive reliability. The explanation is substantive and grounded, not a placeholder.
4. Why theism rather than other non-naturalistic worldviews?
The objection might press further: why not Kantian idealism, panpsychism, or some other non-naturalistic alternative? Several features distinguish theism.
Kantian idealism grounds reliability in the structure of the mind itself, making the fit between mind and world something the mind imposes rather than discovers. This cannot explain why abstract mathematical reasoning, which does not appear to be something we impose on reality, turns out to describe physical reality with astonishing precision.
Panpsychism adds mind-like properties to all matter but does not supply a personal agent who intended truth-tracking minds. Without intention, without a designer who valued truth, the gap between what evolution aimed at (survival) and what we need (reliable abstract reasoning) remains open. Distributing mind throughout matter does not close it.
Theism uniquely combines three features that together solve the problem: a personal agent who can intend truth-tracking, the metaphysical necessity to serve as the ultimate ground of explanation, and an intrinsic identification with Truth itself that gives principled, non-arbitrary reasons to expect minds aimed at truth. No other serious non-naturalistic worldview provides all three together.
+
Philosophers Ruth Millikan and Fred Dretske argue that natural selection establishes a biological function for our cognitive faculties to track environmental features. On this view, the content of a belief, what it is about, is determined by what the belief was selected to detect. This makes truth-tracking a principled part of our evolutionary inheritance, not a lucky accident, and gives naturalism a satisfying answer to P1.
1. Teleosemantics, stated carefully.
Millikan (Language, Thought, and Other Biological Categories, MIT Press, 1984) and Dretske (Naturalizing the Mind, MIT Press, 1995) propose that mental representations acquire their content through their evolutionary history. A belief-state that represents "there is a predator nearby" has that content because ancestral organisms with that type of belief-state were selected for accurately detecting predators. The belief's biological function is to track predators, and this function was established by the history of natural selection.
On this account, evolution does not merely produce behaviors that happen to be useful. It produces cognitive systems whose proper function is accurate representation of specific environmental features. This seems to give naturalism a principled, non-accidental story about truth-tracking.
2. Teleosemantics only grounds tracking within survival-relevant, concrete domains.
Even granting the teleosemantics framework entirely, the tracking it can establish is necessarily limited to whatever features of the environment were relevant to ancestral survival. Selection can only establish a proper function for detecting what ancestors actually needed to detect: immediate physical threats, food sources, potential mates, social dynamics, and similar concrete features of the immediate environment.
But this says nothing about the reliability of our faculties when they reason about entities and questions entirely disconnected from survival: the axioms of set theory, the philosophy of modality, the internal coherence of naturalism as a worldview, or the probability calculus applied to fine-tuning arguments. No ancestral hominid was selected for getting these right. Teleosemantics can establish a proper function for close-range detection of tigers. It cannot establish a proper function for evaluating cosmological arguments.
3. An analogy: designed range does not mean unlimited range.
Think of a thermometer designed and calibrated to measure temperatures between 0 and 40 degrees Celsius, the range relevant to the environment its manufacturer had in mind. Within that range, the manufacturer's guarantee is meaningful: the instrument was built for exactly this purpose. But if you use it to measure temperatures at 500 degrees or at negative 200 degrees, you are outside the designed operating range entirely. The instrument's proper function gives you no basis at all for trusting its readings in those conditions.
Our evolved cognitive systems are something like that thermometer. Teleosemantics may justify a degree of trust within the narrow evolutionary calibration range. Outside that range, within abstract reasoning about metaphysics, mathematics, and worldviews, the proper function story has nothing to say. We are using the thermometer in conditions for which it was never designed.
4. The proper function standard generates serious problems even within its own scope.
Even within survival-relevant domains, teleosemantics faces a deep difficulty. Organisms frequently malfunction. A perceptual system that is working as it was designed to work can still produce false beliefs when conditions are unusual: predators that are camouflaged, food that appears more plentiful than it is, social signals that are misread.
Millikan's own framework acknowledges that proper-function performance is the norm in a statistical or historical sense, not a guarantee for each token performance. Most individual instances of a belief-type may fail to fulfill their proper function even while the type has that function. If proper-function performance is the exception rather than the rule at the token level, then even the local tracking teleosemantics promises is less robust than it initially appears.
5. Teleosemantics does not restore the global reliability needed to trust naturalism itself.
The belief that naturalism is true is not a belief about nearby predators or ripe fruit. It is a wide-ranging metaphysical conclusion arrived at through extended philosophical reasoning. Teleosemantics has no resources to ground the reliability of our faculties when they produce this kind of belief. Even if teleosemantics fully succeeds on its own terms within its limited scope, it leaves the naturalist with no principled basis for trusting the cognitive processes that led to their belief in naturalism in the first place.
+
Philosopher Daniel Dennett argues that Plantinga fundamentally misunderstands how natural selection works. Getting things right about the environment is not merely correlated with survival. It is the very mechanism by which cognitively complex organisms achieve fitness. Organisms that systematically misrepresent reality are selected against. Therefore evolution directly selects for truth-tracking cognition, not merely for behavior that happens to be adaptive.
1. Dennett's argument, stated carefully.
In Darwin's Dangerous Idea (Simon and Schuster, 1995), Dennett contends that Plantinga's tiger scenario misrepresents how selection works in cognitively sophisticated organisms. For such organisms, Dennett argues, "getting it right" about the environment is not an accidental byproduct of fitness. It is constitutive of fitness. A hominid whose perceptual and inferential systems systematically fed it false information about the environment would make systematically bad decisions and be selected against. The pressure of natural selection in cognitively complex creatures is therefore directly a pressure toward accurate environmental representation.
2. Dennett conflates local perceptual accuracy with global abstract reliability.
Dennett's argument is plausible for a limited set of tasks: perceiving real predators, correctly identifying edible plants, accurately tracking the movements of prey and rivals. In these narrow, concrete domains, there is a reasonable case that perceptual inaccuracy costs fitness. Dennett is right that organisms with systematically distorted close-range perception would be selected against.
But "getting it right" in these survival-relevant perceptual domains is an entirely different matter from global abstract reliability of the kind required for evaluating philosophical arguments, assessing mathematical proofs, doing theoretical physics, or determining whether naturalism as a worldview is coherent. Our ancestors were not selected for getting modal logic right. They were not selected for accurately assessing cosmological models. In these abstract domains, the selection pressure Dennett describes simply does not reach.
3. An analogy: excellence in one domain does not transfer to another.
A champion sprinter has legs that are extraordinarily well-calibrated for short-distance running. That calibration is the result of years of specific training and natural aptitude. But it gives no basis for expecting their legs to be reliable for, say, performing microsurgery. Excellence in one domain, even if that excellence was produced by a selection-like process, does not transfer to unrelated domains.
Our evolved perceptual faculties may be excellent at detecting predators and locating food. This does not give us any basis for expecting those same faculties, or the abstract reasoning faculties that grew up alongside them, to be reliable when they reach into metaphysics, philosophy, and the assessment of worldviews.
4. Dennett's own argument illustrates the problem it is trying to solve.
Dennett's conclusion that evolution produces truth-tracking cognition is itself a philosophical argument. It is not reported from a specific study in evolutionary biology. It is an abstract inference about the general tendencies of natural selection, offered in a work of philosophical reflection.
But if our cognitive faculties are reliable only in survival-relevant domains, as the EAAN suggests, then we have no principled basis for trusting precisely this kind of extended abstract philosophical inference. Dennett is using the very kind of reasoning whose reliability is in question to argue that evolution guarantees reliability. The argument uses what it needs to prove.
5. Dennett's reply addresses a narrower target than the EAAN requires.
Even granting Dennett's point in full, it at best explains why evolution might produce reliable perceptual faculties for survival-relevant tasks. It does nothing to establish reliability in abstract theoretical domains, and it does nothing to restore the principled basis for trusting our faculties across all the domains we need to trust when we are assessing worldviews. The EAAN's deepest challenge is that naturalism provides no story about why our faculties should be reliable in exactly those abstract, truth-sensitive domains. Dennett's response does not touch that challenge.
+
Even granting that a God exists, we have no clear reason to think He specifically designed us to have truth-tracking cognitive faculties. God could have created us primarily for worship, obedience, moral formation, or some purpose entirely unrelated to theoretical truth. The probability of reliable faculties given theism is not obviously high, so the contrast with naturalism is less dramatic than the argument claims.
1. The objection, stated carefully.
This challenge, developed in responses to Plantinga's argument such as those found in James Beilby's anthology Naturalism Defeated? (Cornell University Press, 2002), presses the following point. Even if God exists and created us, there are many possible divine purposes that would not specifically require truth-tracking cognitive faculties. God might have designed humans primarily for moral formation, for worship, for aesthetic participation in creation, or for some purpose entirely opaque to us. On this view, P(R/theism), meaning the probability of reliable faculties given theism, is not obviously high, and the contrast with naturalism is therefore less decisive than the argument suggests.
2. Classical theism does not merely assert that God made reliable minds. It grounds that expectation in God's own nature.
The objection treats the connection between theism and truth-tracking minds as a contingent, optional feature that God might or might not have chosen. But on the classical theism that Plantinga defends, the connection is not contingent. It flows necessarily from the nature of God himself.
Aquinas identifies God with Truth itself (Veritas, Summa Theologica I.16.5). God is not a being who happens to value truth among other values. God is the standard and source of all truth. A God who is identical with Truth would not create minds systematically aimed at falsehood or indifferent to truth, any more than a God who is pure goodness would create beings constitutively oriented toward evil. The creation of truth-tracking minds is not one design option among many. It is demanded by who God is.
3. The Logos tradition provides principled grounds, not a bare assertion.
The ancient concept of Logos, which the Christian tradition identifies with the second person of the Trinity, names the rational principle through which all things were made and through which they are knowable. If the cosmos was created through the Logos, rational order is not a surface feature of creation that could easily have been otherwise. It is intrinsic to the creative act itself.
Minds made in the image of a God who is Logos are minds made for truth in a principled sense. Augustine's doctrine of divine illumination, Aquinas's account of the intellect as participating in God's own reason, and Calvin's sensus divinitatis all express this conviction from different angles: human cognitive faculties are oriented toward truth because God, who is Truth, made them so. These are not isolated proof-texts. They are expressions of a coherent, deeply worked-out theological picture of the relationship between the divine mind and the human mind.
4. The EAAN requires only a significant difference in probability, not certainty.
The argument does not require that P(R/theism) be certain or even very high. It requires only that it be significantly higher than P(R/N&E). Even granting some uncertainty about the full range of God's specific design purposes, a God whose nature includes perfect rationality, perfect goodness, and identity with Truth itself has far more principled reasons to create truth-tracking minds than a blind evolutionary process has, since that process has no intentions at all. An intentional, truth-valuing designer is a vastly better explanation of truth-tracking minds than no designer and no intention. The contrast remains substantial even if the theistic probability is not maximal.
5. The theist has converging reasons, not just a single inference.
The expectation of truth-tracking minds on theism is not a single isolated inference that might easily be reversed. It is supported by multiple independent and mutually reinforcing strands of the classical tradition: the doctrine of the imago Dei in the image of a rational God, the identification of God with Truth itself, the Logos theology of creation, and the broader classical claim that God is the ground of all intelligibility. These converging lines of reasoning give the theist robust grounds for expecting truth-tracking faculties, grounds embedded in a comprehensive and coherent theological picture rather than added on as an afterthought.
+
Decades of research by Daniel Kahneman, Amos Tversky, and others has shown that human cognitive faculties are systematically unreliable in many predictable ways: confirmation bias, base rate neglect, availability heuristic, scope insensitivity, and more. If our faculties are already empirically known to be unreliable, perhaps the EAAN's premise is simply confirmed by science. But this seems to cut against theism as much as naturalism: if God specifically designed our minds for truth, why are they so systematically prone to error?
1. The research on cognitive biases is genuine and important.
The work of Kahneman and Tversky (Thinking, Fast and Slow, Farrar Straus and Giroux, 2011) has made a lasting contribution to our understanding of human cognition. People do show systematic errors in probabilistic reasoning, susceptibility to framing effects, overconfidence in narrow domains, and many other well-documented patterns of flawed judgment. None of this is in dispute, and a serious account of human cognition must take it seriously.
2. These biases are domain-specific and correctable, which proves the faculties are broadly reliable.
The crucial point is that cognitive biases are not evidence of global, unrepairable unreliability. They are patterns of systematic but identifiable and correctable error in specific domains, particularly rapid judgment under uncertainty and intuitive probabilistic reasoning.
Consider: the very fact that we can identify these biases as biases, measure their magnitude, map their conditions, and teach people to correct for them proves that our faculties are broadly reliable. You cannot detect a systematic error without a reliable standard of comparison. A thermometer that was entirely broken could not be used to identify which of its readings were inaccurate. The extraordinary success of cognitive psychology as a scientific discipline in documenting human error depends entirely on the reliability of the scientific and rational faculties doing the documenting.
3. Correctable, domain-specific error is exactly what the theistic picture predicts.
The EAAN's positive conclusion (P5, C2) is not that theism guarantees perfect, error-free cognition. It is that theism provides a principled basis for expecting broadly reliable, truth-aimed cognitive faculties. Minds created in the image of an infinite God by finite, embodied creatures would be expected to be genuinely rational and truth-oriented while also being fallible, limited in information, and subject to distortions from emotion, evolutionary history, and cognitive shortcuts.
The pattern of broadly reliable faculties with specific, identifiable, and correctable weaknesses fits the theistic picture naturally. It is what we would expect from beings who are genuinely rational in nature but finite and fallen in condition.
4. The bias data actually supports the EAAN rather than refuting it.
Many cognitive biases follow a telltale pattern: they were locally adaptive in ancestral survival environments but mislead when applied in modern abstract contexts. The availability heuristic, for instance, causes people to overestimate the probability of dramatic and memorable events. For a hominid in a dangerous environment, systematically overweighting vivid memories of threats was probably adaptive; it is better to assume a lion is nearby than to assume it is not. But the same heuristic produces wildly distorted probability estimates when applied to abstract statistical questions.
This pattern is precisely what the EAAN predicts about cognitive faculties shaped by survival pressures rather than truth. On naturalism, we should expect this kind of local-adaptive but globally-unreliable pattern throughout, including in our most abstract philosophical and scientific reasoning. On theism, we should expect the broad reliability required for science to function as a discipline, alongside specific limitations tied to our finitude and our evolutionary history being part of God's means of creating us. The data fits the theistic picture more naturally than the naturalistic one.
+
Philosopher W.V.O. Quine argued that the traditional demand for an external validation of our cognitive faculties is an impossible and confused standard. We are always reasoning from inside our own cognitive situation. There is no Archimedean point outside our minds from which to evaluate whether our faculties are "really" reliable. The EAAN's demand therefore rests on a confused picture of what epistemology can achieve, and the question it raises cannot meaningfully be posed.
1. Quine's naturalized epistemology, stated carefully.
In "Epistemology Naturalized" (in Ontological Relativity and Other Essays, Columbia University Press, 1969), Quine argues that the traditional project of justifying our knowledge from some prior, external, and fully secure foundation is hopeless. We cannot step outside the web of our beliefs to evaluate that web from a neutral vantage point. Epistemology should therefore be "naturalized": pursued as part of empirical science, studying how human cognitive systems actually work rather than demanding a kind of guarantee that no one could ever provide.
On this view, asking "are our faculties reliably truth-tracking given that they are products of evolution?" is asking for exactly the kind of external validation Quine says is confused. We can only evaluate our beliefs using our beliefs. There is no higher court available.
2. Quine's move is self-undermining.
The most powerful response to this position is that it defeats itself. Quine's conclusion, that external validation is impossible and that epistemology should be naturalized, is itself presented as a philosophical conclusion arrived at by reasoning. Quine offers arguments. He expects those arguments to be evaluated and found persuasive by rational readers.
But if we are always inside our cognitive situation with no external standard available, why should we accept Quine's own philosophical argument? It too is a product of the faculties whose reliability cannot be externally validated, evaluated by cognitive faculties whose reliability cannot be externally validated. A position that says "no philosophical conclusions can be externally validated" cannot coherently be offered as a philosophical conclusion we should accept on the basis of argument. It cuts off the branch it is sitting on.
3. The EAAN does not demand an impossible Archimedean point.
Plantinga's argument does not claim to evaluate our faculties from outside our cognitive situation. It argues, from within the naturalist's own framework and using the naturalist's own commitments, that naturalism and evolution together provide no internal principled basis for trust in those faculties.
The request is not for an external guarantee from some neutral vantage point. It is for a coherent internal story, within naturalism, about why these faculties should be trusted in the abstract domains where the naturalist uses them to argue for naturalism. That is a legitimate internal demand. Quine himself cannot escape it: even the naturalized epistemologist needs some account of why their scientific methods are reliable, and if naturalism offers no such account, the Quinean program undermines itself.
4. Theism satisfies the demand Quine says cannot be satisfied.
Quine's naturalized epistemology ultimately accepts that we simply work with the cognitive faculties we have, adjusting our web of beliefs as ongoing experience demands. But this is precisely the move the EAAN challenges: a naturalist who simply works with their faculties without principled grounds for trusting them is doing something rationally unstable.
Theism does not demand an impossible view from nowhere. It offers an explanation within a coherent and comprehensive metaphysical framework for why our faculties are aimed at truth: a rational God whose nature is Truth itself created us with minds in His rational image, so that we might know Him, know the world He made, and know truth. This is an internal explanation grounded in a complete account of reality. It is not a demand for a view from outside all frameworks. It is a competing framework that, unlike naturalism, includes a principled account of why rational inquiry is trustworthy.
