The epistemic abyss of naturalistic evolution

However trustworthy our minds are, naturalistic evolution struggles to know itself


‘Darwin’s doubt’ and the Evolutionary Argument Against Naturalism (EAAN)

common.wikipedia.org, WhiteKnight138alvin-plantinga
Figure 1. Philosopher Alvin Plantinga.

Philosopher Alvin Plantinga (figure 1) developed a penetrating critique of evolutionary naturalism in his “Evolutionary Argument Against naturalism” (EAAN) (Christian philosopher sees no conflict with evolution: What he gets right and what he gets wrong).1 It’s based on a worry Darwin himself had:

“But then with me the horrid doubt always arises whether the convictions of man’s mind, which has been developed from the mind of the lower animals, are of any value or at all trustworthy. Would any one trust in the convictions of a monkey’s mind, if there are any convictions in such a mind?"2

Basically, if naturalism and evolution are true, it’s unlikely that our cognitive faculties are reliable. After all, given naturalism, brain chemistry determines our beliefs, and evolutionary processes are only aimed at producing brain chemistry beneficial for survival and reproduction, not at producing true beliefs. If my brain chemistry produces the beneficial behaviour of running away from a tiger, it doesn’t matter whether that brain chemistry produces a belief that that’s how I escape being eaten or a belief that that’s how I show love to a tiger. The only thing relevant to naturalistic evolution is the beneficial behaviour. As such, assuming naturalistic evolution, there’s no link between beneficial brain chemistry and the truth of the beliefs they produce. It’s just luck if my beneficial brain chemistry happens to produce a true belief.

But let’s say that a given belief produced by our beneficial brain chemistry has as good a chance of being true as it has of being false. If we have just 100 independent beliefs, then the chances of at least 75% of them being true is around 0.00001.3 So, we have reason to believe our cognitive faculties are untrustworthy, given naturalistic evolution.

But if we have a reason to distrust our cognitive faculties, we have a defeater for any belief formed by them—which applies to every belief we have. But this includes belief in naturalistic evolution. Therefore, believing in naturalistic evolution gives us a defeater for belief in naturalistic evolution. It’s a self-defeating belief that can’t be rationally accepted.

Conflicting evolutionary arguments

EAAN has generated many responses.4 The commonest response from naturalistic evolutionists is that they try to show that our belief content is linked to our brain states through evolutionary mechanisms so that they have substantial reliability.5 If that were true, that would allow for the naturalistic evolutionist to rationally affirm naturalistic evolution.

But there’s a problem with this response, one that’s generated by the apologetics of evolutionary naturalists. What is it? Well, cognitive science has shown that predispositions toward teleology, supernaturalism, and religion are hardwired into our cognitive architecture. Now, evolutionary models for the origin of these things have been developed. And evolutionary naturalists have used those models as evidence against teleology, the supernatural, and religion (Children see the world as ‘designed’!). But what’s the crux of the argument? That having an evolutionary explanation for these aspects of our cognitive architecture shows that they are not apt to produce true beliefs.6

However, the naturalistic evolutionist can’t plausibly have both arguments playing in his favour. Why? If one rejects EAAN by arguing that our cognitive architecture is generally reliable, then it significantly raises the probability that the very cognitive architecture which produces our belief in teleology, the supernatural, and religion is also reliable. But that implies that naturalistic evolution is false!

But if one embraces the argument for evolutionary naturalism from the supposed unreliability of our supernatural/teleological/religious beliefs, then it’s highly probable that EAAN also goes through. Thus, naturalistic evolution is impossible to rationally affirm.

Attempted naturalistic ‘escapes’

I see a few main escapes from this epistemic abyss that naturalistic evolutionists could try.

Belief in evolution is reliable, but belief in the supernatural is not?

For instance, they might carve up our cognitive architecture into bits that produce true beliefs and bits that don’t and say that belief in naturalistic evolution is formed by the bits that produce reliable beliefs, whereas those bits that form beliefs in teleology, the supernatural, and religion are not. This, however, would be special pleading of the basest kind.

Alternatively, they might not carve up our cognitive faculties into discrete bits, but they might just generalize. They might say that our cognitive architecture is reliable enough to produce a rational belief in naturalistic evolution while also being unreliable enough to call into question any notion of teleology/supernatural/religion. This, however, would render our cognitive faculties suspiciously ‘fine-tuned’ for belief in naturalistic evolution. That sounds rather like our cognitive faculties were designed to produce that belief! But of course, if they were designed to produce belief in naturalistic evolution, that belief would be false.

Naturalistic evolution is no worse off than any other worldview?

One other escape they might try is to say that, whatever epistemic problems there are for belief in naturalistic evolution because of our cognitive architecture, the problems for theist/supernatural/teleological/religious belief are worse. But why reason in that manner? Apart from some sort of analysis of all the supernatural/teleological/religious alternatives to show that epistemic problems of such worldviews are as big, there’s no reason to think this. Indeed, there may be reasons to deny this.

First, it’s not clear that we would have developed a teleological/supernatural/religious sense if naturalistic evolution were true. After all, many teleological/supernatural/religious worldviews can embrace evolution. They would also include such causes in the full-orbed explanation in a way that would justify belief in the spiritual/religious per se, or perhaps even in specific spiritual claims. But if so, the evolutionary origin of these elements of our cognitive architecture does not mean that they are unreliable.7 It’s only the alleged naturalistic evolutionary origin of our cognitive architecture that makes this likely. At least some supernatural/teleological/religious worldviews allow for our cognitive architecture to be more epistemically robust than naturalistic evolution does.

Moreover, these attempted accounts of the origin of our cognitive faculties are only as strong as the evidence for evolution itself (Design: just a trick of the mind?). Naturalism needs evolution to explain human origins, and thus it needs evolution to explain the origin of human cognition. So, if evolution falls, then these theories fall along with them. However, there are reasons to reject molecules-to-man evolution in general (Genetic entropy: The silent killer), and human evolution in particular (The waiting time problem).

Furthermore, to the extent that there is reason, independent of our automatic belief-forming mechanisms, to think that the supernatural/religious/teleological is true (Philosophical arguments for God), naturalistic evolution is undermined.

And finally, note how serious an epistemic jam the naturalistic evolutionist is in. If EAAN is right, naturalistic evolution cannot be rationally affirmed. If EAAN is wrong, our cognitive architecture provides evidence against the truth of naturalistic evolution. And if naturalistic evolutionists try to ‘split the difference’ (i.e. argue that our cognitive faculties are unreliable regarding belief in the supernatural/religious/teleological but reliable regarding belief in naturalistic evolution), they either beg the question in favour of naturalistic evolution or end up with designed cognitive architecture. Whichever way the naturalistic evolutionist tries to argue, it seems his arguments get shut down quite decisively. In contrast, many non-naturalistic accounts in general, and theistic ones in particular, allow fair play for the problems and limits of our cognitive architecture without sending us into such a deep epistemic abyss as does naturalistic evolution.


To avoid the supernatural, evolutionary naturalists end up tying themselves in knots. It is a conundrum-filled worldview in which mere brain chemistry argues against itself at every turn. Hamlet’s famous words to Horatio are apt, concerning naturalistic evolution:

“There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy.”

Published: 13 October 2022

References and notes

  1. Plantinga, A., Where the Conflict Really Lies: Science, Religion, & Naturalism, Oxford University Press, Oxford, pp. 310–345, 2011. Return to text.
  2. Darwin, C., To William Graham, 3 July 1881; darwinproject.ac.uk/letter/DCP-LETT-13230.xml. Return to text.
  3. Plantinga, ref. 1, pp. 333. Return to text.
  4. Beilby, J. (ed.), Naturalism Defeated? Essays on Plantinga’s Evolutionary Argument Against Naturalism, Cornell University Press, 2002. This book contains responses from 11 philosophers to EAAN. Return to text.
  5. Post, J.F., Review: Naturalism Defeated? Essays on Plantinga’s Evolutionary Argument Against Naturalism, ndpr.nd.edu, 8 July 2002. Post’s summary of the book shows that most chapters at least gestured towards naturalistic attempts to raise the likelihood our cognitive faculties are reliable given naturalistic evolution. Post himself made the same attempt. See also Plantinga, ref 1, pp. 335–339. William Lane Craig also notes that most of the discussion of EAAN has revolved around its first premise: given evolution and naturalism, the probability our cognitive faculties are reliable is low: #415 Plantinga’s Evolutionary Argument against Naturalism, reasonablefaith.org, 29 March 2015. Return to text.
  6. Shermer, M., Why people believe invisible agents control the world, scientificamerican.com/article/skeptic-agenticity, 1 June 2009. Return to text.
  7. Murray, M.J. and Schloss, J.P., Evolutionary accounts of religion and the justification of religious belief; in: Moreland, J.P., Meister, C., and Sweis, K.A. (eds.), Debating Christian Theism, Kindle edition, Oxford University Press, Oxford, pp. 242–258, 2013. Return to text.