Home » Library » Modern Library » The Argument from Reason: C. S. Lewis’ Fundamental Mistakes

The Argument from Reason: C. S. Lewis’ Fundamental Mistakes


[This paper was originally presented as a talk at the “The Self and the Search for Meaning” Philosophy Symposium of the C. S. Lewis Summer Institute at Oxford University from July 28-August 8, 2008.]

C. S. Lewis’ argument against naturalism, found in Chapter 3 of Miracles (arguably his most philosophical work), is perhaps his most famous—not because it convinced many people that naturalism is incoherent, but because of the debate that it inspired. After its original publication in 1947, G. E. M. Anscombe (then a budding philosopher) responded to the argument at Oxford University’s Socratic Club in 1948—while Lewis presided as president—and, arguably, demolished him. His confidence as a philosopher seemed decreased. Although some devoted Lewis fans deny this, the most recently published third volume of his collected letters seems to confirm it. In them he admits to being “obliterated” (Letters 3:35), having lost his “dialectical power” in the wake of the debate (Letters 3:129), and he acknowledges that he refused to do original apologetics ever again, even when asked (Letters 3:651).

After his encounter with Anscombe, Lewis turned to producing fictional works like The Chronicles of Narnia—leading some Lewis fans to view the encounter as a blessing in disguise (like Cook, 2007). However, he did revise the argument in light of Anscombe’s criticisms and published it in a revised edition of Miracles in 1960. It is this revised argument that is my concern. Did the revisions adequately address Anscombe’s concerns and make the argument effective? Did they defend the major conclusion of Miracles and Lewis’ reputation as a philosopher? I will argue that the answer to these questions is “no.” After laying out Lewis’ argument, I will point out that mistakes due to Lewis’ fundamental misunderstanding of evolutionary theory and philosophy of mind derail his argument.

Lewis’ Argument

For those unfamiliar with the debate, a definition of “naturalism” is in order. The naturalist is one who holds that there is no supernatural—that all that exists is “the natural” or “the physical.” It is not necessarily a version of “materialism” (which suggests all that exists is matter), although most naturalists lean that way. But the defining aspects of naturalism, for Lewis, are closer to the defining aspects of determinism:

What the Naturalist believes is that the ultimate Fact, the thing you cannot go behind, is a vast process in space and time which is going on of its own accord. Inside that total system every particular event … happens because some other event has happened; in the long run, because the Total Event is happening. [emphasis original] (Miracles, p. 8)

Lewis argues that if naturalism is true, then human reasoning is not “valid”[1]—no derived conclusion is derived rationally. Why? Suppose that you draw a conclusion and I ask, “Why do you think that?” There are, at least, two kinds of explanation that you could give. First, you might recall the reasoning process that you went through to draw the conclusion; for example, “A, and if A then B, so I concluded B.” That would be a “grounds” explanation. On the other hand, you might explain that your brain went through a certain physical cause-effect process that caused you to have that belief; for instance, “the neurons in my brain did such-and-such and caused me to conclude B.” That would be a neurological “cause-effect” explanation. Lewis argues that since (1) neurological cause-effect processes are nonrational and (2) if naturalism is true, the neurological cause-effect explanation is the fundamental one (the “real” explanation for why you drew that conclusion), the naturalist cannot say that your conclusion was derived rationally—and the same would be true of all conclusions.

Why does Lewis think (1)—that neurological cause-effect processes are nonrational? Because they are not grounds explanations:

[R]easoning has no value as a means of finding truth unless each step in it is connected with what went before in the Ground-Consequent relation…. [I]f what we think at the end of our reasoning is to be [justified][2], the correct answer to the question “Why do you think this?” must begin with the Ground-Consequent because. (Miracles, p. 23)

But why is grounding required for rationality? According to Lewis, because “unless our conclusion is the logical consequent from a ground, it will be worthless and could be true only by a fluke” (Miracles, p. 23). Why would it be a fluke, though? On naturalism, so the thinking goes, our undergoing these processes is the result of natural selection; but if so, the process was initially nonrational—like a simple response to stimuli. It led to truth only accidentally. Yes, the response was beneficial to survival, and so it was selected and passed down, but selection could not have made the process rational.

Why does Lewis think (2)—that if naturalism is true, the neurological cause-effect explanation is the fundamental one? Upon correction from Anscombe, Lewis does admit that “grounds” and “cause-effect” explanations could coexist; they might even “explain” the same phenomenon or conclusion. But he argues that on naturalism the grounds explanation has nothing to do with the conclusion drawn. Why? The cause-effect process is just a part of a piece of a cause-effect process that begins at the Big Bang. Removing the grounds for the conclusion would not prevent the causal process from yielding that conclusion. Thus, the argument goes, on naturalism the cause-effect process must be the fundamental explanation.

So, if naturalism is true, no derived conclusion is derived rationally. Is Lewis ultimately arguing that naturalism is self-defeating? He was arguing that in 1947; it is not so clear that he was arguing that in 1960.[3] Regardless, if naturalism entails that all conclusions are derived nonrationally, naturalism is in trouble.

Lewis’ (Mis)Understanding of Evolution

Lewis’ argument relies on a number of misunderstandings regarding evolution or explanations couched in terms of natural selection, the kinds of explanation that most naturalists provide for why our brains function as they do. For starters, Lewis claims that neurological cause-effect processes selected naturally cannot be rational. Why? Natural selection was not designed to produce true beliefs, so any belief brought about by processes naturally selected could only be true by a fluke—the first such process would have just been a nonrational response to stimuli that happened to get it right. Yes, it would have been useful and thus naturally selected, but since selection could not have made such processes rational, it must be that the neurological cause-effect processes that naturalists think lead them to their conclusions are not rational:

Any thing which professes to explain our reasoning fully without introducing an act of knowing thus solely determined by what is known, is really a theory that there is no reasoning. But this … is what Naturalism is bound to do. It offers what professes to be a full account of our mental behavior; but this account, on inspection, leaves no room for the acts of knowing or insight on which the whole value of our thinking, as a means to truth, depends…. If there is nothing but Nature, therefore, reason must have come into existence by a historical process. And, of course, for the Naturalist, this process was not designed to produce a mental behavior that can find truth. There was no Designer…. The type of mental behaviour we now call rational thinking or inference must therefore have been ‘evolved’ by natural selection…. Once, then, our thoughts were not rational. That is, all our thoughts once were, as many of our thoughts still are, merely subjective events, not apprehensions of objective truth. Those which had a cause external to ourselves at all were (like our pains) responses to stimuli. Now natural selection could operate only by eliminating responses that were biologically hurtful and multiplying those which tended to survival. But it is not conceivable that any improvement of responses could ever turn them into acts of insight…. The relation between response and stimulus is utterly different from that between knowledge and the truth known. (Miracles, pp. 27-28)

There are a couple of errors here. First, Lewis does not understand how evolution would have given rise to the neurological physical processes that lead to true conclusions. Lewis seems to think that a single random mutation (of DNA) caused a human to go through a specific neurological cause-effect process (to have a “thought”) that just happened to lead that human to an advantageous true belief when presented with a specific stimuli—and thus that mutation was naturally selected. But the real evolutionary story, inspired here by V. S. Ramachandran and Sandra Blakeslee (1998), would go something like this:

After a very simple brain evolved, as a cluster of a few nerve cells, there was a slight DNA mutation that gave rise to a slightly larger brain with a resulting more complicated neurological structure. This structure allowed that organism to perform certain beneficial tasks better than its competitors (although other acquired skills were simply neutral). Over a long period of time this happened over and over; each mutation leading to a slightly bigger brain and a more complex neurological structure than the last. This made possible more beneficial abilities—like accurately throwing a spear—and also more neutral ones—like music composition.

Lewis’ mistake can be seen by example. Take spear-throwing. Do we have the ability to accurately throw a spear “by a fluke”? No. It is not as if there were once a bunch of cavemen walking around badly throwing spears until one day, by a fluke, one was born with a brain that went through a neurological cause-effect process that gave rise to the naturally selectable trait of “good aim.” The ability to accurately throw a spear is the result of the gradual process of getting a more complex brain. As it developed, a number of abilities developed with it—motor control, depth perception, and so on—that eventually allowed the caveman to accurately throw a spear. Likewise, it is not as if there was once a collection of cavemen walking around forming false beliefs (or no beliefs) until one day, by a fluke, one was born with the naturally selectable trait of having a brain that went through a neurological cause-effect process that lead to a true belief. As the brain got bigger, a number of abilities gradually developed, including the ability to form true beliefs (which, like spear-throwing, was probably made possible by the combination of other abilities.)

Was the ability to form true beliefs one of the advantageous abilities that allowed for a bigger brain to be naturally selected, or was it just an ability that we developed because we needed bigger brains to throw spears? Probably the former, but either way, contrary to Lewis’ suggestion, the ability to accurately form beliefs is not the result of a specific neurological cause-effect process that was selected because it just happened to lead to true beliefs.

But even if it had been, to say that such processes lead to truth by a fluke is also mistaken. To say as much is to confuse the genetic mutation that makes the process possible with the process itself, as well to think that an absence of a designer for natural selection entails that the development of specific traits is left to chance. Again, an example will help.

The gazelle that gets longer legs as a result of a random mutation of DNA can run faster. Perhaps that mutation was a fluke, in a sense—it was the result of a random pairing of DNA strands. But the fact that its longer legs make it run faster is no fluke. That is what longer legs do. That this particular gazelle got the longer legs is a bit of a fluke, but that gazelles would get longer legs is not. Granted, on naturalism, no one designed natural selection; but given the way that selection pressures and bisexual reproduction work, and the large amount of time that they have to do so, it is nearly inevitable that such a mutation would occur and be naturally selected. It is not a fluke.

In the same way, the person who gets a brain able to go through a “truth-getting-cause-effect-process” as a result of a random mutation of DNA is able to come to form true beliefs better than others. Perhaps that mutation was a fluke—it was the result of a random pairing of DNA strands. But the fact that this process leads to true beliefs is no fluke; that is what such processes do. That this particular person “got” this process is a bit of a fluke, but that humans would acquire the ability to go through such processes is not. Granted, no one designed natural selection, but given the way that selection pressures and bisexual reproduction work, and the large amount of time that they have to do so, it is nearly inevitable that such a mutation would occur and be naturally selected.

Lewis’ mistake here is the same mistake that some modern religious fundamentalists make in understanding evolution. They argue, usually when attempting to get evolution out of the science classroom, that evolution cannot account for the existence of biological phenomena because such things cannot arise “randomly”—as if that is what evolutionary theory suggests that they do. It does not. (According to evolutionary theory, the development of such traits over time is highly likely—that is one reason why the theory is so successful.) Given that Lewis’ argument relies so heavily on a similar (mis)understanding of evolution, his argument has suffered a fatal blow.

Lewis, the Argument from Reason, and Philosophy of Mind

In his original argument, Lewis thought that grounds and cause-effect explanations were mutually exclusive; Anscombe rightly pointed out that both explanations can rightly apply to the same phenomenon. Lewis grants the point, but fails to appreciate the significance of the critique. Lewis suggests that, according to naturalism, even though both kinds of explanation can apply to conclusions, a grounds explanation cannot be the full explanation for a drawn conclusion because one could remove the grounds without subtracting from the cause-effect explanation:

But even if grounds do exist, what exactly have they got to do with the actual occurrence of the belief as a psychological event? If it is an event it must be caused. It must in fact be simply one link in a causal chain which stretches back to the beginning and forward to the end of time. How could such a trifle as a lack of logical grounds prevent the belief’s occurrence or how could the existence of grounds promote it? (Miracles, pp. 24-25)

And, according to Lewis, since cause-effect processes cannot be rational, the naturalist cannot think that any belief is properly grounded.

But those who understand philosophy of mind understand Lewis’ error. For a long time, philosophers suspected that there is a direct link between the mental and the brain; modern developments in neuroscience have solidified this suspicion. If the mind and the brain are not just one and the same thing (many philosophers think that they are), the mental is at least an emergent property of the brain that arises due to its complexity. You can hinder or prevent mental process by hindering or preventing brain processes; and, in general, you cannot affect one without affecting the other. Such a relationship is called a “supervenient relationship.” Mind, we would say, supervenes on brain.

Given this supervenient relationship, the mutual presence of cause-effect and grounds explanations is not just grantable, but inevitable. Grounds explanations supervene on neurological cause-effect explanations. Thus, contra Lewis, it is not the case that you could remove grounds without altering the conclusion drawn by the cause-effect process. To remove grounds one would have to alter the physical process, which would alter the conclusion drawn by the process (or prevent the process altogether). To affect one is to affect the other. You cannot subtract “grounds” without affecting the neurons.[4]

Perhaps we can understand my point most simply this way: Lewis is arguing that only grounding explanations can make conclusions rational, and only mental processes can be grounding explanations. Since he does not think that there is any way that naturalists can maintain that the processes by which they explain their conclusions are mental, he does not think that there is any way that the naturalist can say his/her conclusions are rational. But modern philosophy of mind shows us that the processes countenanced by naturalism can be mental either by being identical to or by giving rise to the mental. So naturalists can do what Lewis thought that they could not do:

It seems never to have occurred to Lewis that the problem to which he devoted so many pages can be avoided simply by asserting that brain processes are causally necessary conditions of acts of rational inference… (Beversluis, 2007, p. 165)[5]

One might think that these theories of mind came along too late for Lewis to account for them in 1960. But the notion that the brain secretes the mental was suggested by Karl Vogt in the 1800s, “identity theory” itself was first called such in 1933 (see Smart, 2007), and interest in identity theory of mind became prevalent in 1956 (see Place, 1956). In fact, Lewis was aware of it in Chapter 1 of Miracles, writing: “The great interlocking event called Nature might be such as to produce at some stage a great cosmic consciousness, and indwelling ‘God’ arising from the whole process as human mind arises (according to the Naturalists) from human organisms” (Miracles, p. 11). It is unfortunate that Lewis did not realize how this theory threatened the argument of Chapter 3.

Lewis Anticipates an Objection

Lewis anticipated something similar to my objections:

The evolutionary product which [the naturalist] has described could also be a power of ‘seeing’ truths [i.e., be a ‘grounds explanation’]…. The Naturalist might say “Well, perhaps we cannot exactly see—not yet—how natural selection would turn sub-rational mental behavior into inferences that reach truth. But we are certain that this in fact has happened. For natural selection is bound to preserve and increase useful behaviour. And we also find that our habits of inference are in fact useful. And if they are useful they must reach truth. (Miracles, p. 32)

Lewis responds:

But notice what we are doing. Inference itself is on trial: That is, the Naturalist has given an account of what we thought to be our inferences which suggests that they are not real insights at all. We, and he, want to be reassured. And the reassurance turns out to be one more inference (if useful, then true)—as if this inference were not, once we accept his evolutionary picture, under the same suspicion as all the rest. (Miracles, p. 32)

Lewis has challenged the naturalist’s notion of inference. When the naturalist presents an argument to answer him, Lewis just replies along the lines of, ‘You are using inference, but whether or not inferences are trustworthy is the issue here. You beg the question by answering my objection with an argument.’

One might think that a similar strategy could be used against my objections. But it cannot; the strategy is fallacious. You could use this strategy to discount, literally, any theory that you like. Take the best theory in the world, and then challenge it by suggesting that it makes trustworthy inference impossible; your argument can be as bad as you like. No matter what answer those defending the theory give, even if they point out how obviously bad your argument is, you can just say, “You are using inference, but whether or not inferences are trustworthy is the issue here. You beg the question by answering my objection with an argument.” A critique that could falsify even that which is true cannot be a valid critique.

In addition, this critique contradicts what Lewis just granted the naturalist the paragraph before: “‘[I]t is incontestable that we do in fact reach truths by inferences.’ Certainly. The Naturalist and I both admit this” (Miracles, p. 31). Since both parties agree that inferences reach truth, if the naturalist uses an inference to answer an objection, Lewis cannot then say that inferences are now “not allowed.” If both parties have granted that inferences reach truth, then they are allowable from the start.


There are modern defenses of Lewis’ arguments, such as Victor Reppert’s (2003) defense, and there are other versions of the argument from reason, such as Alvin Plantinga’s (1993) version. Richard Carrier (2004) critiques Reppert on the Secular Web; Reppert responded recently in C. S. Lewis as Philosopher (Baggett et al., 2008). Paul Draper (2007) also critiques Plantinga’s argument on the Secular Web, where Plantinga (2007) in turn responded. I do not have time to go into all of these exchanges, but am prepared to discuss them. All in all, I think that such arguments suffer from shortcomings similar to those of Lewis. None of this entails that naturalism is true, but I think that attempts to show that it is incoherent fail.


Since presenting this paper in 2008 in Oxford, for the Oxbridge C. S. Lewis Summer Institute “The Self in Search for Meaning,” I have developed my arguments against the argument from reason further. First, in Greg Bassham’s book C. S. Lewis’s Christian Apologetics: Pro and Con (Brill, 2016), I took the con against Victor Reppert’s defense of Lewis’ argument. A couple of years later, when Reppert published an additional reply to my critiques in Philosophia Christi, the journal was gracious enough to let me know and publish my reply (Vol. 20, No. 2, 2019). My half of these exchanges can be found for free on my academia.edu page.


[1] As most philosophers would be quick to point out, Lewis is misusing the term “valid” here. Arguments are valid, reasoning is not; and an argument’s validity is a matter of its form, nothing else. Naturalism will never make modus ponens invalid. Beversluis (2007) argues that this mistake is also detrimental to Lewis’ argument, but I will forgo elaboration on this point.

[2] The original word here is “true,” not “justified.” But, as we all know, the validity of the argumentation behind a conclusion has nothing to do with the truth of that conclusion—there are bad arguments for true things and good arguments for false things. The principle of charity forces me to interpret “true” here as “justified.”

[3] In 1947 the chapter that held the argument was called “The Self Contradiction of the Naturalist.” In 1960 it was called “The Cardinal Difficulty of Naturalism,” although what he says on p. 22 does not match the name change.

[4] That is, unless Lewis was thinking in terms of philosophical zombies—beings that have brains that function just like ours, but have no mental lives. A philosophical zombie has its mind subtracted without affecting the physical processes of its brain. But the metaphysical possibility of philosophical zombies is highly debatable. If the activities of the brain give rise to the mind, then it would seem that they do so necessarily. But even if they do not, in order to defeat Lewis’ argument all that the naturalist needs is an explanation for how brain processes could be rational—and he has that in modern philosophy of mind.

[5] This quote continues: “but denying that the latter are reducible to the former.” I left this out for brevity—it would have raised issues of “reducibility” that I do not have time to address in this paper. It is my contention that Lewis’ argument can be answered whether the mental is reducible to the physical, or not.


Baggett, David, Gary R. Habermas, and Jerry L. Walls (2008). (Eds.). C. S. Lewis as Philosopher: Truth, Goodness and Beauty. Downer’s Grove, IL: InterVarsity Press Academic.

Beversluis, John. (2007). C. S. Lewis and the Search for Rational Religion, 2nd edition. Amherst, NY: Prometheus Books.

Carrier, Richard. (2004). “Critical Review of Victor Repport’s Defense of the Argument from Reason.” The Secular Web. <https://infidels.org/library/modern/richard_carrier/reppert.html>

Cook, Edward. (2007). “‘She Obliterated Me as an Apologist’: Lewis and the Anscombe Legend.” Ralph the Sacred River web log. <http://ralphriver.blogspot.com/2007/01/she-obliterated-me-as-apologist-lewis.html>

Draper, Paul. (2007). “In Defense of Sensible Naturalism,” §4 (“Plantinga’s Faulty Inference“). In God or Blind Nature? Philosophers Debate the Evidence ed. Paul Draper. The Secular Web. <https://infidels.org/library/modern/paul_draper/naturalism.html>

Lewis, C. S. (2000). Miracles. San Francisco, CA: HarperCollins.

Lewis, C. S. (2007). The Collected Letters of C. S. Lewis, Volume 3: Narnia, Cambridge, and Joy, 1950-1963. San Francisco, CA: HarperOne.

Place, U. T. (1956). “Is Consciousness a Brain Process?” British Journal of Psychology, Vol. 47, No. 1 (February): 44-50.

Plantinga, Alvin. (1993). Warrant and Proper Function. New York, NY: Oxford University Press.

Plantinga, Alvin. (2007). “Against ‘Sensible’ Naturalism.” In God or Blind Nature? Philosophers Debate the Evidence ed. Paul Draper. The Secular Web. <https://infidels.org/library/modern/alvin_plantinga/against-naturalism.html>

Ramachandran, V. S. and Sandra Blakeslee (1998). Phantoms in the Brain: Probing the Mysteries of the Human Mind. New York, NY: William Morrow.

Reppert, Victor. (2003). C. S. Lewis’s Dangerous Idea: In Defense of the Argument from Reason. Downer’s Grove, IL: InterVarsity Press.

Smart, J. J. C. (2007). “The Identity Theory of Mind.” In The Stanford Encyclopedia of Philosophy (Summer 2007 edn.) ed. Edward N. Zalta. Stanford, CA: Stanford University. <https://plato.stanford.edu/archives/sum2007/entries/mind-identity/>

Tallon, Philip and Jerry Walls (2005, July 29). “The Lewis-Anscombe Debate: A Dramatization.” Play performed at the Oxbridge C. S. Lewis Conference, Oxford University, Oxford, UK.

Copyright ©2020 David Kyle Johnson. The electronic version is copyright ©2020 by Internet Infidels, Inc. with the written permission of David Kyle Johnson. All rights reserved.

all rights reserved