Home » Library » Modern Library » Causes and Reasons: The Argument from Reason and Naturalism

Causes and Reasons: The Argument from Reason and Naturalism


Rationally Thinking Material Entities?
Reppert’s Arguments
“Drainage” of the Causes—Rickabaugh and Buras’ Argument


Arguments from reason (AfR) are theistic arguments that state that human reason’s capacity to know the truth, to judge, and to draw inferences cannot be trusted if human thinking is regarded as a purely natural (physical, biochemical, or neurobiological) process—if the human mind is a purely physical phenomenon.[1] Since naturalism regards itself as a rational worldview based on the reliability of human reason and scientific knowledge, if it follows from the truth of naturalism that human reason cannot acquire reliable, true knowledge, then naturalism is self-refuting. Consequently, if we are to avoid total skepticism, we must explain human thought and humanity’s ability to know the truth as deriving from divine intervention or creation, whether directly or indirectly.

Although the argument has many forms—in fact, we should not speak of a single argument, but rather of a “family” of arguments—the most famous and best-known one is that of C. S. Lewis. Thus I here expound Lewis’ argument—and some rather well-known criticisms of it—before I turn to more modern and more sophisticated versions of the argument.

After the Second World War, Lewis—who was originally known to the general public for his fiction, most notably The Chronicles of Narnia—was clearly the best known and most successful Christian apologist in Britain in the 1950s, defending Christianity not only in books but also in his famous BBC lectures.[2] The argument from reason is only part—albeit a very important and original part—of Lewis’ apologetics. He first explained it in 1947 in Miracles (Lewis, 1947).

The argument from reason is intended to support the claim that naturalism is a self-refuting theory. Why? Because according to naturalism, everything in the world is the result of physical causes, including the process of rationally inferring conclusions from premises. But physical causes, and processes determined by physical causes, are devoid of any kind of consciousness or purpose. Thus, according to naturalists, human thought is the result of an irrational process. If naturalism is right, when I infer something on rational grounds (the argument goes), then I merely believe or imagine that it is the content of the premises that leads me to the conclusion that I reach. In reality, it is the neurons and the neural pathways in my brain working in ways determined by the laws of nature that bring about my arrival at this particular conclusion. And since there is no ‘rationality’ or ‘purpose’ in physical processes, I must accept that the operation of rationality in my mind is determined by irrational processes. Consequently, I cannot trust the rationality of my own reasoning; nor can I trust the arguments that support naturalism. But if naturalism is wrong, I can assume that the ‘ground-consequent relation’ is ‘perceived’ by some rational agent within me, the soul or spirit, independently of the physical processes in my brain (even if I assume their presence).[3]

Lewis’ argument was challenged in a lecture at the Socratic Club of Oxford University in 1948 by the then-very-young and little known conservative Catholic English philosopher G. E. M. Anscombe. The essence of her critique was to question the presumption that a natural causal process cannot be both an operation of natural causes and, at the same time, a logical argument that reaches a certain conclusion from certain premises. Why should we assume that the physical causal processes that act on our minds necessarily influence them in a direction contrary to rationality? Although Lewis attempted to respond to the critique, the public perception was that Anscombe had defeated Lewis’ argument—and Lewis drew this conclusion, too. In the second, revised edition of Miracles in 1957, he replaced the term ‘irrational’ with ‘non-rational,’ even if not entirely consistently, and no longer claimed that his argument proved the self-refuting character of naturalism. He softened this to the weaker claim that the operation of human reason is very difficult to understand under the naturalistic worldview, and that this constitutes a cardinal difficulty for naturalism (Lewis, 2001).[4]

It is easy to understand the misunderstanding behind the original Lewisian argument. When rationality is influenced by some causal factor outside of rationality (for example, when one’s thinking about religion is determined by the religious beliefs of his/her sociological environment), that is usually detrimental to rationality: it makes it difficult or impossible for the human mind to grasp the truth. Therefore, if human thought is entirely determined by causal factors outside of reasoning under naturalism, then this certainly supports the argument that naturalism should not be relied upon as a basis for trusting the capacity of human thought to recognize the truth. The ontological possibility overlooked in Lewis’ argument, however, is that human thinking is itself a causal, physical process, but one that is very different in character from all other causal processes in nature. We now turn to newer versions of the AfR that do not regard human rationality or thinking as a per definition noncausal, nonnaturalistic process, but rather try to analyze some of rationality’s specific features in a nonnaturalistic way.

Today, philosophers of religion are primarily concerned not with the Lewisian form of the argument, but with the version of the argument developed by the contemporary American Christian analytic philosopher Alvin Plantinga, the so-called evolutionary argument against naturalism (EAAN) (Plantinga, 1994). Plantinga is not arguing that the mechanisms of human thought are unreliable in the light of naturalism, but that the process that led to the evolution of human thought—unguided biological evolution, according to naturalism—can reasonably be assumed to result in unreliable cognitive abilities. Therefore, we can only have confidence in human cognition if we presume that the process of humanity’s biological evolution was guided by a being (God) who can be presumed with high probability to have intentions to create intelligent creatures capable of recognizing the truth.[5]

Even though today’s debates tend to revolve around this argument, an American Christian philosopher, Victor Reppert, has attempted to revive and modernize the original Lewisian form of the argument (Reppert, 2003). The title of his book, C. S. Lewis’ Dangerous Idea, is an obvious reference to Daniel Dennett’s famous 1995 work on evolution, Darwin’s Dangerous Idea, which draws naturalistic conclusions from the theory of biological evolution.

Rationally Thinking Material Entities?

Before we examine Reppert’s argument in more detail, we must address one of the fundamental objections to the argument, which (of course) has been revived by Reppert’s critics, too. This objection is based on the fact that processes similar to rational human thinking exist outside of the human mind, and in the nonrational segments of the human mind, too. The most trivial example of the former is the digital computer. (Computer technology has not only progressed dramatically since Lewis’ time, but has also completely transformed our way of life and society. This is why the objection carries so much weight to today’s audiences.)

Digital computers are increasingly capable of performing sequences of operations modeled on the reasoning processes of conscious human thought, and it is at least possible that they will eventually come to perform all of the functions of human consciousness; but no one thinks that they have an immaterial ‘soul’ or ‘spirit.’ In addition, some critics of the argument from reason also point out that we human beings perform a significant part (perhaps most) of our reasoning and veracious inferential operations unconsciously. There are processes closely (even if not exactly) resembling human thought that are certainly determined by natural causes. Why should we then assume that conscious thought processes reveal something ‘completely different,’ a reality independent of the material world? (Johnson, 2015a, p. 98; Johnson, 2015b).

Argument from reason proponents reply (I think rightly) that digital computers have not been created by mechanical natural causes (those without consciousness or purpose), but by humans with consciousness and purpose. Of course, if we start from naturalism, this does not change the fact that the operation of digital computers is also determined by physical causal mechanisms operating in the material universe; but then the argument based on digital computers is superfluous. If we assume for the sake of argument that the thought processes of human beings could be the result of some nonphysical factor, then we can argue by extension that the functioning of artificial ‘thinking’ machines also has an explanation outside of nature (albeit indirectly). A similar (if less obvious) consideration applies to unconscious thought processes. These processes are very often automated ‘runs’ of reasoning originally consciously performed in the brain; moreover, we know that these processes only occur in human beings who have been bred and raised by ancestors capable of conscious thought, roo. It is a matter of controversy whether higher animals can draw conclusions, too, even unconsciously. But if they can, then this ability could also be explained by positing that they, too, have some sort of soul (Swinburne, 1997, p. 23).[6]

Reppert’s Arguments

Reppert’s argument from reason has been broken down into six subarguments both by him (Reppert, 2009) and his critics (Carrier, 2004). All six arguments seek to prove that human thought/reasoning processes are not reducible to physical, causal processes in the brain. These are:

  1. The argument from intentionality
  2. The argument from truth
  3. The argument from mental causation
  4. The argument from the psychological relevance of logical laws
  5. The argument from the unity of consciousness in rational inference
  6. The argument from the reliability of our rational faculties

We will now examine all of Reppert’s arguments separately—with the exception of the sixth. (The reason for this omission is that the sixth argument deduces the unreliability of human cognition from the fact that, according to naturalism, the functioning of our minds serves primarily as an adaptation to the environment rather than knowledge of the objective truth. This argument, however, is developed by Plantinga with much more sophistication and thoroughness in his own antinaturalist argument mentioned above, so it would be superfluous to discuss it here.)

Each of Reppert’s arguments employ the same strategy: he tries to show that the phenomenon of human mental life is not reducible to naturalistic concepts, then concludes that it cannot, therefore, be explained by naturalistic causes, and finally claims that the operation of the immaterial ‘soul’ or ‘spirit’ in man provides a better explanation for the phenomenon in question.

According to the argument from intentionality, human reasoning leading to conclusions from premises cannot ‘correspond’ to or be ‘explained’ by biochemical processes in the brain, first and foremost, because the thoughts in the conclusion have a property that parts of the brain, neurons, or brain events lack. This property is ‘aboutness’: a thought represents (or is directed toward) something outside of itself (‘intentionality’), but a neuron, an electrical discharge in a neural pathway, or sets of these are not ‘about’ anything. The 19th-century philosopher Franz Brentano reintroduced the concept of intentionality into European philosophy and believed that intentionality (“the mark of the mental”) proved that materialist theories of the mind were false. Today there are many philosophers who believe either that intentional properties of the mind (or parts of the mind, or neurons) are analyzable in purely physical terms (eliminative materialists), or that terms signifying intentional concepts (“belief,” “desire,” “about,” etc.) have only a purely instrumental value: they provide a way of speaking which is useful for predicting the behavior of some physical systems (human beings) whose laws of working we still don’t know enough about to describe and predict human behavior in purely physical terms. Among philosophers attracted to a physicalist ontology, few have accepted the outright eliminativist materialist denial of the reality of beliefs and desires. Nor have many of them found it easy to answer the puzzling question raised by the instrumentalist position: how can the intentional idiom make useful predictions if it fails to describe and explain anything real?

Given these difficulties, a naturalistic/physicalistic account of intentionality should instead postulate second-order properties (properties supervening on other properties) as “bearers” of intentionality: both human brains (or parts of human brains, or brain events, or neurons) and some physical things (for example, thermometers or age-revealing tree rings) have some common properties supervening on their physical features, and these common properties make representation possible. This naturalist account of intentionality shouldn’t postulate a logically necessary relation between physical properties and intentional properties. For naturalism can concede the logical possibility of mental substances (souls) with intentional properties; naturalists only claim that this view is an implausible, unlikely account of the workings and properties of the human mind.

1. Reppert’s argument from intentionality is that we cannot make sense of the relation of ‘correspondence’ between brain states/events and the facts that they represent as necessary for the existence of intentionality, since brain states are quite different from the external realities that they are supposed to represent.

Well, are brain states/events, or a series of them, really completely ‘different’ from the external reality that they represent? Let us first think about how a representational device represents a segment of reality. Suppose that we have a map of the Himalayas. It is obvious that the map is composed of different material than the Himalayas, it is obvious that it is much smaller than the Himalayas, and it’s obvious that there are many features of the Himalayas—such as the color of the snow, plants, or fauna—that a map cannot represent. But this does not change the fact that the Himalayas do resemble the map in some respects. What is to the right or left of an object on the map is also to the right or left of it in reality. What is part of another entity larger than itself on the map—for example, a city in a country—is also part of it in reality. The map also shows the altitude in some form, usually with colors. And so on. Think of an airport signal system, where flashing lights indicate the arrival and departure of planes. Of course, both the flashing lights and the electrical mechanism giving rise to them are fundamentally different pieces of reality (e.g., much smaller ones made of different materials) than the aircraft. But differences in the speed of the aircrafts, for example, can correspond to the length of the pauses between successive flashes of light. In principle, stronger, larger flashes of light could indicate larger aircraft. If our antinaturalist opponent replies that these representational devices, like digital computers, were created by conscious, purposeful beings, we can invite him to consider the following analogy: birds fly between two rocks in a desert, casting shadows on the sand. The shadows in the sand are obviously of a very different material composition than that of the birds, but the shadows’ relative sizes, the speed of their appearance, and the direction of their movement will resemble these features of the birds. So the situation here is the same as with the map or the airport signal system, although the shadows are an entirely natural phenomena that did not result from conscious human design.

These analogies prompt the question: why should there fail to be similarities between brain states/events and objects/events in the outside world? The location, speed, frequency, or intensity of electrical discharges in the brain’s neural pathways can map the spatial location of an object in the outside world—e.g., it is likely that the appearance of an object to the left of us is first registered by events in the left hemisphere (such as its direction, the speed of its movement, or its sequence—in the same way that the shadows on the sand can map the flight of the birds, or the flashes of the airport signal system can map the appearance of planes. Our brain states can also indicate the significance of the object in question for us. A Roman Catholic believer probably has a detectably different and more intense reaction in his/her brain at the sight of the Eucharist than at the sight of a comparable wafer in terms of color, shape, or size.

In response to his critics, Reppert (2005) argues that the representational character of representational brain states/events cannot be grounded in their ‘similarity’ to the objects or events that they represent—because this similarity is too remote. He uses the analogy of a piece of a pancake that happens to resemble the shape of the state of Minnesota failing to make that piece a ‘map’ of Minnesota (Reppert, 2005, p. 8). The problem, however, is that it’s not clear: is a piece of pancake not a map because it isn’t similar enough to the state of Minnesota, or is it because we have decided—we have such linguistic conventions—that a piece of pancake is not a map? In the first case, what would happen if, as a result of some strange, improbable, but certainly not inconceivable physical process lines and points appeared on the piece of pancake that corresponded to Minnesota’s rivers and cities, colors appeared that corresponded to its topography, and so on? Most of us would probably admit after a while that the pancake is indeed a map of Minnesota. But in the second case, if representation is simply a matter of linguistic convention, then representation is not an objective relation, so anything can represent anything. But that would just pull the rug out from under the argument from reason, since the whole argument is based on the assumption that (certain) mental contents of our minds do ‘objectively’ represent something.

However, Reppert-defender Darek Barefoot says something else in defense of the antinaturalist notion of representation. He argues that when we think of an ‘atom,’ for example, we are not thinking of a specific atom (e.g., a specific ‘piece’ of the Himalayas, an airplane, or a bird), but of all of the atoms that have ever existed or will exist. But the ‘atom’ representation—the concept or image in my mind—is obviously neither causally related to all of the atoms in the universe (as in the above cases where there was an obvious causal relationship between the external object and the representation) nor similar to them. The appropriate naturalist response to this is that the ‘atom’ image in my mind is in some respects very similar to all of the atoms in the world. And a direct causal relationship with the object of representation is not absolutely necessary for representation. When an individual object is represented by something in my mind, the representation does not necessarily stand in a direct causal relationship with the object represented. For my image of all of the atoms in the world to be representative, it is not necessary for my mind to stand in a causal relationship to all the atoms in the world. It is sufficient that something—for example, the atomic model that I saw in my school textbook—has formed in me an image of the atom that is similar in some respects to all of the atoms in the world. This problem was already posed by John Locke and George Berkeley, and as is well known, these classical authors gave their own answer: according to Locke, we are able to represent kinds in our minds by general ideas obtained by abstraction, whereas according to Berkeley, we are able to represent them by individual ideas representing the prototype of a kind of thing.

2. Reppert’s next argument is that while statements can be true or false, brain states/events cannot be. However, if we accept that the relation of ‘correspondence’ can be analyzed in naturalistic terms, then we can also accept—provided that we endorse the correspondence theory of truth—that the truth or falsity of a statement consists in the existence or absence of a correspondence relation. And our belief about the truth or falsity of a statement is a belief about the presence or absence of this relation (or a brain state corresponding to such a belief).

3. In a logical inference premises determine what the conclusion will be, but in a very different way than causes determine consequences in nature. This is the part of Reppert’s ‘mental causation’ argument (which is somewhat misleadingly labeled since in the philosophy of mind, ‘mental causation’ concerns how mental states/events have physical consequences. In the present context, however, it is the relation between two mental contents (the premises and the conclusion of an inference) that Reppert considers to be impossible to analyze and explain in naturalistic terms.

According to Reppert, the beliefs constituting the basis of our inferences also differ from our brain states because, even if brain states can represent facts about the external world, it is not clear how they can represent fictitious entities or false propositions. Even if my true belief (and my corresponding state of mind) that “There’s a lion charging at me from the left” can be said to ‘correspond’ to the fact that there is now a lion charging at me from the left (and that fact is what triggered that belief), the question remains as to what corresponds to my belief that “The wise old character in Star Wars, Master Yoda, is eight hundred years old,”or my (false) belief that “A giant reptile lives in Loch Ness.”

However, naturalism can deal with these difficulties. Fictional beliefs refer to the mental states (and corresponding physical facts) of other people, and false beliefs refer to imagined but possible facts in the external world (or the mental states of those from whose communications the false belief originated).

Reppert also points out that human inferences, as thought processes unlike physical processes, cannot be subject to strict laws. Richard Swinburne, a similarly committed dualist, but not a proponent of the AfR, makes a similar claim: he argues that mental states and relations cannot be determined by brain states/events alone because, even if we were to match every mental state to a brain state, the mere correlation is not ‘explanatory.’ Real integration of mental states into the causal web of natural processes would require systematically elaborated psychophysical laws; and the search for such laws is almost hopeless since the brain is ‘something very different’ from anything else that science studies. This argument originates from Donald Davidson (1970).

This line of thought ignores the fact that a significant proportion of natural (physical) events and processes are also causally determined, but these causal relations cannot be cast in the form of natural laws. The deductive-nomological model of scientific explanation has come under serious attack in recent decades. Like the concept of natural laws, science probably cannot do without the concepts of causality, causal sequences, and the propagation of causal effects (Salmon, 1984). We cannot account for, with separate laws, what happens when two sets of causal sequences determined by the laws of nature (e.g., the structure and state of the body of John Smith at a given moment, or the effect of alcohol on the human body) meet at a point in space and time (e.g., when John Smith got drunk). But even so, Smith’s drunkenness is a perfectly understandable and natural fact, and we don’t need to look for any nonphysical explanation.

Barefoot gives further reasons why the mental states that supervene on our brain states, or the statements that express them, cannot be the starting points for our inferences. He argues that sensory impressions are caused by the influence of objects in the external world on our senses, and can therefore be explained by natural causes, whereas the abstract content of propositions is ‘perceived’ by introspection, an internal process which, at least prima facie, is not obviously the result of natural causal processes (though Barefoot admits that this abstract content may still be such a result). However, the existence of introspective knowledge, independent of sensory impressions, was already denied by thinkers such as St. Thomas Aquinas, hardly a naturalist. For example, in his debate with his Augustinian opponents, Aquinas rejected the idea that the soul knows itself directly. Even Aquinas believed that all of our abstract thought presupposes a turn to the imagination (conversio ad phantasmata). At any rate, it is an open question whether anyone has ever had the kind of internal ‘perception’ of an abstract concept, completely independent of sensory impressions, to which Barefoot refers. The fact that children learn about concrete sensory experiences first, and about abstract concepts only later (if at all)—and only through their sensory experiences—certainly speaks against the idea that we have the kind of internal mental perception that Barefoot imagines.

Barefoot’s second argument runs as follows: if naturalism is true, then we cannot draw rational conclusions because the linguistic form of the statements that are the starting point of these inferences (whether sequences of sounds, or letters written on paper) is completely contingent and bears no relation to the abstract content represented by the statements. This is perfectly true, but it is not an argument against naturalism; rather, it is an argument against the possibility of any meaningful human thought or communication. The words and grammatical rules of human language are obviously neither logically nor causally related to the content of the ideas communicated by language. When we ‘agree’ on the meaning of words (usually tacitly and not consciously), we do not stipulate that a word should always have the same natural quality as its reference (for example, the color ‘green’), but that it should always have the same place among other words in the system of linguistic conventions (Schlick, 1972, pp. 110-111). But this is true even if these thoughts are not physical or natural.[7] If the contingent character of the relationship between language and the content conveyed by language precludes the communication of meanings, then no one can infer anything (except in themselves). But in this case the argument from reason—in virtue of consisting of sentences intended to convey certain contents—will obviously not be intersubjectively valid, either.

4. According to Reppert, “if we accept the laws of logic—which we must do if we are to arrive at a proposition from another by rational inference—then we must also accept the existence of some non-physical, non-spatial and non-temporal reality” (Reppert, 2003, p. 81). For the laws of logic extend not only to our natural world (as the laws of nature do), but to all possible worlds, including worlds in which no physical objects exist at all.

However, from the philosophical truth that the laws of logic apply to all possible worlds (including nonphysical, nontemporal, nonspatial worlds), it doesn’t follow that our knowledge about the laws of logic is some sort of “intuition” or insight into the common structure/nature of these possible worlds. It is at least possible that these other worlds (including those “nomologically congruent” worlds that have the same natural laws as our actual world, and those incongruent worlds that don’t) are worlds that we can conceive. The common laws of all possible worlds are the laws of logic, and the laws of logic are the laws of human thinking, or rather that of human language or basic human linguistic conventions. I don’t regard this “logical positivist style” analysis of the nature of logic compelling or even unproblematic, but I think it is a possible philosophical option, and I don’t see what sort of knock-down argument Reppert could offer against it.

The elements of our experience of the world can be ordered in our mind in a way that is completely different from the order that prevails in our world, including in such a way that these elements (or at least some of them) have no place in space and time (setting aside good philosophical arguments against the coherence of bodiless or aspatiotemporal persons). But certain rules apply to all such imaginable worlds, just as you can play all sorts of other games with the pieces in a chess game, but not just any game, because the possibilities of the games are limited, for example, by the fact that there are only thirty-two pieces in chess.

The rules of logic determine what conceivable combinations of the elements of the world are possible—even when they are not ‘put together’ within the framework of space and time. Therefore, in order to see logic as a basic set of rules for all human thought, we do not have to assume an abstract super-reality ‘beyond’ the limits of space and time, nor should we think that this reality has a basic ontological structure more fundamental and universal than the laws of natural science.

5. According to Reppert, when we draw conclusions from certain premises, it is an operation that the mind performs as a ‘unified’ agent; whereas scientific research on the brain ‘breaks down’ its functioning into natural parts and subfunctions. But Reppert ignores the fact that a system (e.g., the human digestive system) can perform certain tasks as a whole even if it is made up of parts that perform partial functions; and lots of connections have been discovered between different parts of the brain and their functions in recent decades by cognitive science.

Now that we have run through each of Reppert’s arguments, we’ve seen that all of the various features of the human mind that he adduced to support his position could be analyzed in naturalistic terms.

“Drainage” of the Causes—Rickabaugh and Buras’ Argument

Two younger contemporary philosophers who acknowledge their indebtedness to Reppert, Brandon Rickabaugh and Todd Buras, have developed the argument from reason in a novel and (I believe) more successful way (Rickabaugh & Buras, 2017).

According to Rickabaugh and Buras, for Lewis’ argument to be successful, he would have to rule out the possibility that human reasoning is a normal causal process in the brain. They formulate the principle Lewis defends as follows:

Exclusion principle: It is true of any belief b that if it has a “complete mechanistic explanation,” it “cannot also have any rational explanation” at the same time (Rickabaugh & Buras, 2017, p. 388).

Mechanistic explanations are based on the physical conditions at a point in space-time and the laws of nature. For example, a 2.5 kg iron ball dropped from the roof of a ten-story building will fall to the ground in a certain amount of time because of the laws of universal gravitation, and will fracture the skull of the person walking there because of the laws of biology. Rational explanation, on the other hand, moves from the content of certain premises towards certain conclusions. For example:

P1. My neighbor has recently started wearing a swastika armband.
P2. Nazis used to wear swastikas.
P3. My neighbor has no connection to Eastern religions that use the swastika as a religious symbol.
P4. As far as I know, they aren’t shooting a film in the house, so the neighbor can’t be wearing a swastika because he is an extra in a historical film.
Conclusion: My neighbor is a Nazi.

Rickabaugh and Buras argue that whatever the case is with Lewis, they can establish the exclusion principle. Let us see how.

According to them, if naturalism is true then every physical event has a complete physical explanation. This means that when I draw the above conclusion from my neighbor’s behavior, the drawing of it is a physical (brain) event that is completely explained by my prior physical (brain) processes at the time when I accepted the premises and drew the conclusion.

To establish the exclusion principle, we have to assume that both the premises and the conclusion are mental contents that supervene on the brain states/events corresponding to them. This raises the question of whether the supervenience implies token or type identity. Buras and Rickabaugh opt for token identity: a mental content corresponds to a particular brain state/event. In this case, the brain facts corresponding to the acceptance of the premises—together with the laws of nature governing the functioning of the brain—will by natural/nomological necessity bring about the facts corresponding to the acceptance of the conclusion.

But then the content of the premises has no causal effect on the acceptance of the conclusion. Brain states and causal effects in the brain would trigger acceptance of the conclusion in exactly the same way if we had different premises or even no premises at all, and we would accept the content of the conclusion without any argumentation. After all, if we allow that mental states—in this case, our apprehensions of the propositions in the premises—can be realized in multiple ways, then they can be realized by multiple types of physical states. But if we allow multiple realizability, then the essential mental properties of the mental state type in question will not be causally sufficient—for they can be realized by different physical/neural states, even those with completely different physical properties, and their causal sufficiency cannot be grounded in the essential property of the mental state (the apprehension of the proposition), which is common to all realizations. In other words, if the same content can be realized by several neural states with different properties, then the content as content will not be causally sufficient.

And this conclusion—that the content of the premises has no relevance to the acceptance of the conclusion—is obviously highly counterintuitive. Therefore, we can only decide that the premises (which are obviously mental contents) themselves have some effect on the nature of the conclusion, and since this ‘effect’ cannot ‘propagate’ through causal mechanisms, naturalism is false: it is not true that only physical causal processes take place in the brain, it is not true that brain events/states have only physical explanations, and therefore when we draw the conclusion of an argument, this does not have a complete physical explanation.

Rickabaugh and Buras’ argument shows that mental states (our beliefs containing the premises of arguments) have their own causal efficacy. If Rickabaugh and Buras reject on rational grounds the possibility that mental causal factors can be physical factors (with capacities/efficacy different from that of all other physical factors), then the way is open to a nonnaturalistic explanation of human rationality.

However, on a naturalistic worldview the path from premises to a conclusion in an argument can be a physical process. What kind of physical process might this be? As we saw earlier, representations can represent not only physical facts, but also other representations—other mental states or events. Inference can thus be understood naturalistically as a special sort of representation: representation by a special sort of representative physical entities—parts of the brain, groups of neurons, brain states, or brain events. The conclusion of an inference (and the physical entity corresponding to it) represents the contents or parts of the contents of other representations (the premises, and their “bearers” in the brain).

Perhaps we can clarify this point with an analogy. Imagine, as we did earlier, that flashing lights indicate each plane arriving at an airport, the direction from which they are coming, and at what intervals they follow each other. Now suppose that there are two different airport signal systems: one for arrivals from the north, and one for arrivals from the south. Suppose further that a mirror is installed at a suitable point to reflect the light from both signal systems. So if we look at the mirror, for example, we can see that today there were 12 arrivals from the north and 13 from the south, making for a total of 25 arriving planes. In observing such we have drawn a conclusion from premises. And we could do this because there was a physical entity that represented the physical facts represented by the other two entities (the signal systems). In making inferences, our brains can work in this way; one part of the system can represent the states/events of many other parts of the system and their contents—representing the facts of the external world. When we reach a conclusion from premises, there is a complete physical explanation for this, one that differs from the explanation we would give if we reach the same conclusion from other premises, or if we just believe the conclusion without any argumentation. Thus inference—just like representation—can be described in a fully naturalistic way, and so doesn’t require a nonnaturalistic (theistic/spiritualistic) explanation.


There may be features of the human mind that are conceptually irreducible to physical features. But if so, Reppert and other AfR proponents haven’t succeeded in revealing such features. None of the varieties of the AfR considered here show that the working of human mind couldn’t have a naturalistic explanation.


[1] Arguments from reason are not to be confused with arguments from consciousness. Arguments from consciousness (e.g., Swinburne, 2004, p. 192) claim that there is no rational explanation for the existence of consciousness if naturalism (or materialism) is true; but this sort of argument does not claim that if there were such an explanation, it would render human cognition unreliable.

[2] His book based on these radio broadcasts was published with the title Mere Christianity (Lewis, 2014).

[3] In principle, the argument could also support some spiritual worldview quite different from Christian theism—e.g., Taoism—but I will not address this possibility because contemporary defenders of the argument use it to defend Christian-style theism.

[4] See “What Lewis Really Did to Miracles” for an overview of the debate (Smilde, 2011).

[5] By now there is a vast literature, pro and con, on EAAN. For a trustworthy companion to the wider debate, see Troy Nunley’s 2005 PhD dissertation “A Defense of Alvin Plantinga’s Evolutionary Argument against Naturalism.”

[6] This conclusion is endorsed by one of the best-known interactionist substance dualists of our time, Richard Swinburne.

[7] Richard Carrier, one of Reppert’s critics, argues that the first elementary units of language were originally sound-imitating words; for example, one of the English vernacular names for ‘pig’ was the equivalent of the characteristic sound made by pigs (oinking), while the children’s vernacular name for defecation was derived from the sounds made during defecation. If this is the case, then it provides a naturalistic explanation for the meaning of the words. But even if this theory is incorrect, its falsehood would not affect our claims (Carrier, 2004, p. 61).


Barefoot, Darek. (2007, June 25). “A Response to Richard Carrier’s Review of C. S. Lewis’ Dangerous Idea.” The Secular Web. <https://infidels.org/library/modern/darek-barefoot-dangerous/>.

Bassham, Gregory. (Ed.) (2015). C. S. Lewis’s Christian Apologetics: Pro and Con. Leiden, Netherlands: Koninklijke Brill.

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

Davidson, Donald. (1970). “Mental Events.” In Experience and Theory (pp. 79-101). Ed. Lawrence Foster and J. W. Swanson. Amherst, MA: University of Massachusetts Press.

Haldane, John Burdon Sanderson. (1940). “Why I am a Materialist.” In Science and Life: Essays of a Rationalist (pp. 27-35). Ed. J. Maynard Smith. London, UK: Pemberton Publishing Co. (Originally published 1940.)

Haldane, John Burdon Sanderson. (1954). “I Repent an Error.” The Literary Guide (April): 7, 29.

Johnson, David Kyle. (2015a). “Naturalism Undefeated.” In C. S. Lewis’s Christian Apologetics: Pro and Con (pp. 91-104). Ed. Gregory Bassham. Leiden, Netherlands: Koninklijke Brill.

Johnson, David Kyle. (2015b). “A Reply to Victor Reppert.” In C. S. Lewis’s Christian Apologetics: Pro and Con (pp. 113-120). Ed. Gregory Bassham. Leiden, Netherlands: Koninklijke Brill.

Lewis, Clive Staples. (2014). Keresztény vagyok [Mere Christianity]. Budapest, Hungary: Harmat Kiadó. (Originally published 1942.)

Lewis, Clive Staples. (2001). Miracles: A Preliminary Study. San Francisco, CA: HarperCollins.

Lewis, Clive Staples. (1947). Miracles: A Preliminary Study. London, UK: G. Bles.

Nunley, Troy. (2005). “A Defense of Alvin Plantinga’s Evolutionary Argument against Naturalism.” [Doctoral dissertation, University of Missouri—Columbia]. <https://mospace.umsystem.edu/xmlui/handle/10355/4184?show=full>.

Plantinga, Alvin. (1994). Naturalism Defeated. Unpublished manuscript. Philosophy Faculty, Calvin University. <https://web.archive.org/web/20161113125031/http://philofreligion.homestead.com/files/alspaper.htm>.

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

Reppert, Victor. (2005, March 7). “Updated Reply to Carrier.” Dangerous Idea blog. <http://dangerousidea.blogspot.com/2005/03/updated-reply-to-carrier.html>.

Reppert, Victor. (2009). “The Argument from Reason.” In The Blackwell Companion to Natural Theology (pp. 344-390). Ed. William Lane Craig and J. P. Moreland. Malden, MA: Blackwell.

Rickabaugh, Brandon and Todd Buras. (2017). “The Argument from Reason and Mental Causal Drainage: A Reply to Peter van Inwagen.” Philosophia Christi Vol. 19, No. 2: 381-399.

Salmon, Wesley. (1984). Scientific Explanation and the Causal Structure of the World. Princeton, NJ: Princeton University Press.

Schlick, Moritz. (1972). “Pozitivizmus és Realizmus.” In A Bécsi Kör Filozófi´ja [Philosophy of the Vienna Circle] (pp. 93-133). Ed. Ferenc Altrichter. Budapest, Hungary: Gondolat Kiadó.

Smilde, Arend. (2011). “What Lewis Really Did to Miracles: A Philosophical Layman’s Attempt to Understand the Anscombe Affair.” Journal of Inklings Studies Vol. 1, No. 2 (October): 9-24.

Swinburne, Richard. (1997). The Evolution of the Soul (2nd ed.). Oxford, UK: Clarendon Press.

Swinburne, Richard. (2004). The Existence of God (2nd ed.). Oxford, UK: Oxford University Press.

Copyright ©2024 by Szalai Miklós. This electronic version is copyright ©2024 by Internet Infidels, Inc. with the written permission of Szalai Miklós. All rights reserved.