[I am planning to expand this older essay into a new essay on naturalist epistemology, as part of my blog series “Thinking about the ‘Metaphysics’ in Metaphysical Naturalism,” on my philosophy blog Civitas Humana. As such, some of the content below may be revised in the newer version, which I post a link to here, once it is finished. -MWF 9/29/15]
It is not uncommon on Facebook and other social media to come across the following meme, quoting Christian apologist and “former atheist,” C.S. Lewis:
Frankly, if C.S. Lewis really thinks that atheism or naturalism entails the human mind and our reasoning faculties developing in this way, then we might as well equally mischaracterize Christianity with the following counter-meme:
Needless to say, C.S. Lewis’ quote above is a straw man, and I highly recommend Jeff Lowder’s “Let’s Attack a Straw Man, C.S. Lewis Style!” for a detailed analysis of how Lewis mischaracterizes both atheism and naturalism in many of his writings .
While C.S. Lewis’ poor ability to think in the statement above makes a good case for the non-existence of god by his own reasoning, the thinking skeptic also has the tools to address this objection. By analyzing how the human mind actually operates and has originated (not at all by accident), and how it bears relation to our environment, it can be demonstrated that reliable truth-finding cognitive faculties would develop in an atheistic and naturalist universe. This analysis will show not only that the naturalist explanation does not have to rely on a massive ad hoc leap, which assumes the existence of a non-demonstrable problem solver, but is likewise far more parsimonious.
Apologetic epistemological arguments take a variety of forms, such as theologian Alvin Plantinga’s evolutionary argument against naturalism (EAAN), which is rebutted by J. Wesley Robbins in “Evolutionary Naturalism, Theism, and Skepticism about the External World,” Stephen Law in “Plantinga’s Evolutionary Argument Against Naturalism Refuted,” and Paul Churchland in “Is Evolutionary Naturalism Epistemologically Self-Defeating?,” as well as apologist Victor Reppert’s formulation of the argument from reason (AfR), which is rebutted by naturalist Richard Carrier in “Critical Review of Victor Reppert’s Defense of the Argument from Reason.” Yet the common objection remains largely the same: atheism or naturalism is ultimately untenable, since, even if they were true, both would entail a lack of the reliable cognitive faculties to know that they are true. Accordingly, apologists claim that it is unreasonable to be an atheist or a naturalist, since allegedly these would be self-defeating views. In order to refute this objection, the skeptic need only provide an ontology of knowledge that is both compatible with and rendered probable under atheism and/or naturalism.
To begin with, the apologetic claim is already prima facie wrong. Human cognitive faculties are demonstrably reliable, when such faculties have enabled us to predict whole solar events, to build cities and computers, and even taken us all the way to the Moon. We can furthermore observe through evolutionary science that human DNA, which includes the coding for our mind, evolved through a slow process of unguided mutation and natural selection. Nevertheless, this process has still produced the successful results listed above. Even if apologists were correct in assuming that this outcome would be unlikely in a atheist or naturalist universe, you cannot tell a lottery winner that it is impossible to win the lottery, when they have the ticket and winnings right in front of you. However, this outcome does not even need to be too improbable.
How could an atheist or naturalist universe give rise to reliable cognitive faculties with good probability? Rather than throw up one’s hands and assume that it just has to be a miracle performed by a supernatural designer, as C.S. Lewis does, serious philosophy actually engages the data and looks for the process that actually explains the clear results in front of us. Far from being improbable, evolutionary development, when it has the right environment and enough time to produce minds as complex and sophisticated as the human mind, will almost always select such cognitive faculties to be reliable and truth-finding. In fact, that is the very reason why those cognitive faculties exist.
Atheism, Naturalism, the Human Mind, and Reliable Knowledge
The Mind as Virtual Model: First, what is the relation between our mind and the world? The mind operates as a virtual model that recreates our environment in intelligible terms, with features such as a colors and shapes, which are all reducible to patterns of matter and energy in the external world, but that the mind abstracts and conceptualizes. In this way the reality we experience is very much a virtual reality, but this virtual reality in the mind is not one entirely detached from actual reality. As philosopher J. Wesley Robbins explains in “Evolutionary Naturalism, Theism, and Skepticism about the External World,” under a naturalist “pragmatist view” of the mind, the mind’s beliefs are “only identifiable with reference to what in their holder’s environment caused them to form that belief.” One of the key features of this mind is that it holds “aboutness” in regards to the world. Every moment of our waking experience our mind is reconstructing the world around us and both subconsciously and consciously forming beliefs regarding it. These beliefs develop and change based on experience and analysis. It is fair to say that these beliefs can be both “true” and “false,” but how does the mind operate to sort out the true beliefs from the false ones?
Reason as a Truth-Finding Language: Reason is a form of analytical knowledge that makes propositions about the world. Reason, as an expression of language, is syntactical in nature, in forming relationships between various terms and concepts, and categorizing them. We perform logical operations through reason, by deducing conclusions from premises that are based on such relationships and their categorized concepts. In this way, reason allows us to articulate our beliefs in clear propositions that can meaningfully correspond to the world, and also operate in clear relationships and nuance between each other. The result of this process is the ability to form complex theories about reality.
How does reason get us to the truth? In part, because logically valid statements are in one sense tautologically true. Reason derives from the simple axioms of the law of identity (A=A), the law of the excluded middle (Either-Or), and the law of non-contradiction. These axioms entail that certain propositions will have conclusions that are “by definition” true. Furthermore, one can also argue that these axioms are not purely tautological assumptions, but also are demonstrably true features of our world. If cats can be black or white (Either-Or), and there is a cat that is all black, then every part of the cat is black (A=A), and none of the cat is white (non-contradiction). This derives from a uniformity of features that we simply observe in our universe.
While reason can provide logically “valid” truths, in that conclusions follow from premises, it does not always provide “sound” truths. This is why the armchair logic used by may presuppositional apologists does not provide compelling arguments for the existence of a deity. In order for an argument to be logically valid, a conclusion must simply flow from its premises, but this does not entail that the premises themselves are correct. A sound argument is one for which both the premises and the conclusion are true. But how do we know if our premises are true?
Sensory Organs and Empiricism: Propositions are useless if they do not in someway correspond to reality. Our direct involvement with reality, however, is not linguistic, but perceptory. We learn of the world around us through sight, hearing, taste, touch, etc. Reason provides us with a language for abstracting and categorizing these experiences. Reason also allows for us to construct relationships where we can infer and predict certain conclusions that are not directly observable. This is why not all science is laboratory science. Empiricism also involves drawing probable inferences from limited data.
Empiricism provides us with the tool for knowing whether the premise of a proposition is true. Hence, the synthesis of reason and empirical observation provides a basis for sound arguments, where not only valid conclusions are reached, but sound conclusions can also be drawn by empirically testing the premises. Through both of these methods, we are able to form complex and testable theories about the world, of which atheism and naturalism are both.
The Selection of Cognitive Faculties: Not all life forms possess the capacity for reason or have sensory organs, but even very basic animals do. Even a cockroach from sensing a change of motion and light in front of it can detect a larger animal there and run away from it. Thus, reason is not at all an unlikely feature to evolve in complex life. Reason merely provides an analytical tool for processing information and responding in logical ways. If the host does not operate in logical ways, such as the roach perceiving the motion and change of light but then not running away, then the host will quickly perish from the hazards in its environment. Accordingly, there is strong selection for hosts with reliable reasoning faculties that can draw true conclusions about their environments.
Likewise, the more ways in which your senses can perceive your environment, the more ways that you can learn of and respond to it. The evolution of the eye is a good example of this. Early eyes could do little more than tell the host “lights on” or “lights off.” This is a useful capability to have, as opposed to having no ability to sense light, but it is not the most useful organ imaginable. Accordingly, the eye developed to further communicate “light from above” and “light from below,” and eventually learned to recognize various wavelengths of light (colors) and to focus light, etc. All of this occurred through direct correspondence between the development of sensory organs, the mind, and the host’s environment, per a pragmatist view of the mind. Accordingly, reliable sensory organs would develop not purely by accidence, but by their reliability to perceive true features about their environment.
Mind-Body Physicalism: Another point that should be made is that all of the features of the mind can be explained in purely physical terms. As I explained in my post about the definition of naturalism, naturalism, by definition, does not entail physical materialism, but physical materialists are a subset of naturalism, to which I belong. Through the discoveries of neuro-science, everyday more and more features of the mind are being explained physically. We can isolate certain memories, tell through brain scans when a host is thinking about a certain abstract object, and identify which features of the brain perform which operations. Furthermore, physical trauma to the brain damages or destroys these functions. People who have experienced severe brain damage can even lose the ability to recognize patterns such as faces, which shows that such patterns are not immaterial Platonic forms, but concepts formed by patterns in a physical mind.
Thinking Robots: As we learn more about neuroscience, we are discovering more and more how the mind physically operates and would be selected to have truth-finding faculties. But from another direction, new discoveries in computer science are also confirming naturalism! As Richard Carrier explains in his article, “Ten Years to the Robot Apocalypse,” computers are now being shown more and more to be capable of having all of the same cognitive faculties that humans have. Just as our biological minds form virtual models of the world, computers are being developed that form similar virtual models through mechanical minds.
Apologists in the past have argued that computers are purely passive objects. Computers can allegedly only draw true conclusions if you, the supernaturally truth-knowing human, program them properly. But new developments in computer science are showing that computers can learn truths about themselves and their environment, even without being programmed to know them. Carrier discusses one such experiment demonstrating this:
“In the Lipson-Zykov experiment they gave a robot a basic Bayesian learning program and four working legs, but told it nothing about itself, not even that it had legs, much less how many or how they worked. In no time (sixteen trials) it figured out it had four legs, how they were oriented, how they moved, and how to use them to get around efficiently. It built a model of itself in its digital brain and tested hypotheses about it, revised the model, and so on, until it had a good model, one that was, it turns out, correct. Then it could use that model to move around and navigate the world.”
If set in competition with each other, computers with reliable reasoning and perceptory functions would overwhelmingly defeat computers without or with less reliable functions. In like manner, humans minds, which are really just biological computers, were selected to have accurate truth-finding faculties, despite evolving through unguided mutation and natural selection. Contrary to C.S. Lewis’ intuition, such a result would be by far the most probable outcome.
The Selection of Accurate Beliefs: Simply because a host has reliable cognitive faculties does not entail it will have accurate beliefs. After all, early humans had reliable sensory perception organs that showed the Sun and Moon moving around the Earth, and could logically make the inference that the Earth was at the center of these two bodies. But that belief was false. How, then, can we get accurate beliefs about the world? The answer is that beliefs are not formed in a vacuum. They build on each other and evolve as new data is accumulated and new arguments are offered. Ideas go through a sort of natural selection (dubbed memetic evolution), just as biological organisms are selected for their physical traits.
Ideas, through methods like science, can be selected for their truth value, whittling out false beliefs, so that only true beliefs remain. To be sure, false beliefs can be selected as well, but as Darrel Ray shows in The God Virus, these beliefs will reflect viral tendencies that reveal that they are not based on truth value, and thus can be discarded. But ideas that are selected through critical processes, such as the scientific method, will be selected to be more robust in their truth value. Add a couple thousand years of human progress, testing the universe with scientific instruments, and performing rigorous peer review, and we will get an accurate picture of reality.
Here is just one description of a naturalist process that would lead to accurate beliefs, proposed by professional epistemologist Konrad Talmont-Kaminski (“Evolution, Cognition and Value: the Ingredients for a Naturalist Philosophy,” pp. 38-39):
“The open-endedness afforded by the theory of bounded rationality goes beyond the traits of individual heuristics and consists in the capacity to develop new heuristics … In the case of evolution this is made possible by the mechanisms of variation, selection and retention. With each new paleontological ﬁnd, our knowledge of how various adaptations arose is becoming ever more complete, in some case already providing us with a veritable slide-show of the intermediate forms that led to supremely functional outcomes. In the case of reasoning, the mechanisms are only now being revealed through work in a variety of the sciences of cognition as well as work in evolutionary explanations of human behaviour (for an introduction to the latter, see Laland & Brown 2002). The basic idea is that, while some heuristics may be in some way hardwired into our brains through evolution … there is not a set range of heuristics that cognition is limited to relying upon. People are capable of using their limited knowledge of their environment to identify problems with existing heuristics and to develop new heuristics that extend their abilities. Paramount examples of this capacity ﬁll the history of scientiﬁc methodology. Thus, the double-blind method that has become standard for testing pharmaceuticals, is the result of successive developments in experimental methods that took place during the twentieth century and were a reaction to the realisation that it is not enough to simply provide a control group which is not subject to the treatment being tested. The initial use of single-blind testing, in which the researchers knew whether they were administering the placebo or the treatment under examination, also turned out to be insufficient in some cases. The reason is that the placebo effect is subtle enough to affect outcomes due to the attitude of the researcher interacting with the subject, even though that person may not be consciously altering their behaviour. The vital point is that knowledge of the environment plus existing heuristics is sufficient to develop new heuristics. There is no need for a general heuristic-generating mechanism.”
Warranted Naturalist Belief: As shown in the discussion above, we have every theoretical reason to suspect that reliable reasoning faculties and perceptual organs would develop under metaphysical naturalism, which is a form of atheism. Furthermore, the beliefs produced through these faculties are not static, but are selected through processes like the scientific method that whittle them down for their truth value. If, with these faculties and the knowledge they produce, we continually only find evidence for a physical, natural universe, and if such faculties and knowledge in no way lead us in the direction of supernaturalism or theism, then I would would say that we can have the warranted inference that metaphysical naturalism, and thus atheism, are probably true. Does this mean that we will necessarily reach 100% certain “truth” that atheism and naturalism are correct? I wouldn’t say that overtly, but most naturalists consider themselves to be fallibilists. One can be a fallibilist and still have warranted belief.
Furthermore, in a physical materialist hypothesis about reality, natural evolution would select for all of the abilities to know about this reality. If there were something beyond the physical universe that exerted no pressure on the development of our mind, it would make sense that we are unable to perceive or know about. But this is all very strange if a supernatural being, allegedly from beyond our physical universe, designed our minds. Why wouldn’t such a being design us to perceive things outside of the physical universe, especially if he wanted us to know about him? Instead, if physical materialist naturalism is true, we have the exact results that we would expect.
Problems for Theistic Theories of Knowledge
I explained above, however, that an atheist and/or naturalist theory of knowledge is not only capable of being articulated in a way that renders (fallibilist) epistemic warrant as both consistent with and probable under atheism and/or naturalism, but is also far more parsimonious than theistic theories of knowledge. Apologists, in assuming that god just magically solves every problem in life, often do not consider the ramifications for our cognitive faculties being designed by an intelligent agent, rather than through selective natural processes.
The Ultimate Boeing 747 Gambit: To begin with, there is no explanation for how the intelligent agent itself obtained reliable cognitive faculties. With naturalist minds, a long process of evolution can be demonstrated to show how reliable cognitive faculties would eventually and gradually form. But there is no such explanation for god’s mind. A deity’s reliable cognitive faculties would just “exist,” without any cause, in a perfectly accidental arrangement of fine-tuning. How probable would that be?
This problem is posed for theism by Richard Dawkins in an argument that has come to be termed The Ultimate Boeing 747 Gambit. The typical apologetic response is to claim that god is not “complex,” but rather a “simple” mind (drawn from the notion of divine simplicity). Since god is completely non-demonstrable, theologians can of course just make up on their couch any new attribute about god that they please. As I have explained in a previous essay, however, the idea of a “simple” mind is not just improbable, but largely incoherent. It is thus disputable whether such a “simple” mind is even conceptually (not just metaphysically) possible.
Without such appeals to simplicity, a greater problem exists for explaining how god’s cognitive faculties just happen to be reliable (in order to explain how god then made human cognitive faculties reliable), than whatever problems there are for just explaining how human cognitive faculties became reliable on their own through selective processes. To use an analogy, the apologists are presupposing the unexplained fine-tuning of a super computer in order to merely explain how a calculator could develop. The naturalist, in contrast, has a far more parsimonious theory that does not require appeals to more unexplained complexity.
God as Cartesian Demon: Even if god’s cognitive faculties just happen to be tuned right, however, a further problem is involved with the notion of our cognitive faculties being dependent upon an “intelligent” designer. Blind nature does not work very well as a Cartesian demon that creates an entirely false illusion of reality. While imperfect cognitive faculties may be implied by naturalism, this does not mean that natural forces would very easily program us with an entirely false picture of reality. Such complete and total deception normally requires conscious design. In order to have completely unreliable cognitive faculties, therefore, that are programmed to consistently form false beliefs, one would almost need an intelligent designer to make it that way. Accordingly, in theism god could act as a Cartesian demon to program us with an entirely false picture of reality.
The apologist will counter that god would never design us to have unreliable cognitive faculties. But here is the problem: if all of our capacity for knowledge depends on god, we could never know if god is deceiving us in this way! Our belief that god would never deceive us could merely be caused by god deceiving us. If all of our knowledge is dependent on god, and our understanding of what god would or would not do is also dependent on god, we could never know whether god actually designed us to have reliable cognitive faculties, or simply programmed us with false beliefs, including the false belief that god would never deceive us. Given our inability to resolve this issue with any certainty, it can thus be argued that theistic theories knowledge are ultimately untenable, so that the original objection is reversed.
The Existence of Superstitions: While naturalism does make it probable that we will have “reliable” knowledge, it does not guarantee that we will have “perfect” knowledge. Accordingly, our minds will still form many false relationships and create superstitions that are only slowly discovered to be false. Consider just one example: bloodletting. Early doctors observed that bloodletting brought down fevers, and so incorrectly concluded that it helped the patient. It was only upon further investigation, however, that this was proven to be a false conclusion, since bloodletting actually causes other problems that harm the patient.
Now, this is an understandable mistake for a naturalist view of the mind, since accurate knowledge under naturalism is a painstaking process with many bumps along the way. But it does not make sense if god allegedly programmed our minds to know the truth. Why would god allow humans for hundreds of years to have a false faith in bloodletting that harmed countless patients? One cannot use the old freewill defense, because these doctors were often well-intentioned and genuinely wanted to help their patients. And yet god allowed them to persist in a destructive superstition for centuries. Why would god so callously allow humans to persist in such ignorance, if he allegedly designed our minds to know the truth? Such problems are inexplicable if god actually exists (at least, a non-evil god), but they are easily at harmony with a naturalist theory of knowledge.
“God” as a False Belief that Would Be Selected: Another irony is that if there is one false superstition that our minds would be selected to believe in under naturalism, it is actually belief in god and the supernatural. Early humans lived on a planet teeming with life, much of which was hostile and dangerous. Accordingly, early humans had to compete with other animals (and sometimes other humans) to survive, which selected our minds to detect agency and to seek out intelligence that threatened us. An accidental side effect of this, however, was that our minds became programmed for agency over-detection (discussed further by Stewart Guthrie in Faces in the Clouds). Humans personified many things that are actually purely mechanical — harvest cycles, lighting, natural disasters, etc. — and we connected such phenomena with divine beings and other supernatural concepts. This problem still exists today when theists personify things like the Big Bang and the origin of life, which are likewise almost certainly impersonal and natural events, and instead attribute them to god. It has taken a long time for humans to root out this superstition, which is why a purely naturalist universe is only now being proven in the modern age.
So, while there are many problems with theistic theories of knowledge, in a naturalist theory of knowledge it is perfectly explicable why the false superstition of god would commonly exist. Humans have projected a personal agency that we call “God” onto an otherwise blind and impersonal universe, because of an understandable selection for agency over-detection. Modern science and philosophy are only now finally uprooting this false superstition.
The ultimate irony is that upon identifying some of the problems with a theistic theory of knowledge, it is actually god that is C.S. Lewis’ milk jug! What would be the odds that a finely-tuned, ultra-intelligent being, with perfectly reliable cognitive faculties, would merely “exist” by chance, as the uncaused creator of the universe, in order to give us reliable cognitive faculties? It would be like me upsetting a milk jug and hoping that it splashes out a map of London! In contrast, under a naturalist theory of the mind and knowledge, a slow, demonstrable, and gradual process can be used to explain how such cognitive faculties would eventually develop not by accident, but as the probable result of natural processes.
C.S. Lewis’ milk jug argument strikes me as a rather indolent approach to epistemology. No matter how silly the Bible seems, no matter how completely invisible and non-demonstrable god is, you can’t say that it is not true, because you can’t know anything unless it is true! Apologists make such arguments in an effort to deflect all criticism, and seem to be less concerned with solving actual philosophical problems, rather than making sure that a problem always persists in order that their god can be shoved through that gap. Nevertheless, there are secular answers to these issues that can solve every alleged problem posed by apologists, in a genuine pursuit of philosophy that is not about preserving mystery about the universe (and thus room for god), but actually explaining it.
 For an additional rebuttal to C.S. Lewis’ argument from reason, I recommend chapter 6 of philosopher John Berversluis’ C.S. Lewis and the Search for Rational Religion.