In a 2012 debate entitled The Great Debate: Has Science Refuted Religion, physicist Sean Carroll made a rather bold claim in his opening remarks (at about the 12:30 mark):
"The argument is finished. The debate is over. We've come to a conclusion. Naturalism has won. If you go to any university physics department, listen to the talks they give or the papers they write--go to any biology department, go to any neuroscience department, any philosophy department, people whose professional job it is to explain the world, to come up with explanatory frameworks that match what we see--no one mentions God. There's never an appeal to a supernatural realm by people whose job it is to explain what happens in the world. Everyone knows that the naturalist explanations are the ones that work."
Naturalism: A Critical Analysis edited by William Lane Craig and J.P. Moreland is essentially a direct challenge to Carroll's claim. The book is a collection of essays by academic philosophers in university departments that do not "know that naturalist explanations are the ones that work." They instead level a host of ontological, epistemological, ethical, and theological arguments against the veracity of naturalist explanations. In this review I will attempt to explain what some of those arguments are.
Naturalism is a bit of a slippery thesis so there is no one official version of it; nevertheless, I think Paul K. Moser and David Yandell capture the two major pillars of naturalism in the first essay of the book entitled "Farewell to Philosophical Naturalism." They offer these two dimensions of naturalism:
"Core ontological naturalism: every real entity either consists of or is somehow grounded in the objects countenanced by the hypothetically completed empirical sciences (that is, in the objects of a natural ontology).
"Core methodological naturalism: every legitimate method of acquiring knowledge consists of or is grounded in the hypothetically completed methods of the empirical sciences (that is, in natural methods)" (p.10).
These two definitions capture both what naturalism says exists, and how naturalism says we come to know anything. What exists are basically physical objects and forces--the stuff of physics. What we know is basically whatever the physical sciences tell us is so. In fact, the centrality of the physical sciences in the naturalist view has led some (including Moser and Yandell) to use the term scientism to capture the view. Moser and Yandell include the phrase "hypothetically completed" in their definitions to describe the empirical sciences because naturalists do not suggest that "science now" is representative of all that exists and all that can be known. Instead, naturalists acknowledge the incompleteness of today's scientific understanding, yet wish to maintain that the final word on reality will be in some sense "scientific" or "science-like." Naturalist's tend to agree that today's scientific picture is a reasonable approximation of what tomorrow's scientific picture will look like. Since today's empirical sciences do not use "supernatural" entities to do any explanatory work (e.g. God), naturalists are unified in rejecting the existence of anything supernatural. Furthermore, since the empirical sciences do not use "faith," "divine revelation," or "pure reason" to detect truths about the world, naturalists do not take those methods to be reliable truth-detecting tools. So, in short, naturalists tend to define their position as (1) a commitment to the existence of only entities postulated by empirical sciences and (2) a commitment to only the methods of the empirical sciences.
Objections to Methodological Naturalism
The thesis of methodological naturalism (also known as epistemological naturalism) seems to suffer the most from this criticism: it is self-defeating. To see how it is self-defeating, consider this proposition: "Methodological naturalism is true." Now, ask yourself, by what empirical science was this proposition verified? Was it verified by a chemistry experiment?--by a particle collider?--through a telescope?--a microscope?--how exactly was the thesis of methodological naturalism grounded by its own standard of knowledge: that of empirical science? The answer seems to be that it was never verified by any empirical science. If this is right, then methodological naturalism fails its own test. To believe the proposition of methodological naturalism, then, is to entail that one disbelieve it. It is similar to the status of the proposition, "This sentence is false." If one believes the sentence is true, then one should believe the sentence is false. The sentence undermines itself and is therefore self-defeating. Moser and Yandell write as much in their essay:
"[Methodological naturalism] is not itself a thesis offered by any empirical science....[It is not] represented in the empirical scientific work of either physics, chemistry, biology, anthropology, psychology, or any other natural or social empirical science....The self-defeat of [methodological naturalism] would result from its failing to be included in or approved by its own proposed single standard for methodological...integrity: the (hypothetically completed) empirical sciences" (pp.10-13).
The only way to rescue methodological naturalism would be to weaken the strength of its claim. Instead of saying something like "every method of knowledge is grounded in the empirical sciences," it could say, "most methods of knowledge are grounded in the empirical sciences." But, of course, such a move would open the door to non-empirical methods of knowledge--like a priori truths--and remove the teeth from the naturalist's position. If this revised version of methodological naturalism were the version Sean Carroll asserts has "won," it would be a trivial victory.
Objections to Ontological Naturalism
Moser and Yandell wield a similar criticism to that of methodological naturalism against ontological naturalism. Namely, ontological naturalism is not itself a thesis of any of the empirical sciences. There is no experiment a scientist can run that yields the conclusion "X, Y, and Z, and only X, Y, and Z, exist." Moser and Yandell say this:
"Neither individually nor collectively do [the empirical sciences] offer theses about every real entity or every legitimate epistemological method....the empirical sciences, individually and collectively, are logically neutral on such matters as the existence of God, the reliability of certain kinds of religious experience, the objectivity of moral value, and the reality of thinking substances" (p.11).
Philosopher Edward Feser (who did not contribute to this volume) has offered a clear analogy to this point. He suggests that a metal detector is really good at detecting metallic objects; however, the metal detector is no good at detecting anything else. If we think of the empirical sciences as analogous to a metal detector (or a group of different kinds of detectors), it does not follow from the fact that they might detect X, Y, and Z that only X, Y, and Z exist. It might be the case that A, B, or C also exist, but the detectors of the sciences simply are not calibrated to detect them. It would thus be a logical mistake to reason from the existence of X, Y, and Z as detected by the empirical sciences to the conclusion that nothing other than X, Y, and Z exists. Claims about all of reality are, as Moser and Yandell say elsewhere, monopolistic claims; but, the empirical sciences are not equipped to make such monopolistic claims. Instead, the empirical sciences make specific claims within limited domains. To reach the more fundamental, all-encompassing level of reality, one must use meta-physics, which is in the realm of a priori philosophy rather than empirical science. This philosophical realm is, however, barred to the naturalist due to his methodological commitments. Thus, it seems as though ontological naturalism falls short of finding justification within its own naturalistic commitments.
Aside from this logical criticism of ontological naturalism, there are perhaps good empirical objections to it as well. Charles Taliaferro and Stuart Goetz each write chapters pertaining to the existence of immaterial minds which, if correct, falsify ontological naturalism by virtue of the established existence of something non-material. Taliaferro writes chapter 6 entitled "Naturalism and the Mind," and Goetz writes chapter 7 entitled "Naturalism and Libertarian Agency." For the sake of brevity, I will not reprise these arguments in much detail. In short, Taliaferro argues on the one hand that the three major naturalistic strategies for dealing with the mind (eliminative materialism, identity materialism, and non-reductive materialism) are each flawed; on the other hand, he argues that dualism (i.e. that immaterial minds exist) is a good explanation for what the mind is and why it has the properties that it does. Furthermore, he suggests that dualism is at home in an overall theistic, non-naturalistic, world view. Goetz argues for dualism as well but he does so on the basis of libertarian free will. Goetz suggests that the best explanation for the appearance of free will is that there are non-physical minds who make undetermined choices for reasons or purposes. Since material objects are not capable of making choices for reasons--material objects are only supposed to obey determinate laws--the existence of free will entails the existence of something non-physical: namely, immaterial conscious agents. If either of Taliaferro's or Goetz's essays are successful, dualism is established and ontological naturalism is therefore falsified.
A third criticism of ontological naturalism comes from Robert C. Koons in chapter 3 entitled, "The Incompatibility of Naturalism and Scientific Realism." Scientific realism is basically the proposition that the sciences yield objectively true facts about an externally existing world. In other words, the theories and theses of the sciences reliably produce knowledge about the external world. Koons' thesis is partially captured in this excerpt:
"I will argue that nature is comprehensible scientifically only if nature is not a causally closed system--only if nature is shaped by supernatural forces (forces beyond the scope of physical space and time)" (p.50).
His argument proceeds by pointing out how the empirical sciences recognize a kind of preference for simplicity in their theories. This is to say, the natural world appears to obey elegant, parsimonious, and symmetrical laws rather than crude, profligate, or asymmetrical laws. Provided Koons' observation here is accurate (which seems reasonable given the history of scientific theorizing), it follows that simplicity is a kind of reliable indicator of the objective features of reality. In other words, scientists are better-equipped to produce true theories about the world if they rely on simplicity as a guide. But, Koons argues, in order for simplicity to be a reliable guide toward truth, it must be the case that simplicity was systematically or causally imposed upon or woven into the natural world. It is not good enough for one to suggest that the simplicity in the world came about "by chance." For if the simplicity is only there by happenstance, that would undermine the reliability of the preference for simplicity in scientific theorizing (i.e., simplicity would not be a reliable indicator of truth). The natural world itself, however, as a closed system, could not be the cause of its own alignment with simplicity. The simplicity must have come from outside the closed natural world, which is to say, it must have come from a super-natural source. Obviously, if Koons is right and simplicity in science requires a supernatural source, ontological naturalism would be false.
A fourth objection to ontological naturalism comes from cosmological considerations. William Lane Craig produces his now-familiar Kalam Cosmological Argument in chapter 9 of the book entitled "Naturalism and Cosmology." The argument is quite simple yet potent:
(1) Whatever begins to exist has a cause of its existence.
(2) The universe began to exist.
(3) Therefore, the universe has a cause of its existence (p.244).
Premise (1) is defended by the principle of sufficient reason. If some entity did not have to exist, or some event did not have to happen, then the entity or event is metaphysically contingent. If some entity has to exist, or some event has to happen, then the entity or event is metaphysically necessary. Metaphysically necessary entities or events do not require causes or explanations, but metaphysically contingent entities or events do require causes or explanations. Thus, insofar as something begins to exist, it cannot be the case that the thing must exist; for if it must exist, it would never "begin" to exist. It would instead simply "always" exist. So, premise (1) derives from the fact that a contingent entity requires some explanation of its existence. In this case, the coming into being requires a cause which explains why that being came into existence.
The defense of premise (2) occupies the majority of Craig's chapter. He argues that the evidence in support of the standard big bang model of the universe adds significant weight to the veracity of premise (2). Insofar as the big bang indicates that the universe began to exist a finite time ago, then it appears that the universe is a contingent entity that requires a cause or explanation of its coming into existence. Thus, premise (3) follows logically from (1) and (2). It generally does not require much more argument to show why the truth of premise (3) causes trouble for naturalism. On ontological naturalism, the spatio-temporal universe comprises all existant entities. If the entire spatio-temporal universe required a cause of its own existence, such a cause must exist outside of that universe, for it is incoherent to assert that the universe caused itself.
The bulk of Craig's work in his chapter is to consider the various strategies naturalists have offered to avoid premise (2). I will not try to summarize these different strategies but simply record that Craig's objections to them struck me as promising, and it is no surprise that he concludes that they all ultimately fail to unseat the standard big bang model. Consequently, if Craig's Kalam Cosmological Argument is correct, a supernatural cause of the universe must exist and ontological naturalism is therefore false.
The final essay in the volume was written by William Dembski. He titles his chapter "Naturalism and Design" and argues against ontological naturalism on the basis of design in nature. He develops a criterion for design detection--what he calls the specification-complexity criterion--and explains how this criterion can eliminate the adequacy of naturalist appeals to chance and necessity for some natural systems. Insofar as naturalism prohibits the existence of design in nature (because naturalism only permits physical forces to operate via chance and necessity), the detection of design in nature falsifies naturalism. Dembski writes the following in his closing paragraph:
"[N]aturalism allows only certain sorts of fundamental causes (chance and necessity). Those causes are (demonstrably) incapable of generating specified complexity. But nature exhibits specified complexity, especially in biology. Therefore naturalism is false" (p.277).
Dembski spells out the term "specified complexity" in some detail in his chapter, but the essence of the idea is fairly easy to grasp. Consider the sentence "Naturalism has some problems." We can say that this sentence is "specified" because the arrangement of symbols matches up to independently given patterns: namely, the patterns of syntax and semantics of the English language. The sentence in question is specific to the patterns of the English language whereas the following arrangement of symbols is not: "ijfoe$lan*gi@elanuejkep/,2ken aknd]=`po5}/fnmkl." The latter sequence came about by my randomly mashing buttons on the keyboard and does not match up to the patterns of English (nor, I suspect, the patterns of any other language). The sentence "Naturalism has some problems" is also complex in the sense that the arrangement of the symbols is fairly improbable. The probability of that exact sequence coming about by my randomly mashing buttons on the keyboard is fairly small. The sequence "too," however, is much less complex than our naturalism sentence and, even though it is specific to the English language, it is not complex enough to eliminate the chance occurrence of it by random button mashing. Thus, neither specification nor complexity by itself is sufficient to eliminate chance and invoke design; instead, both specification and complexity are required on Dembski's account to eliminate chance and invoke design.
Dembski then applies his criterion to the bacterial flagellum motor found in the biological world and argues that the motor passes the specification-complexity criterion and therefore the biological system could not have arisen by naturalistic chance or necessity; instead, it must have arisen by design. The objection to naturalism is then plain: if a biological system arose by design, the design must have come from an intelligent agent operating outside the boundaries of purely deterministic physical law. The existence of such an intelligent agent then falsifies ontological naturalism.
There are a few other arguments against naturalism in the book that I have not discussed but should mention. Dallas Willard levels an argument against naturalism to the effect that knowledge is impossible in the naturalist's world. J.P. Moreland and Michael Rea both highlight technical problems for naturalism with respect to understanding material objects and their properties. And John E. Hare argues that there is an incompatibility between naturalism and morality. These essays are interesting in their own right and my omission of them here is not meant to say anything about their quality.
Overall, the arguments presented against naturalism in this volume at best falsify naturalism, and at worst cast a sizable degree of doubt on Sean Carroll's assertion that naturalism has "won." The book does demand a pretty robust familiarity with philosophical jargon and some intellectual discipline, but it rewards those who invest the effort with a deep grasp of the difficulties of the naturalist view that are otherwise unknown or ignored by its public advocates.
 Alex Rosenberg is a philosopher who has both advocated the use of the label "scientism" and subscribed to it himself. See his 2011 book An Atheist's Guide to Reality: Enjoying Life Without Illusions and my Amazon review of his book called "How Fermions and Bosons Get It Wrong."