REPLACING METHODOLOGICAL NATURALISM

Richard Dawkins, during a recent interview about his new book, The God Delusion, proclaimed “[T]he big war is not between evolution and creationism, but between naturalism and supernaturalism.”¹ I agree with Dawkins to the extent that naturalism, whether we are talking about metaphysical naturalism or methodological naturalism, is one of the primary barriers to fruitful dialogue between science and religion and to the interdisciplinary synthesis of knowledge in general. This is because naturalism seems to prevent scientific discussion of many important topics, including: human freedom, morality, purpose in nature, and God.

Although this problem is not new, Alvin Plantinga discussed it in some detail ten years ago in an article titled Methodological Naturalism?, it has not been resolved and it has not gone away.² The recent battles over supernatural causation and the definition of science, as seen in the case of the Kansas Board of Education and in the District Court case in Pennsylvania, Kitzmiller v. Dover, are evidence of this.³ In addition, as if to add fuel to the fire, some scientists are now openly calling for attacks on religion and the supernatural. For example, at a recent conference on science and religion called Beyond Belief: Science, Religion, Reason and Survival, which was held at the Salk Institute in California, Steven Weinberg, a Nobel laureate in physics, said “Anything that we scientists can do to weaken the hold of religion should be done and may in the end be our greatest contribution to civilization.”⁴ In The God Delusion, Dawkins is clear that anything supernatural is the object of his attack: “I am attacking God, all gods, anything and everything supernatural, wherever and whenever they have been or will be invented.”⁵

The time is ripe, then, to revisit the issue of naturalism. In this paper I focus on methodological naturalism with the goal of demonstrating why the scientific community should abandon it and replace it with a new methodological principle. I accomplish this by performing four tasks. First, I analyze different formulations and justifications of methodological naturalism that have been put forth by scientists and philosophers of science. Second, I show how all of these formulations and justifications have serious problems and therefore the principle of methodological naturalism should be abandoned. Third, I argue that a new methodological principle needs to take its place. I propose and defend such a principle, which I call the principle of methodological neutralism, with two goals in mind. The first goal is that it will be acceptable to the scientific, religious, and philosophical communities. The second goal is that it will allow for greater dialogue (or at least the possibility of greater dialogue) between science and religion and for greater interdisciplinary synthesis in general. Fourth, and finally, I reply to some objections that might be raised against my view.

Before we can move to our first task, we must clarify briefly what naturalism is. Naturalism is a metaphysical view that denies the existence of supernatural entities. Usually this view amounts to a kind of materialism and therefore it denies the existence of non-material beings such as God. Some scientists hold that naturalism is a necessary condition of science. For example, Arthur Strahler, a geologist, said: “The naturalistic view is that the particular universe we observe came into existence and has operated through all time and in all its parts without the impetus or guidance of any supernatural agency. The naturalistic view is espoused by science as its fundamental assumption.”⁶

Is Strahler correct? Should naturalism be a necessary condition of science? Philip Johnson, who is part of the Intelligent Design movement, has argued that the answer is “no” for several reasons.⁷ First, naturalism is not a scientific view based on empirical evidence. Second, naturalism is an unproven philosophical view that persons accept or reject as a matter of choice. As Johnson puts it, naturalism is “a dogmatic statement about the nature of the universe.”⁸ But dogma does not belong in science.

However, not everyone holds that naturalism is a necessary condition of science. For example, Robert T. Pennock, a philosopher of science, agrees that “If science assumed [metaphysical naturalism] … then Johnson’s charge of scientific dogmatism might have some merit.”⁹ Instead, Pennock agrees with Michael Ruse, another philosopher of science, that science only assumes methodological naturalism.¹⁰ Methodological naturalism is an epistemological principle that governs how science is practiced. It prohibits the use of supernatural explanations in science. However, unlike metaphysical naturalism, it does not make any claims about the existence or non-existence of supernatural entities. As Ruse explains: “[I]n no sense is the methodological naturalist … committed to the denial of God’s existence. It is simply that the methodological naturalist insists that, in as much as one is doing science, one avoid all theological or other religious references.”¹¹

But should methodological naturalism be a necessary condition of science? Against Ruse, Plantinga has argued that the answer is “no.”¹² Who is correct? To answer this question we must first answer a more basic question: What is the proper understanding of science? Unfortunately, there is no universally accepted definition of science. In fact, some philosophers of science have argued that all known attempts to distinguish science from non-science have failed.¹³ This makes our task more difficult and we cannot get around this difficulty by merely stipulating that science should be defined to include (or to exclude) methodological naturalism. Such a stipulation would beg the question at hand. Therefore, if we wish to justify the use of methodological naturalism in science, a different type of argument must be given. For example, Niall Shanks, a philosopher of science, argues that methodological naturalism is “an inductive generalization derived from 300 to 400 years of scientific experience.”¹⁴ Inductive arguments, however, do not demonstrate their conclusions with certainty; therefore this is not enough to justify its use as a necessary condition of science.

Although I will spend some time discussing realist and antirealist conceptions of science, I do not have enough space here to defend in detail what I think properly constitutes science. Instead, the bulk of my efforts will be aimed at demonstrating why methodological naturalism should be abandoned by the scientific community. To accomplish this requires that we examine methodological naturalism more closely. Let us turn to that task now.

I. Varieties of Methodological Naturalism

Unfortunately, one problem that complicates our task is that there is some division within both the scientific and philosophical communities on the topic of methodological naturalism. There are differences both in the terminology used and in the definitions put forth. There are also differences with respect to the role the principle plays in science and in the arguments that have been used to justify its use in science.

For example, with respect to terminology, Eugenie Scott, an anthropologist who works for the National Center for Science Education, calls the principle ‘methodological materialism.’¹⁵ In contrast, Nancy Murphy, a professor of Christian Philosophy, calls the principle ‘methodological atheism.’¹⁶ However, both of these terminological choices have disadvantages. In the case of Scott, the word ‘materialism’ might be misinterpreted by some as excluding electro-magnetic fields, space-time, and other things scientists discuss. In the case of Murphy, scientists would need to have a clear understanding of God in order to understand what ‘atheism’ meant. Unfortunately, there are many different philosophical and theological conceptions of God. Which one would scientists pick and why? To avoid all of these problems, I use the phrase ‘methodological naturalism’ exclusively throughout the paper.

Despite these terminological differences, Scott and Murphy agree on the general meaning of the principle. Scott explains it this way: “[S]cience acts as if the supernatural did not exist.”¹⁷ Murphy understands it to mean that “scientific explanations are to be in terms of natural (not supernatural) entities and processes.”¹⁸ They also agree that the use of the principle in science is legitimate despite being members of different disciplines (science and philosophy) and different religious persuasions (Scott is an atheist and Murphy is a Christian). To determine if they are correct we need to examine the role that the principle plays in science.

When we examine Scott’s view of methodological naturalism more closely we see that the prohibition against the supernatural is both a priori and necessary. It is a priori because she defines science this way: “By definition, science cannot consider supernatural explanations.”¹⁹ And it is necessary because any discipline that rejects the principle is not scientific as the following passage makes clear: “Defining science as an attempt to explain the natural world using natural processes and mechanisms allows us to say to creationists like Henry Morris that ‘God did it’ is not science.”²⁰ I will demonstrate later on that her understanding of the principle is incompatible with a realist conception of science.

Not all scientists, however, agree with Scott. For example, Massimo Pigliucci, a professor of ecology and evolution who is not sympathetic to intelligent design, holds that methodological naturalism is provisional and a posteriori. It is a posteriori because it is arrived at due to lack evidence: “Since there is no evidence of any [G]od or supernatural design in the universe, the scientifically-informed conclusion has to be that there is none.”²¹ It is provisional because he claims “falsification of the naturalist paradigm is indeed possible.”²² In other words, if the naturalist paradigm were ever falsified science as a discipline would continue but without the principle of methodological naturalism. I will argue later on that we do not have to wait until the naturalist paradigm is falsified in order for scientists to abandon methodological naturalism. I will also argue that the correct scientific stance on the supernatural should be neutrality, not denial of existence—even if it is a provisional denial—as Pigliucci claims.²³

There have also been related disagreements within the scientific community about God and the supernatural. For example, some scientists such as the late Stephen Jay Gould, a paleontologist and evolutionary biologist, have argued that science and religion are completely separate. “Non-overlapping magisteria” (NOMA) was the phrase he used.²⁴ His point was that scientists, speaking as scientists, cannot comment on God and supernatural. In opposition to Gould, Dawkins, an evolutionary biologist and an ardent atheist, has declared that “[T]he existence of God is a scientific hypothesis like any other.”²⁵

Clearly, the scientific community is not speaking with one voice to the public and this is not helpful to the ongoing cultural and legal battles concerning science and religion. Indeed, in Kitzmiller v. Dover, Judge John E. Jones III ruled that Intelligent Design was “a religious view … and not a scientific theory” because, among other reasons, Intelligent Design failed “to meet the essential ground rules that limit science to testable, natural explanations.”²⁶ The requirement that science only use natural explanations is precisely the injunction of methodological naturalism. What we must determine is if the use of methodological naturalism in science is justified. Let us turn, then, to the task of evaluating the arguments that have been put forth for prohibiting the supernatural in science.

II. Prohibitions of the Supernatural

As I mentioned earlier, Scott calls the principle ‘methodological materialism.’ This suggests that we should understand the natural as matter and the supernatural as non-matter. However, in order to prohibit the supernatural, which is understood as the non-material in this case, we would need to have a clear conception of what matter is so we could negate it. The problem is that recent discoveries in science, for example quantum theory and dark matter, have made matter as a general concept more of a mystery than a well-formed concept. For example, dark matter is invisible and at least some of it is non-baryonic and thus of a composition unknown to us.²⁷ When dark matter is combined with dark energy, which is also of a composition unknown to us, we realize that we do not understand about ninety percent of the universe.²⁸ The lesson here is that scientists cannot know what they are going to discover about reality prior to investigation. I call this the principle of discovery and methodological naturalism violates this principle because it does not allow for the possibility of discovering the supernatural.

The principle of discovery is related to what I call the principle of evidence, which states that scientists in their search for truth should follow the evidence wherever it leads. Methodological naturalism violates this principle because no matter what evidence we might gather we are never allowed to follow it to a supernatural cause. Recently, Antony Flew, a philosopher famous for his atheism, stated that he now believes in some kind of God based on scientific evidence about the origin of life and the complexity of nature. Had Flew strictly followed the principle of methodological naturalism he could never have reached this conclusion. However, Flew stated that his “whole life has been guided by the principle of Plato’s Socrates: Follow the evidence, wherever it leads.”²⁹

The principle of discovery and the principle of evidence are both part of the view that science is a type of realism. Realism, roughly speaking, is the view that science aims at discovering objective truths about reality, where reality is understood as that which exists independently of our minds. Realists recognize that achieving this aim is difficult. Humans have limitations and often make mistakes. Theories are always underdetermined by data and thus are never proven true in an absolute way; instead, they are always tentative and thus subject to future revision. Even so, real progress has been achieved. For example, the transition from Newtonian physics to Einstein’s relativity has brought us closer to the truth.

Realism in one form or another has been the dominant view of science for most of history and it is currently the dominant view among philosophers of science.³⁰ What is essential to realism is that our theories must conform to reality in order to be true. If we gather evidence that conflicts with a theory we must modify or abandon that theory. This is the principle of self-correction. But methodological naturalism potentially jeopardizes this principle as Del Ratzsch, a philosopher of science, explains: “[I]f part of reality lies beyond the natural realm, then science cannot get at the truth without abandoning the naturalism it presently follows as a methodological rule of thumb.”³¹

Methodological naturalism is opposed to realism because it violates the principles of discovery, evidence, and self-correction. Scientists opposed to realism are called antirealists. Methodological naturalism is closest to the idealist kind of antirealism. This is because in idealism reality must conform to ideas instead of ideas conforming to reality. Methodological naturalism is guilty of idealism because the interpretation of evidence and the construction of theories must conform to a naturalistic framework since supernatural explanations are prohibited.

Unless scientists are willing to abandon realism, they have only two options: (1) abandon methodological naturalism, or (2) argue that there is no conflict between methodological naturalism and the principles of discovery, evidence, and self-correction because there is something else about the supernatural that justifies its exclusion from science. I will argue for the first option later, but let us examine the arguments of those who have chosen the second option.

A. The Supernatural is Not Empirical

One argument for excluding the supernatural from science is the claim that the supernatural is not empirical. According to Strahler, “supernatural forces, if they can be said to exist, cannot be observed, measured, or recorded by the procedures of science—that’s simply what the word ‘supernatural’ means.”³² The problem with this argument is that Strahler simply assumes that the existence of supernatural causes cannot be arrived at indirectly through empirical means. But why assume that? As Ratzsch explains:

Within the scientific context, all that is required for something to be a legitimately empirical matter is for it to have appropriately definable, theoretically traceable empirical consequences or effects or connections. Those connections can be exceedingly indirect, and they typically are not direct consequences just of the theoretical matters in question, but only of those matters in conjunction with a variety of other principles (sometimes referred to as “auxiliary” or “bridge” principles). It is by that means that even such exotica as quarks and the deep past get included within the empirical realm. But although a supernatural being could obviously have untraceable effects on nature, surely it cannot be claimed that a supernatural being simply could not have traceable effects upon empirical matters.³³

B. The Supernatural is not Testable

Another reason for excluding the supernatural from science is the argument that claims concerning supernatural causes are not testable. One understanding of testability is controllability. Scott makes this point by saying, “you can’t put God in a test tube.”³⁴ But, as Ratzsch remarks, by this logic we should also exclude things like supernovas and the Big Bang from science, since we cannot produce and control them in a lab.³⁵

A second understanding of testability is falsifiability, a point made famous by Karl Popper.³⁶ According to Popper, a theory, hypothesis or assertion was scientific only if empirical data could show it to be false. Using this, one could argue that the existence of God is not a scientific question since no empirical data could show it to be false.

However, there are several problems with using falsifiability to exclude the supernatural from science. First, falsifiability, if strictly adhered to, has the unwelcome result that almost every existential claim becomes unscientific. Roger Penrose, a physicist, gives the example of Dirac’s monopole theory to show that falisifiability, is too stringent a criterion:

[Dirac’s argument was that] the mere existence of a single magnetic monopole somewhere in the cosmos could provide an explanation for the fact that each particle in the universe has an electric charge that is an integral multiple of some fixed value (as is indeed observed). The theory which asserts that such a monopole exists somewhere is distinctly un-Popperian. That theory could be established by the discovery of such a particle, but it appears not to be refutable, as Popper’s criterion would require; for, if the theory is wrong, no matter how long experimenters search in vain, their inability to find a monopole would not disprove the theory! Yet the theory is certainly a scientific one, well worthy of serious consideration.³⁷

Second, William A. Dembski, a philosopher and part of the Intelligent Design movement, has argued that, strictly speaking, no scientific theory is empirically falsifiable in a definitive way because one can always add auxiliary hypotheses to harmonize the discordant data with the theory.³⁸ Penrose gives an example that illustrates Dembski’s point.³⁹ The example is supersymmetry in modern physics. Supersymmetry predicts the existence of superpartners for all observed fundamental particles of nature but so far none have been found. The reason given for not finding them is that very large amounts of energy are needed to create them and the current technology in our particle accelerators cannot generate enough energy. But suppose we built more powerful accelerators and we still did not detect superpartners. According to Penrose, “It could (and probably would) be argued that there had simply been too much optimism about the smallness of the degree of the symmetry breaking, and even higher energies would be needed to find the missing superpartners.”⁴⁰ And that is Dembski’s point: auxiliary hypotheses can always be added to harmonize the discordant data.

As a result of the above, Dembski argues that the main point of Popper’s criterion is not about demonstrating falsehood. Instead, it is about eliminating theories because of new evidence. Dembski uses the word ‘refutable’ to denote this difference in meaning. He argues that if refutability, instead of falsifiability, is a necessary condition of science then Intelligent Design should not be considered unscientific because it is refutable:

If it could be shown that biological systems that are wonderfully complex, elegant and integrated—such as the bacterial flagellum—could have been formed by a gradual Darwinian process … then intelligent design would be refuted on the general grounds that one does not invoke intelligent causes when undirected natural causes will do. In that case Ockham’s razor would finish off intelligent design quite nicely.⁴¹

Barbara Forrest, a philosopher of science and a critic of Intelligent Design, seems to have acknowledged Dembski’s point about refutability when she said “[S]hould life be genuinely created in the laboratory from the non-organic elements which presently comprise living organisms, this discovery would add tremendous weight to philosophical naturalism.”⁴² In other words, the creation of life in a lab would make the appeal to supernatural causation to explain life superfluous.

However, while intelligent design could be refuted in this way, Dembski claims that Darwinian evolution is irrefutable.⁴³ This is because even if all of the currently known Darwinian mechanisms fail to account for the complexity of life Darwinists will merely respond that there must be some undiscovered mechanism that will eventually explain it. But why assume that? For some, it is because they assume methodological naturalism. If only natural causes are allowed in scientific explanations then we are forced to explain the complexity of life through them. Methodological naturalism forces evolution to be interpreted naturalistically. But this leads to a third problem.

If scientists assume methodological naturalism as a necessary condition of science then the naturalistic interpretation of evolution is neither falsifiable nor refutable and therefore should be rejected as unscientific. Scientists can only avoid this objection by either rejecting refutability as a necessary condition of science, which is very unlikely, or by holding that methodological naturalism is only a provisional principle and thus subject to abandonment. As we have already seen, some scientists, such as Pigliucci, have taken the second option.

Fourth, and finally, in a way somewhat similar to Dembski’s refutability, Ratzch mentions another way that the supernatural can be empirically eliminated. He notes that in the history of science non-empirical philosophical prescriptions such as the requirement that proper scientific explanations must be deterministic have been abandoned because of empirical data. The outcome of various experiments lead to quantum mechanics, which made the scientific community abandon the non-empirical requirement of determinism. Thus it is possible that the same can happen with the supernatural:

[G]iven the nature of science and its operations, and the interconnectedness of the empirical and the nonempirical with the scientific context, the nearly inescapable conclusion is that the empirical can not only come to some sort of grips with some nonempirical matters, but can trigger changes in scientifically incorporated nonempirical positions, and can do so in circumstances and by means which make it proper to say that the nonempirical positions in question were empirically at risk. And being empirically at risk is all that is required for something to be both falsifiable and testable.⁴⁴

C. The Supernatural Violates Natural Laws

Another reason for excluding the supernatural from science is the argument that the supernatural violates natural laws. Pennock raises this objection:

Lawful regularity is at the very heart of the naturalistic world view and to say that some power is supernatural is, by definition, to say that it can violate natural laws. … Controlled, repeatable experimentation … would not be possible without the methodological assumption that supernatural entities do not intervene to negate lawful natural regularities.⁴⁵

There are, however, several problems with this view. First, if we were to hold that the laws of nature were purely deterministic (something along the lines of what Pierre-Simon Laplace had in mind) then we would exclude from science many things considered scientific such as quantum mechanics and the social sciences. Indeed, some have argued that the discovery of quantum theory, which shattered the Laplacian view, allows for the compatibility of physical science and divine intervention.⁴⁶

Second, let us suppose, for sake of argument, that quantum theory is false and thus the laws of nature are purely deterministic. Even in this extreme scenario, William P. Alston, a philosopher, has argued that divine intervention would be compatible with physical science—except if we were to make the unwarranted assumption that the universe is a closed system of laws:

If we suppose that divine intervention in a physical process would involve a violation of physical law, it is because we are thinking of physical laws (of a deterministic form) as specifying unqualifiedly sufficient conditions for an outcome. … [For example:] A man standing upright in the middle of a deep lake without sinking would be a violation o