Methodological Naturalism and the Supernatural (1997)
with post-conference notes (updated 4-7-1997)
[The following paper was presented on Saturday, February 22, 1997, 9:15 a.m., at Naturalism, Theism, and the Scientific Enterprise - an interdisciplinary conference at the University of Texas, Austin.
Please consider this paper a work-in-progress. Although its reception at the NTSE conference was almost uniformly positive, there are definitely at least problems with it that require correction, points that need expansion, and distinctions that need to be drawn. Recent input from Robert Schadewald and Kevin L. O'Brien has given me more reason to believe that my major thesis (or at least one part of it, since I am sure I have not differentiated all of my claims properly) is through-and-through flawed. Expect a revision of this paper sometime in the future. In the meantime, read it with the balance of openness and cautious skepticism that is due any philosophical paper.]
Supernatural forces, if they exist, cannot be observed, measured, or recorded by the procedures of science - that's simply what the word "supernatural" means. There can be no limit to the kinds and shapes of supernatural forces and forms the human mind is capable of conjuring up from "nowhere." Scientists therefore have no alternative but to ignore "claims" of the existence of supernatural forces and causes. This exclusion is a basic position that must be stoutly adhered to by scientists or their entire system of processing information will collapse. To put it another way, if science must include a supernatural realm, it will be forced into a game where there are no rules. Without rules, no scientific observation, explanation, or prediction can enjoy a high probability of being a correct picture of the real world.
Most contemporary scientists would agree with the statement above, made by geologist Arthur Strahler in his massive critique of scientific creationism, Science and Earth History. According to the standard scientific picture, the following propositions expressed by Strahler are true:
- Methodological naturalism leaves no room for appeals to the supernatural.
- Science must follow the procedures of methodological naturalism to accomplish its aims.
Hence, the consensus seems to be that there is not, and cannot be (even in principle), any overlap between science and the supernatural. Like their scientist counterparts, many philosophers also support the two propositions . For instance, anyone who is addicted to the Internet may have noticed Strahler's quote in one of my personal WWW documents . However, I have been reconsidering my commitment to the two propositions, and now believe I have adequate grounds to tenatively reject both. I do not claim to have resolved anything with certainty, but I do have a few thoughts and ideas that I would like to field for general inspection and evaluation.
First, I will attempt to argue that methodological naturalism is in principle capable of leading to both the falsification and the confirmation of a large number of supernatural hypotheses (to the extent that science can actually "falsify" or "confirm" anything, that is). Then I will argue that there are supernaturalistic methodologies that have the same virtues as methodological naturalism, and that it is even in principle possible for a supernatural methodology to fulfill the aims of science better than methodological naturalism can. Finally, I will consider and answer in the negative the question of whether, given the way I have answered the first two questions, scientists would be better off abandoning methodological naturalism.
I. Definitions and Presuppositions
Before I start critiquing the propositions, I want to set forth very brief and minimal descriptions of the way I define the important terms used in this paper - this will set forth in plain sight most of the core presuppositions upon which the points I make in this paper depends.
First, I take methodological naturalism to be the practice of adhering to the kind of methodology a metaphysical naturalist devoted to fulfilling the aims of science would adhere to. Methodological supernaturalism, correspondingly, is the practice of adhering to the kind of methodology a metaphysical supernaturalist devoted to fulfilling the aims of science would adhere to. Astute readers will note that these two categories are not altogether mutually exclusive - a point which will resurface later on in the paper.
Second, I take science to be an enterprise that endeavors (1) to uncover as many phenomena present in the universe as possible, and (2) to provide for those phenomena explanations that correspond with what really exists "out there." It should be clear from this that this paper presupposes a scientific realist interpretation of the goals of science, which I will not attempt to argue for here. I will leave it to others to determine how other interpretations of the objectives of science might affect the arguments I make (my suspicion, however, is that they will not have a significant effect).
Third, I take the supernatural to designate the standard things like gods, fairies, ghosts, necromancy, divination, and so forth, all of which are united by the fact that they exploit materials or processes the basis of which extends or derives from beyond the physical universe. It is worth noting that the word "supernatural" is not, contra Strahler, analytically the same thing as "beyond the domain of science" - to equate the two, one must show that anything that exploits or results from transcendent materials or processes is beyond the domain of science, which is not a simple matter of definition.
II. Can Methodological Naturalism Say Anything About the Supernatural?
Proposition 1 expresses the standard view of the scientific community that methodological naturalism renders science and the supernatural entirely isolated from one another. Science presumably can say nothing at all about the supernatural, whether positive or negative. Even if supernatural forces are constantly at play around us, methodological naturalism would presumably forbid scientists from ever acknowledging them. In this section, I will argue that this view is false - trivially in that methodological naturalism can falsify supernatural hypotheses, and not so trivially in that methodological naturalism can also confirm supernatural hypotheses.
At this point, it would be a good idea to try to explicate the nature of methodological naturalism a little more thoroughly. I have claimed that methodological naturalism is the practice of adhering to the kind of methodology a metaphysical naturalist would employ if he wanted to fulfill the basic aims of science. Recall also that I take the basic aims of science to be the uncovering of phenomena in the universe, and the attempt to provide for those phenomena explanations that correspond to the way the universe actually is. So how would a metaphysical naturalist try to fulfill these aims? By doing exactly what practing scientists currently do: he would use emprical observation as a means of discovering things about the world, and would attempt to provide plausible naturalistic explanations for the things he observes.
Now at first glance, this may seem to entirely destroy my project - if the methodological naturalist must attempt to provide naturalistic explanations for the things he observes, then doesn't that automatically bar him from even considering the possibility of a supernaturalistic explanation, much less accepting it as the most plausible one? I would argue that, initial appearances to the contrary, methodological naturalism does not rule out any explanation a priori. To see why this is, let us recall that methodological naturalism is defined by the way a metaphysical naturalist trying to advance science would act. I think it is relatively clear that such a person would have to be a fallibilist with respect to even his metaphysical beliefs; remember that he wants to be as sure as he can that his explanations correspond to the way the world really is. Since he does not have all possible data, he cannot be sure that there is not some kind of evidence for the supernatural out there, so he would not want to trap himself in a routine that would ignore even blatant evidence for the supernatural in favor of a less plausible naturalistic hypothesis. The metaphysical naturalist who wishes to fulfill the aims of science cannot rule out the possibility that his metaphysical views may eventually be shown to be wrong - hence, he must be open to some degree to supernatural explanations. However, this degree is likely to be a very slight one, which is what gives methodological naturalism its naturalistic flavor - the methodological naturalist, in acting like a metaphysical naturalist devoted to science, will, while being open to the slight possibility of evidence for the supernatural, consider naturalistic hypotheses on average more parsimonious than supernaturalistic ones, and hence will give them more benefit of doubt. In short, his methodological commitments can tell him to examine all possible naturalistic explanations for a phenomenon first, and can assign higher prima facie probabilities to such explanations on average, but the least plausible naturalistic explanation will tend to have a lower prima facie probability than the most plausible supernaturalistic explanation.
Let me try to illustrate this with a few examples:
Let's imagine that instead of the scientific data we currently have, that we have uncovered the following facts: (1) every stratum of rock contains exactly the same fossils at every level, even in the very lowest strata; no species is represented in the fossil record that is not alive today; and no evidence of intermediates between any two taxa shows up anywhere; (2) radioisotope dating shows that every stratum of rock in the earth, and every piece of rock in the solar system has the same age - namely, about a few thousand years; (3) every species - even anatomically similar ones living in the same region - possesses genetic material composed of radically different chemicals and relying upon vastly different coding systems. This is the kind of scenario scientific creationists probably dream about, and rightfully so - what kind of naturalistic explanation would explain such data as elegantly as a supernatural creation hypothesis? A methodological naturalist faced with such data would, I think, end up having to choose between something along the lines of a supernatural creation hypothesis and the naturalistic hypothesis that the atoms in the universe configured themselves into their current form through a massive thermodynamic fluctuation. In the face of such a dilemma, I don't think a methodological naturalist would have any misgivings about choosing the supernatural explanation. Add to all of this data the sudden appearance of the glowing word "God" in the sky over every populated area, and I think even Strahler would accept a supernatural explanation without fearing that he is destroying science.
This, of course, was a toungue-in-cheek example designed to show the in principle possibility of methodological naturalism confirming a supernatural hypothesis. In reality the universe does not show such blatant, unambiguous signs of creation, which is why scientists currently judge naturalistic explanations to be prima facie more parsimonious on average. But it is not beyond the realm of possibility that this state of affairs might change - there are still some areas of research left where supernaturalistic explanations may ultimately turn out to be more plausible than naturalistic ones. I believe cosmology is good example of such a area - currently, our best cosmological theories seem to contain a number of ad hoc constants that are stuck in various places to make a fit with the data. If, after many decades of reasearch, no naturalistic explanation as to why those unlikely constants - providing the exact conditions necessary for a universe with life in it - are there instead of others can be formulated, I believe the idea of supernatural creation will begin to seem more and more plausible as time goes by. This is not to say that I believe this is how things will end in actuality - quite the contrary in fact - but I think I have to conceed the possibility that they might end up that way, and methodological naturalism will not serve as a barrier to it. The same kind of scenarios can be drawn up for other currently troublesome areas of research, such as origin-of-life studies.
Of course, if naturalistic methodology can lead to the confirmation of supernaturalistic hypotheses, it can also lead to the falsification of supernaturalistic hypotheses. In fact, this is what has happened to a number of models offered by scientific creationists. Far from being unfalsifiable, the hypothesis that the earth was created 6,000 years ago by a supernatural entity has been falsified by the discovery of so much data that points to the old age of the earth. Likewise, the notion that a supernatural being directly created every "kind" seperately is given the lie by excellent transitional sequences. This is, I think, trivial, but there seems to be a lot of confusion about it, because scientists often assert that hypotheses like these creation models are unfalsifiable. However, such assertions on their part are themselves a bit confused - I think what scientists mean when they label hypotheses like special creationism unfalsifiable, is to draw attention to the fact that proponents of supernaturalistic hypotheses frequently fail to admit any data that disconfirms their hypotheses, and hence act unscientific themselves. But it is perfectly clear that a large number of hypotheses that appeal to the supernatural make sufficient emprical predictions to be falsified.
III. Can Science Be Based Upon Methodological Supernaturalism?
Proposition 2 expresses the standard view of the scientific community that science must be based upon methodological naturalism if there is to be any hope of fulfilling the aims of science. In this section, I will attempt to argue that there are a few different aspects in which methodological supernaturalism can, at least in principle, but also currently in practice, offer the same virtues that scientists tend to ascribe to methodological naturalism alone.
First, let's consider a very trivial case. Let us suppose that the universe is, in fact, a supernaturalistic one. In fact, it contains a God that is directly concerned with human affairs, and is helpful and loving to the point that it will answer clearly, fully, and honestly, and question submitted to it in earnest prayer. In this universe, methodological naturalism would not be able to compete in the slightest degree with methodological supernaturalism. The best way to acquire new data and formulate correct explanations for them would be to use the method of praying earnestly to God, and accepting the explanations he offers as the real picture of the universe. So obviously, it is not true that there is no conceivable way for the aims of science to be fulfilled except through methodological naturalism, because a supernaturalistic universe may be more amenable to methodological supernaturalism.[a]
A supernaturalistic methodology, of course, need not rely upon a non-empirical means of discovery of phenomena, as long as it assigns on average a higher prima facie probability to supernaturalistic explanations than it does to naturalistic ones. The methodological naturalist may object to this type of methodology on grounds that, in tending to seize upon supernatural forces and entities, the methodological supernaturalist is likely to posit far more entities in the universe than actually exist. This objection, however, only holds force if there is already good reason to believe that the universe is naturalistic - the methodological supernaturalist may just as easily say that the methodological naturalist, in lending too little prima facie possibility to the existence of supernatural forces and entities, will end up positing far fewer entities in the universe than actually exist. And it cannot be the case that the hypothesis that accounts for observations with fewer entities is always the more parsimonious one, otherwise naturalists should be far more hospitable to the suggestion by Wheeler and Feynman that there may be only one electron in the universe, which appears to be many because it zig-zags back and forth through time as well as space. A more pressing worry is whether such a supernaturalistic methodology would be a "science stopper"; say, for instance, some methodological supernaturalist appeals to the psychic powers of invisible dwarves as an explanation for why various atoms stick together - wouldn't this prevent any progress in chemistry from every being made, because it buries all interesting phenomena under the first stupid hypothesis one dreams up of? Not necessarily, I think. If the methodological supernaturalist remains devoted to empiricism as a means of collecting data, I would expect him to eventually end up with a whole pantheon of different dwarves that correspond to the different types of chemical bonds that can exist, and this could actually pave a path for fruitful future research and technological developments in the same way that a naturalistic understanding of chemical bonds does. Just because one posits a supernatural force as the explanation of a phenomenon, does not mean that one ceases to examine the phenomenon and refinine her explanations.
Finally, I want to note the existence of a class of supernaturalistic methodologies that overlap almost perfectly with methodological naturalism. These, I will classify under the name of "methodological deism" - the practice of employing the kind of methodology a deist would use to fulfill the goals of science. Deists are people who believe that there is a God who set everything in motion in accordance with natural law and no longer interferes. So such methodologies would entail the seeking of naturalistic explanations when dealing with anything except the origin of the universe, and the seeking of supernaturalistic explanations when dealing with the origin of the universe. So for a person working in any field except for cosmology, methodological deism and methodological naturalism will be identical, because both rely upon empirical investigations and an initial search for naturalistic explanations in just about every subject investigated. Hence, the two methodologies have exactly the same virtues in most areas, and in the realm of cosmology methodological deism will have the same potential virtues that the previous class of supernaturalistic methodologies we examined have. I note the existence of methodological deism to underscore exactly how divorced methodological naturalism is from dogmatic commitment to metaphysical naturalism - methodological naturalism is almost exactly the same type of methodology a certain class of through-and-through supernaturalists would use to investigate and explain the universe.
IV. Should Scientists Abandon Methodological Naturalism?
In the last two sections, I have made a number of claims that are far more hospitable to supernaturalism than most scientists currently seem comfortable with. Especially given the last section, in which I claimed that science can be based upon supernaturalistic methodologies, does that mean that I think scientists should abandon methodological naturalism? In this section, I will argue that there is no need for them to change methodologies.
So far, the history of science seems to have shown a progression from supernaturalism to naturalism. Scientists would probably not give naturalistic explanations greater prima facie plausibility than supernaturalistic ones if the data encountered in the past did not largely suggest that the universe is naturalistic. Naturally, we cannot know that there are no supernatural entities lying in wait just around the next corner - perhaps the trend of moving humanity farther and farther from the center of existence will some day reverse itself. In any case, I hope I have shown in the second part of this paper that should the data start to suggest the existence of the supernatural, methodological naturalism will not ultimately stand in its way, although it will probably slow its acceptance down. Should such a shift start, and itself become a trend, methodological naturalism will likely gradually be supplanted by some variant of methodological supernaturalism as the data influences our standards of parsimony, which in turn dictates to us which methodology we use.
Given the fact that methodological naturalism does not cause us to run the risk of overlooking blatent signs of the supernatural, given that it has as many virtues as any supernaturalistic methodology is likely to have in a universe that yields the data ours does, and given the fact that the scientific community is already entrenched in the specific symbols and procedures of methodological naturalism, there seems to be no reason at present to ask that scientists take up a new brand of methodology. To do so would only serve to divide the scientific community into two sections that will have extreme difficulty communicating with one another; who knows what the social and political ramifications of such a schism would be? It seems more reasonable to try to make the scientific community shift as a whole, by operating within the framework of methodological naturalism, and seeking out the evidence that no scientist's standard of parsimony can permit a naturalistic explanation for. Until such evidence is found, there is no reason to demand a shift - if the universe is naturalistic, then we are in good shape already. If the universe is supernaturalistic, then it is only a matter of time before the dominant methodology of science shifts of its own accord.
 Strahler, Arthur N. 1987. Scienceand Earth History. Buffalo: Prometheus.
 At least, this is the impression I have received from my experiences at UC Irvine and the University of Illinois at Chicago - perhaps it is just a peculiarity of these two departments, or perhaps my impressions are distorted, but most philosophers I have met are friendly towards traditional scientific attitudes.[b]
 Vuletic, M.I. 1996. Frequently Encountered Criticisms in Creationism vs. Evolution: Revised and Expanded. http://www2.uic.edu/~vuletic/cefec.html I intend to wait for feedback from the NTSE conference participants before making any revisions to the possibly culpable section of this document.[c]
[a] Ironically for me, this "trivial" case does not work in the unqualified form in which I have stated it. For a project based on revelation to be consistent with science, those engaged in the project must be fallibilists - an attitude rarely adopted by proponents of revelation. However, if a person manages to mesh revelation and fallibilism (and I see nothing in principle impossible about this) then the resultant attitude should be consistent with the scientific attitude.
[b] I now believe my impressions were distorted - further interaction with other philosophers at UIC has demonstrated to me that modern philosophers seem to adopt metaphysical naturalism as their default position, but have varying beliefs about the consequences and scientific necessity of methodological naturalism.
[c] The changes have now been made.
Last updated: Sunday, 23-Jan-2011 16:42:07 CST