Science Doesn’t Need Methodological Naturalism

This entry is part 1 of 6 in the series Science Doesn't Need MN

(Update 3/29: Please regard this series as a first draft with important revisions yet to come.)

It’s virtually a given that science cannot operate apart from methodological naturalism. Eugenie Scott of the National Center for Science Education provides a definition for the term as she expresses some of the devotion that is felt for it:

Science neither denies nor opposes the supernatural, but ignores the supernatural for methodological reasons. The history of science has shown that progress comes from logical and empirical study rather than reference to revelation or to inner psychological states. That’s how we play our game; his basketball won’t work on our baseball field. The essence of science is empiricism and control of variables, and if there is an omnipotent God, it certainly can’t be controlled like temperature or humidity. Science has made a little deal with itself: because you can’t put God in a test tube (or keep it out of one), science acts as if the supernatural did not exist. This methodological materialism is the cornerstone of modern science.

(Materialism and naturalism mean essentially the same thing in this context.)

In this blog series I intend to argue that this is wrong; that methodological naturalism (MN) is not essential to modern science. In this first part I will lay out three primary reasons scientists and philosophers invoke for adhering to MN. In the process I am opening up discussion especially for the purpose of hearing whether there are other major justifications for MN that I have missed. (I predict that some readers will want to jump in right now to refute my arguments. I suggest you wait until the next part in the series when I actually begin to make them.)

1. Testability
The first common justification for MN is testability: if God is involved in some natural phenomenon, we could never know it scientifically, for there is no scientific test that could detect his actions. We’ve already seen Eugenie Scott’s views on that. She is hardly the only one. Speaking as witness at the Dover Intelligent Design (ID) trial, Robert Pennock (philosopher of science at Michigan State University) said,

Someone who says well, we have to consider the possibility of supernatural interventions might say well, you know, God was in there or some supernatural designer was in there changing the bits inside the computer.

Well, you know, we don’t know if that’s true, and no scientist can ever know if that’s true. That’s not a testable proposition. So in that sense we can never rule that out. That’s part of what it means to be a methodological naturalist.

Elsewhere he has also said,

For a hypothesis to be scientific it must be inter-subjectively testable, and fit within the framework of law-governed cause-effect relations. This is the core of what it means to be a natural object and to be amenable to scientific investigation. Agreeing to be constrained by this sort of epistemological approach as the means of gathering public knowledge about the empirical world is just what it is to be a methodological naturalist…. In proposing a theistic science Johnson claims to be expanding science to supernatural possibilities undreamed of in this philosophy, but what he and other so-called Creation scientists are really doing is reducing God to a scientific object, placing God in the scientific box.

2. Rationality


Alongside testability, scientific rationality supposedly requires MN. Prominent Internet Infidel Mark I. Vuletic put it this 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.

At Dover, Robert Pennock voiced his concerns over “God or some supernatural being … in there fiddling with the gates.” He also writes,

Without the binding assumption of uninterruptible natural law there would be absolute chaos in the scientific worldview.

Arizona State University Physics Professor Lawrence M. Krauss quoted J.B.S. Haldane in the Wall Street Journal Opinion pages:

My practice as a scientist is atheistic. That is to say, when I set up an experiment I assume that no god, angel or devil is going to interfere with its course; and this assumption has been justified by such success as I have achieved in my professional career. I should therefore be intellectually dishonest if I were not also atheistic in the affairs of the world.

Then Krauss went on to speak for himself and an atheist compatriot:

Though the scientific process may be compatible with the vague idea of some relaxed deity who merely established the universe and let it proceed from there, it is in fact rationally incompatible with the detailed tenets of most of the world’s organized religions. As Sam Harris recently wrote in a letter responding to the Nature editorial that called him an “atheist absolutist,” a “reconciliation between science and Christianity would mean squaring physics, chemistry, biology, and a basic understanding of probabilistic reasoning with a raft of patently ridiculous, Iron Age convictions.”

The atheistic British philosopher A.C . Grayling has recently been quoted with this concern:

to say nothing of local suspension of the laws of nature for arbitrary reasons e.g. in answer to personal prayer, which makes a nonsense of the idea that the world or the deity is rationally comprehensible:

3. It Works


Finally, as Eugenie Scott points out, MN works:

Most scientists today require that science be carried out according to the rule of methodological materialism: to explain the natural world scientifically, scientists must restrict themselves only to material causes (to matter, energy, and their interaction). There is a practical reason for this restriction: it works. By continuing to seek natural explanations for how the world works, we have been able to find them. If supernatural explanations are allowed, they will discourage – or at least delay – the discovery of natural explanations, and we will understand less about the universe.

Three Reasons…
These then are the best justifications I know of for MN as a rule for science:

  1. The supernatural is not testable.
  2. Admitting the supernatural into science would undermine scientific rationality.
  3. MN works.

To preview where I am headed, I plan to argue that whatever these three reasons’ validity may be, none of them individually or taken together is sufficient to make MN a rule for science; and that there exists another and better basis upon which science can proceed productively, practically, and rationally.

Series Navigation (Science Doesn't Need MN):

Why Science Doesn’t Need Methodological Naturalism (2) >>>Why Science Doesn’t Need Methodological Naturalism >>>

Comments 7
  1. Greg

    My thoughts are quite unorganized at the moment, but what I see as the defining factor is whether a willful being is in the mix. If you allow concerns of God as a willful being into science, then the “hard” sciences like physics, chemistry, biology, will end up about as predictive as psychology.

  2. Holopupenko


    Greg’s largely got it right. We run “controlled” experiments for, among other things, to eliminate intentional or accidental influences by rational agents.

    But I have a bigger beef. The definition of science writ large is “mediate intellectual knowledge obtained through demonstration.” This means philosophy and theology are sciences in their own right. What distinguishes one science from another is the formal objects they study. If one is a physicist one studies physical beings in motion. If one is a biologist one formally studies life. If one is a theologian, one employs philosophical reflection to study revealed knowledge about God and God Himself, but not in the sense of God being just another object among objects.

    This is the problem Plantinga faces: he doesn’t make a formal distinction between objects, so he has to increase the bounds of all the individual sciences under a meta-umbrella by which to criticize methodological naturalism. (Dembski rides this horse into the ground.) Moreover, the “three primary reasons scientists and philosophers invoke for adhering to MN” you provide are a little bit of a red herring because you’re speaking from the perspective of the philosophy of science… whose formal object, by the way, is epistemic methodology. The philosophy of nature is much better equipped to demarcate bounds of study without sacrificing the symphony of truth.

    While I might criticize Eugenie Scott for the scientism I see creeping in between her lines, as a stand-alone statement, she’s correct. However, where she would be incorrect (and as unscientific as DL) would be to claim the “rules of the game” of the MESs are the “best” rules. That’s the question to pose to her.

  3. Bryan

    I eagerly await your next posts. The problem I have with these objections is (for one) that they seem to be based on an equivocation of operational science (i.e. chemistry, physics) and the historical sciences, which as Popper had suggested are more like metaphysical frameworks than actual sciences, concerning the infinite and unobservable past. So I see the objections as largely irrelevant.

  4. Crude

    My own contribution would be this:

    Putting aside how poorly naturalism is defined, not all of what are arguably naturalistic claims are testable. And rationality itself (either of minds or of the constituents of the universe) is not a ground rule of naturalism. Nor is testability. And insofar as anything is praised because it “works”, that opens the door to far more than naturalism.

    So my own take would be different: Science, arguably, has various limitations both in terms of scope and method. But it’s nonsense and spin to suggest that either the methods or the scope are tied to “naturalism”. (Really, that’s partly evident with Greg’s objection. It’s not as if the being in question has to be supernatural to be a problem – a “natural” being of sufficient power can lead to the same problems.)

  5. JAD

    I have no problem with MN as long as it’s just that MN. In other words, empirical science is the study of natural causation, or the chain of cause and effect, and trying to understand the mechanistic side of the world. However, that does not require that I assume that mechanism is all that there is, or that we can explain everything mechanistically. That is a metaphysical assumption (materialism or naturalism) not a methodological one.

    So when Scott writes this:

    Science has made a little deal with itself: because you can’t put God in a test tube (or keep it out of one), science acts as if the supernatural did not exist. This methodological materialism is the cornerstone of modern science.

    She is actually smuggling a metaphysical view of the world (her view) into science. For example, why is it necessary to act or pretend that “the supernatural [does] not exist“?

    Just admitting that we don’t know scientifically how this or that happened would be a better response. That does involve any kind of metaphysical assumption.

    What a thought apply agnostic thinking to science.

  6. Bill R.

    I am interested to see where you are going with this argument. I would say I am sympathetic, but cautious, towards attempts to demonstrate that MN is not necessary for science.

    My own two cents…. As someone who is finishing up a PhD in chemistry, I’ve done many experiments (from dozens to thousands, depending on how you define “experiment”). Most have succeeded in giving me useful information about the processes involved (whether the outcome of the experiment was positive or negative), but some have failed to give interpretable results for any number of reasons – power spikes in the electrical system, starting materials with unknown contaminants, etc. Because those experiments are dead and buried in the past (and in my lab notebook), science will never be able to explain, with 100% confidence, how or why they went wrong. In particular, science will never be able to rule out the occasional intervention of God, angels, demons, etc. as the cause for those failed experiments. Even with that possibility open, however, my labmates and I can still do science. Hence, as scientists, we need not assume that the supernatural realm never interferes with whatever portion of the chain of natural cause-and-effect we study in our experiments, only that supernatural intervention occurs rarely enough for us to statistically identify what is normal (i.e. “natural”).

  7. Pingback: In the News (Mar. 24, 2011) | The Apologist

Comments close automatically after 120 days. Comment numbering may be incorrect due to a temporary bug.