Methodological Naturalism — Scope and Limits

MN as Constitutive of Science

Abstract

On this view, the restriction of science to natural causes is not an optional policy but part of the discipline's constitution: a hypothesis that invokes supernatural agency is not thereby refuted, but it is thereby removed from science's jurisdiction. The position received its canonical legal-philosophical statement at the Kitzmiller v. Dover trial, where Robert Pennock testified "that Intelligent Design, in its appeal to supernatural mechanisms, was not methodologically naturalistic, and that methodological naturalism is an essential component of science" (SEP 'Religion and Science' §1.2). Its most eloquent historical articulation, however, is Victorian: Thomas Henry Huxley's account of the agnostic method, which grounds the exclusion of supernatural explanation not in a definition but in an ethics of evidence.

Formal statement

1. Science is a practice individuated by its method: explanation of phenomena by natural entities and law-governed processes, testable against public evidence.
2. Supernatural hypotheses invoke agents whose actions are not law-governed or independently testable in the required way.
3. Therefore supernatural hypotheses cannot function as scientific explanations (whatever their standing elsewhere).
4. Therefore MN is constitutive of science, while remaining silent on ontology: it "does not make any statements about whether or not supernatural entities exist."

Key evidence / textual basis

The philosophical anatomy is given twice over in the corpus. Papineau's SEP entry notes that among philosophers of religion "'methodological naturalism' is the view that religious commitments have no relevance within science: natural science itself requires no specific attitude to religion, and can be practised just as well by adherents of religious faiths as by atheists or agnostics" (SEP 'Naturalism' §2.1). De Cruz's entry supplies the standard two-part taxonomy: "Naturalists draw a distinction between methodological naturalism, an epistemological principle that limits scientific inquiry to natural entities and laws, and ontological or philosophical naturalism, a metaphysical principle that rejects the supernatural" (SEP 'Religion and Science' §1.2).

The primary historical source is Huxley. Recalling his coinage, he writes: "So I took thought, and invented what I conceived to be the appropriate title of 'agnostic.' It came into my head as suggestively antithetic to the 'gnostic' of Church history, who professed to know so much about the very things of which I was ignorant" (Huxley, "Agnosticism" (1889)). The method behind the label is the point: "Agnosticism, in fact, is not a creed, but a method, the essence of which lies in the rigorous application of a single principle... Positively the principle may be expressed: In matters of the intellect, follow your reason as far as it will take you, without regard to any other consideration. And negatively: In matters of the intellect do not pretend that conclusions are certain which are not demonstrated or demonstrable" (Huxley, "Agnosticism" (1889)). In "Agnosticism and Christianity" he sharpens it into a norm as much ethical as epistemic: "it is wrong for a man to say that he is certain of the objective truth of any proposition unless he can produce evidence which logically justifies that certainty" (Huxley, "Agnosticism and Christianity" (1889)).

Huxley also supplies the operational rule that makes supernatural testimony scientifically inert — a proportionality canon he deploys against the Gadarene demoniac narrative: the rule of common sense is "to require corroborative evidence in exact proportion to the contravention of probability by the thing testified" (Huxley, "Agnosticism" (1889)) — a consciously Humean standard (see David Hume and Miracles and the Laws of Nature). To the reply that science itself rests on faith he answers that "faith is not necessarily entitled to dispense with ratiocination because ratiocination cannot dispense with faith as a starting-point" (ibid.).

Institutionally, the view was engineered as much as argued: the X-club, "founded in 1864 by Thomas Huxley and friends," had as "its explicit aim... to promote a science that would be free from religious dogma" (SEP 'Religion and Science' §1.2).

Leading proponents

• Thomas Henry Huxley — coined "agnostic"; in corpus (Science and Christian Tradition, Collected Essays V). His agnosticism is the Victorian prototype of MN: not a denial of the supernatural but a refusal to admit it without evidence proportioned to its improbability.
• Robert T. Pennock — Kitzmiller expert witness; argued MN is "an essential component of science." Primary works (Pennock 1998, 2000) not in corpus; cited via SEP.
• Barbara Forrest — the standard statement of the MN/ontological-naturalism distinction (Forrest 2000); not in corpus; cited via SEP.

Strongest counter-arguments

First, the definitional move is contested even by neutral parties. The SEP's design-arguments entry observes that the prohibition on supernatural agency, "commonly known as methodological naturalism — is often claimed (mistakenly, some argue) to be definitive of genuine science" (SEP 'Teleological Arguments' §2.2) — the parenthesis registering that the claim of definitiveness is itself philosophically disputed, not settled. Demarcation criteria have a poor track record in philosophy of science; the same entry notes that "'established' limitations on science have been overturned in the past."

Second, constitutive MN threatens to make naturalism unfalsifiable science-wide. If no possible empirical finding could ever count scientifically in favor of design or special divine action, then science's uniform verdict for naturalistic causes is partly an artifact of the rule, not a discovery — the core of Johnson's charge, examined under the third view below.

Third, the Huxleyan evidential ethic proves less than the definitional claim. Huxley himself concedes of the Gadarene demons: "I admit I have no a priori objection to offer... I am unable to show cause why these transferable devils should not exist" (Huxley, "Agnosticism" (1889)). His exclusion of the supernatural is probabilistic, not definitional — precisely the second view below.

Responses

Defenders reply (i) that the definitional point is institutional rather than metaphysical: science as a public, cumulative practice requires testability, and "supernatural agent" hypotheses characteristically lack the constrained expectations testability demands; (ii) that unfalsifiability-of-naturalism worries confuse the rule with its outputs — MN leaves the theological question genuinely open, which is why theists practice science without contradiction (SEP 'Naturalism' §2.1); (iii) that Huxley's evidentialism and Pennock's constitutivism converge in practice even if they diverge in theory: on either, no extant supernatural hypothesis meets the bar.

Assessment

Assessment: Strong — this is the operative consensus of working science and was decisive at Kitzmiller; but its philosophical foundation (definition vs. track record) is genuinely contested, and its best historical exemplar, Huxley, is more naturally read as holding the pragmatic view.

MN as Provisional / Pragmatic

Abstract

The middle position holds that science's preference for naturalistic explanation is real, rational, and overwhelmingly well-earned — but a posteriori and defeasible, not definitional. Science excludes supernatural hypotheses the way a physician excludes exotic diagnoses: because they have persistently failed, not because the rulebook forbids them. This reading fits the actual history (early-modern science was not methodologically naturalistic; the norm hardened in the nineteenth century), avoids the demarcation quagmire, and — importantly for the theist — leaves a principled door open: if the evidence were strong enough, a design inference would be assessable rather than inadmissible.

Formal statement

1. The naturalistic constraint on science is a methodological policy, and policies require justification.
2. Definitional justifications fail: MN is historically recent, and demarcation criteria are notoriously unreliable.
3. The actual justification is inductive: naturalistic explanation has succeeded, and supernatural explanation has persistently receded ("purported gaps have been closed by new scientific theories").
4. Therefore MN is a rationally held but defeasible preference — and its continued success is itself evidence relevant to the ontological question.

Key evidence / textual basis

The historical premise is documented in the SEP: "Methodological naturalism is a recent development in the history of science, though we can see precursors of it in medieval authors such as Thomas Aquinas who attempted to draw a theological distinction between miracles, such as the working of relics, and unusual natural phenomena, such as magnetism and the tides" (SEP 'Religion and Science' §1.2). Newton, Kepler, Hooke, and Boyle appealed to supernatural agents; "still, overall there was a tendency to favor naturalistic explanations in natural philosophy" (ibid.).

The inductive premise is stated in the design-arguments entry: "Purported gaps have been closed by new scientific theories postulating means of natural production of phenomena previously thought to be beyond nature's capabilities. The most obvious example of that is, of course, Darwin's evolutionary theory and its descendants" (SEP 'Teleological Arguments' §2.2). The same entry concedes the flip side that makes the preference defeasible in principle: "The position that there are causal gaps in nature is not inherently irrational — and would seem to be a legitimate empirical question" (ibid.).

The evidential-weight thesis is Draper's: the preference for naturalistic causes "may have been encouraged by past successes of naturalistic explanations, leading authors such as Paul Draper (2005) to argue that the success of methodological naturalism could be evidence for ontological naturalism" (SEP 'Religion and Science' §1.2). Note the structure: if MN's success is evidence for ontological naturalism, then MN is not a definition — definitions do not accrue evidential weight. Ruse's Creationism entry draws the same picture from the theistic side, cataloguing the options open to "the would-be methodological naturalist" who is also a Christian, and observing that "many significant theologians of our age think that, with respect to miracles, science and religion have no conflict" (Ruse, SEP 'Creationism' §6).

Huxley, read carefully, belongs here as much as to the first view: his canon is a sliding evidential scale, not an exclusion rule, and he insists that "with scientific Theology, Agnosticism has no quarrel... The scientific theologian admits the Agnostic principle, however widely his results may differ from those reached by the majority of Agnostics" (Huxley, "Agnosticism and Christianity" (1889)). His war was with "Ecclesiasticism" — belief mandated regardless of evidence — not with theology as inquiry.

Leading proponents

• Paul Draper — MN as an epistemological principle whose success carries evidential weight (Draper 2005); in corpus via SEP citation only.
• Thomas Henry Huxley — on the evidentialist reading of his agnostic method (in corpus).
• Michael Ruse — MN and evolution as a "package deal" held on scientific rather than definitional grounds; author of the SEP Creationism entry (in corpus as reference source).

Strongest counter-arguments

First, from the constitutive side: a merely pragmatic MN gives away too much — it implies court cases like Kitzmiller were wrongly decided on demarcation grounds, and it obliges scientists to keep re-litigating supernatural hypotheses that have failed for centuries. Second, from the theistic side: the "track record" argument is circular if the record was compiled under the very rule at issue — explanations invoking design were never allowed to compete, so their absence from the win column shows little (Johnson's complaint, below). Third, Draper's evidential extension cuts both ways: if MN's success is admissible evidence for ontological naturalism, then by parity a genuine, stable explanatory gap (e.g., the origin of life, or cosmic fine-tuning — see The Fine-Tuning Argument) would be admissible evidence against it; the naturalist cannot both claim the evidential credit and refuse the evidential risk.

Responses

Pragmatists reply (i) that Kitzmiller can be defended on the narrower ground that this particular program (ID) had failed scientifically and was religiously motivated, without needing MN-as-definition; (ii) that the circularity charge fails historically — the early-modern period did allow supernatural explanation, and it lost on performance, not by rule (SEP 'Religion and Science' §1.2); (iii) that accepting evidential risk is a feature, not a bug: it is precisely what makes the position non-dogmatic, and the naturalist is content to let fine-tuning and abiogenesis be argued on the merits.

Assessment

Assessment: Live — arguably the most defensible philosophical account of what scientific practice actually does, and increasingly common among philosophers of science on both sides of the theism debate; its cost is that it surrenders the clean demarcation line that institutional actors (courts, school boards) want.

MN Rejected: Theistic Science

Abstract

The third position holds that MN, however innocent in principle, functions in practice as metaphysical naturalism with a methodological alibi — and that theists should refuse it. Its legal-rhetorical architect is Phillip Johnson, who argued that methodological naturalism "slides into" the metaphysical kind; its philosophical heavyweight is Alvin Plantinga, who has argued both that religious doctrines may legitimately inform scientific practice and that the real conflict is "not between science and religion, but between science and naturalism"; its scientific program is the intelligent-design movement of Behe and Dembski. The position is philosophically serious and institutionally defeated: its Kitzmiller loss and the mainstream verdict on ID leave it under pressure, though Plantinga's independent critique of naturalism remains live.

Formal statement

1. MN restricts science to naturalistic explanations regardless of evidence.
2. A restriction that operates regardless of evidence is a philosophical commitment, not a scientific finding.
3. For the theist, who has independent grounds for believing God acts in nature, adopting that commitment methodologically means systematically ignoring what she takes herself to know.
4. Therefore the theistic scholarly community is entitled to a science that uses all it knows — a "theistic science" conducted from Christian commitments (terminology flagged below) — and design should be admissible as an explanatory category.

Key evidence / textual basis

Johnson's distinction is set out in the SEP Creationism entry: methodological naturalism is "the scientific stance of trying to explain by laws and by refusing to introduce miracles," whereas "Naturalism is a metaphysical doctrine... According to naturalism, what is ultimately real is nature" (Johnson 1995, 37–38, quoted at Ruse, SEP 'Creationism' §5). Johnson's "theistic realist" believes God "always has the option of working through regular secondary mechanisms," but that "many important questions — including the origin of genetic information and human consciousness — may not be explicable in terms of unintelligent causes, just as a computer or a book cannot be explained that way" (Johnson 1995, 209, quoted ibid.). His central claim is that the methodological position "slides into" the metaphysical one, so that the methodological naturalist ends up the theist's opponent whatever her private beliefs (ibid.).

Plantinga's dissent from MN is registered in Papineau's entry: against the consensus that religious commitments have no relevance within science, "Some think that religious doctrines do make a difference to scientific practice, yet are defensible for all that (Plantinga 1996)" (SEP 'Naturalism' §2.1). His broader argument relocates the conflict: on the SEP's summary, "Alvin Plantinga (2011) has argued that the conflict is not between science and religion, but between science and naturalism" (SEP 'Religion and Science' §1.3). The vehicle is his Evolutionary Argument Against Naturalism (first framed in 1993): naturalism, he contends, is epistemically self-defeating, since if naturalism and evolution are both true it is improbable that our cognitive faculties would be reliable. On divine action he holds that randomness is "a physicalist interpretation of the evidence": God may "guide the course of evolutionary history by causing the right mutations to arise at the right time" (Plantinga 2011: 121, quoted at SEP 'Religion and Science' §3.1). Ruse records that Where the Conflict Really Lies (2011), based on Plantinga's 2005 Gifford Lectures, extends the critique to Darwinian theory itself, "arguing that the evidence in its favor is scanty," and "gives a very sympathetic reading of the thinking of Michael Behe" (SEP 'Creationism' §12).

{{UNSOURCED: Plantinga 1996 (the paper cited at SEP 'Naturalism' §2.1) and Where the Conflict Really Lies (2011) are not in corpus; his positive proposal — a science conducted from within the resources of the Christian community, commonly labelled "theistic science" or "Augustinian science," contrasted with a religiously neutral science — is here reported only as glossed by SEP entries, and the labels themselves await verification against the primary texts. Primary texts needed for the detailed statement of the proposal and the "deep concord, superficial conflict" thesis.}}

The ID wing supplies the would-be scientific content: "ID advocates propose two specialized Rs — irreducible complexity (Behe 1996) and specified complex information (Dembski 1998, 2002)" (SEP 'Teleological Arguments' §4.2). And even a critic-neutral source notes the fallback position: "even if the case is made that ID could not count as proper science, which is controversial, that would not in itself demonstrate a defect in design arguments as such. Science need not be seen as exhausting the space of legitimate conclusions from empirical data" (ibid.).

Scripturally, proponents appeal to Rom 1:20 (bib) as warrant that design is perceivable in nature, and to Ps 19:1-4 (bib) as the datum a theistic science would not be forced to bracket; Col 1:16-17 (bib) is cited on both sides of the intra-Christian debate (see Key Scripture below).

Leading proponents

• Alvin Plantinga — MN critique (Plantinga 1996), EAAN, guided-mutation compatibilism; primary anti-MN works not in corpus (see gap flag).
• Phillip E. Johnson — Darwin on Trial (1991) and Johnson 1995; the MN-slides-into-metaphysics argument; not in corpus, quoted extensively in SEP 'Creationism'.
• William Dembski — specified complex information as a design-detection criterion (Dembski 1998, 2002); not in corpus; cited via SEP.
• Michael Behe — irreducible complexity (Darwin's Black Box 1996); not in corpus; cited via SEP.

Strongest counter-arguments

First, the slide thesis is asserted, not demonstrated. Ruse presses this directly: "as Johnson himself notes, many people think that they can be methodological naturalists and theists. Methodological naturalism is not a religion equivalent" (SEP 'Creationism' §5). Johnson's response to the accommodationist options is, in Ruse's judgment, "remarkably little" — sneer rather than argument, as when he calls theistic evolution a "makeshift compromise" between "supernaturalism in religion and naturalism in science" (Johnson 1995, 212, quoted ibid.).

Second, the scientific program has not delivered. The design-arguments entry catalogues the objections along three lines: "that ID advocates have simply gotten the relevant science wrong," that even correct science does not yield the empirical support design advocates claim for it, and that demonstrably better alternative explanations of the same phenomena leave the ID cases without force (SEP 'Teleological Arguments' §4.2). Ruse's own verdict on creationism and "its offspring, Intelligent Design Theory" is blunt: scientifically worthless, philosophically confused, theologically blinkered (SEP 'Creationism' §12) — a partisan assessment, but one aligned with the Kitzmiller outcome.

Third, the intra-Christian objection. Classical theism's primary/secondary causation framework (Aquinas; cf. William Carroll's neo-Thomist argument that "God is not a cause in the way creatures are causes, competing with natural causes," SEP 'Religion and Science' §3.1) implies that a science of secondary causes is theologically proper, not a concession to naturalism. On this Thomist line, theistic science misconstrues divine action as one more cause within the world — a theological error older than the debate (see Origin of the Universe on Aquinas and the eternity of the world).

Responses

Proponents reply (i) that Plantinga's EAAN, whatever one thinks of ID, stands independently: if unguided evolution undermines the reliability of our faculties, MN's alliance with evolutionary naturalism is self-referentially unstable (SEP 'Religion and Science' §1.3); (ii) that the fallback holds — even if ID is not science, design inferences may be rational conclusions from empirical data, and the constitutive-MN theorist has no non-question-begging reply (SEP 'Teleological Arguments' §4.2); (iii) that the Thomist objection, if sound, cuts against interventionism generally, not against the entitlement claim — Plantinga's point is about the epistemology of the Christian scholar, not about occasionalist physics.

Assessment

Assessment: Under pressure — the ID program lost decisively in court and has not gained scientific standing; Johnson's slide thesis is widely judged undemonstrated. But Plantinga's philosophical critique of MN's justification, and the EAAN, remain live topics in the literature and cannot be dismissed with the movement's institutional defeat.