The no-miracles argument (NMA) holds that the best explanation of the empirical success of mature scientific theories is that they are approximately true. Frost-Arnold (2010) argued that methodological naturalists committed to strict parity between philosophical and scientific methods should reject this argument, because realism as an explanans offers neither novel predictions nor genuine unification, the twin standards scientists demand of an acceptable explanation. The paper's conclusion leaves open how, or whether, the realist can meet this challenge, closing with the substantive question of what a scientifically acceptable explanation is, and in particular whether some version of realism meets that standard: either some version of scientific realism can be reformulated to satisfy predictive novelty and genuine unification, or naturalists must abandon strict parity with scientific methods, treating the argument as a reductio of that naturalism. A positive answer must additionally confront two difficulties the paper itself flags: the disunity and occasional incompatibility of mature theories (e.g., quantum mechanics versus general relativity; divergent entity concepts across subfields), which the paper invokes against the 'giant conjunction' defense of the NMA, and the circularity of glossing a science's 'maturity' in terms of empirical success. The dilemma remains actively debated but unresolved in the literature through 2025, including the direct replies, the 'pragmatic turn' debate, critiques of naturalized metaphysics, and the selective-realism and pessimistic-induction literature.

Background

Scientific realism is the view that our best scientific theories are approximately true and that the unobservable entities they posit (electrons, genes, black holes) exist mind-independently. Its classic argument is the no-miracles argument (NMA), associated with Hilary Putnam: the empirical success of science would be a miracle or cosmic coincidence if our successful theories were not approximately true, so their approximate truth best explains their success. The NMA is an inference to the best explanation (IBE), the pattern of reasoning from an explanandum (the phenomenon to be explained, here the success of science) to the explanans (the hypothesis doing the explaining, here realism).

Two framing developments define the problem. First, Larry Laudan's 'pessimistic meta-induction' surveys historically successful theories (e.g., the caloric and aether theories) whose central claims were later abandoned, and inductively concludes that our current best theories are probably false too. Realists usually reply by restricting the induction to theories that are 'mature' and 'non-ad hoc' (not merely constructed to fit known data), but by common admission the notion of maturity remains vague, and spelling it out in terms of predictive success risks circularity. Second, many proponents of the NMA endorse a methodological naturalism on which philosophy of science may use no methods that science itself does not use, so that philosophical theses are held to the same standards of acceptability as scientific hypotheses.

The source of this problem, Greg Frost-Arnold's 'The No-Miracles Argument for Realism: Inference to an Unacceptable Explanation' (Philosophy of Science 77(1), 2010), pressed an internal difficulty for that combination of views. Scientists, the paper observes, usually reject explanations whose explanans neither generates novel predictions nor unifies apparently disparate established claims. The stock example of such a pseudo-explanation is the 'dormative virtue': opium puts people to sleep because it has a sleep-inducing power, a relabeling of the explanandum that predicts nothing new and unifies nothing. Genuine unification, in the tradition of Philip Kitcher's account (an explanation unifies by deriving many phenomena from few argument patterns, as Newton's mechanics derived both planetary motion and terrestrial free fall) and Margaret Morrison's (a unified theory predicts or explains unexpected phenomena), is a demanding standard. Frost-Arnold argues that realism, as it appears in the NMA, is precisely a dormative-virtue-style explanans: it generates no new predictions and unifies no disparate established claims, so naturalists who insist on strict parity between philosophical and scientific methods should reject their own favorite argument for realism. Kitcher's (2001) 'Galilean Strategy', which defends realist inference patterns by exhibiting their use and success within science, is the most prominent alternative naturalistic route that the paper also engages.

The paper's conclusion identifies but does not resolve the substantive questions about the relation between philosophical method and scientific practice: what a scientifically acceptable explanation is, whether some version of realism meets that standard, or whether the argument instead works as a reductio of the strict-parity naturalism it targets. This dilemma constitutes the problem recorded here; the precise analyses such a vindication would require (of unification and of maturity) and the disunity of mature theories are difficulties the paper itself flags along the way rather than co-equal open questions posed at its close.

Problem Statement

The problem, posed at the close of Frost-Arnold (2010), is whether scientific realism, as deployed in the no-miracles argument, can be vindicated under the strict-parity methodological naturalism that its own proponents endorse: can the realist explanation of scientific success satisfy, without circularity, the same standards of predictive novelty and unification that scientists demand of an acceptable explanation? In the paper's own closing formulation, the substantive questions are what a scientifically acceptable explanation is, and in particular whether some version of realism meets that standard. The dilemma has two horns:

(1) A positive vindication. Supply precise, operational criteria for what counts as a scientifically acceptable explanation, and in particular for when an explanans genuinely unifies heterogeneous domains of inquiry in the way scientists themselves count as unifying (paradigm: Newton's unification of terrestrial and celestial motion), as opposed to merely verbal, dormative-virtue-style pseudo-unification; and show that some version of scientific realism meets those criteria. Such a vindication must confront the difficulties the paper itself raises against realist defenses: how the disunity and sometimes incompatibility of mature scientific theories (for example the unresolved tensions between quantum mechanics and general relativity, or divergent concepts of central theoretical entities across subfields, which the paper uses to defeat the 'giant conjunction' defense of the NMA) can be accommodated within any global account that infers truth or approximate truth from empirical success; and how the 'maturity' restriction on the realist thesis can be characterized without circularity, given that explications of maturity in terms of empirical success or novel predictive success appear to beg the question against the anti-realist. For each difficulty, the positive answer must offer an argument sufficient to show that it does not defeat the vindication: the disunity and incompatibility of mature theories must be shown not to undermine the global realist inference, and the 'maturity' restriction must be given a characterization that is non-circular for the purposes of the argument. The answer must not merely concede either difficulty as an open worry, but it is not required to settle the general problem of theory disunity or to give a fully general account of maturity.

(2) A negative answer. A sustained argument that no version of realism can meet the twin standards under strict parity, together with the consequences the paper's conclusion identifies: that methodological naturalists must abandon strict parity between philosophical and scientific methods in order to justify realism, treating Frost-Arnold's argument as a reductio of that naturalism, or else reformulate realism in some other way.

References

  1. Greg Frost-Arnold (2010), "The No-Miracles Argument for Realism: Inference to an Unacceptable Explanation", Philosophy of Science 77(1): 35-58. DOI: 10.1086/650207, https://doi.org/10.1086/650207
  2. Hilary Putnam (1979 [1975]), Mathematics, Matter and Method (Philosophical Papers, Volume 1, 2nd edition), Cambridge University Press. DOI: 10.1017/CBO9780511625268, https://doi.org/10.1017/CBO9780511625268
  3. Larry Laudan (1981), "A Confutation of Convergent Realism", Philosophy of Science 48(1): 19-49. DOI: 10.1086/288975, https://doi.org/10.1086/288975
  4. Philip Kitcher (1981), "Explanatory Unification", Philosophy of Science 48(4): 507-531. DOI: 10.1086/289019, https://doi.org/10.1086/289019
  5. Margaret Morrison (1990), "Unification, Realism and Inference", The British Journal for the Philosophy of Science 41(3): 305-332. DOI: 10.1093/bjps/41.3.305, https://doi.org/10.1093/bjps/41.3.305
  6. Philip Kitcher (2001), "Real Realism: The Galilean Strategy", The Philosophical Review 110(2): 151-197. DOI: 10.1215/00318108-110-2-151, https://doi.org/10.1215/00318108-110-2-151
  7. Moti Mizrahi (2012), "Why the ultimate argument for scientific realism ultimately fails", Studies in History and Philosophy of Science Part A 43(1): 132-138. DOI: 10.1016/j.shpsa.2011.11.001, https://doi.org/10.1016/j.shpsa.2011.11.001
  8. Paul Hoyningen-Huene (2013), "The Ultimate Argument Against Convergent Realism and Structural Realism: The Impasse Objection", in EPSA11 Perspectives and Foundational Problems in Philosophy of Science, Springer, pp. 131-139. DOI: 10.1007/978-3-319-01306-0_11, https://doi.org/10.1007/978-3-319-01306-0_11
  9. Greg Frost-Arnold (2014), "Can the Pessimistic Induction be Saved from Semantic Anti-Realism about Scientific Theory?", The British Journal for the Philosophy of Science 65(3): 521-548. DOI: 10.1093/bjps/axt013, https://doi.org/10.1093/bjps/axt013
  10. Seungbae Park (2017), "The Unificatory Power of Scientific Realism", Disputatio 9(44): 59-73. DOI: 10.2478/disp-2017-0027, https://doi.org/10.2478/disp-2017-0027
  11. K. Brad Wray (2018), Resisting Scientific Realism, Cambridge University Press. DOI: 10.1017/9781108231633, https://doi.org/10.1017/9781108231633
  12. Anjan Chakravartty (2017), "Scientific Realism", in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (substantive revision of 12 June 2017), https://plato.stanford.edu/entries/scientific-realism/
  13. Jesús P. Zamora Bonilla (2019), "Realism versus anti-realism: philosophical problem or scientific concern?", Synthese 196(10): 3961-3977. DOI: 10.1007/S11229-015-0988-6, https://doi.org/10.1007/S11229-015-0988-6
  14. Florian J. Boge (2021), "Incompatibility and the pessimistic induction: a challenge for selective realism", European Journal for Philosophy of Science 11(2), article 46. DOI: 10.1007/s13194-021-00367-4, https://doi.org/10.1007/s13194-021-00367-4
  15. Sam Baron (2021), "Unification and mathematical explanation in science", Synthese 199(3-4): 7339-7363. DOI: 10.1007/s11229-021-03118-3, https://doi.org/10.1007/s11229-021-03118-3
  16. Seungbae Park (2022), "In Defense of the No-Miracles Argument", in Embracing Scientific Realism (Synthese Library), Springer, pp. 1-22. DOI: 10.1007/978-3-030-87813-9_1, https://doi.org/10.1007/978-3-030-87813-9_1
  17. Seungbae Park (2022), "Critiques of the Pessimistic Induction", in Embracing Scientific Realism (Synthese Library), Springer, pp. 23-43. DOI: 10.1007/978-3-030-87813-9_2, https://doi.org/10.1007/978-3-030-87813-9_2
  18. Rasmus Jaksland (2023), "A trilemma for naturalized metaphysics", Ratio 36(1): 1-10 (first published online 2022). DOI: 10.1111/rati.12344, https://doi.org/10.1111/rati.12344
  19. Sandy C. Boucher and Curtis Forbes (2024), "The pragmatic turn in the scientific realism debate", Synthese 203(4), article 111. DOI: 10.1007/s11229-024-04528-9, https://doi.org/10.1007/s11229-024-04528-9