Hume on Miracles: The Argument and Its Critics

Summary

David Hume's argument against miracles, presented in Section X of An Enquiry Concerning Human Understanding (1748), is the most influential modern philosophical challenge to claims of prophetic revelation. Hume argued, in compressed form, that the evidence for a miracle could never reasonably outweigh the evidence against it, because the testimony for a miracle is necessarily exceptional while the laws-of-nature evidence against it is established by uniform experience. Three centuries of responses — from Hume's contemporaries to Bayesian reconstructions in the twenty-first century — have produced one of the most extensively debated arguments in philosophy of religion. The framework engages Hume's argument seriously while noting that the strongest contemporary responses substantially weaken its force, particularly when miracle claims are treated not in isolation but as part of a cumulative case.

Hume's Argument: Statement

Hume's argument is presented in two parts. The first part offers a general epistemological principle about miracles; the second applies the principle to actual historical miracle claims.

Part I: The general principle

Hume's core inference can be reconstructed as follows:

A wise person proportions belief to evidence.
A miracle is, by definition, a violation of the laws of nature.
The laws of nature are established by uniform experience — the strongest evidence available.
Therefore evidence against a miracle is, in any case, the strongest possible: uniform experience against the violation.
Evidence for a miracle is testimony. Testimony can vary in reliability but is intrinsically weaker than uniform experience.
Therefore: "no testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous than the fact which it endeavours to establish."

The argument's force is not that miracles are impossible — Hume is careful not to claim impossibility — but that testimony can never reasonably establish them. The argument is epistemic: it concerns the conditions under which belief in a miracle would be rational.

Part II: The application

Hume's second part argues that, even granting the general principle's permissiveness, no actual miracle claim has met the bar. He offers four arguments: (a) miracle witnesses tend to be of insufficient number, integrity, or social standing; (b) human nature is disposed to exaggeration and credulity, especially regarding the marvellous; (c) miracle reports predominantly come from "ignorant and barbarous nations"; (d) miracle claims of different religions undermine each other, since each tradition's miracles are evidence against the others' claims.

The four sub-arguments are uneven. Argument (a) is largely empirical; argument (b) appeals to general psychology; argument (c) has been criticized as ethnocentric; argument (d) raises a genuine logical question about religious-plurality of miracle claims.

The Underlying Logic: Balancing Probabilities

Hume's argument is, in modern terms, a probabilistic argument. Belief in a miracle requires that the probability of the testimony being true exceed the probability of the miracle being false. Since the latter is, on Hume's reading, fixed at near-zero by uniform experience, the former would need to be fixed at near-certainty — which testimony, given human fallibility, can never reach.

Three features of this logic are important for understanding the responses.

First, the argument depends on treating the base rate of miracles as essentially zero, established by experience. Critics will challenge this base-rate calculation.

Second, the argument depends on treating testimony as intrinsically weaker than direct uniform observation. Critics will challenge this as well.

Third, the argument treats miracle claims in isolation. The Bayesian framework allows considering background hypotheses (the existence of God, the prior probability of revelation) that affect the likelihood ratios. The framework's cumulative case enters here.

Major Responses: Western Tradition

Earman's Hume's Abject Failure

John Earman's Hume's Abject Failure: The Argument Against Miracles (2000) is the most influential recent reconstruction and critique. Earman, himself non-theist, argues that Hume's argument as commonly understood does not work. His central points:

Hume's argument confuses the probability of the testimony being true with the probability of the miracle being true. These can come apart: testimony may be more probable than the underlying event if the testimony's reliability is high.
Hume's "uniform experience" principle, taken strictly, would rule out belief in any sufficiently rare event, not just miracles. Many established scientific findings concern events whose individual instances are rare or unique.
The Bayesian analysis of testimony shows that multiple independent witnesses of moderate reliability can establish a low-prior-probability event, contrary to Hume's suggestion.

Earman's conclusion is not that miracles have been established but that Hume's argument fails to show they cannot be.

The McGrews' Bayesian Reconstruction

Timothy and Lydia McGrew, in The Cumulative Case for the Resurrection of Jesus (2009) and related work, develop a detailed Bayesian analysis of how testimony for a low-prior event can produce high posterior probability. They argue that the cumulative weight of independent corroborating testimony, when multiple witnesses each have moderate reliability, can overcome very low priors. The argument is contested but represents the most technically developed contemporary response.

Swinburne's Principle of Credulity

Richard Swinburne, especially in The Existence of God (1979, 2nd ed. 2004) and The Resurrection of God Incarnate (2003), develops a broader epistemic framework. The "principle of credulity" holds that, absent special reason to doubt, things are as they appear to be. Applied to testimony: absent special reason to doubt testimony, testimony should be believed. This reverses Hume's default: the burden falls on the skeptic to defeat the testimony, not on the believer to overcome a presumption against it.

Swinburne also develops a probability framework in which background information (the existence of God, the prior probability that God would reveal himself) affects the likelihood of any specific miracle claim. The strategy is to argue that, given theism, the probability of revelation-type miracles is non-negligible.

Larmer's Specific Case

Robert Larmer's work, especially The Legitimacy of Miracle (2014), challenges Hume on the definition of miracle as "violation of natural law." Larmer argues that natural laws are descriptive regularities, not exception-forbidding metaphysical principles. On this view, a miracle is not a violation of natural law but an event produced by an agent whose causal contribution does not enter the description that natural laws codify. The definitional move substantially changes the argument's structure.

Major Responses: Islamic Tradition

Although Hume's argument is post-classical, the Islamic tradition had developed extensive defenses of miracle testimony in response to the philosophical critiques of Greek rationalism, in dialogue with internal challenges from the falāsifa, and in response to Christian-Muslim polemics. Three figures are particularly relevant.

al-Bāqillānī (d. 1013), in al-Bayān and Iʿjāz al- Qurʾan, develops the doctrine that the Qurʾan's iʿjāz (its literary inimitability) is itself a continuing miracle, available to direct present examination rather than depending on past testimony. This sidesteps Hume's testimonial problem by anchoring the miracle claim in a present text. See Maslik 6 articles for development.

al-Ghazālī, in Tahāfut al-Falāsifa and especially in his treatment of causation, argues against the necessary connection of natural causes. If natural laws are God's customary action rather than necessary metaphysical structure, the conceptual barrier to miracles is dissolved at its root. Ghazālī's occasionalism does what Larmer's definitional move does in modern philosophy: it shows that miracles are not "violations" of anything.

Ibn Rushd, defending the falāsifa, articulated a more measured position in Tahāfut al-Tahāfut: the regularities of nature are real but not absolute; God's customary action can be suspended for specific purposes; the question is evidential, not metaphysical. Ibn Rushd's position anticipates significant elements of the modern Earman position without sharing its non-theistic frame.

The Framework's Engagement

The framework engages Hume's argument seriously, with three observations.

First, Hume's argument is strongest when miracle claims are treated in isolation. The cumulative-case approach the framework develops does not consider individual miracles in isolation; it considers the entire prophetic phenomenon against the background of the philosophical, cosmic, human, innate-religious, and textual evidence (Masāliks 1-4 and 6). Within the cumulative case, the prior probability of revelation events is not the bare-zero of Humean uniform experience but conditioned by the broader case.

Second, Hume's argument is weakest where the framework works. The framework's central evidential mark for the Qurʾan's prophetic origin is not a past miracle requiring testimony but a present text available to direct examination — the iʿjāz. The Bāqillānī move neutralizes the Humean testimonial problem at its source. See theories-of-ijaz.

Third, Hume's fourth sub-argument (mutual undermining of religious miracles) is the most difficult and demands the most careful response. The framework's approach is not to claim that Islamic miracles refute other religions' miracles but to identify what features specifically support the Qurʾanic claim (see four-marks-of-prophecy and the Maslik 6 articles).

The framework does not claim Hume's argument is decisively refuted. It claims that the argument's force is substantially reduced once the cumulative case is in view, and that the contemporary critical literature (Earman, the McGrews, Swinburne) has identified serious problems with Hume's formulation.

What This Article Can and Cannot Establish

This article does not claim to resolve the miracles debate. It contributes:

A careful presentation of Hume's argument, the strongest version (rather than the popular caricature).
A map of the major responses and where they cut against Hume's formulation.
The framework's specific point of engagement: that the Humean argument is at its strongest against isolated miracle testimony, and at its weakest against the iʿjāz- based case of the Qurʾan (where a present text replaces past testimony).

Connections to Other Masalik

Maslik 5 (this maslik): Hume's argument concerns the possibility of revelation and the evidence for it. See also possibility-of-revelation and four-marks-of- prophecy.
Maslik 6 (Textual): the framework's strongest response is to relocate the evidential center from past testimony to present text (iʿjāz). See quranic- inimitability and theories-of-ijaz.
Maslik 1 (Philosophical & Metaphysical): the background prior probability of revelation depends on arguments for theism from Maslik 1.

Key Distinctions

Hume's definitional challenge ("miracle = violation of natural law") vs. alternative definitions (Larmer, Ghazālī)
Uniform experience vs. Bayesian background information
Isolated miracle claim vs. cumulative case
Past testimony vs. present text (the iʿjāz relocation)
Hume's dismissive psychology of testimony vs. rigorous Bayesian analysis of testimony (Earman, McGrew)
Mutual undermining of religious miracles (Hume's fourth point) vs. specific evidential features of any one claim

Major Proponents (of Hume's argument or its descendants)

David Hume — Enquiry Concerning Human Understanding, Section X (1748)
Bertrand Russell — Why I Am Not a Christian (1927) and related essays
Antony Flew — early career writings against miracles (Flew later abandoned this position)
J. L. Mackie — The Miracle of Theism (1982); the most sophisticated mid-twentieth-century development of the Humean line

Major Critics (defending the rationality of belief in

miracles)

John Earman — Hume's Abject Failure (2000)
Timothy and Lydia McGrew — Bayesian reconstruction of cumulative testimonial evidence
Richard Swinburne — The Existence of God (1979/2004); The Resurrection of God Incarnate (2003)
Robert Larmer — The Legitimacy of Miracle (2014)
C.S. Lewis — Miracles (1947); accessible but philosophically substantive defense
al-Bāqillānī, al-Ghazālī, Ibn Rushd — classical Islamic anticipations