The Existence of God: an argument for God
Can miracles or history show that God exists?
The argument from miracles holds that good testimony can show that events beyond nature's power have occurred, and that God best explains them. Richard Swinburne and Timothy McGrew defend it. David Hume argued that such testimony must outweigh the uniform experience behind nature's laws, and that none has; other critics doubt an unexplained event would point to God.
- 15 minute read
- Last reviewed:
The argument, step by step
Here the argument's defenders are writers who hold that testimony can show a miracle occurred and that a miracle would point to God, such as Richard Swinburne; its critics include David Hume and those who defend his case, and writers such as Antony Flew and Christine Overall who doubt that a miracle would point to God. Points that answer Hume sit in the defenders' column, whoever makes them.
The argument, as its defenders state it:
- Premise 1There is good evidence, usually the testimony of witnesses, that events have occurred that nature alone could not produce.
- Premise 2If such an event occurred in a religious setting, God probably caused it.
- ConclusionTherefore, God probably exists.
1Premise 1There is good evidence, usually the testimony of witnesses, that events have occurred that nature alone could not produce.
What both sides acceptWriters define a miracle in more than one way. Hume called it a violation of the laws of nature; Augustine held that a miracle is contrary not to nature but only to what we know of nature; and Timothy McGrew and Robert Larmer, the authors of the Stanford Encyclopedia's entry, who themselves defend arguments from miracles, define it as an event that exceeds what nature can produce. The question at this step is whether testimony could ever give good reason to believe that such an event occurred.12
For the argument
For the argumentEnough independent witnesses could make a miracle probable
Charles Babbage argued in 1837 that if independent witnesses who tell the truth more often than not can be found, enough of them could always outweigh the improbability of a miracle. John Earman argues the same on grounds of probability in Hume's Abject Failure (2000). If Hume meant that testimony could never establish a miracle, writes David Corner, the author of the Internet Encyclopedia's entry, he would appear to be wrong.21
Reply, against the argumentReal witnesses are rarely independent
Arif Ahmed argues, in a 2015 paper, that this reply assumes the witnesses' reports are independent of one another, an assumption that real cases plausibly always violate. Corner adds that the number of witnesses needed might be very large, and that one writer's report that a crowd saw a miracle is not the testimony of the crowd itself.21
Reply, for the argumentAhmed's own conclusion is limited
McGrew and Larmer reply that Ahmed grants, later in the same paper, that a similar case can be made under some conditions even when witnesses are not independent, and that he concludes his arguments do not supply the "everlasting check to all kinds of superstitious delusion" that Hume hoped for.2
For the argumentHume's standard would make history and science absurd
In the nineteenth century, satires applied Hume's caution about extraordinary testimony to secular history; the best known, by Richard Whately (1819), "establishes" that Napoleon never existed. Their point, as McGrew and Larmer put it, is that Hume and his followers turn a sensible demand for more evidence into a barrier no other field of history could use without absurdity. Corner raises a similar worry for science: reports of giant squid were once dismissed, yet the animals were found.21
Reply, against the argumentA marvel is not a miracle
Hume distinguished a marvel, an event unlike anything in one's experience, from a miracle, an event contrary to it. Someone from a hot country who doubts reports of ice reasons correctly, he held, though what is reported is only a marvel. Corner writes that Hume needs this distinction so that his case against miracles does not also block the progress of science, and that whether he succeeds is a matter of some controversy.1
For the argumentHume's argument assumes its conclusion
C. S. Lewis and other critics of Hume argue that he begs the question: to judge a report by ordinary experience is to assume that nature ran its usual course, which is to assume that no miracle occurred. Some also hold that Hume treats the probability of a miracle as zero from the start.13
Reply, against the argumentThe charge cuts both ways
Corner answers that asking a skeptic to set aside ordinary experience begs the question in the other direction, since experience fails to apply only if a miracle in fact occurred. Robert Fogelin and others answer that Hume asks whether it can be rational to believe miracle reports, not whether miracles happen. Hume grants that a miracle is conceivable, and that an extraordinary event, such as eight days of darkness over the whole earth, could be believed on strong enough testimony.13
Against the argument
Against the argumentTestimony must outweigh uniform experience
David Hume argued in 1748 that a wise person proportions belief to the evidence. The laws of nature rest on uniform experience, so the evidence against any miracle is as strong as experience can give. Testimony could establish a miracle, he held, only if its falsehood would be more miraculous than the event it reports.42
Reply, for the argumentDifferent from experience, not contrary to it
Richard Price answered in 1777 that a miracle is better described as an event different from experience than as one contrary to it. On this view, as McGrew and Larmer explain it, experience would tell against a reported miracle as strongly as Hume supposes only if everyone had watched the very place and time and testified that nothing happened.2
Reply, against the argumentHume sets out how high the bar is
Some of Hume's defenders read his argument as a measure of the evidence a miracle report needs. William Vanderburgh, echoing Fogelin, holds that Hume shows how very high the bar is that a report must clear, and Peter Millican offers a sympathetic reconstruction of Hume's case within his wider philosophy.2
Against the argumentNo miracle report has met the standard
In the second part of his essay, Hume held that no miracle has been attested by enough witnesses of good sense, learning, and integrity; that people love to tell and hear of wonders, all the more in the cause of religion; and that such stories flourish chiefly among people with little learning. Paul Russell and Anders Kraal, writing on Hume's view of religion, say his main aim was to discredit the historical miracle reports that are supposed to give the major established religions their authority.43
Reply, for the argumentHume's conditions cut both ways
John Leland (1755) argued that Hume's conditions are not necessary, and that witnesses of high standing could more easily spread an imposture. George Campbell (1762) answered that religious zeal can make a person resist a rival religion's miracle as readily as accept one's own. McGrew and Larmer add that miracle stories abounded in Hume's own century, and that the scientists Isaac Newton and Robert Boyle were well-known defenders of the miracles Christianity reports.2
Against the argumentRival religions' miracles cancel each other out
Hume argued that in religion whatever is different is contrary, so the miracles reported for each religion count against the others. He described cures attributed to the emperor Vespasian, and cures reported in the early 1700s at the tomb of the Abbe Paris, a Jansenist (a follower of the Dutch theologian Cornelius Jansen), as in many ways as well attested as one could wish, and left unstated the conclusion that his readers should reject the New Testament's miracles too.425
Reply, for the argumentA new faith is harder to found on false miracles
William Adams answered in 1767 that it is easy to establish false miracles with the help of a false religion, but hard to establish a false religion with the help of false miracles. On this view, stories told to an established, friendly audience are not parallel to those reported at a faith's founding. Hume's critics also argued that his accounts of these cures distorted the historical sources.2
Against the argumentHistorians cannot conclude that a miracle happened
F. H. Bradley (1874) and Ernst Troeltsch (1913), with his principle of analogy, made something like Hume's principle a rule of critical history, and Flew (1966) argues that historians can read the past only by assuming it ran by the same regularities as the present. Bart Ehrman (2003) argues that historians can establish only what probably happened, and a miracle is by definition the least probable of events.2
Reply, for the argumentProbability before and after the evidence
McGrew and Larmer reply that Ehrman's argument confuses how probable a miracle is apart from the evidence with how probable it is given the evidence. Believers in miracles, they add, generally do not deny that nature has dependable regularities, since a miracle is by its nature an exception to them.2
2Premise 2If such an event occurred in a religious setting, God probably caused it.
For the argument
For the argumentIn its religious setting, a miracle confirms a message
Historically, many writers in the debate have granted that when an event's religious meaning is plain, and the teaching it confirms is not otherwise objectionable, God worked it. Joseph Butler (1736) called miracles one of the "direct and fundamental proofs" of Christianity.2
Reply, against the argumentTestimony can show a fact, not its cause
Baden Powell (1859) argued that testimony can establish an extraordinary fact, but not that God caused it, which is a matter of opinion resting on other grounds. Gregory Dawes (2009) argues that it is hard to meet the standard needed to attribute a particular event to God's action, though he does not treat this as an absolute barrier.2
Reply, for the argumentThe details of a case can rule out other causes
McGrew and Larmer grant that testimony is not the source of evidence for a supernatural cause, but answer that the description of an event, and still more its setting, might make natural explanations non-starters. Whether it does depends on the details of each case.2
Against the argument
Against the argumentA perfect God would not break his own laws
Baruch Spinoza (1670) held that God's will is the same as the laws of nature, so that a miracle is impossible. Voltaire (1764) argued that a God of infinite wisdom would have made the world as good as it could be from the start, so that a miracle would show a lack of foresight or of power.2
Reply, for the argumentMiracles are how a revelation could be confirmed
William Paley answered that a revelation could be made in no other way anyone can conceive: only by suspending regularities no human can break could God put an unmistakable seal on a message. McGrew and Larmer add that Spinoza's argument is unlikely to persuade anyone who does not already identify God's will with nature's laws.2
Against the argumentAn unexplained event may have an unknown natural cause
Flew argues that a miracle can serve as evidence for a revelation only if it can be identified as a miracle without appealing to that revelation. Yet an event whose natural cause is not yet known looks no different from one with a supernatural cause, and Corner adds that it might also have no cause at all.1
Reply, for the argumentGod's action is a personal explanation
Defenders answer that it is unreasonable to reject a supernatural explanation in the hope that a natural one will turn up. Defenders such as Swinburne treat God's action as a personal explanation, which explains an event by an agent's intentions, as we explain what people do.1
Reply, against the argumentAn appeal to God's will predicts nothing
Corner, drawing on Patrick Nowell-Smith (1955), answers that ordinary personal explanations yield predictions, while ascribing an event to God's will yields none. A naturalist is therefore justified, he argues, in holding out for a natural explanation, or in calling the event inexplicable.1
Reply, for the argumentA miracle may be a basic action
Corner himself suggests that a miracle may be a basic action of God, like the way a person simply moves a finger. On that view, ascribing a miracle to God makes no claim about its cause, so a believer need not show that it has no natural cause.1
Against the argumentA miracle would count against God
Overall (1985) argues that a miracle would be evidence against God: if order in nature points to God, a breach in that order points away; disputes over miracles hinder science and philosophy; and a God who did intervene would be obliged to do so more often and more evenhandedly.2
Reply, for the argumentEach of Overall's arguments has an answer
McGrew and Larmer reply that her first argument rests on a fallacy about probability, that her second is arguably false, since such disputes have not noticeably slowed science, and that her third tells only against a theology few believers in miracles hold. Larmer, one of the entry's authors, answered Overall in 1988, and the two continued the exchange in print, with replies from Overall in 1997 and 2003 and from Larmer in 2004.2
Against the argumentCalling on God makes a miracle no more likely
Hume held that ascribing a miracle to an almighty being does not make it any more probable, since we can know the attributes and actions of such a being only from our experience of the usual course of nature.4
3ConclusionTherefore, God probably exists.
What both sides acceptHow much a reported miracle could show depends on what else a person has reason to believe. McGrew and Larmer write that judging a miracle claim is sensitive to how probable it is that God exists, so it turns on the wider case for and against God. Corner suggests that whether miracle reports are credible turns on the larger question of whether to hold a supernaturalistic or a naturalistic view of the world.21
For the argument
For the argumentOne piece of a larger case
Swinburne has developed probability-based arguments for miracle claims within a wider case for God, the most elaborate such work, in McGrew and Larmer's judgment. Timothy and Lydia McGrew (2009) set out a more modest version, which aims to show only that the evidence adds considerably to the case for a miracle.2
Reply, against the argumentOnly the whole of the evidence counts
Graham Oppy (2006) objects that probability arguments of this kind are of no interest unless they rest on all the relevant evidence available.2
Reply, for the argumentScience weighs evidence the same way
McGrew and Larmer answer that the objection would count equally against comparing likelihoods in scientific reasoning, where such comparisons are everywhere, and that a successful modest argument may shift the burden of proof onto those who deny the miracle.2
Against the argument
Against the argumentMiracle reports cannot establish the view they presuppose
Corner argues that appeals to miracle reports will persuade only those already inclined to believe, perhaps because they are sympathetic to a supernaturalistic view of the world, and that reports whose credibility depends on presuming that view cannot establish it. A failure of the argument from miracles, he adds, would show at most that belief in miracles needs independent reasons for rejecting naturalism.1
Where does the argument come from?
Arguments from miracles were more popular in the eighteenth century than they are today, Corner writes, and the apologetic interest behind them, the wish to defend the rationality of belief in God, has strongly shaped the modern idea of a miracle.1 The principal argument against the credibility of miracle reports, McGrew and Larmer write, comes from Hume.2 Butler's The Analogy of Religion (1736) and Paley's A View of the Evidences of Christianity (1794) argued for Christianity in part from its miracles, and many of the points against such arguments had already been aired in the half century before Hume wrote.2
David Hume (1711 to 1776) published "Of Miracles" in 1748; it is Section X of An Enquiry Concerning Human Understanding.642 In it he wrote, in his eighteenth-century words: "A wise man, therefore, proportions his belief to the evidence."4 The essay, McGrew and Larmer write, provoked a tidal wave of responses, chief among them those of Leland (1755), John Douglas (1757), Campbell (1762), Adams (1767), and Price (1777).2 Its influence reached biblical scholarship: Bradley and Troeltsch made something like Hume's principle a rule of critical history, and many later studies of Jesus and the New Testament accept it, explicitly or implicitly.2
Much recent work uses probability theory. Swinburne, beginning with The Concept of a Miracle (1970), pioneered its use on questions in the philosophy of religion, and Earman's Hume's Abject Failure (2000) attacked Hume on the same grounds.2 Fogelin's A Defense of Hume on Miracles (2003) and Millican's work answered Earman.2 McGrew and Larmer note that Hume's critique has itself come under heavy fire, and is now viewed in some quarters as needing defense.2
This page covers the general question of what testimony to a miracle can show. Each faith's own case for its truth, which for some faiths includes miracles such as Jesus's resurrection or, in Islamic teaching, the Quran's inimitability, is set out on its own page: the strongest case for Buddhism, the strongest case for Christianity, the strongest case for Hinduism, the strongest case for Islam, the strongest case for Judaism, and the strongest case for the Latter-day Saint faith.27
Across traditions
Several faiths have their own teaching on miracles and extraordinary powers and on what they can show, some of it centuries older than Hume's essay.78
In the Long Discourses, a collection of early Buddhist teachings, the householder Kevaddha asks the Buddha to have a monk display psychic powers, so that the town of Nalanda will grow even more devoted. The Buddha twice declines, and at the third request names three kinds of demonstration: psychic power, revealing another's thoughts, and instruction. A person without faith, he explains, would say that the first two were done by spells, and he is repelled by such displays. Instruction teaches others how to think and focus the mind, what to give up, and what to achieve (Long Discourses 11).9
Much of the debate on this page took shape among Christian writers, who also disagreed among themselves. Augustine (354 to 430) held that miracles are made possible by hidden potentials God placed in nature, and Thomas Aquinas (about 1225 to 1274), the Catholic theologian, that a miracle goes beyond the order usually seen in nature without being contrary to nature in any absolute sense.11011 Protestant writers, among them Charles Leslie (1697) and Douglas (1757), offered tests for telling the miracles of the apostles from later miracles reported in Catholic tradition, and Hume, McGrew and Larmer note, expected his Protestant readers to reject the later cures he described.2
Patanjali's Yoga Sutras, which traditional Hindu schools accept as the authoritative source on yoga, a path of meditation, deal in their third chapter mainly with super-normal powers that can come to a practitioner whose mind is in extreme states of concentration.12 Edwin Bryant, who wrote the Internet Encyclopedia's entry on the text, writes that many in ancient India seem to have taken up yoga-like practices in quest of such powers, and that the chapter can be read as Patanjali's warning against being side-tracked from the real goal of yoga.12
In Islamic thought, a prophet's miracle is a mu'jiza, an Arabic word for that which shows the impotence of others: a challenge to do the like, which others are taken to be unable to meet. Many theologians and philosophers alike treated such miracles as a decisive sign of a true prophet, and the Quran's own inimitability is often cited as the most convincing proof of Muhammad's prophethood.7 They disagreed about how miracles happen. Avicenna (Ibn Sina, about 980 to 1037) explained a prophet's miracles naturally, through the prophet's exceptional mind acting on other bodies.78
Al-Ghazali (about 1056 to 1111), in the seventeenth discussion of The Incoherence of the Philosophers, set out to make the philosophers grant that the miracles the Quran reports are possible.8 Either God, who creates events by habit, sets the habit aside, he argued, or the miracles may have natural causes that human science does not yet know.8 He did not hold that miracles should be the main ground of belief in the Quran: the true proof, he held, is the purifying effect on the heart of careful study of the Quran and the traditions about Muhammad.7
In Judaism, Judah Halevi (died 1141) has the Jewish sage of his Kuzari hold that the miracles of the Exodus, such as the plagues of Egypt and the parting of the sea, were public events that needed no proof by reason for those who saw them.13 Later generations know of them, the sage holds, through "concurrent tradition": a report passed down through many separate chains of transmitters, whose number and separation rule out distortion by conspiracy. Ehud Krinis, who wrote the Stanford entry on Halevi, notes that in Muslim culture, knowledge passed down this way was counted the most reliable of all.13
Latter-day Saint scripture speaks directly to signs and faith. A revelation in the Doctrine and Covenants teaches that faith does not come by signs, but that signs follow those who believe, and come by faith and by God's will rather than as people please (Doctrine and Covenants 63:9-10).14 The Book of Mormon teaches that God has not ceased to be a God of miracles, and that he ceases to work them among people when they fall into unbelief (Mormon 9:19-20).14
Where does the debate stand?
Philosophers remain divided over Hume's argument and over what a miracle could show. Swinburne and the McGrews defend arguments from miracles, Larmer argues that a wide range of objections in principle to believing miracle reports all fail, and Earman rejects Hume's reasoning on grounds of probability.2 Fogelin, Millican, and Vanderburgh defend Hume's case against such critics, and Ahmed argues that the many-witnesses reply to Hume rests on an assumption that real cases plausibly always violate.2 McGrew and Larmer conclude that evidence for a miracle cannot by itself establish or rule out God's existence, though it may be one piece of a larger case.2 Corner holds that miracle reports cannot by themselves establish a supernaturalistic view of the world, but that belief in miracles may still be reasonable for someone with independent reasons to reject naturalism.1
Common questions
What is Hume's argument against miracles?
David Hume argued in 1748 that a wise person proportions belief to the evidence. Since the laws of nature rest on uniform experience, testimony could establish a miracle only if its falsehood would be more miraculous than the event. He held that no actual miracle report has met that standard, and that rival religions' miracles cancel each other out.
Did Hume say that miracles are impossible?
Hume does not say that. He grants that a miracle is conceivable, and Robert Fogelin and others read him as asking whether it can be rational to believe miracle reports. An event such as eight days of darkness worldwide, he allowed, could be believed on strong enough testimony. Some critics hold that he treats a miracle's probability as zero.
Would a miracle prove that God exists?
Writers differ. Historically, many granted that God must have worked a miracle whose religious meaning is plain and whose message is not otherwise objectionable. Timothy McGrew and Robert Larmer hold that evidence for a miracle cannot by itself establish or rule out God's existence, and Antony Flew that an unexplained event may have an unknown natural cause.
Read next
Sources
- David Corner. Miracles. Internet Encyclopedia of Philosophy. Accessed 2026-11-20.
- Timothy McGrew and Robert Larmer. Miracles. Stanford Encyclopedia of Philosophy, 2024. Accessed 2026-11-20.
- Paul Russell and Anders Kraal. Hume on Religion. Stanford Encyclopedia of Philosophy, 2024. Accessed 2026-11-20.
- David Hume. An Enquiry Concerning Human Understanding, Section X: Of Miracles. Hume Texts Online (davidhume.org), 1748, Section X, Parts 1 and 2. Accessed 2026-11-20.
- David Simpson. Pascal, Blaise. Internet Encyclopedia of Philosophy. Accessed 2026-11-20.
- C. M. Lorkowski. David Hume: Religion. Internet Encyclopedia of Philosophy, opening section. Accessed 2026-11-20.
- Jon McGinnis and Rahim Acar. Arabic and Islamic Philosophy of Religion. Stanford Encyclopedia of Philosophy, 2023. Accessed 2026-11-20.
- Frank Griffel. al-Ghazali. Stanford Encyclopedia of Philosophy, 2026. Accessed 2026-11-20.
- Bhikkhu Sujato. Translations of the early Buddhist discourses. SuttaCentral, 2026. Accessed 2026-10-28.
- Dale Tuggy. Theories of Religious Diversity. Internet Encyclopedia of Philosophy. Accessed 2026-11-20.
- Robert Pasnau. Thomas Aquinas. Stanford Encyclopedia of Philosophy, 2022. Accessed 2026-11-20.
- Edwin Bryant. The Yoga Sutras of Patanjali. Internet Encyclopedia of Philosophy, opening section and section 1. Accessed 2026-11-20.
- Ehud Krinis. Judah Halevi. Stanford Encyclopedia of Philosophy, 2026, sections 1 and 6.2.1. Accessed 2026-10-28.
- The Book of Mormon (1920 edition), the Doctrine and Covenants and the Pearl of Great Price (1921 editions). The Church of Jesus Christ of Latter-day Saints, 1920. Accessed 2026-10-28.