Apologetics · FaithLogic
Are Miracles Possible? — Hume’s Argument and Why It Fails
Before any evidence for the resurrection can be weighed, one objection has to be cleared: that miracles are impossible in principle, so no amount of testimony could ever establish one. This is David Hume's argument, from 1748, and it is still the background assumption behind most dismissals of the resurrection. It is also, on close reading, one of the weaker famous arguments in philosophy — and some of its sharpest critics are not Christians.
Hume’s Argument
"Of Miracles" (Enquiry Concerning Human Understanding, Section X)
Hume's reasoning:
1. A miracle is a violation of a law of nature.
2. Laws of nature are established by "firm and unalterable" uniform experience.
3. So the evidence against any miracle is, by definition, as complete as evidence can be.
4. The testimony for a miracle is always weaker than that.
5. 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."
In practice Hume treats that last condition as never met, and adds practical arguments: miracle reports cluster among the "ignorant and barbarous," witnesses are unreliable, and rival religions' miracles cancel each other out.
1. A miracle is a violation of a law of nature.
2. Laws of nature are established by "firm and unalterable" uniform experience.
3. So the evidence against any miracle is, by definition, as complete as evidence can be.
4. The testimony for a miracle is always weaker than that.
5. 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."
In practice Hume treats that last condition as never met, and adds practical arguments: miracle reports cluster among the "ignorant and barbarous," witnesses are unreliable, and rival religions' miracles cancel each other out.
Why It Fails
1. It is circular
Hume defines a "law of nature" as what uniform experience establishes, and then says the experience is uniform. But whether experience is uniform is exactly the question — if a genuine miracle has ever occurred, experience is not uniform. To assume it is, in order to rule out the miracle, is to assume the conclusion. As critics have put it: Hume has smuggled "miracles never happen" into the definition of the evidence, then discovered it there.
2. The probability reasoning is broken (John Earman)
John Earman — a philosopher of science, not a Christian — wrote a book-length analysis titled Hume's Abject Failure (2000). His verdict: Hume's argument, taken as a piece of probabilistic reasoning, "is a shambles."
Bayes' theorem shows the flaw. The probability that a report is true depends on the prior probability of the event and on how likely the evidence would be if the event had not happened. If several independent witnesses, with nothing to gain and much to lose, all report the same surprising event, the probability that they are all mistaken or lying can drop below the low prior of the event itself. Multiple independent testimony can rationally establish an event that is very improbable in advance — this is routine in courts, forensics, and history. Hume's blanket "testimony can never suffice" is simply not how evidence works.
Bayes' theorem shows the flaw. The probability that a report is true depends on the prior probability of the event and on how likely the evidence would be if the event had not happened. If several independent witnesses, with nothing to gain and much to lose, all report the same surprising event, the probability that they are all mistaken or lying can drop below the low prior of the event itself. Multiple independent testimony can rationally establish an event that is very improbable in advance — this is routine in courts, forensics, and history. Hume's blanket "testimony can never suffice" is simply not how evidence works.
3. It assumes there is no God
Hume's "violation of a law of nature" only makes sense if nature is a closed system. If a Creator exists (see the arguments for God), then a miracle is not a violation — it is the author of the system acting within it, the way a novelist can write an event into a story that its internal rules wouldn't produce. The prior probability of a miracle is not near-zero if (a) God exists and (b) God has a discernible reason to act — for instance, to vindicate a messenger. The whole argument only bites for someone who already assumes naturalism, in which case the resurrection was never going to be on the table anyway.
4. "Science has disproved miracles" — C. S. Lewis's reply
Science describes what nature does when left to itself. A miracle is the claim that in one instance nature was not left to itself. Cataloguing the regular course of events cannot, even in principle, show that no agent ever acts on it — any more than studying the normal rules of billiards tells you whether a hand ever reaches in and moves a ball. Lewis (Miracles, 1947): the question of whether miracles occur is a philosophical and historical one, not one that laboratory science is equipped to settle.
5. Hume’s "more miraculous" test actually favours the resurrection
Hume says to believe a miracle only if its falsehood would be "more miraculous" than the event. Apply that to the resurrection. If it did not happen, you must explain: the empty tomb that Jesus' opponents never refuted by producing a body; multiple independent appearance traditions; the transformation of terrified deserters into people who died for the claim; the conversion of the persecutor Paul and the skeptic James; and the sudden existence of a movement, in Jerusalem, weeks later, with resurrection at its centre. A naturalistic account has to make all of that arise from a mistake or a lie that no one ever exposed and many died refusing to recant. Whether that is "less miraculous" is exactly the historical argument — which is the point: it is an argument to be had, not one Hume can close in advance.
The Honest Position
What the objection gets right, and what it doesn’t
Right: extraordinary claims do call for strong evidence. A single anonymous report of a miracle, centuries later, with obvious motive, is reasonably discounted — and most such reports are. Craig Keener's two-volume Miracles (2011) documents thousands of modern claims precisely to show they should be examined case by case, not dismissed by decree.
Wrong: the in-principle move — "no evidence could ever count" — is not defensible. It is circular, it misuses probability, and it presupposes the atheism it is meant to support. Once "miracles are impossible" is off the table, the resurrection becomes what it should always have been: a historical question, to be settled by the evidence.
Wrong: the in-principle move — "no evidence could ever count" — is not defensible. It is circular, it misuses probability, and it presupposes the atheism it is meant to support. Once "miracles are impossible" is off the table, the resurrection becomes what it should always have been: a historical question, to be settled by the evidence.
💡 Are Miracles Possible — Reasonable Summary"Hume argued that a miracle is a violation of a law of nature, that laws rest on uniform experience, and so that testimony can never outweigh the evidence against a miracle. The argument fails four ways. It is circular — whether experience is 'uniform' is the very question. Its probability reasoning is broken: the philosopher John Earman (not a Christian) titled his analysis 'Hume's Abject Failure' and showed by Bayes' theorem that multiple independent testimony can rationally establish a very improbable event, as it does in courts and history every day. It assumes naturalism — if a Creator exists, a miracle is not a 'violation' but the author acting, and its prior probability is not near-zero if God has reason to act. And 'science disproves miracles' confuses describing nature's regular course with proving no agent ever acts on it (Lewis). Hume's own 'which is more miraculous' test, applied to the resurrection, turns into the historical argument itself: the empty tomb never refuted, the appearances, the martyred witnesses, the converted enemies. The objection rightly asks for strong evidence; it wrongly claims no evidence could ever count. Remove that, and the resurrection is a historical question again."
Related Sections
• Does God Exist? — if a Creator exists, a miracle is not a "violation."
• The Resurrection — The Most Verified Event in Ancient History — the historical case, once "impossible" is off the table.
• The Resurrection — Four Minimal Facts Every Scholar Accepts.
• Prophecy and How Typology Works — the other main line of verifiable divine action.
• The Resurrection — The Most Verified Event in Ancient History — the historical case, once "impossible" is off the table.
• The Resurrection — Four Minimal Facts Every Scholar Accepts.
• Prophecy and How Typology Works — the other main line of verifiable divine action.
This is part of the FaithLogic study library — the biblical, historical and manuscript case that Jesus is God, with every claim linked to its source. Open the interactive version for search, cross-references, read-aloud, and the full library.