Hume’s Critique of Miracles

Michael Gleghorn examines Hume’s influential critique of miracles and points out the major shortfalls in his argument. Hume’s first premise assumes that there could not be miracles and his second premise is based on his distaste for the societies that report miracles. As a Christian examining these arguments, we find little of value to convince us to reject a biblical worldview saying that God can and has intervened in natural history to perform miracles.


One of the most influential critiques of miracles ever written came from the pen of the skeptical Scottish philosopher David Hume. The title of the essay, “Of Miracles,” originally appeared in Hume’s larger work, An Inquiry Concerning Human Understanding, first published in 1748. This was the Age of Enlightenment, a time in which skepticism about miracles was becoming increasingly widespread among the educated elite.{1} So what were Hume’s arguments, and why have they been so influential in subsequent scholarly discussions of this topic?

download-podcastHume essentially “presents a two-pronged assault against miracles.”{2} He first argues that “a miracle is a violation of the laws of nature.” But since “a firm and unalterable experience has established these laws, the proof against a miracle,” he says, “is as entire as any argument from experience can possibly be imagined.”{3} In other words, given the regularity of the laws of nature, Hume contends that miracles are exceedingly improbable events. But this is not all. He also argues that since miracle reports typically occur among uneducated, barbarous peoples, they are inherently untrustworthy and, hence, unworthy of our belief.{4}

Now clearly, if Hume is correct, then this presents a real problem for Christianity. For Christianity is full of miracles. According to the New Testament, Jesus walked on water, calmed raging storms, healed diseases, exorcised demons, and brought the dead back to life! But if miracles are really as utterly improbable as Hume maintains, and if reports of miracles are completely lacking in credibility, then it would seem that the New Testament’s accounts of miracles are probably unreliable and that Christianity itself is almost certainly false!

So how compelling are Hume’s arguments? Should believers be quaking in their boots, fearful that their most cherished beliefs are a lie? Not at all! As philosopher of science John Earman observed in a scholarly critique of Hume’s arguments, Hume’s essay is not merely a failure; it is “an abject failure.” He continues, “Most of Hume’s considerations are unoriginal, warmed over versions of arguments that are found in the writings of predecessors and contemporaries. And the parts of ‘Of Miracles’ that set Hume apart do not stand up to scrutiny. Worse still, the essay reveals the weakness and the poverty of Hume’s own account of induction and probabilistic reasoning. And to cap it all off, the essay represents the kind of overreaching that gives philosophy a bad name.”{5} Now admittedly, these are strong words. But Earman argues his case quite forcefully and persuasively. And in the remainder of this article, I think the truth of his remarks will become increasingly evident.

Hume’s Argument from the Laws of Nature

What are we to say to Hume’s argument that “a miracle is a violation of the laws of nature” and that “the proof against a miracle…is as entire as any argument from experience can possibly be imagined”?

First, we might question whether miracles should be defined as violations of the laws of nature. According to Christian philosopher Bill Craig, “An examination of the chief competing schools of thought concerning the notion of a natural law…reveals that on each theory the concept of a violation of a natural law is incoherent and that miracles need not be so defined.”{6} Thus, we might object that Hume’s definition of a miracle is simply incoherent. But this is a debated point, so let’s instead turn our attention to a more pressing matter.

When Hume says that the laws of nature are established upon “a firm and unalterable experience,” is he claiming that the laws of nature are never violated? If so, then his argument begs the question, assuming the very thing that needs to be proved. It would be as if he argued this way:

• A miracle is a violation of the laws of nature.

• Experience teaches us that the laws of nature are never violated (i.e. that miracles never occur).

• Therefore, experience teaches us that miracles never occur.

Such an argument is clearly fallacious. Hume would be assuming “as a premise for his argument the very conclusion he intends to prove.”{7} But this is probably not what Hume intended.

As Earman observes, Hume’s view rather seems to go something like this: “When uniform experience supports” some lawlike regularity “that is contradicted by testimony,” then one must set “proof against proof,” and judge which of the two is more likely. The result of this new formulation, however, is that “uniform experience does not furnish a proof against a miracle in the sense of making the . . . probability of its occurrence flatly zero.”{8}

This is an important point. After all, there is a great deal of human testimony that solemnly affirms the occurrence of miracles. Thus, the only way that Hume can maintain that the uniform experience of mankind is against the occurrence of miracles is by assuming that all miracle reports are false. But this assumption, as we’ll see, is completely untenable when miraculous events are attested by numerous, independent witnesses.

Hume’s Argument Against the Reliability of Human Testimony

In Part II of “Of Miracles,” David Hume argues that there has never been the kind of testimony on behalf of miracles which would “amount to entire proof.”{9} He offers four reasons for this claim.{10}

First, no miracle on record has a sufficient number of intelligent witnesses, of good moral character, who testify to a miraculous event that occurred in public and in a civilized part of the world. Second, human beings love bizarre and fantastic tales, and this irrationally inclines them to accept such tales as true. Third, miracle reports are usually found among barbarous peoples. And finally, the miracle reports of different religions cancel each other out, thus making none of them effective for proving the truth of their doctrines.

What should we say in response to these arguments? While all of the points have merit, nevertheless, as Bill Craig observes, “these general considerations cannot be used to decide the historicity of any particular miracle.”{11} The only way to determine if a miracle has actually occurred is by carefully examining the evidence. How many witnesses were there? Are they known to be honest, or are they generally unreliable?

These questions are particularly important when one considers the cumulative power of independent witnesses for establishing the occurrence of some highly improbable event like a miracle. By “independent witnesses” I simply mean witnesses whose testimony to an event comes from firsthand experience and is not dependent on the testimony of others.

As Charles Babbage demonstrated in his Ninth Bridgewater Treatise, if one can find enough independent witnesses to a miraculous event, who tell the truth more often than not, then one can always show that the occurrence of the miracle is more probable than not.{12} Craig explains the matter this way: “If two witnesses are each 99% reliable, then the odds of their both independently testifying falsely to some event are only . . . one out of 10,000; the odds of three such witnesses being wrong is . . . one out of 1,000,000.” “In fact,” he says, “the cumulative power of independent witnesses is such that individually they could be unreliable more than 50% of the time and yet their testimony combine to make an event of apparently enormous improbability quite probable in light of their testimony.”{13}

So while Hume’s arguments should make us cautious, they cannot prevent human testimony from plausibly establishing the occurrence of miracles. And the only way to determine if the testimony is plausible is to carefully examine the evidence.

Hume and Probability Theory (Part 1)

Hume argues that since miracles run contrary to man’s uniform experience of the laws of nature, no testimony can establish that a miracle has occurred unless “its falsehood would be more miraculous than the fact which it endeavors to establish.”{14} Although Hume makes it sound as though establishing one miracle would require an even greater miracle, all his statement really amounts to, as John Earman rightly notes, is that no testimony is good enough to establish that a miracle has occurred unless it’s sufficient to make the occurrence of the miracle more probable than not.{15}

But in Hume’s view this is virtually impossible. No testimony is really ever sufficient to establish that a miracle has occurred. And this is problematic. For it can be perfectly reasonable to accept a highly improbable event on the basis of human testimony. In fact, we do it all the time.

Suppose the evening news announces that the number picked in the lottery was 8253652. As Craig observes, “this is a report of an extraordinarily improbable event, one out of several million.”{16} If we applied Hume’s principle to such a case, it would be irrational for us to believe that such a highly improbable event had actually occurred. So something is clearly wrong with this principle. But what?

The problem, says Craig, is that Hume has not considered all of the relevant probabilities. For although it might be highly improbable that just this number should have been chosen out of all the possible numbers that could have been chosen, nevertheless one must also consider the probability that the evening news would have reported just this number if that number had not been chosen. And this probability is “incredibly small,” for the newscasters would have no reason to report just this number unless it had, in fact, been chosen!{17}

So how does this relate to the question of miracles? When it comes to assessing the testimony for a miracle, we cannot simply consider the likelihood of the event in light of our general knowledge of the world.{18} This was Hume’s mistake. Instead, we must also consider how likely it would be, if the miracle had not occurred, that we would have just the testimony and evidence that we have.{19} And if it is highly unlikely that we would have just this evidence if the miracle had not occurred, then it may actually be highly probable that the miracle did, in fact, occur. Even if a miracle is highly improbable when judged against our general knowledge, it may still turn out to be highly probable once all the specific testimony and evidence for the miracle is taken into account.{20}

Hume and Probability Theory (Part 2)

There’s still another problem with Hume’s critique, namely, that he never actually establishes that a miracle is highly improbable in light of our general knowledge of the world. He simply assumes that this is so. But the problem with this becomes evident when one reflects upon the fact that, for the Christian, part of what’s included in our “general knowledge of the world” is the belief that God exists. What’s more, as believers we have at our disposal a whole arsenal of arguments which, we contend, make it far more plausible than not that this belief is really true.

But notice how this will influence our estimation of the probability of miracles. If belief in God is part of our general knowledge of the world, then miracles will be judged to at least be possible. For if an all-powerful God exists, then He is certainly capable of intervening in the natural world to bring about events which would never have occurred had nature been left to itself. In other words, if God exists, then He can bring about miracles! Thus, as Bill Craig observes, whether or not a miracle is considered highly improbable relative to our general knowledge of the world is largely going to depend on whether or not we believe in God. So the question of God’s existence is highly relevant when it comes to assessing the probability of miracle claims.{21} While those who believe in God may still be skeptical of most miracle reports, they will nonetheless be open to the possibility of miracles, and they will be willing to examine the evidence of such reports on a case-by-case basis.

To conclude, although Hume’s critique of miracles is one of the most influential ever written, it really doesn’t stand up well under scrutiny. Indeed, John Earman concludes his devastating critique of Hume’s arguments by noting his astonishment at how well posterity has treated Hume’s essay, “given how completely the confection collapses under a little probing.”{22} Although Hume was doubtless a brilliant man, his critique of miracles is simply unconvincing.


1. William Lane Craig, Reasonable Faith: Christian Truth and Apologetics, 3rd ed. (Wheaton, IL: Crossway Books, 2008), 248.

2. Ibid., 250.

3. David Hume, “Of Miracles,” in An Inquiry Concerning Human Understanding, ed. Charles W. Hendel (Indianapolis: Bobbs-Merrill, 1955), 122.

4. Ibid. See Hume’s discussion in Part II of his essay.

5. John Earman, Hume’s Abject Failure: The Argument against Miracles (Oxford: Oxford University Press, 2000), 3.

6. Craig, Reasonable Faith, 261.

7. Irving M. Copi, Introduction to Logic, 2d ed. (New York: Macmillan, 1961), 65.

8. Earman, Hume’s Abject Failure, 32

9. Hume, “Of Miracles,” 124.

10. See ibid., 124-41.

11. Craig, Reasonable Faith, 277.

12. This sentence is a paraphrase of a statement from Babbage’s treatise cited in Earman, Hume’s Abject Failure, 54.

13. Craig, Reasonable Faith, 272, n. 26.

14. Hume, “Of Miracles,” 122-23.

15. Earman, Hume’s Abject Failure, 41.

16. Craig, Reasonable Faith, 270.

17. Ibid., 271.

18. Jason Rennie, “Epistemology and the Resurrection: An Interview with William Lane Craig,” in Sci-Phi Show Outcasts, 2006, available in the “Interviews” section at (note: this page is accessible by members only. We urge you to register free of charge to access this and many excellent resources. The link is down the page underneath the “Closer to Truth” links.)

19. Craig, Reasonable Faith, 270.

20. Rennie, “Epistemology and the Resurrection.”

21. Ibid. See also the discussion in Craig, Reasonable Faith, 274-76.

22. Earman, Hume’s Abject Failure, 71.

© 2010 Probe Ministries