Belief in supernatural miracles is a prerequisite to believing in a number of religious faith traditions. David Hume has presented an argument against the rationality of believing in miracles. His argument goes like this:

1. A miracle is “a violation of a law of nature” and “contrary to experience.”
2. If so, then the prior probability* of a miracle is very low.
3. Therefore, the prior probability of a miracle is very low. (1,2 modus ponens)
4. If the prior probability of a miracle is very low, then we should only believe in the miracle if the evidence for it is extremely strong.
5. So, we should only believe in a miracle if the evidence for it is extremely strong. (3,4 modus ponens)
6. The evidence for a miracle is never extremely strong.
7. So, we should never believe in any miracle. (5,6 modus tollens)

*prior probability is the probability of some event before evidence is taken into account.

I think there is more than one premise in this argument in which we can find fault, but for now, I will focus on premise 4. There may be an intuitive appeal to the idea that events that happen very rarely require extraordinary evidence, but I think we can jettison this intuition with some simple counterexamples.

On January 8th, 2022 the Dallas Cowboys defeated the Philadelphia Eagles with a score of 51-26. This is a fact that I am reporting to you. I can also tell you that that is the only game in the history of the NFL to end in that exact score. Considering that the NFL has existed for almost 100 years now, this means that the prior probability of any game ending in a score of 51-26 is exceptionally low. However, it seems like it is reasonable for you to believe that it happened merely because I told you it did. I am some faceless person on the internet who you have no reason to trust. My word should count as very little evidence. Despite this, it is still reasonable for you to believe what I have said about the Cowboys and Eagles game. After all, why would I lie about that? This example shows that some events, even with a very low prior probability, can be reasonably believed in based on little evidence, let alone extremely strong evidence. For this reason, I find fault in premise 4 of Hume’s argument as I have outlined it.