Is It Possible To Prove That Miracles Have Happened?

New Testament scholar Bart Ehrman likely is best known as the author of Misquoting Jesus, a book that attempts to show that the Gospels are not reliable, and that the supernatural events were added long after the original accounts were recorded. I have read his book and have also heard some of Erhman’s arguments in one of his debates with Mike Licona, author of The Resurrection of Jesus: A New Historiographical Approach. During this debate, Erhman contended that it is impossible to prove that miracles have happened in history. Of course, I believe the argument is flawed, and I thought the error was interesting enough to write about it.

Ehrman’s claim is that the general methods we employ to determine what has happened in history can never be used to show that miracles occurred. His argument was something like the following:

Because the events in history cannot be observed directly or re-created, all of our knowledge of historical events is probabilistic by definition. In other words, if there is evidence that points to a particular event, the stronger and more numerous the evidence, the more likely it is the event happened. However, because the evidence is merely cumulative, we can never claim to know with 100% certainty that a particular event happened; we can only know beyond a reasonable doubt, or in some cases with even less confidence. I agree so far.

Miracles, Erhman continues, are by definition not only improbable events, but the least probable event in any given situation. This definition could be debated, as ultimately it is circular reasoning. He is assuming that a miracle, a supernatural event, cannot happen because he “knows” that supernatural events cannot happen. If there is a God who intervenes in the world, however, it would not make sense to declare miracles improbable at all. However, for the sake of his argument, let’s continue to assume that a miracle is the least probable event in any situation.

Finally, Erhman declared that if the historical method is used to show whatever was the most probable event in a given case, and a miracle is by definition the least probable event, the historical method never can be used to show that a miracle occurred (because a miracle would never be the most probable event). Note that he is not saying miracles have not occurred; he is merely saying that it would be impossible to know beyond a reasonable doubt that a miracle occurred in history.

It is with this conclusion that I disagree. While on the surface his argument may sound reasonable, even if you don’t know the problem with his logic, I suspect your intuition tells you something is wrong.

The error in his logic turns out to be a subtle difference in what he means by the word “probable” in the one premise versus the same word in the conclusion. We must ask on what basis we determine what is most probable in each case.

To show what I mean, suppose I asked you how probable it is that an airplane flying overhead would crash into a car driving on the highway. No doubt we all would agree this is extremely unlikely. But does this mean that such an event could never be shown to have happened? Imagine if investigators found the smashed remains of a car under a pile of airplane parts. Suddenly this improbable event appears to be much more probable.

In one sense, this is an improbable event, and in another sense, it is very probable. Both conclusions are correct. In the first case, we were talking about the overall probability that this particular event will happen to a particular person at a particular time. In the second case, we’re no longer talking about the same type of probability; now, we are discussing not the likelihood of the event itself, but rather, the most probable interpretation of the evidence in front of us. Once there is evidence to consider, the evidence overrides any previous ideas we had about what was probable.

This illustrates the problem with Erhman’s logic. If he were to observe the evidence that an airplane had crashed into a car, in order to be consistent with his earlier statement, he would have to refuse to take the evidence into account, simply because he had already determined that such an event is unlikely to happen. This is the surest way to guarantee that we draw our conclusions based primarily on our own biases. He has determined ahead of time what he thinks is most likely to happen, and then interprets all new evidence in the light of his initial assumption. If he has decided that it’s unlikely that the Son of God would become a human, be crucified, and rise from the dead, he has declared that no amount of evidence can change his mind. I agree that, in a general sense, such a thing may be improbable (although, if you think about it, nearly every event in history is improbable in a way). But if we have the type of evidence that we do (such as the reliability of the manuscript copies, the lack of motive for the Gospel writers to lie, the empty tomb, the sudden explosive growth of Christianity despite persecution, etc.), rather than bothering with the general probability of such a thing in the absence of any evidence, we need to consider the evidence itself. At that point, it becomes clear that the accounts of supernatural events in the Gospels are very likely true from a rational, objective historical perspective, since that is the best interpretation of the evidence.

Ultimately, Erhman’s logic betrays his anti-supernatural presupposition, and causes him to be closed-minded and unable (or unwilling) to follow the evidence where it leads. When investigating history, we ought to withhold judgments regarding what is most probably only after, not prior to, observing the evidence. I would very much like to ask him what he thinks about Albino tigers, bullets being stopped by Bibles, or a Royal Flush in Poker.