How Historically Probable is the Resurrection?

April 22, 2011

The question of what claims historians can make about the resurrection is a thorny one. If you’ve ever heard Bart Ehrman debate this issue, you’ve probably heard him make an argument along these lines:

1) Miracles, by definition, are highly improbable events.

2) Historians can only establish what most probably happened in the past.

3) Therefore, historians will never be able to tell us that a miracle probably occurred.

The problem with this argument is that #1 and #2 are talking about two very different kinds of probability.

#1 is talking about probability in the sense of the relative frequency with which an event occurs.  Statistically, how often does event X occur over time? Ehrman is obviously correct about #1. Miracles are highly improbable events. That’s why we call them “miracles”. Statistically, most dead people stay dead.

#2 is talking about Bayesian probability – probability as a measure of a state of knowledge. It’s about the degree of confidence that one can have about a hypothesis as a function of various factors, including its prior probability and the evidence for it. Ehrman is correct about #2 as well. In order to gain access to the past, we must rely on the evidence that’s been left behind, whether it be eyewitness testimony, artifacts, photographs, etc. Our data will always be incomplete. Thus, history (especially ancient history) will never be able to say that something happened with 100% certainty. In that sense, historians can only say that an event “probably occurred”.

But it is demonstrably wrong to say that, because of these limitations, historians can only conclude that an event probably happened if the event has a high prior (frequency) probability.

In reality, historians routinely conclude that improbable events probably occurred. They are able to do so because they don’t limit themselves to considerations of frequency probability. That’s a proper starting point for weighing competing explanations, but the craft of the historian is all about evidence. And sometimes the evidence leaves little doubt that a highly improbable event did, in fact, occur.

In 1990 heavyweight champion Mike Tyson faced James “Buster” Douglas for what was supposed to be little more than practice for the unstoppable Iron Mike. Las Vegas bookies put the odds at 42-to-1 in Tyson’s favor. [Spoiler alert if you recorded the fight, but still haven’t watched it] Douglas K.O.’d Tyson in the 10th round. This was a highly improbable event. Yet I doubt you’d find any historian unwilling to affirm that it probably occurred. One could think of all sorts of improbable events that are, nonetheless, historically secure.

Ehrman is a smart man and I’m sure he gets this. His strategy is to argue that miracles will always be the “least probable” explanation. That is, there will always be more probable natural explanations for an event. If he’s talking strictly about frequency probability, he is correct. But to make his point, he needs to go further and argue that, even after a Bayesian analysis which factors in the evidence, the miracle explanation will still inevitably be the least probable.

And it is that contention which, I say, cannot be stated so categorically. What happens if you factor in the evidence and find the naturalistic explanations hopelessly inadequate to account for the data? An explanation with a minuscule prior probability can become the most probable explanation where the evidence strongly supports it and appears to rule out all alternatives. The problem is that such evidence is hard to come by.

Earlier this week (in a post discussing Brian LePort’s review of Michael Licona’s new tome, The Resurrection of Jesus: A New Historiographical ApproachJames McGrath said the following:

It seems to me that the attempt to find ways to get history to investigate and pronounce on miracles is misguided. It doesn’t seem that any methodological adjustments will allow one to claim that it is more probable that a miracle occurred than that either a miracle story was fabricated, or a mundane event took on supernatural interpretations in the course of retelling. History deals in probabilities, and miracles are by definition improbable.

McGrath’s final sentence (if I may be so bold) suffers from the same confusion about probabilites as Ehrman’s formulation. But in large measure, I think McGrath is right. Our data with respect to historic events is typically quite limited. As a result, we will rarely (if ever) have sufficient data to be able confidently to exclude the naturalistic explanations.

But we should be clear on what the problem is for miracle hypotheses. It’s not that miracles are by definition improbable and therefore historians, who only deal in probabilities, couldn’t ever affirm one. The problem is that our data sets are invariably inadequate to rule out the more probable naturalistic explanations.

But what if that wasn’t the case? Here’s a thought experiment: Suppose that Barack Obama were to suffer a tragic accident wherein he lost a leg. The accident happens in a very public setting and is viewed by millions in person and on television. Doctors try in vain to reattach the leg, but are unsuccessful. With the proceeds going to charity, the leg is purchased by a Las Vegas casino and is permanently displayed next to a grilled cheese sandwich bearing the image of Christ. In the months that follow, Obama is clearly seen by millions to have only one leg. He vacations in Hawaii and is photographed repeatedly wearing only a Speedo. It is self-evident that the leg is gone.

Now suppose that one day Obama goes to church and receives prayer for healing. Suppose he comes bursting out of the church with two good legs, claiming to be remarkably healed. Countless doctors examine him and agree that he now has two legs. DNA tests confirm that both legs are his and that this is indeed President Obama. He vacations again in Hawaii and the Speedo pics portray him clearly in all his bipedal glory. The old leg is still keeping company with the grilled cheese sandwich. Skeptics test the DNA from the old leg and find that it too matches Obama’s DNA.

What now? Must the historian still insist that the miraculous explanation is the least probable explanation?

If you rule out supernatural events categorically, well then yes, the miracle is still the least probable explanation. But only because you’ve decided that miracles have a frequency of zero. If you allow for the possibility of miracles, then it would be possible for a historian to affirm that a miracle probably occurred (at least in theory) where the data set was so compelling as to allow the historian to rule out all competing natural explanations.

There is merit to the idea that history, as a science of sorts, ought not to dabble in the supernatural. And if this is Ehrman’s position, I would probably agree. But we should not treat this methodological limitation of the discipline as though it were an argument against miracles. If historians have decided to exclude the miraculous from consideration in their discipline, the fact that historians cannot affirm that a miracle probably occurred does not mean that a miracle probably did not occur. It simply means that the discipline of history is, by definition, incapable of affirming the miraculous.

There is evidence for the resurrection of Jesus, found largely in the evidence that persons in the first century claimed and seemed genuinely to believe that they had experienced the risen Christ. But it’s probably safe to say that the evidence is nowhere near the threshold that would allow even a miracle-permitting historian to say that the event is “historically probable”. To answer the question posed in many an Ehrman debate, no, historians cannot prove the resurrection, but neither should we expect them to.
[Disclaimer: I am neither a statistician nor a mathematician. If I have butchered Bayesian probability theory, please forgive me (and please let me know).]



  1. So much of the 20th century (and continuing) historical Jesus scholasticism seems predicated on a false assumption of “miracles don’t happen” that then necessitates a complete, blithe, and condescending disregard for any theory that allows for even the possibility that miracles as recorded in the Bible are historically probable. As much fundamentalistic, presumptive thinking goes on there as in any Young Earth “Adam rode the dinosaurs” group of theorists.

    Blech! to both sides!

  2. I have just found this blog, it looks good. I like your approach.

    I agree with you that it is important to make the distinction between the approach that assigns miracles (or whatever) a zero probability, which makes them impossible to accept under any circumstance. And the approach that assigns them an immesurably small probability.

    OP said: “But to make his point, he needs to go further and argue that, even after a Bayesian analysis which factors in the evidence, the miracle explanation will still INEVITABLY be the least probable. And it is that contention which, I say, cannot be stated so categorically.”

    This can be demonstrated mathematically. He is saying that the evidence will never overcome the low probability, under any circumstance. but this is made false by your obama example.

    • Thanks Boz. Glad I’m not the only one who thinks that’s a distinction worth drawing.

  3. Re the Resurrection:

    On one hand is the vagarities of all forms of testimony – differing angles of sight, ways of understanding what may or may not have happened, etc. Testimony is all you really have — the testimonies of the New Testament, the reports of how those who testified were affected by it (earliest church history), and a few reasonable mentions in other sources of the existence and position of a few of the people who were reportedly involved in what transpired (Pilate, John the Baptist, Herod, even Jesus’ brother James, but not Jesus). We must remember that the New Testament itself is not only told from many angles (accurately reflecting the truth about testimony itself), but is a set of very biased and partisan documents, explicitly written and preserved with a very specific purpose in mind. That I very much believe them to be right about that purpose does not resolve the historical questions.

    There is the other end of it: outside of the Greco-Roman politically-based sphere of Rome’s writer/historians and Josephus (who usually had their own rather vigorously-applied opinions), most actual events of the day have little or no ‘straight-up’ record. Even some of what’s taught about the early Roman Empire in school is built on shifty sources. Christians are not the only ones who claim more than than they know.

    • I agree that it’s important to remember that the gospels weren’t written as dispassionate objective histories. The authors were clearly attempting to persuade their audiences to believe. John says as much (20:31).

      We certainly do have a scarcity of sources for most 1st century events. So the weakness of the historical evidence for the resurrection shouldn’t come as any surprise. But even if there’s good reason for the weakness, it’s a weakness nonetheless. I get nervous when apologists try to claim that because there’s more gospel fragments than texts about Caesar (or whomever they choose) we can therefore be historically certain about the resurrection. The weakness of another historical claim does not improve the strength of one’s own claim.

  4. Things happen that are highly improbable. The difficulty in some accepting evolution is exactly that. Improbable<impossible. Though with evolution there is evidence ongoing that it continues. Some improbable phenomenons continue to occur on a limited basis. I think the difficulty in accepting the miracles of Jesus and the resurrection is the fact that they either aren't reproduced or not in a sense that would lead one to believe that they are exclusive to Christianity.

  5. Lets say we do accept miracles as viable historical explanations. Sure! The problem becomes the justmeandmystoryisms. We have to be honest about how terrible the 1st century historical record is, and how mediocre the preservation of details is surrounding early Christian experience. The most significant, well documented phenomena, such as church growth and ecumenical structure, are easily explained by natural causes. So we’re left with a bare bones fragment of data preserved in writings separated by decades of unknown manuscript history. Even the Pauline texts, most important among non-anonymous sources, only have manuscripts surfacing in the 3rd century onward (and later for many patristic sources). Bottom line: if thats enough to support resurrection from the dead, we would certainly accept more strained supernatural claims by others who have provided significantly more witnesses and volumes more contemporary documentation.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: