Is the Resurrection Unbelievable?

A short adventure in probability theory

Among other responses to Ayaan Hirsi Ali’s recent “coming out” as Christian in Unherd magazine, Michael Shermer has added his .02 on his Substack here. Richard Dawkins has since tweeted it out as “a brilliant riposte.” It’s a long piece that ambitiously tries to cover a lot of bases. I won’t try to cover all those bases in reply, but I thought there might be value in taking up just one point: the probability of Jesus’ resurrection.

Michael wrote me a nice e-mail about five years ago when I was just emerging as a writer, after my take on the Jordan Peterson vs. Sam Harris debates went viral. He could tell I was a Christian, so he wanted to share a little of his own writing on why he wasn’t convinced by the evidence for the resurrection. We went on to have a couple short cordial exchanges, nothing extensive, but I appreciated the friendly gesture. At the time, I was still using a pen name, so I just engaged on the substance without revealing this was actually an area of heavy philosophical research focus for both my parents. But now that the cat’s out of the bag, I’ve been thinking for a while about how to introduce people to some of this work in an accessible way. So, as a beginning, while the topic is a little bit current, I thought I would dust off something my father wrote as a gentle introduction to a question whose answer can get a bit technical, depending on how deep you want to go with it: Is the resurrection unbelievable?

Most people expressing skepticism about Christianity, whether they’re laymen or academics, will say yes. When it comes to the claim that Jesus rose from the dead, they will say of course, just about any explanation of the claim seems “more probable” than Jesus actually rising from the dead. After all, isn’t a miracle always the least likely possible explanation for anything, by definition?

Well, yes and no, and that’s what my dad is going to unpack here. He originally wrote this little piece as a lay-level response to this lay-level piece by philosopher Larry Shapiro, whom Shermer has relied on heavily over the years in framing his own arguments against the resurrection. Since Shermer doesn’t add much beyond Shapiro, it will be relevant to Shermer’s recent Substack. I’m mostly preserving Dad’s original words here, but I’ve made some light cuts and additions with his permission. You won’t find this article anywhere else unless you know exactly what to look for with help from the Wayback machine, so I’m happy and proud to make it a little more accessible (although I’m still throwing a paywall on it for now, because I haven’t run any exclusive content for a while, and on account of I don’t have a day job, I need to keep giving y’all a reason to give me money so I can pay rent and stuff!)

So with that, over to Dad.

Is the Resurrection Unbelievable? by Dr. Timothy McGrew

We are dealing here with probabilities and explanatory reasoning, so it is just as well for us to get clear on a few relatively uncontroversial points from the theory of knowledge at the outset. First, all probabilities are numbers between 0 and 1 (including those endpoints), or as we are more apt to say in ordinary language, between 0 and 100 percent. Logical impossibilities (“Some circles have corners”) and obvious absurdities (“All politicians are honorable”) take values down at the low end of the scale; logically necessary claims (“All bachelors are unmarried”) and obvious facts (“The population of Earth on June 10, 2018 is greater than 400”) take values at the high end.

One of the ways to move a belief B rationally from a low probability to a higher one is to learn information that would be far more probable (in the sense of a ratio) if B were true than it would be if B were false and to adjust your probabilities – a process known as conditionalization – accordingly. We do this sort of reasoning all the time in everyday life, though we rarely stop to think about its structure. If I think that a house is abandoned but, upon entering it, I find a fresh cup of tea steeping, my discovery almost certainly changes my mind. I do not have to think that it was highly likely that someone would be making tea if the house were inhabited; I just have to recognize, at least tacitly, that however low that probability may be, it is many orders of magnitude greater than the probability that a fresh cup of tea would be steeping there if the house were uninhabited.

The tea example illustrates another point. It often makes a great difference how the evidence is described. I might truthfully but misleadingly describe what I found in the house as “something cooling down,” and that fact, described in that fashion, would do little or nothing to change anyone’s opinion that the house is abandoned. Almost any piece of evidence loses its force if it is described in a sufficiently fuzzy manner, just as a case composed of many pieces of data may look thin if most of them are omitted. This is one of the problems we must guard against when we are evaluating hypotheses on the basis of a body of evidence.

As long as the probability of B is a real number greater than 0, it is possible in principle for evidence to come in that will raise its probability as close as one likes to 100 percent. (You specify how low the starting point is and how close you want us to come to 100 percent, using real numbers greater than zero; I’ll describe the probabilistic properties that the evidence would need to have to do the job.) For matters of fact outside of one’s direct experience, reasonable belief is a matter of proportioning one’s convictions to the net weight of the evidence and requires, at a minimum, that the probability of B in the light of all of the relevant evidence be greater than 50 percent.

Hume’s contention regarding miracles, couched in this terminology, is twofold; that (1) there is no miracle claim for which the evidence actually available suffices to make the miracle believable, and that (2) in principle, because of the kind of claim that it is, no evidence could. Shapiro’s article focuses on a version of the former, more modest claim with respect to one particular miracle claim. Since I believe that stronger claim is hopelessly misguided, I think it was wise for Shapiro to limit himself here to the more modest one, though as we shall see, something like the bolder claim is hiding in the wings.

Like Hume, Shapiro primes the pump by introducing an example where he expects to engage with his audience’s sympathies. Given only the description he offers, no one would reasonably conclude that Sally had been abducted by aliens; there are certainly much better explanations available for her wild tale. This point can be broken down into three sub-points:

1. Sally’s testimony is the only evidence we have relevant to the question of her abduction by aliens.

2. Given what we know, the hypotheses that Sally is lying or that she is out of her mind are intrinsically more probable than the hypothesis that Sally was actually abducted by aliens.

3. Either of those hypotheses, were it true, would do at least as good a job of explaining the fact that Sally has told this particular tale.

If the case with respect to the resurrection of Jesus were relevantly similar, then Shapiro’s example would be a fair parallel, and our unwillingness to believe Sally ought to be paralleled by our unwillingness to believe that Jesus rose from the dead. We will return to this alleged parallel later.

Shapiro introduces a second example – that of testimony regarding a winning lottery ticket – in order to head off an objection. It is improbable that Sally will win a reasonably large lottery; yet under normal circumstances we might accept her testimony that the ticket she holds is indeed a winner. Why shouldn’t we do the same thing with regard to a reported miracle?

Shapiro’s first response to this parallel is that we should not accept Sally’s word for it until her claim has been verified. I do not find this objection particularly persuasive. If I know that Sally is reasonably cautious and that the report she has heard of the winning number is a reliable one, there is no obvious reason for me to remain skeptical.

But his second response is more interesting: since someone or other must win, the claim we are being asked to believe, though admittedly improbable, is not an unprecedented one, not something that “our ordinary understanding of the world tells against.” In view of this difference, Shapiro suggests, we ought to treat a miracle report with more skepticism than a report about the lottery ticket.

There are two problems with this response. First, the distinction Shapiro is drawing has nothing to do with probabilities. The lottery, for all he has said to the contrary, may be as big as one pleases, and the probability of Sally’s holding the winning ticket as low as one likes. If probabilities are doing the heavy lifting in terms of the conditions for rational belief, then why is he moving away from them in the analysis of this case – apart from the understandable desire to avoid this particular criticism?

Second, the claim that “our ordinary understanding of the world tells against” miracle claims is open to several serious objections. First, the appeal to our ordinary understanding of the world as a bulwark against changing our opinions is an awkward sort of argument to make. The history of science offers us all sorts of unprecedented phenomena that we have come to accept in the light of the evidence. How absurd it must have seemed, when the idea was first mooted, that a living organism, cut in half, could regenerate into two organisms; or that a touch from a piece of ice might kindle a fire; or that a metal contraption weighing more than 200 tons could be held aloft by the pressure of ordinary air, provided that the contraption moves through it quickly enough; or that the table on which my laptop is perched is almost entirely empty space; or that radiant energy is quantized. Yet most of us now take these ideas for granted. Our “ordinary understanding of the world” is perhaps not so potent a talisman against the unfamiliar as all that.

Even more importantly, this whole way of casting the issue involves a fundamental conceptual mistake regarding the common conception of a miracle. Christians no less than their atheist friends believe that there is a way that the universe behaves when it is left to itself. In the ordinary course of events, virgins are not pregnant and dead men stay dead. The central Christian claims are not attempts to contest these generalizations; they presuppose them. That is why the Christians have been shouting from the rooftops for nearly twenty centuries that the resurrection was a miracle. The evidence for how the universe behaves when it is left to itself, be it ever so strong, must not be mistaken for evidence that it always is.

When we leave the intuition pumps and come to the discussion of Jesus’ resurrection, Shapiro is most concerned that the deck be stacked in a very particular way. He tells us that it is “well established” that none of the writers of the Gospels were eyewitnesses of the crucifixion, and he assures us that all unbiased historians agree that there are irreconcilable inconsistencies in the accounts of the discovery of Jesus’ empty tomb. From the supposed fact of these errors in the Gospels, Shapiro wants to draw the conclusion that “there is some probability (and it doesn’t have to be terribly great) that the gospel accounts of Jesus’s resurrection are false.”

Shapiro throws out these claims quite cavalierly, with the snide jab that only “those in the business of defending Christianity” could possibly disagree with him. Like many, he falls into the unfortunate habit of dismissing any attempt to harmonize the gospels as an exercise in desperate Christian apologetics. In fact, there is nothing distinctly desperate or Christian about attempting to harmonize primary sources in the course of doing history. Historians do so all the time. Shapiro might counter that historians aren’t normally working with documents that contain miracle claims, but then this would reveal what is really lurking in the background of Shapiro’s claim—i.e., an automatic bias against the gospels merely because they include miracles. Shapiro would, of course, be free to harbor that bias. It just shouldn’t be confused with good historical praxis.

The double standard becomes obvious on even a brief survey of the sort of discrepancies that crop up in classical sources. Take the murder of Julius Caesar, which we have attested in the works of Suetonius (Julius 78-82), Plutarch (Caesar 58-66 and Brutus), Appian (Civil Wars 2.117), Cassius Dio (Roman History 44.19-22), and Seneca (Epistles 83). Were we minded to pick at discrepancies in the accounts after the manner of some New Testament scholars, what a wide field for caviling these documents would afford us! What was the name of the first conspirator to approach Caesar? Cimber Tullius? Tillius Cimber? Tullius Cimber? Metillius Cimber? Atillius Cimber? Was his pretense to make a request, or was it to acknowledge a favor? Did someone seize Caesar’s cloak or not? If so, what did he do with it? Was the first blow to Caesar’s neck? or to his shoulder? or to his back? To riff on Bart Ehrman’s favorite line that “It all depends on which gospel one reads,” we might say that it all depends on which historian one reads.

Yet, notwithstanding all of these discrepancies, no sane historian doubts the central claim of the reports: Caesar was murdered in the Forum by a group of conspirators. The public murder of an unpopular emperor is no doubt a more common event than the resurrection of a Jewish Rabbi, and we may therefore justly demand more evidence for the latter than for the former. But minor variations in the accounts are not doing significant work here, since we find such minor variations everywhere we turn in historical accounts of events that unquestionably happened. Whatever weight this argument may have rests squarely where Hume himself placed it: on the prior improbability of a miracle claim.

Granted, there could be circumstances under which the argument from error or contradiction carried force against the gospels’ testimony. We don’t simply want to strip-mine the gospels for “nuggets” of truth while waving our hands over the details as if they don’t matter. The details do matter, particularly in the specifics of Jesus’ resurrection appearances. We don’t just have vague reports that the disciples “saw Jesus.” We are told that they touched him, spoke with him at length, and shared meals. If the gospels are indeed a product of eyewitness testimony, this would suggest a variant on C. S. Lewis’s famous trilemma: Either the disciples were lying, or they were mad, or they were telling the truth.

Naturally, Shapiro would dispute that the gospels are the work of witnesses, or those close to witnesses, but this would require him to show a good deal more work. If, for instance, he could convincingly show that the gospel writers were not just imperfect but habitually in the business of inventing or embellishing their narratives of Jesus’ life, then this would reasonably cast doubt on the central miracle claims. A witness in court might tell us that someone’s car was blue when it was really green, but as long as we trust that he is earnestly trying to provide an accurate account to the best of his less than perfect recollection, we still consider him generally reliable. We will be much less inclined to trust him if he is a proven fabulist. We might also mistrust him if his errors were so numerous and so basic that he simply couldn’t be relied upon to know what he’s talking about. A full and careful defense of the gospels against these charges goes beyond the scope of a short piece, but resources are readily available for anyone inclined to dig further.

Shapiro wants to draw our attention to the existence of alternative explanations for the data, and in particular to two of them: that Jesus’ disciples staged a resurrection in order to take advantage of extant messianic expectations, and that the story of Jesus’ resurrection was a rumor that emerged many decades after his death. Neither explanation does the sort of work he needs it to do. Take the idea of hijacking the messianic hopes of the Jewish people, which attempts to seize the “Liar” horn of our trilemma. First-century messianic expectations and Jewish national and religious aspirations were a mismatch for what the Christians were saying about Jesus in multiple ways. The Jews were looking for a military leader; Jesus (so his followers wrote) maintained that his kingdom was not of this world. Their expectations were bound up with national pride; Jesus (so we read in the Gospels) told them that people would come from all quarters of the world to sit down with Abraham, Isaac, and Jacob in the kingdom of God. The Jews taught that God must be worshiped in Jerusalem; Jesus taught that God might and should be worshiped everywhere. The Jews held to elaborate traditions regarding the keeping of the Sabbath; Jesus, according to their reckoning, broke them frequently. And there is no reason to think that devout first-century Jews thought that a resurrection – much less one that had no immediate national and military consequences – was to be expected of the messiah. It is hard to imagine a conspiracy more poorly designed to exploit first-century expectations than the one here hypothesized.

The idea that the story of Jesus’ resurrection was a rumor started decades after his death fares, if possible, even worse. To be sure, it requires a robust defense of the gospel accounts to uphold that testimony as reliable and rational. But it’s immediately obvious that the story was in the air shortly after Jesus’ crucifixion. Saul of Tarsus, better known by his Greek name of Paul, was converted within scant years thereof. He knew full well the arguments that had persuaded Annas and Caiaphas and Gamaliel that Jesus was a deceiver of the people, and he assented to them with fervor. He could interrogate not fragmentary documents or transient rumors but living people. And in his first letter to the church at Corinth, Paul records for us the terms in which he himself was catechized in the mid 30s:

I delivered to you as of first importance what I also received:

that Christ died for our sins in accordance with the Scriptures, that he was buried, that he was raised on the third day in accordance with the Scriptures, and that he appeared to Cephas, then to the twelve. Then he appeared to more than five hundred brothers at one time, most of whom are still alive, though some have fallen asleep. Then he appeared to James, then to all the apostles.

There is, therefore, simply no time for the hypothesis of late legendary development. It stands in direct conflict with everything we know about the origins of Christianity from the primary sources. And the first-century documents that describe those origins, including the letters of the indefatigable apostle Paul and the history of the origin of the church in the Acts of the Apostles, afford us both an adequate cause for the effects we see in the late first and early second centuries and a case study in interlocking evidence, as the undesigned coincidences among the various documents render the hypotheses of fraud or legendary development preposterous. (This is to say nothing of the numerous external confirmations of Acts, showing the care and seriousness with which its author took the task of preparing his reportage.)

Shapiro assures us that his alternative hypotheses are not fantastically improbable. On the contrary, in the light of our data, that is exactly what they are. They can be advanced with some plausibility only so long as we keep the data in soft focus, in parallel to the case where Sally’s tale is the only evidence we have. But in point of fact, the evidence in this case is vastly richer and more powerful than his analogy suggests.

The only excuse for adopting such poor shifts rather than the historic Christian explanation of the evidence would be that a resurrection is an even more improbable option. “I simply cannot accept,” Shapiro writes, “that this probability is greater than that for any of the explanations I offered above, or for many others even less probable.”

Indeed, he cannot, but the reason for that inability lies not in any deficiency in the evidence but rather in an immovable philosophical determination not to allow a divine foot in the door. A determination which has no rational basis, though, from a human standpoint, it is eminently understandable. To quote the ever-quotable David Berlinski, “If we are obliged to accept absurdities for fear of a divine foot, imagine the prodigies of effort that would be required were the rest of the divine torso found wedged at the door, and with some justifiable irritation demanding to be let in.”