In this post, I’d like to comment on some issues that have been raised by readers over at Professor Joshua Swamidass’s Peaceful Science forum, in response to my article on Michael Alter’s book, The Resurrection: A Critical Inquiry, which convincingly demolishes Resurrection apologetics. But before I do that, I’d like to discuss the Christine Ford case, and its relevance to the evidence for the Resurrection.

Part A: The relevance of the Christine Ford case: memory is unreliable, liars are difficult to spot, and facts trump everything

Left: Dr. Christine Blasey Ford, professor of psychology at Palo Alto University. Official video by the United States Senate Committee on the Judiciary, as posted to www.judiciary.senate.gov. Image courtesy of Wikipedia.
Right: Senator Susan Collins, of Maine. Image courtesy of United States Congress and Wikipedia.

I imagine that most of my readers would have followed the confirmation proceedings for Brett Kavanaugh with great interest, whatever their political persuasion, and I’m sure everyone has their own opinion about the accusation of rape that was made against him by Dr. Christine Ford and two other women. In this post, I’d like to focus on Dr. Ford’s accusation, as it received the most media coverage. I have to say I found Christine Ford’s testimony very powerful and deeply moving, when I watched it online. But at the end of the day, I was forced to conclude that her accusations were almost certainly false. What changed my mind? Facts. Put simply: the elements of her story which could be checked, turned out to be false. Senator Susan Collins highlighted the most damaging points in her speech justifying her vote to confirm Judge Kavanaugh:

• “none of the individuals Professor Ford says were at the party has any recollection at all of that night,”
• “Professor Ford [herself] testified that not a single person has contacted her to say, ‘I was at the party that night,'”
• “not a single person has come forward to say that they were the one that drove her home or were in the car with her that night,” and
• “even though she left that small gathering of six or so people abruptly and without saying goodbye and distraught, none of them called her the next day – or ever – to ask why she left – is she okay – not even her closest friend, Ms. Keyser.”

Summing up, Senator Collins declared:

The facts presented do not mean that Professor Ford was not sexually assaulted that night – or at some other time – but they do lead me to conclude that the allegations fail to meet the “more likely than not” standard. Therefore, I do not believe that these charges can fairly prevent Judge Kavanaugh from serving on the Court.

As I don’t want to bore readers with specifics, I will simply refer them to the following articles, especially the first article by Victor Davis Hanson, an American military historian, recipient of the National Humanities Medal in 2007 by President George W. Bush, and a former professor of classics at California State University, Fresno:

(1) One Ford Narrative Too Many by Victor Davis Hanson (American Greatness, October 7, 2018).

(2) Rachel Mitchell: Ford Has No Case by Rod Dreher (The American Conservative, October 1, 2018).

(3) The Three Lies of Christine Blasey Ford by J. R. Dunn (American Thinker, October 2, 2018).

(4) White House Finds No Support in FBI Report for Claims Against Kavanaugh by Rebecca Ballhaus, Michael C. Bender, Kristina Peterson and Natalie Andrews (Wall Street Journal, October 4, 2018).

So what really happened? One intriguing possibility is suggested here. Another recent theory, which doesn’t require any outlandish assumptions, can be found here.

Despite the discrepancies in Dr. Ford’s testimony, many people were inclined to believe her story for four reasons:

(i) a willingness to ascribe the discrepancies in her account to memory-damaging trauma, based on junk science (rebutted here);

(ii) a misplaced faith in the accuracy of memory – especially the notion that rape victims accurately remember the faces of their attackers (rebutted here);

(iii) a misplaced trust in human nature – in particular, the myth that women seldom lie about rape (rebutted here, here and here); and

(iv) a misplaced confidence in their own ability to spot liars (rebutted here).

The people who credited Dr. Ford’s account thought they had very powerful reasons for believing it to be true, which made them willing to downplay the discrepancies in her story as mere details: to their way of thinking, critics who harped on these points were nitpickers. They were wrong. The reasons which seemed overwhelming to them turned out to be based on bad science. As Emily Yoffe points out in her article, The Bad Science Behind Campus Response to Sexual Assault (The Atlantic, September 8, 2017), trauma does not render a rape victim incapable of recalling their assault or describing it coherently or chronologically. Hence it could not have contributed to Dr. Ford’s chronological confusion, which was exposed in Arizona prosecutor Rachel Mitchell’s damning report.

Dr. Ford also stated that she was 100 percent certain that Kavanaugh was her attacker, but victims’ memories of their assailants’ faces can be terribly mistaken, on occasions, even when they make a conscious effort to commit them to memory:

In 1984, a young college student was raped at knife-point in her apartment. A few days later, she identified her alleged assailant from a series of police photos as Ronald Cotton. In a poignant New York Times article, “I Was Certain, but I Was Wrong,” she wrote that she “knew this was the man. I was completely confident. I was sure.”

Her confidence was due to the fact that during the assault she “studied every detail of the rapist’s face…to make sure that he was put in prison…to rot.” Yet she was wrong. Three years later, when confronted in court with her actual attacker (as determined via DNA), she stated, “I have never seen him in my life. I have no idea who he is.’”

On the basis of her compelling and apparently credible testimony, Cotton was twice sentenced to a life in prison. Thus, because we are told to believe young women without question, a young man spent 11 years of his life behind bars for a crime he did not commit.

Another commonly cited reason for believing Dr. Ford was that false accusations of rape are rare; the vast majority, it was implied, are true. However, the myth that nearly all rape reports are true turns out to be an unjustified inference from the fact that only a small percentage can be shown to be false. But by the same token, only a very small percentage of sexual assault reports can be shown to be true, as blogger Francis Walker points out to devastating effect, in an article titled, How To Lie And Mislead With Rape Statistics: Part 2 (Data Gone Good, January 27, 2015), which concludes that “no one knows — and in fact no one can possibly know — exactly how many sexual assault reports are false.” In the vast majority of cases, argues Walker, we simply don’t know who’s telling the truth:

At the end of the day, my point is this – If someone tries to advocate a particular policy based on the fact that “only 2-8% of rape reports are false” an appropriate response might be “Sure, but only 1-8% are true.”

Finally, many people believed Dr. Ford because they found her testimony moving and credible. But other viewers had very different impressions. So what does that tell us? Regarding our supposed ability to spot liars, research shows that even people whose job it is to professionally detect lies — judges, police officers, customs agents – are no better than other people at spotting them: their level of accuracy is around 54 percent (which is barely better than chance). Studies show that these individuals believe themselves to be better than chance at spotting liars, but the same studies also show that they aren’t. What that means is that the millions of people (including officers of the law) who watched Christine Ford and Brett Kavanaugh testify had no reliable way of spotting who (if either of them) was lying.

The moral of the story is that facts trump everything, and that in assessing the truth of highly contentious claims, we should always be governed by our heads, rather than our hearts. Emotion can lead us badly astray.

This moral can be equally applied to the Resurrection accounts. We might find them deeply moving, but that should not sway us in our judgement as to whether or not they are true. We need to coolly and calmly assess the evidence.

Part B: Responding to Professor Swamidass’s Criticisms of my debunking of Resurrection apologetics, based on the arguments in Michael Alter’s book

In September, I wrote an article for The Skeptical Zone, titled, Michael Alter’s bombshell demolishes Christian apologists’ case for the Resurrection, in which I explained why, after reading Alter’s book, The Resurrection: A Critical Inquiry, I believed that Christian apologists’ case for the Resurrection had been effectively demolished. A central claim in the argument I put forward was that the Gospel accounts of Jesus’ passion and crucifixion are highly doubtful, on no less than 17 key points. The relevance of the Christine Ford case should now be obvious: as we saw, Dr. Ford’s testimony was undermined by the damaging fact that the key points of her story which were capable of being checked, turned out to be false.

Professor Joshua Swamidass kindly discussed my views on a thread over at Peaceful Science, titled, Alter’s Case Against The Resurrection, which was opened by Dr. Patrick Trischitta. I’d like to address the main criticisms voiced by Professor Swamidass against my arguments, in addition to several criticisms put forward by physics graduate Daniel Ang on the same thread.

Objection 1: It’s illicit to multiply probabilities, in an attempt to discredit the Gospels’ Passion narratives

In my article, Michael Alter’s bombshell demolishes Christian apologists’ case for the Resurrection, I argued against the historical reliability of the Passion narratives in the Gospels, on the grounds that the Passion narratives of Jesus’ last 24 hours are historically doubtful, on at least seventeen points:

It turns out that the Gospel accounts of Jesus’ passion and crucifixion are highly doubtful, on no less than 17 points. (Yes, you read that right.) One or two points would be bad enough, but perhaps acceptable: after all, improbable things happen every day, and it would be surprising if the historical details of Jesus’ crucifixion contained nothing out of the ordinary. But 17 highly improbable occurrences over a 24-hour time period strains credulity.

Professor Swamidass objected that in putting forward this argument, I am guilty of multiplying probabilities illicitly:

Multiplied a large number of maybes together and everything looks impossible.

For the record: nowhere in my article do I argue that one can multiply these seventeen probabilities together. That would only work if they were all independent events, but some of them are not. Rather, the methodological claim I’m making is a very modest one, which all historians would subscribe to: you should not consider a work to be historically reliable if it contains a large number of highly dubious assertions within a short time period. That’s all.

So it’s the relatively short time period that makes the argument a particularly telling one. Seventeen highly improbable occurrences in 24 hours is a bit fishy, to say the least.

Nevertheless, some readers may be inclined to ask, out of idle curiosity: how many of these seventeen improbable occurrences are actually independent events? Let’s enumerate them:

a. Was the Last Supper a Passover meal, and did Jesus tell his disciples to drink blood?
b. Did Jesus die on the Jewish Passover?
c. Do the Gospels accurately represent Jesus trial before the Jewish Sanhedrin?
d. Was Pontius Pilate reluctant to convict Jesus?
e. Judas’ betrayal of Jesus and subsequent death
f. The chief priests’ mockery of Jesus on the Cross
g. The story of the good thief: fact or fiction?
h. Jesus’ last words on the Cross: fact or fiction?
i. Did Jesus’ mother and the beloved disciple stand at the foot of the Cross?
j. The three hours of darkness: fact or fiction?
k. The earthquake at Jesus’ death: fact or fiction?
l. Was the Veil of the Temple torn in two?
m. Were Jewish saints raised at Jesus’ death?
n. Blood and water from Jesus’ side?
o. Was Jesus buried in a new rock tomb?
p. Was there a Guard at Jesus’ tomb?
q. The women visiting Jesus’ tomb on Sunday: does the story add up?

Clearly f is dependent on b, which is to some degree dependent on a. Claim c also bears some relationship to b, as does q. However, claims d, e, g, h, i, j, k, l, m, n, o and p are each capable of standing in isolation: one could accept any one of these claims while rejecting the others, and vice versa – although some points raised in h would weaken claim i, and a rejection of i also undermines claim n.

All in all, I think it would be fair to say that there are at least ten logically independent occurrences on this list. Make of that what you will.

Objection 2: Multiple improbable events can happen to any of us, even over a short time period. Why not Jesus, too?

Lottery ticket. Image courtesy of Open Geofiction.

In response to my argument above, physics graduate student Daniel Ang advanced the objection that many improbable things may happen to someone, even over a short time period, and he offered the following example:

…[I]t’s common for many improbable things to happen during a 24 hour time period if you’re just taking into account prior probabilities (which are difficult to assign rigorously anyway). Today I attended an Indonesian festival in Boston (something very rare, only happens once a year). A temperature sensor on my experimental apparatus broke (having never broken in the last 2 years). I spent two hours thinking about information theory (never did this ever – I am a physicist, not information theorist). That’s already three improbable occurrences in the life of one person on a relatively mundane day. By these standards, my diary entry would be regarded as a mythological fabrication. This is just not the way to do history.

In reply, I pointed out that I wasn’t arguing from prior probabilities; rather, I was making use of posterior probabilities (a point I’ll expand on, below).

I also noted that the number of improbable events on the day mentioned by Ang was just three: a fairly small number. I then added:

1. Please see my point above about the number of improbable events. The likelihood of 17 (or even 10) such events over a 24-hour period is much lower than that of three such events.

2. When all we have available to us are copies of four biographies written 30 to 60 years after Jesus was crucified, then we really have to fall back on probabilities, when assessing the credibility and reliability of the sources. How else can we proceed? We cannot interview the authors of the Gospels or the witnesses to the alleged events.

3. Using your logic, I could defend any unlikely alleged revelation on the grounds that improbable events happen all the time. I don’t believe any trained historian would buy that defense.

Additionally, many of the improbable events described in the Gospels appear to be written for an evangelistic purpose – e.g. the earthquake, the sky turning dark, the story of blood and water, and so on. They were written to show that Jesus is the Son of God. So if I were a historian and I found a large number of stories about a famous person which all related to a single day in the life of that person, and which were all highly improbable on historical grounds, as well as a plausible motive for why these stories might have been created about that person in the first place, I would be rather skeptical. (NOTE: Readers will note that I haven’t listed Jesus’ Resurrection as one of these improbable events, as I see no way of assessing the historical probability of such an event. All I have claimed is that apologists’ arguments claiming to show the high probability of the Resurrection are fallacious.)

A final observation I’d like to make is that none of the three improbable events chronicled by Daniel Ang are the kinds of important events that a historian would include in the biography of a famous person. (Remember: the Gospels are first-century biographies of Jesus.) Attending a festival (without creating any kind of incident there), a broken temperature sensor, and thinking about information theory: these are hardly earth-shattering events. Minor improbable events may be more frequent than we realize. However, at least half of the seventeen highly improbable events I identified in the Passion narratives were events of major importance. Let’s err on the side of caution and say eight. If someone were to claim that eight highly improbable events of major importance happened to them in the space of a single day – say, winning a $100,000 lottery ticket, a death in the family, a job promotion, a meeting with a long-lost friend, a car crash, a house burning down, an earthquake and a UFO sighting – should we believe them? Or should we be skeptical? You decide.

Objection 3: Am I guilty of a misuse of probability theory?

Bayes’ theorem spelt out in blue neon at the offices of Autonomy in Cambridge. Image courtesy of mattbuck and Wikipedia.

Professor Joshua Swamidass also suggested that my article, Michael Alter’s bombshell demolishes Christian apologists’ case for the Resurrection, contained an elementary mathematical fallacy, which he described in his thread, Alter’s Case against the Resurrection:

The fallacy is mistaking P(Story1) for P(Story1|Evidence) vs. P(Story2|Evidence). The P(Story1) will always be low for just about every true story. P(Story1|Evidence) vs. P(Story2|Evidence), on the other hand, can point to clear winners and losers.

With the greatest respect, Professor Swamidass has misconstrued my argument. I am not arguing that because P(Story1) is low, we can discount Story1. Nor am I arguing that because because P(Story2) is greater than because P(Story1), we should believe Story2 rather than Story1, where Story 1 describes a highly unusual occurrence and Story2 describes a fairly common or regular occurrence. Rather, what I am arguing is simply that the alleged evidence for Story1 isn’t strong enough to swing the balance of probability in Story1’s favor. In mathematical language: I am saying that [P(Evidence|Story1)/P(Evidence|Story 2)] is greater than 1 but not very high, whereas P(Story2)/P(Story1) is very high, so we should still favor Story2 over Story1.

For mathematically minded readers, here’s an explanation of my argument. Using Bayes’ Theorem, we can calculate P(Story1|Evidence) and P(Story2|Evidence) as follows:

P(Story1|Evidence) = P(Evidence|Story1) * P(Story1)/P(Evidence)

P(Story2|Evidence) = P(Evidence|Story2) * P(Story2)/P(Evidence)

This means that the ratio of P(Story1|Evidence) to P(Story2|Evidence) is equal to
P(Evidence|Story1) * P(Story1) divided by P(Evidence | Story2) * P(Story2), or in other words:
[P(Evidence|Story1)/P(Evidence|Story 2)] divided by P(Story2) / P(Story1).

Hence if [P(Evidence|Story1)/P(Evidence|Story 2)] is greater than 1 but less than P(Story2)/P(Story1), we should favor Story2 (which appeals only to regular occurrences) over Story1 (which invokes highly unusual and antecedently improbable occurrences).

In my last post, I identified no less than 17 highly improbable occurrences in the Gospel accounts of Jesus’ last 24 hours. I’d like to look at three of these, to illustrate how Bayes’ Theorem plays out. I have chosen these three cases, because they illustrate all three kinds of highly improbable occurrences in the hours leading up to Jesus’ death:

(i) occurrences which are psychologically unlikely, as they involve people acting out of character, or in a way which defies legal or social norms;

(ii) occurrences which are physically unlikely, as they involve rare natural occurrences; and

(iii) miraculous occurrences.

All three of these are antecedently unlikely; nevertheless, if the evidence is good enough, it would be reasonable to credit them. But is it? Let’s take a look.

Nikolai Ge, Christ and Pilate (“What is truth?”), 1890. Public domain. Image courtesy of Wikipedia.

1. Pilate’s alleged reluctance to condemn Jesus to death. This, I argued, is psychologically unlikely, given what we know of Pilate’s character. Here, Story1 is that Pilate was indeed reluctant to condemn Jesus, while Story 2 is that he sentenced Jesus to death without compunction. In my post, I described several incidents in the life of Pilate which illustrated his cruelty. Clearly, killing Jesus would not have been out of character for him. Given this background information about Pilate’s character, we might reasonably conclude that the ratio P(Story2)/P(Story1) is pretty high – say, 10 to 1.

But now, we have to consider some new, additional evidence: the testimony of the Gospels, which all agree that Pilate was curiously reluctant to order Jesus to be crucified. In order to show that Story1 is more creditworthy than Story2, one has to show that the ratio of P(Evidence|Story1) to P(Evidence|Story2) is even greater than 10 to 1. Obviously if Story1 is true, then it is hardly surprising if the Gospels record it; hence P(Evidence|Story1) is very high. In fact, I’m happy to concede it’s around 1.

What about P(Evidence|Story2)? In order to obtain a ratio of greater than 10 to 1, one would need to show that P(Evidence|Story2) is less than 0.1. However, I don’t think it is, and I don’t think a neutral historian would, either. Even if Pilate were a cold, vicious killer, I can think of one very practical reason why the Gospels might want to whitewash his character by portraying him as reluctant to condemn Jesus: doing so would shift the blame for his death away from Rome, reducing the friction between the Roman Empire and the religion of Christianity, and making Christianity appear less of a threat to the established social order. That would have suited the early Christians, as it would have reduced the likelihood of persecution. Mark’s Gospel (which may well have been written in Rome, far away from events happening in Jerusalem) is generally agreed to have been the earliest Gospel, from which Matthew and Luke are known to have copied large slabs. (Even John may have been familiar with Mark’s Gospel.) Even if Story2 is true and Pilate had no compunction about sentencing Jesus to death, I would say there’s at least a 20% chance that Mark (for reasons best known to himself) would have chosen to portray Pilate as reluctant to have Jesus crucified, and that the other Evangelists either accepted his account as true or deliberately chose not to dispute it (perhaps because they wanted to downplay the threat posed by Christianity to Roman society), and that they propagated Mark’s account (adding some extra touches of their own). In other words, P(Evidence|Story2) is at least 0.2, while P(Evidence|Story1) equals 1, meaning that the evidence favors Story1 five times more strongly than Story2. But on the other hand, the antecedent probability of P(Story2) is ten times that of P(Story1).

If I’m right, then the ratio of P(Story2|Evidence) to P(Story1|Evidence) is equal to 2:1, making the Gospel account of Pilate being reluctant to condemn Jesus only half as likely as the claim that Pilate sentenced Jesus to death without compunction, when all the evidence is considered.

I should mention that the Gospels also record other highly doubtful events in connection with Jesus’ Passion and Crucifixion, which are even more unlikely, on psychological grounds, as they involve violation of legal and/or social norms prevalent at that time: for instance, the claim that a male disciple of Jesus (who was crucified as a political criminal) would have been allowed to stand at the foot of the Cross, or that Roman soldiers would have disobeyed Pilate’s express order that the legs of the three crucified criminals (including Jesus) were to be broken. The same logic can be applied in these cases, to even more devastating effect, as we are no longer dealing with an individual acting out of character (which is odd enough), but rather, with the violation of a socially imposed rule, custom or practice. On Bayesian grounds, these events are far less likely than the one I analyzed above.

2. The earthquake at Jesus’ death. In my previous post, I cited recent seismological evidence for a magnitude-6 earthquake in Palestine within a year or two of Jesus’ death. Suppose we grant that the quake occurred at some time during a three-year period. Then the chance that it occurred on the very day of Jesus’ death is about 1 in 1,000, and the chance that it occurred during the three-hour period between noon and 3 p.m. (when Jesus died) on that day is 1 in 8,000, or roughly 1 in 10^4. Thus if Story1 is the claim (found in Matthew’s Gospel but not in Mark, Luke or John) that there was an earthquake at or shortly before Jesus’ death, and Story2 is the claim that the earthquake occurred at a different time or on a different day, then P(Story1) is about 10,000 times lower than P(Story2).

Next, we have to consider the evidence of the Gospels. This favors Story1, but by how much? Let us grant that if Story1 is true, then the probability that at least one of the Gospels would record the earthquake at Jesus’ death is pretty high: close to 1. In other words, P(Evidence|Story1) equals 1. What about P(Evidence|Story2)? I can think of three good reasons why Matthew might have included this story, even if it didn’t happen: (i) his memory deceived him, and he mistakenly associated his recollection of the earthquake with his recollection of Jesus’ death; (ii) he wasn’t personally present at Jesus’ death, but he relied on accounts given by other people whose memories of the two events were faulty and who mistakenly conflated them in their minds, years afterwards; and (iii) growing up in a culture where the deaths of famous people were often said to be heralded by natural portents, he would have been naturally disposed to believe that Jesus’ death was accompanied by signs at least as great, and having heard reports of an earthquake occurring around that time, he may have taken it to be the sign of Divine wrath that was manifested at Jesus’ death. So I think it is fair to assume that P(Evidence|Story2) equals 0.5. In other words, since P(Evidence|Story2) is 0.5, while P(Evidence|Story1) equals 1, the evidence favors Story1 twice as strongly as it does Story2. But on the other hand, the antecedent probability of P(Story2) is 10,000 times that of P(Story1).

Hence the ratio of P(Story2|Evidence) to P(Story1|Evidence) is equal to 5,000:1, making Story2 much more credible than Story1, and rendering Matthew’s story of the earthquake massively unlikely. And that’s without even considering the very strange fact that the other Gospels don’t even mention the earthquake at Jesus’ death.

3. The three hours of darkness before Jesus’ death. I argued in my previous post that the three hours of darkness could not have been a solar eclipse (which would have lasted less than eight minutes) or a sandstorm (which would have driven everyone indoors), and that it could only have been a supernatural occurrence (i.e. a miracle), if it happened at all. This is our Story 1, while Story 2 is that nothing astronomically out of the ordinary happened before Jesus’ death. So how would a historian with no bias against the supernatural calculate the prior probability of Story 1? By setting a ceiling for it – namely, the prior probability of any miracle (let alone a miracle involving continual darkness) occurring over a three-hour period of recorded history. Now, given that recorded history begins around 3,500 B.C., it follows that the number of days from the dawn of history until the crucifixion is about 1.3 million, so the number of discrete three-hour intervals during this period would have been about 10 million. Hence using Laplace’s sunrise argument, we might naively estimate the prior probability that the three-hour period preceding Jesus’ death was miraculously dark to be 1 in 10 million, at the very most. (Needless to say, I’m being very generous here.)

To surmount this antecedent unlikelihood, we need powerful evidence which is readily explained by a miraculous darkness, but which cannot be readily explained otherwise. But the only evidence we have is that of the three Synoptic Gospels, written thirty or more years after the Crucifixion. However, even if nothing miraculous happened, it would not be surprising if the Gospels recorded a miraculous occurrence: as I showed in my post, similar occurrences are recorded as occurring at the deaths of other famous people, too, in Jesus’ day. We might reasonably conclude that P(Evidence|Story 2) is about 0.1 – in other words, even if nothing happened in the sky at Jesus’ death, there was still a 1 in 10 chance that a story of some supernatural occurrence would make its way into the Gospels, anyway. Hence the ratio P(Evidence|Story1)/P(Evidence|Story 2) is not very high: about 10 to 1. Since P(Story2)/P(Story1) is 10,000,000 to 1, this means that for a neutral historian, carefully weighing the evidence, Story 2 is favored over Story 1 by a factor of 1,000,000 to 1.

Alternatively, if one rejects attempts to calculate the prior probability of miracles on account of the inscrutability of the Divine will, one could still argue that since legends of eclipses at the deaths of famous people were known in Jesus’ day, the legend hypothesis is an adequate explanation for the story of three hours of darkness appearing in three of the four Gospels. Thus we have no good reason to invoke a miracle.

Part C: Further objections from Professor Swamidass and Daniel Ang

4. Not so improbable after all?

Daniel Ang wrote that he didn’t think the seventeen events I listed in the Passion narratives were highly improbable, after all:

The main disagreement between us is that you think for many incidents in the Gospels, P(S1) is incredibly small. For all of the 17 incidents that I’ve given some thought, P(S1) is at worse 50-50 to me.
[Note: S1 in this quote refers to Story 1, or the accounts given in the Gospels – VJT.]

I responded:

Seriously? The probability of a highly illegal night trial at Caiaphas’ residence is 50%? The probability that Pilate would be reluctant to condemn a man accused of advocating insurrection and non-payment of taxes to death is 50%? The probability that the Romans would have allowed a male disciple or a member of the family of an accused political criminal to stand at the foot of the cross is 50%? The probability of Jewish saints being raised to life at someone’s death is 50%? The probability of a guard being placed at the tomb of a crucified man, to make sure he doesn’t rise again, is 50%? Seriously???

5. Even if these improbable events didn’t happen, how relevant are they?

Daniel Ang also raised another objection – namely, that a number of the seventeen improbable occurrences I cited in the Gospel narratives of Jesus’ Passion and Crucifixion were of relatively peripheral importance, and that consequently, they could not be used to impugn the historical reliability of the Gospels:

Secondly, you have not answered the question of whether it’s justified to classify 17 improbable occurrences equally. As a lot of people in this thread have argued, things like Communion and the good thief, even if improbable or straight out wrong, are immaterial to the overall reliability of the Gospels.

As I noted in my article on Michael Alter’s book, the Passion narratives in the Gospels are an ideal litmus test of the Gospels’ historical accuracy, since (a) Jesus’ Passion is related in all four Gospels; (b) the four Gospels exhibit a relatively high degree of agreement in their narratives of Jesus’ final 24 hours; and (c) the events described therein are historically checkable. In other words, what matters here is not whether the improbable occurrences I mentioned are central to the story of Jesus’ life or merely peripheral: rather, what matters is whether they are good candidate events for assessing the historical accuracy of the Gospels. The seventeen events I cited all form part of the litmus test I described above, because they all took place during a time period that was heavily covered by all four Gospels: the last 24 hours of Jesus’ life.

6. A fairer test of the Gospels’ historical reliability?

Image of the Cover of the famous Carolingian Gospel Codex Aureus of Sankt Emmeram. Made in ca. 870 at the Palace of Holy Roman Emperor Charles the Bald. Image courtesy of Bayerische Staatsbibliothek München and Wikipedia.

Another objection which could be made to my argument is that it is overly focused on the historicity of the 17 doubtful incidents recorded in the Passion narratives in the Gospels. Perhaps a fairer test would be to examine the text of the Gospel narratives themselves, and see how much is historically plausible, how much is doubtful, and how much is beyond the ken of historical investigation. Fortunately, I have already done this exercise. I would invite readers to have a look at Section F of my original article on Michael Alter’s book: How the Gospel Passion Narratives Evolved Over Time.

It turns out that the “common core” of material which is found in all four narratives is nearly all historically reliable, with the exception of Pilate’s reluctance to condemn Jesus to death and his granting Joseph of Arimathea permission to bury Jesus’ body in a tomb: these two claims are highly doubtful. If we look at the additions to the “common core” made by Mark (and used by Matthew and Luke), we find that about half is historically plausible, while half is doubtful. Nearly everything added to the narrative by Matthew is either historically doubtful or pertains to supernatural events which lie beyond the ken of historical investigation. Luke’s Gospel builds on Mark’s in a very different way from Matthew: Luke is more cautious, and his additions are a mix of historically plausible corrections to Mark, historically implausible attempts to absolve the Roman authorities of blame in Jesus’ death, Luke’s own “spin” on Jesus’ death on the Cross, and some supernatural material of his own. John’s Gospel omits several dramatic passages found in Mark, Matthew and Luke, such as the three hours of darkness, the tearing of the Temple veil in two, the mockery of the passersby and of the thieves, but adds other, equally dramatic (and no less doubtful) passages of its own: notably, the passages relating to events allegedly witnessed by the Beloved Disciple, and embellishments relating to Jesus’ burial. The Resurrection narratives contain an abundance of supernatural appearances whose details are unverifiable and which therefore lie beyond the ken of history.

To sum up: one would be totally mistaken to infer that the Gospels are historically unreliable. On the contrary, once one takes the trouble to sift out the dross, one finds that a significant amount of historical gold remains. But the question we need to address is: is the nucleus of historically reliable material that one finds in the Gospel Passion narratives sufficient to build a strong historical case for the Resurrection? I would contend that it is not. And the biggest obstacle to such an enterprise is this: we simply don’t know whether, or how, Jesus was buried – or for that matter, who by, or when, or where. You can’t prove a resurrection without a burial.

7. Could new evidence boost the Gospels’ historical reliability?

Daniel Ang also asked me about an argument used by philosopher Alvin Plantinga to criticize evidentialist apologetics – namely, the argument from dwindling probabilities, which has been rebutted by Drs. Tim and Lydia McGrew:

By the way @vjtorley are you aware of Plantinga’s 2006 argument for dwindling probabilities against historical evidentialist apologetics? I just came across it and it sounds like a more sophisticated form of the “multiplication by priors” that you are implicitly using in this review.

My reply was very brief:

The overall thrust of the argument that Dr. Tim McGrew formulated (and that Dr. Lydia McGrew has also publicly defended) is that as the evidence for religious belief stacks up, and new evidence emerges, the probability of God’s existence will also be boosted. But that’s irrelevant to Michael Alter’s argument, since he explicitly lists the existence of the God of the Hebrew Bible as one of his background assumptions. You can’t get any fairer than that.

The Shroud of Turin: modern photo of the face, positive left, digitally processed image right. Image courtesy of Dianelos Georgoudis and Wikipedia.

Still, it occurred to me afterwards that the probability that the Gospel accounts of Jesus’ passion, crucifixion and burial are historically accurate could be boosted by the emergence of subsequent evidence for the events described in the Gospels.

The Turin Shroud is an excellent example of the kind of evidence I had in mind: a piece of cloth, purporting to be Jesus’ burial shroud, which appears to bear the imprint of a man who was scourged, crucified, and pierced in the side, in the manner described in the Gospel narratives. If the Shroud were authentic, it would dramatically undermine Michael Alter’s recent debunking of Resurrection apologetics. But is it authentic?

The answer appears to be: almost certainly not. Most readers will be familiar with the 1988 carbon dating of the Shroud, which placed it between 1260 and 1390 A.D. Shroud believers typically respond by citing a 2005 paper by Shroud expert Ray Rogers in the journal Thermochimica Acta, which alleged that the sample dated by experts wasn’t from the original Shroud, but from a piece of cloth that was patched onto the Shroud later when it was repaired after being damaged by a fire in 1532. However, most experts appear to be unconvinced by Rogers’ claims:

A Skeptical Response to Studies on the Radiocarbon Sample from the Shroud of Turin by Raymond N. Rogers, Thermochimica Acta 425:189-194, 2005 by Steven D. Schafersman, Ph.D., Consulting Scientist and Administrator of The Skeptical Shroud of Turin Website, February 8, 2005.

Private Internet Debate Challenges Ray Rogers’ Thermochimica Acta Paper: An exchange of views between Mark Antonacci and Dr. Thibault Heimburger.

Claims of Invalid “Shroud” Radiocarbon Date Cut from Whole Cloth by Joe Nickell (Skeptical Inquirer, March 2, 2005)

What’s more, two additional studies since the publication of Ray Rogers’ 2005 paper have backed up scientific claims that the Shroud dates from the Middle Ages:

2 studies show that the shroud of Turin in fact dates from the Middle Ages (Sciences et Avenir, April 22, 2015)

For a skeptical overview of the controversy relating to the Shroud, I would invite readers to have a look at Dr. Steven Schafersman’s highly informative slideshow and Professor Jerry Coyne’s articles:

A Skeptic’s View of the Shroud of Turin: History, Iconography, Photography, Blood, Pigment, and Pollen by Steven D. Schafersman (December 2003; updated January 2005). [This is probably the best skeptical source, overall.]
Pope Francis endorses the fake Shroud of Turin by Jerry Coyne (Why Evolution is True, November 29, 2014)
The Shroud of Turin: why religion is a pseudoscience by Jerry Coyne (Why Evolution is True, February 20, 2014)

Here’s another article that may be useful for readers who can access it (I can’t):

The Origins of the Shroud of Turin (History Today, November 11, 2014)

Finally, the most damning evidence against the Shroud (to my mind) is that it’s anatomically incorrect, as Gregory Paul has shown in this paper:

The Shroud of Turin: The Great Gothic Art Fraud — Because If It’s Real the Brain of Jesus Was the Size of a Protohuman’s! (The Secular Web, 2010-05-06)

Finally, here’s a balanced summary of arguments for and against the Shroud:

Shroud of Turin – 1st century relic, or medieval forgery? 10 arguments for and against by Sandra Rimmer.

To sum up: it’s beyond reasonable doubt that the Shroud of Turin is a fraud. In no way does it authenticate the Gospel accounts.

I know of no other archaeological evidence which has a direct bearing on the disputed events in the Gospel Passion narratives. If any reader knows of such evidence, I would invite them to present it on this thread. And now, over to you.