Inside the Numbers of the Talpiot Tomb
April 10, 2008 — JimMark Elliott and Kevin Kilty have written an important essay on the topic by that title and you can download it (in PDF) here. It’s very much worth a read- especially by those who continue to assert claims for the Tomb and its contents which simply cannot be sustained. With thanks to Mark for passing word along of it.
UPDATE: Randy Ingermanson has also recently done further work on the stats. You can read his study here. Thanks to Randy for the tip.
Being worse at math than Gerhard von Rad, all such arguments are outside my realm of experience and capability. So, I’ll just wait for the Second Coming and when Jesus doesn’t return to visit the family plot at Talpiot, I’ll know even more certainly that it’s not the right place.
April 10, 2008 at 3:28 pm
Ultimately, without Mary Magdalene, there is no statistically significant evidence that this tomb belonged to Jesus. That being said, this does not mean you cannot find a modestly large probability value.
Ultimately, the number hinges upon the assumptions one makes before even beginning to look at the evidence. Depending upon which ones you take, you can go as low as <0.001% or as high as 48%.
This is why I disagreed with the following comment from Elliott and Kilty:
“We think it is unwarranted to include assumptions concerning the existence of a rock cut tomb and an ossuary inscribed Judas son of Jesus in calculating name frequencies located in the Talpiot Tomb. [Randy] Ingermanson has not demonstrated that his negative assumptions about these issues actually exist.”
Four responses are appropriate:
(a) Assumptions about the likelihood of Jesus being interned (ultimately) in a tomb and the likelihood he had a son are inevitable. So also are many other assumptions. You cannot avoid them. Elliott and Kilty made their own assumptions on these matters last year. They may not call them assumptions, but that’s what they are.
(b) These are assumptions. If Ingermanson could demonstrate that they “actually exist,” they would cease to be assumptions!
(c) Ingermanson does not really make “negative assumptions.” Minimally, this is a grossly oversimple characterization of the sophisticated and impressive work he has completed.
(d) Ingermanson did not make his assumptions in a vacuum. Instead, he consulted with scholars and cited scholarly articles by academics like Pfann, Rollston, and Magness. In fact, he explicitly says again and again that he defers to the judgments of the experts.
After a long time reflecting on this subject, I have come to lament the introduction of statistical analysis in this topic. Back when the Talpiot discussion started, I assumed that statistical analysis would be of probative value.
Upon reflection, I think this might have been an error. I would certainly say that the introduction of statistical analysis has not, on balance, clarified this issue. This is not to say that statistical analysis is not possible here. It is only to say that it is much more complicated than it appears at first blush.
Ultimately, the problem has nothing to do with the math itself. It has to do with how the author applies it and how the reader interprets it. If either is not careful, the math can give a false sense of precision.
On that note, I would encourage people to read Randy Ingermanson’s latest essay.
lib.stat.cmu.edu/aoas/99g/supplement.pdf
His essay is the only one I have read that handles this complicated situation with sufficient care. And I say that as somebody who took a few stabs at this subject himself.
Even if you disagree with Ingermanson’s conclusions, look at his methodology carefully. This monumental effort was his response to the problem of all the assumptions that must be taken, and therefore our lack of solid factual knowledge about Jesus. Ingermanson judged this to be the required amount of labor that is required to produce a probability estimate. I think he is on the right track.
And therein lies the problem with the mathematical discussion to date. He might not be correct, but he’s the only one I have seen doing the kind of work that is required to answer this kind of statistical question.
April 10, 2008 at 7:05 pm
Mark Elliott and I have read Randy Ingermanson’s latest paper, and we feel it is a valuable and impressive work. Indeed, he has made a very careful analysis. We only disagree with two of his points. First, he compares his statistical results directly to ours even though there are several good reasons why the two are not exactly comparable, including that we made only a single calculation meant as an example. More detailed analyses are coming from us. Second, while we examine name evidence only, Ingermanson includes very speculative factors such as Jesus’ paternal status and whether or not he came to rest in a rock-cut tomb. In effect his analysis adds two variable “controls” to the outcome. Perhaps this is “handling the situation with sufficient care,” but it is also possible that it simply adds to the feeling that Cost expresses about the declining usefulness of probability in this instance.
Finally, we disagree with Cost, himself, about the Mary Magdalene issue. We do not agree that only the name Mary Magdalene provides evidence regarding identification of this tomb. It seems reasonable to Mark and I that simply having a sufficient correspondence between ossuary inscriptions and names of persons in Jesus’ family is interesting evidence in its own right. If it turns out somehow that the Mariamne inscription can be tied to Mary Magdalene then we would consider it strong additional evidence that this is the tomb of Jesus. Perhaps even overwhelming evidence.
There is a strong suggestion in Cost’s remarks that this issue is settled and no one may disagree with Ingermanson because of the care of his analysis. With all due regard to Ingermanson’s efforts, we doubt this is a closed case.
April 10, 2008 at 10:18 pm
“There is a strong suggestion in Cost’s remarks that this issue is settled and no one may disagree with Ingermanson because of the care of his analysis. With all due regard to Ingermanson’s efforts, we doubt this is a closed case.”
I made no such suggestion.
April 10, 2008 at 11:26 pm
I agree with Kevin’s statement above that says (if I understand correctly) that we are not that far apart. I believe we agree on all the following points:
1) We agree that Bayes’ Theorem is an excellent way to approach the problem.
2) We agree that the “Mariamenou” inscription is not to be automatically equated with Mary Magdalene. (It seems that Andrey Feuerverger also now agrees with this position.)
3) We agree that the inscriptions “Jesus son of Joseph,” “Maria,” and “Yoseh” are names that one might well expect to find in a Jesus family tomb and therefore provide positive evidence for the tomb hypothesis.
I am not fully certain I understand the points of disagreement that Kevin lists. He and Mark Elliott have written a paper that estimates a probability of 48% for the authenticity of the tomb (along with an alternate probability of 6% if one changes an assumption about Yoseh). I did not mean to imply that this paper is their final word on the subject. They have produced an ESTIMATE, much as I have produced one. I will be eager to see their future work, but right now, that one estimate is the only number of theirs that I could have mentioned in my paper, and I did not want to ignore their work because it gave me some ideas for my most recent calculations. No offense was intended!
As for the two historical issues, I am not certain why these should be called “speculative.” I know of no historian who believes that Jesus was as likely to have had a son as a randomly chosen man of Jerusalem. Nor do I know of any historian who would argue that Jesus was as likely to be buried in a rock-cut tomb as a randomly chosen man of Jerusalem. If Keven can give me a list of historians who believe these, I’ll be happy to talk with them and modify my calculations as appropriate to reflect the current thinking of historians and archaeologists. But I believe that my calculations are in line with current thinking on these issues. The folks who signed the Duke University statement took a far harder line than I allowed in my calculations!
Let me reiterate that I personally don’t think the matter is settled yet. I believe there is a lot of work still to be done, and that it is worth doing. That is why I’ve gone on record as saying that I disbelieve the “virtually nil” statement of the scholars who signed the Duke statement. The only way to do a calculation that gets a “virtually nil” result is to assume as fact the bodily resurrection and ascension of Jesus. In that case, the p-value is zero. But my calculations, like Kevin’s and Andrey Feuerverger’s, assume the opposite.
I look forward to seeing any new calculations by Kevin and Mark (and also Andrey, if he is still willing to pursue this topic that attracts so much scorn).