[Cross-posted to HNN: Revise & Dissent]
I’ve been working through the statistical analysis of the Tomb of Jesus data provided by the Discovery Channel (PDF download). They argue that the chances of their tomb not being the tomb of Jesus are one in six hundred. I looked through it and found two major errors. One is simple to see if you’re a historian, and the other should be blindingly obvious to a statistician.
The odds were calculated by Prof Andrey Feuerverger, Professor of statistics and mathematics at the University of Toronto. He was given names found on four ossuaries, boxes for holding skeletal remains, from a tomb in Jerusalem. The names were in English translation Jesus son of Joseph, Mary, Jose, a dimuntive of Joseph and Mariamne. It’s argued that Mariamne is Mary Magdelene and Jose is a diminutive of Joseph, which makes hime Jesus’s brother. What Prof. Feuerverger did was look at how often these names turn up in this historical record and then calculate the probability that these would turn up in a tomb somewhere.
He started by saying:
son of Joseph
Then he divided by four to allow for historical bias. Why four is never explained. This gives a probabilty of finding the names in one tomb as 1/608,000. Obviously there was more than one tomb in Jerusalem in the first century CE. Prof Feuerverger uses a figure of 1000. This reduces the odds of finding a tomb which could be said to belong to Jesus to 1 in 608. Generally social scientists get interested around odds of 1 in 20, so this looks impressive.
Unfortunately I can only assume Prof. Feuerverger was given absolutely no historical data. Look at the names in the tomb again. Jesus, Mary, Joseph – he’s not really essential but let’s leave that for now – Mariamne?
One argument of the documentary will be that Jesus and Mary Magdelene married and had a son. Another ossuary in the tomb is marked Judah, son of Jesus. But would a tomb really need Mary Magdalene in it to be considered the Tomb of Jesus? This borrows a popular idea from the Da Vinci Code, but the problem with the Da Vinci Code as a historical document is that it’s a work of fiction. The idea that Mary Magdalene was married to Jesus comes from the last supper and a rather feminine looking character on Jesus’s right hand. Unfortunately Da Vinci didn’t leave captions on the painting, so is this a woman?
While Da Vinci didn’t leave captions, he did do more than one painting. Look at the picture of this woman below:
The woman is in fact a bloke. The painting was called John the Baptist, and sure enough the traditional interpretation was that John the Baptist was in the painting on Jesus’s right rather than Mary Magdalene. So Mariamne shouldn’t be included in the calculations. What happens if you do that. The probability that the finding are due to chance fall to 1 in 3.8. So if you pay even a small amount of attention to the historical context the case suddenly becomes a lot less compelling. Prof. Feuerverger can be forgiven this, he’s a statistician, not a historian.
Unfortunately there’s an even more basic error. Look again at the names. Who would you expect to find in the Tomb of Jesus? Draw up a list. Every reader might have a different list, but I’m willing to bet that you all included Jesus in that list. Yet odds of Jesus son of Joseph being mentioned in the tomb are quoted at 1/190. Are they really that high?
Let’s imagine for the sake of argument that it wasn’t Jesus son of Joseph inscribed on the ossuary. Instead say it’s Brian. Now, if Brian’s name was on a ossuary and the name Jesus wasn’t found in the tomb would you really argue this was the tomb of Jesus? By definition Jesus’s name must appear, else no-one would make the documentary. What about Jesus, son of Joseph? That’s rarer, but Jesus/Yeshua is a very common name in Palestine in this period. They’re all over the place, so to have a convincing Tomb of Jesus at the very least the name Jesus son of Joseph must appear, to avoid confusing it with all the other Jesuses. What are the odds that a tomb containing an ossuary inscribed Jesus, son of Joseph would have the name Jesus, son of Joseph in it? One in one. Without that name appearing nothing would have been claimed.
The reason I find this so worrying is that it’s a well known problem in statistics. It has its own name, the Prosector’s Fallacy. Wikipedia has a nice example. Suppose you win the lottery, then you must have cheated. How do I know this? The odds of winning the lottery are one in 14 million in the UK. So the chances of you winning without cheating are so tiny they can be ignored. But couldn’t I say that about any lottery winner? It’s only because you’ve won the lottery that I’m saying it about you.
Equally the Discovery Channel are saying the odds that this tomb having the ossuary marked Jesus, son of Joseph in it are 1/190. But the only reason they’re saying that about this tomb is that they already know that it has the ossuary in it. Anyone can ‘predict’ the winners of a horse-race after the race has ended.
Taking this into account the odds of finding a tomb which could pass for a Tomb of Jesus aren’t 600 to 1. They go to 50 to 1 in favour of finding a tomb of someone who could pass for Jesus. Do you really need any of Jesus’s brother’s in the tomb to convince someone you’ve found a tomb of Jesus. Does it make sense to say “Historical Bias = 4” ?
The big mystery isn’t that there’s an ossuary marked Jesus, son of Joseph. It’s that there aren’t more of them. Sure enough Amos Kloner, who wrote the paper this speculation is based upon, knows of at least two other ossuaries with the name Jesus, son of Joseph. As he says, it’s a common name.
Now if someone can find the ossuary of Brian, that will be much more impressive.