Is Bayesian Statistics The Key To A More Scientific History?

In his 2012 book Proving History: Bayes's Theorem and the Quest for the Historical Jesus, historian Richard Carrier puts forward Bayesian statistical analysis as a new breakthrough in the study of history. He even goes so far as to say that any method that cannot be reduced to this theorem is invalid and should be abandoned. In On The Historicity of Jesus, he proceeded to apply this approach to the question: did Jesus exist? and came to the answer that there is at best only a 32.36 probability that He did. Is he right?

 

I would have to read about Bayesian studies and what that is about to understand what it is, and why would anyone want to say something does not exist if it does not fit a particular theorem.....That to me, does not feel right....but again, I need to know why...

I do believe someone by the name of Jesus did exist.

 

There's an old saying: "Figures don't lie, but liars can figure." It seems to me that Bayesan approaches are particularly vulnerable to that problem. Dawkins warned us that Bayesian approaches are only as good as the original numbers fed in. When these are purely subjective estimates instead of measured quantities, the results are vulnerable to the GIGO principle (Garbage In Garbage Out) which is well-illustrated by Carrier’s execution. As somebody said "The probability of meeting a T-rex in your street is 50%.. You either meet him or you dont"

Bayesian approaches to religion attained a high level of notoriety when Stephen Unwin used it to prove the existence of God (or at least a 67% probability that He exists.) Theologian Richard Swinburne picked up the ball and ran with it, using Bayesian conditional probabilities to prove not only the Existence of God, but the Incarnation and Resurrection of Jesus. So I guess it’s only fair that an atheist should use it for the other side. Like the Christian Bayesians, though, Carrier relies on his own subjective estimates of the various conditional probabilities involved. Although he acknowledges that “To ascertain these probabilities with the kind of vague and incomplete data typically encountered in historical inquiry “ requires special efforts, Yet he is confident he’s up to the task. Given his reputation as a militant atheist, I think there’s reason to be skeptical, especially in light of some of the questionable “facts” that he cites.

I think there's potential for tightening up the methodology to reduce the problem of human "guesstimates". Maybe assignment of values could be done by panels of experts as a control on individual biases. Of course, where religious topics are involved, it probably wouldn't help if the panels consisted of all Christians, or all atheists, and they should have good reputations for objectivity and peer reviewed articles of their own. Rigorous peer review might accomplish the same purpose, but that, of course, has its limitations, considering the peer-reviewed stuff that does get published. Maybe both. But so far the claims tht Bayes's theorem is the Grail of historical analysis have yet to be proven.

 

To me it seems like a pointless and very reductionist way to approach the study of history, and from what you wrote above, one that's deeply flawed on a methodological level.

More generally, I don't think statistical models are really of much use in the study or theory of history. There may be some applications of course within certain areas where data is available which is fully reliable. So one can look at trade, economics etc in that way.

But to come up with a result like there's a 32.36 probability of Jesus existing is really baloney in my opinion. I can't see any value at all in such results.

 

Existing seems to be likely, after all his words have survived for an amazingly long time. I don't doubt Buddah or Mohammed existed.

Son of God? Clearly that's another matter. If you don't buy it, be glad you live in this century.