I have been reading your review over the weekend and I think I understand most of it now
. One thing that surprised me was that if I didn't know any better I would have thought I had simply structured my review around section 2 of yours... as far as I can tell we point out exactly the same things:
1) The problem of inflation of errors in BT (section 2 beginning)
2) That h and ~h are defined as specific hypothesis and not simply as "existence or not" and that they are not (as would be implied by the notation and subsequent computations) each others negation
3) That the Rank-Raglan prior computation appears to ignore this
4) ...and we are both puzzled by the quote on p245 that it does not matter how Jesus came to be a RR hero because it never happened to historical figures
5) we both immediately think that the principles can be applied to other reference classes to give inconsistent results
Hahaha! I actually thought the same thing initially: that you were working from my review. But my comments on Bayes Theorem were also based on others with BT experience whom had looked at Dr Carrier's usage and had come to the same conclusions as you, so it isn't really surprising we both reported the same problems.
I think I agree with Carrier that we tend to argue in Bayesian terms anyway, so BT can be used for questions of history. But one important comment I included in my review was by a poster called arcseconds, whom said that Carrier should have shown BT working on non-controversial examples first, to show some kind of calibration had been done, before using it on something as controversial as the historicity of Jesus. Carrier has "Proving History", but he barely covers any actual situations, to demonstrate "proof of concept" when it comes to questions of history.
A point about the Gospels you highlight:
Since the Gospels, Paul and other early epistles talk about a crucified Jesus Christ whom was a descendent of David, I believe that even if we discount nearly all of the Gospels' content as unverifiable, then the mere existence of the same character appearing in the epistles of Paul and in the Gospels needs to be evaluated. But this, again, is an argument for another day.
There are two things that worries me. First that it seems like Carrier ends up discarding everything in the gospels as historically irrelevant, just with the exception of the RR-related information which turns out to be hugely historically relevant.. That seems a bit too convenient.
Secondly that the gospels is an excellent illustration of how the specific form of ~h helps explaining the evidence. Since we assume in ~h that Christians believed or taught a historical Jesus (at some point) it would seem that given ~h we can explain *any* amount of historical details about Jesus found in christian writings because ~h already assumes that was what Christians actually believed.
Yes, it is a bit like defining h
to include "Paul is not interested in the historical Jesus, so won't include details about him", then pointing to Paul as evidence for h
. 100% expected on minimal historicity! But I'm not sure what the alternative is. Shouldn't h
be tailored to the evidence?
timhendrix wrote:I think it is a bit odd Carrier haven't responded directly to your review considering how many other reviews of much worse quality he has discussed, but perhaps it is a time issue.
I don't expect Carrier to respond to my review. I really am an amateur on the topic, so I don't think he even needs to (though I'd be interested to know whether I am a liar or whether I am insane as well! But I guess since I don't have a PhD, I'd only be a liar
). But I expect that when experienced scholars look at Carrier's work, they will see the same problems as I did, much like you and I saw many of the same problems with Carrier's use of BT. Carrier is going to be hammered for his ill-use of BT, his weird definition of 'euhemerism', his misreading of 'the Ascension of Isaiah' and Plutarch's 'Isis and Osiris', etc. And rightfully so.