A variety of early online reviews have appeared of my new book On the Historicity of Jesus (including Amazon reviews, to which my responses, if any, will appear there in appended comments). I will blog a series on them this week. If you know of any reviews I don’t cover by the end of the first week of July, post them in comments (though please also remark on your own estimation of their merits).
One of the early reviews posted will be published in the Journal of Religious History later this year, by Raphael Lataster, a doctoral student in religious studies and a historicity agnostic. His review is accurate and positive. But he states one criticism:
My only real criticism is that the minimal mythicist theory fits the evidence so perfectly which some may see as suspicious. This could be because the theory is simply true, or because it has been carefully crafted for this purpose, and suffers from a lower prior probability as a result (cf. apologists who inadvertently damage their hypotheses by inventing evidentially-unsupported excuses to counter the evidences of evil and hiddenness, in arguing over God’s existence). It is up to historicists, however, to show that this theory is inherently implausible.
What he means is that he is uncertain how much I have gerrymandered the theory (see Proving History, index “gerrymandering”). That’s a valid concern, and one might not notice the care I took to avoid it. The basic idea is that if we have theory h and a collection of excuses c, the addition of c may make h fit the evidence well, but it necessarily reduces the prior probability of h, unless the contents of c are highly probable independently of h or e (the evidence h must explain). And that means b, the background evidence, must on its own entail that c is highly probable. I explain the concept further in PH.
To see how I dealt with this, two examples will serve as models. You can then see how I used the same technique throughout the book, especially following the implicit guide I provide in OHJ, pp. 606-16, where I show how b features in assessing the prior probability of any assumptions relied upon in h or in h’s connection to e. An obvious example that I didn’t bother to mention but which illustrates the concept is that b includes the fact that Paul is biologically human. Thus any assumptions about what Paul could or couldn’t do that make e probable on h (assumptions that are not themselves in h and thus would be c) are highly probable, because “Paul is a biological human” is highly probable, and any assumption a that is entailed by being biologically human is therefore as highly probable (technically more probable, if a would also be true of people who aren’t human, but the increase would be trivial, because the probability he wasn’t human is trivial).
So to see this at work, look at how I treat the five elements of the minimal mythicist theory in OHJ, pp. 52-55 (which I explain the relevance of again, crucially for the point I am making now, on pp. 246-48), and thus explain their high priors relative to the prior probability of mythicism. Because there the question is not what the prior probability of those elements is independently of mythicism but what they would be if mythicism were true; their actual independent priors would then equal that probability times the prior probability of mythicism as a whole. For example, if the prior for mythicism were 1% and the relative prior of an element I employ is 99% then the independent prior probability of that element would be 1% x 99% = 0.99% or basically still just about 1%, or technically that plus its prior on historicity if it can also be true on historicity, although I designed the hypothesis so few elements of it would be, so I could get a proper dichotomy to test (that is actually required by the logic of Bayes’ Theorem, since it must be the case that P(h|~h) = 0).
In that section, I show that each element I adopt is almost certainly true if mythicism is true, because the alternative (another mythicist theory with a different element) is too improbable to credit. Therefore those elements share nearly the whole prior probability space for mythicism and therefore do not reduce the prior probability of the theory I am testing (except too trivially to show in the math, as I there explain).
The second example is my treatment of the question of the flesh created for the cosmic Jesus being Davidic and Jewish, OHJ, e.g. p. 581. I show that independently of whether mythicism is true or not, the scriptural requirement that the last messiah (the one who would eternally rule) must be of such flesh is so overwhelmingly strong that it would be effectively impossible (i.e. extremely improbable) that any Jewish innovator constructing a cosmic Jesus would do so without working it in. Because otherwise their scheme would not fulfill prophecy and thus would be deemed false. To sell it, it had to fit prophecy. So we could have predicted in advance that a final cosmic Jesus would be imagined to have been Davidic and Jewish. Therefore, we do not have to gerrymander that as an assumption. It is an assumption entailed by the background evidence. It therefore does not reduce the prior (by any significant amount).
As for those, so for all others. It is possible I overlooked something, but I was very careful, so if I did, someone else will have to find it. This would not include, of course, occasions where I speculate but don’t use that speculation in my argument–if I don’t use it in determining any probabilities, it is not relevant to my probabilities, and thus not relevant to my conclusion. Lataster would not make that mistake (he knows how logic and Bayes’ Theorem work), so I only mention it here to forestall others who might. Criticizing my conclusion by criticizing a speculation I make but don’t use is called a straw man.
For a complete list of my responses to critiques of OHJ, see the last section of my List of Responses to Defenders of the Historicity of Jesus.