James McGrath has reviewed my book Proving History. We’ve argued before (e.g. over claims Bart Ehrman made), so there is backstory. But his review is unexpectedly kind and praising at points, and he likes the overall project of explaining the underlying logic of history as fundamentally Bayesian and making productive use of that fact. He does conclude with some select criticism, though, and that is what I will respond to here.
General Points of Agreement
His general point he summarizes thus:
I felt like I had already read two very different books. One of them is lucid, focused, methodologically rigorous, and full of powerful insights. The other reads like a case study in what happens when someone ignores their own sound methodological advice in practice when turning their attention to the figure of Jesus, an exercise which has led even some of the brightest minds in history to write some absolute nonsense, and many more to write things that were merely unpersuasive. I have read many other books where the same criticism could be made that I will here make of Carrier: the methodology is brilliant but the application is problematic.
As a generalization, I wholeheartedly agree with him. But do I make the mistakes he alleges? No. And showing why I do not makes for a good teaching example on how to understand and apply formal logic and Bayesian methodology to history (and to Jesus studies in particular). So I’m quite grateful for his thoughtful review.
The first general point McGrath already well understands, which is that even though Bayesian method can be misused or mistakenly applied (as he rightly says, “garbage in, garbage out”) the great value of it is that it makes this easy to detect, demonstrate, and critique (as McGrath says, “Carrier makes a strong case that Bayesian reasoning can help expose when this is taking place”). In other words, Bayesian methodology makes progress more accessible by making bad arguments more visible, and providing an effective way to expose or correct them.
I fully grant this applies to everyone and welcome its use to correct my own errors, in order to revise my conclusions toward what is genuinely more defensible. I discuss this fact several times in the book, as one of the method’s greatest merits. And indeed, a main mission of my next book (On the Historicity of Jesus Christ) will be to start this kind of productive debate, rather than to end it (in the matter of whether we can really know Jesus existed as a real historical man). But alas, McGrath never once actually uses any Bayesian argument in his critique. Thus he abandons the very tool he just agreed we need to deploy.
McGrath also correctly gets my take on mythicism generally. As he says (and I say much more in support of this point than he singles out for example):
If the internet mythicists of today were to take this advice to heart, they might work on trying to actually publish scholarly works in appropriate venues, instead of complaining about their being dismissed in a manner that Carrier rightly says is appropriate when people without relevant expertise try to bypass peer review and scholarly discussion and promote their ideas directly to a public ill prepared to assess them.
Particularly as there really is plenty of crazy coming from mainstream scholars even in mainstream venues (I’ve mentioned my bafflement at Robert Eisenman’s bizarre theories about the origins of Christianity before, for example, and I needn’t mention that pretty much no one in the field buys them, either), so it’s not like mythicism would really be so out of place there (it’s just dogmatically opposed with more fury; theories like Eisenman’s apparently get a pass because at least they still have a historical Jesus in them).
Dennis MacDonald’s theory that Mark is a literary emulation and transvaluation of Homer is among those McGrath might reject (as it’s often considered a radical thesis, though IMO it’s probably substantially correct, even if not in every detail), yet I’m sure even he would extend it intellectual respect, as it was argued through proper channels and methods. That is indeed what mythicists need to do. And I am working on doing that.
Ignorance vs. Embarrassment
McGrath starts by confusing ignorance with provisional knowledge:
For instance, the fact that we do not have a complete and comprehensive background against which to determine what would have been embarrassing to early Christians is by Carrier’s own statements irrelevant, since a Bayesian approach assesses probability based on the evidence we have. If new evidence comes to light, we change our analysis if necessary, but we ought to proceed based on the evidence we have, as Carrier acknowledged elsewhere in the book.
This confuses two different principles, not knowing something and believing we know something (when we happen, unbeknownst to us, to be wrong), which I tease out in Chapter 2 with examples about Julius Caesar shaving or having an imperial moonbase (among other things: pp. 26-29), discussing Axiom 5, “any argument relying on the inference ‘possibly, therefore probably’ is fallacious.” Related is also Axiom 3, “overconfidence is fallacious; admitting ignorance or uncertainty is not” (p. 23). And I discuss this specific problem for Jesus and the origins of Christianity extensively under “The Problem of Ignorance” on pp, 129-34 (and there citing other scholars in agreement with me), and it’s a principle even Bart Ehrman himself affirms (while simultaneously ignoring his own advice).
Ignorance cannot be replaced with assertions of knowledge. If we do not know what was embarrassing to all first century Christians, we do not know what was embarrassing to all first century Christians. Full stop. Anything further violates Axioms 3 and 5. Sometimes we do have evidence (direct or background), and then we are not ignorant (even if we might be wrong, we must work with the information we have). I give examples of this for the criterion of embarrassment, and discuss what is necessary to make such an argument work, throughout pp. 158-69. Otherwise, our assumptions just fail to be generalizable, e.g. the usual intuitions you hear from Jesus scholars would entail the castration of Attis should have been too embarrassing for his followers to invent, yet they obviously invented it anyway. Thus, evidently our intuitions on this point just suck. We therefore cannot rely on them.
This does not mean no argument can be made. I detail just how to make one work. But that it can be done in principle is not the same thing as it having actually been done. I await a truly sound argument from embarrassment to the historicity of Jesus. Once a scholar knows how to do that non-fallaciously, if there is still something left over that still yet points toward a historical Jesus, that would be progress in my view. Hence I welcome it. Proving History lays out how that can be done. If it can. And that depends on the actual evidence we have. So we’ll see.
McGrath also misses the whole point of my argument on pp. 162-68, where I mathematically demonstrate that apparently embarrassing stories in the Gospels have a low prior probability of being true (exactly contrary to our intuition). He reacts to this by saying:
Carrier claims that anything included in an early Christian work must have been worth including and therefore could not have been embarrassing. But by this token, any detail in any text seemed worth including to its author, and could therefore be deemed suspect.
That’s not my argument. I carefully explain that my argument applies to (1) friendly (2) propagandistic texts (3) that lack internal indications of embarrassment–that’s three criteria a text has to meet in order to fall to the analysis I provide (see pp. 162-63; note also my preliminary demarcation between sober history and propaganda on pp. 160-61; and my discussion of actual examples of demonstrated embarrassment in the Gospels on pp. 131-34).
Hence McGrath falsely generalizes a finding I reach for that subset of texts, to all texts whatever. That’s incorrect reasoning. Moreover, he fails to interact with my mathematical demonstration of the point: the only way he can conclude I am wrong, is to find where I go wrong in that math. But the math is pretty solid. So the conclusion is well established: the Gospels would not likely include any embarrassing material unless it served a purpose that would just as easily warrant its fabrication in the first place (a general point I had already discussed on pp. 134-37).
By contrast, McGrath refers to my remark elsewhere that “the absence of any likely or discernible reason for the story as we have it to have been made up (by the historian or his source),” which is one of several criteria (note: one of several; it’s not always going to be sufficient on its own) that we can apply to warrant a priori trust in a report (p. 253). Yet note that my section about the Gospels explains why there are several reasons for stories in them to be invented. Thus, we are not in a condition of “the absence of any likely or discernible reason for the story as we have it to have been made up (by the historian or his source),” so McGrath is misapplying the latter rule. In fact I show that we have a general reason to expect all the stories in the Gospels had such a reason for their composition (or that most did: since my argument is probabilistic), even when we do not know what it is specifically.
A mundane example (not connected to embarrassment) is the number of fish counted in the miraculous catch in John 21:11 (the number is 153). Only a rank fundamentalist would believe that this actually happened, and Peter actually counted out 153 fish, and that this meaningless detail was transmitted over half a century of oral storytelling for some indiscernible reason. No, there was some reason for this number (and indeed the whole story) to be fabricated. That is obvious even without knowing what that reason specifically was (I know what some of those proposed reasons are; my point is that knowing any of them isn’t even necessary to reach this conclusion). And as in this case, so in all cases of alleged embarrassment–as I argue and demonstrate at length.
McGrath has been burned on this before (see recap), so he clearly struggles to parse his words this time, trying to thread the byzantine needle of dogma he wants to maintain. I’ve already discussed the evidence extensively in The Dying Messiah Redux (fully revised and updated) so I won’t reiterate the evidence and argument laid out there. McGrath says this time:
[Carrier] also continues to discuss whether “messiahs” (i.e. anointed figures like kings and high priests) could die – something that no one to my knowledge has ever disputed – rather than focusing on the smaller set (again, as his Bayesian approach requires) of the specific type of anointed one that the earliest Christians claimed, namely a Davidic anointed one, the figure expected to restore and rule the kingdom of his forefather David. There is no text from this period which suggests that the individual of Davidic descent who was expected to restore the dynasty of David to the throne was also expected to fail before achieving that by getting executed by foreign rulers.
Notice how ultra-specific he has to make his definition of “messiah” to create any kind of objection to what I actually argue. He claims this is what a Bayesian approach requires, but that’s exactly backwards. Creating hyper-specific definitions in order to avoid the consequences of generalizations you don’t like is not sound reasoning. In fact McGrath makes no effort to construct an actual Bayesian argument for his conclusion contrary to mine. And that’s the poster child for abusing Bayesian reasoning: claim your argument is Bayesian, and then do nothing whatever to show how or why it’s Bayesian. Don’t do that.
McGrath is also confusing what I argue in the book (that theories like this have to be properly examined and ruled out before being dismissed) with what I have argued elsewhere (that this theory is probably true). He apparently has no real argument against what the book says, and instead insists the problem has been sorted, when in fact it has not. We have several texts showing that the Jews did indeed imagine dying messiahs presaging the end times. That’s all the Christians needed. Because all one has to do is combine that, with the idea of restoring a new world order (as they believed Jesus would do in the future), to get Christianity. And as I’ve explained before, we simply cannot rule out this possibility. Not on present evidence, which is far too scarce (as I and other scholars have pointed out: again, see my remarks above).
In Bayesian terms, as I’ve explained before, the probability that someone would invent the kind of messiah McGrath is talking about is zero (because everyone would notice the total lack of military triumph over the Romans, therefore such a thing could not be invented), which entails that if Christians invented their messiah, the probability that he would be a messiah whose victory is invisible, like the messiah who dies in the timetable of Daniel 9:24-27 (which “predicts” the last messiah’s death before an imminent end of the world) in order to perform a celestial temple-cleansing (as in Hebrews 9), is virtually 100%. Indeed, an imaginary dying-and-rising messiah is precisely the kind of messiah Jews were likely then to invent, if they were to invent any, given the growing multicultural fashion for saviors of just such a kind.
McGrath seems to want to make an argument from prior probability, such as that Jews tended not to do things like that, but that is where the problem of ignorance kills his argument: we have no business claiming to know what Jews did or did not tend to do, given that there were dozens of sects at the time, and we know barely anything about even a handful of them, much less all of them. There is simply no data for McGrath to build a reference class that gives him the kind of “low prior” he seems to want.
This would be the wrong way to go about it anyway, since (a) we have to start with a prior probability based on the best data-set we have for building a reference class (Proving History, pp. 229-53, esp. 233), and (b) the evidence establishes the Jews of this period were actually very innovative, and minor sects very iconoclastic (as in, ready to abandon, reinterpret, or turn upside down views pressed by the elite orthodoxy), so the prior probability of an innovation of this kind isn’t going to be low. We have to look elsewhere to argue against it (if we are bent on doing so).
Thus, McGrath misses the point when he concludes:
The only way to argue that this was more likely invented is to say that anything in a text could be invented – which is true, but it is not a valid piece of Bayesian reasoning. This is true of all texts, in theory, and an “explanation” that can fit any evidence has little or no explanatory value in and of itself.
This is all wrong. This is in no sense “the only way” to argue the gospel was more likely invented. And I certainly never endorse the argument “could be, therefore was.” In fact I explicitly denounce that reasoning (Axiom 5, pp. 26-29). To the contrary, I explain what “the only way to argue” for ahistoricity is on pp. 117-19 and 204-205, and McGrath does not mention what I say there at all.
McGrath instead goes on to say:
The question is whether it is more likely that a group who followed a messianic claimant of the sort we know appeared in this period in history made sense of his failure by giving his execution of positive interpretation, or whether it is more likely that a group invented a purely fictional failed Davidic anointed one and set about trying to persuade their contemporaries to believe in him. To every scholar with a substantial background in early Christianity or ancient Judaism, the choice seems a no-brainer.
This is precisely the problem: scholars refuse to properly examine the alternatives with a proper method because of their fallacious, methodologically-unsound, dogma-prone assumptions. It’s “a no-brainer.” Therefore we don’t even have to make any valid argument against it. We just, uh, “know it in our bones” or something. That’s the exact opposite of Bayesian reasoning.
The question indeed is precisely what McGrath describes: what best explains the evidence (all the evidence) we have, that an actually executed man inspired it all (as the canonical Gospels claim), or revelations of a self-sacrificing archangel inspired it all (as Hebrews 9 and the earliest credible redaction of the Ascension of Isaiah claim)? That cannot be answered from the armchair, much less with dogmatic assumptions about the ideological bullheadedness of Jews, who (I guess we’re supposed to believe) could never innovate their way out of a paper bag, much less a real socio-theological crisis like the Roman occupation.
McGrath shows his enslavement to dogmatism most explicitly when he says:
If Carrier’s method can take such relatively straightforward mundane statements [that Jesus was born of a woman, under the law, and of the seed of David], and claim that they more likely refer to things that transpired in a celestial realm, then Carrier will have showed that his method does nothing to add rigor to historical study.
Here McGrath asserts dogmatically that such a conclusion can never be demonstrated. Before he has seen any argument regarding it. This is the opposite of Bayesianism. If it is a priori impossible for McGrath to be wrong about this, then he is cutting himself off from ever discovering his own incorrect conclusions. No method could ever persuade him, because, well, he can’t possibly ever be wrong. That’s dogmatism, not scholarship.
The first rule of being a Bayesian is: you can always be wrong. McGrath has failed to learn so basic a lesson as that. Bayesianism tells us that all arguments can be reduced to a dispute over just three numbers. Until he is willing to actually engage such a dispute, his assertions of certainty are unfounded, and exemplify all that is wrong with Jesus studies today.
Nazareth vs. Scholarship
McGrath makes the following bad arguments regarding the problem of the origin of Jesus being called a Nazarene (besides not engaging with any of the actual arguments and scholarship I discuss on the point, pp. 142-45, which is already an alarming failure to properly engage with my work, and another example of his dogmatism), which I have made easier to reference by numbering them:
On the one hand, [Carrier's] use of Rene Salm’s self-published armchair opinions (p.142) is at variance with  his emphasis on both the need for relevant expertise (Salm has no qualifications or experience relevant to either the archaeological or the linguistic matters on which he comments, and regularly gets things wrong), and  the appropriateness of ignoring fringe claims with little probability of being correct when doing Bayesian calculations.  Carrier’s own linguistic deficiencies become apparent when he treats Irenaeus’ account of the symbolic meaning given by Gnostics to words as indicative of the actual meaning of those words (p.142). While there is a legitimate line of investigation about the relationship between the various forms of Nazareth and terms allegedly associated with it (in particular Nazorean),  without the relevant background in Semitic linguistics, the matter is not going to be given a plausible explanation, and because of his lack of background and expertise in this particular area, Carrier seems unable to discern when Salm is offering him bunk packaged to look to the non-expert eye like scholarship.
Let’s take these in order:
 First, McGrath ignores my explanation in Axiom 12 (p. 36) that we can cite non-scholars when they make arguments we (as experts ourselves) verify are worth examining. That Salm gets some things wrong is not a valid argument to dismiss everything he has said on this point, not least because his argument has appeared in a peer reviewed field-relevant academic publication (Bulletin of the Anglo-Israel Archaeological Society 26 , with rebuttals by his peers), which McGrath just got done insisting mythicists should do. So here, Salm does what McGrath wants mythicists to do, but then that’s apparently not good enough, and so even his properly submitted peer reviewed work is to be dismissed without even examining it. McGrath is the one not living by his own principles here. My use of Salm is entirely consistent with mine.
Second, McGrath strangely fails to mention that I back Salm’s work up with that of other peer reviewed articles and bona fide experts, including Eric Laupot, J.S. Kennard, and Robert Price, as well as actual ancient evidence (none of which McGrath mentions or addresses). Moreover, McGrath fails to mention that I explicitly state that “I do not agree with all the theories of either Salm, Kennard, or Laupot” (p. 142). In particular, I believe, and have said this repeatedly in public, that Salm is wrong to conclude there was likely no Nazareth in the time of Jesus, and I mention no such argument in my discussion in Proving History. So if McGrath is trying to impute that to me, that’s dirty pool, and a plain violation of Axiom 12 (pp. 34-37), deploying the very “baggage fallacy” I explicitly warned readers against. McGrath should be a better student than that.
 McGrath says that some unstated theory (note that he does not say what) is a “fringe claim” that has “little probability of being correct when doing Bayesian calculations,” but then does no Bayesian calculations to show this. Once again he is abusing Bayes’ Theorem by claiming it gives us a conclusion, without showing how, or even giving any indication that he knows how. He just dogmatically asserts that the conclusion must surely go his way, even though he makes no attempt to find out. Indeed, he doesn’t even state what the claim is that he can somehow show by some undescribed Bayesian analysis has “little probability.” That is not scholarship. It’s argument by dogmatic assertion.
 McGrath betrays his failure to read endnotes when he says “[Carrier] treats Irenaeus’ account of the symbolic meaning given by Gnostics to words as indicative of the actual meaning of those words.” I wrote (on p. 142) that Ireanaeus “believed” Nazaria meant Truth in Semitic, and in the associated endnote I state (n. 49, p. 315): “We do not know of any such word; more likely the actual derivation was from something else, like natsar, as perhaps “keeper of secrets” (i.e., the mysteries; by derivation from, e.g., Isaiah 48:6 and 42:6), which Christians proclaimed (and thus equated with) ‘the truth’.” Thus I very explicitly did not treat Irenaeus’ observation as “indicative of the actual meaning of those words.”
 But I did cite the Gospel of Phillip as confirming what Irenaeus thought, a fact McGrath also fails to mention, despite the fact that this is the only claim I cited Salm for, and it is unmistakably true. So McGrath’s red herring fallacy, claiming Salm gets facts wrong, therefore he got this fact wrong, looks borderline dishonest to me. There is not a single thing I cite Salm in defense of that isn’t demonstrably correct. Evidently that ruins McGrath’s storyline, so he won’t think to tell you that. Instead, McGrath says Salm offered me “bunk.” Well, no. What I cited him for is fact, not bunk. And I got the linguistic details right, despite McGrath claiming I didn’t.
By contrast, McGrath commits an obvious fallacy when he argues from “the ‘prophecy’ Matthew appeals to here [predicting Jesus would be a Nazarene] is not found in any known Scripture” to the conclusion that Matthew invented that prophecy. That’s a non sequitur. Matthew would sooner invent a prophecy naming a town, not an epithet that has no evident connection to a town, and then try to pass that off as predicting a town (and Matthew, a Torah-observant Jew making arguments directed at fellow Jews, would not be so stupid as to try to pass off an invented scripture, much less expect such a thing to be a convincing argument to them). We know for a fact the Christians used scriptures we no longer have, and scriptures we do have but with variant readings and verses we don’t have (the epistle of 1 Clement is full of examples of scripture citations we can’t identify, for example; and many more examples can be found throughout the second century church fathers). Thus, McGrath cannot argue from “he cited a scripture we don’t have” to “no such scripture existed.”
It’s ironic that he follows this ignorant fallacy with the conclusion that I don’t know what I’m talking about, when obviously he’s the one who is not up on basic field-relevant facts like this.
Jesus as a Judeo-Pagan “God”
McGrath mentions having other criticisms, but doesn’t give any specifics, so there is nothing more to respond to, as regards the merits of Proving History.
But some of his half-formed criticisms seem to be allusions to things not in Proving History. At one point, for example, he writes about my theories regarding the reference to James “the brother of the Lord” in Galatians, even though I don’t discuss that anywhere in Proving History. And though McGrath says of this issue that “Carrier falls back on arguments that do not reflect the rigorous Bayesian logic he calls for historians to use,” he does not demonstrate this, or even explain what he means. In fact I have toyed with Bayesian models for answering this question. So what am I failing to do, by his estimation? It’s not like he couldn’t build his own actual Bayesian argument for whatever conclusion he wants to maintain here. That’s the only way to make progress. Whereas making vague, unsupported claims of methodological errors that are never identified is of no use to anyone.
Similarly, at another point, he remarks that “[Carrier] even continues to make reference to the long-discredited idea that Jesus was thought of as ‘God’ or ‘a god’ by his earliest followers,” which I can only assume is a parroting of that claim in Ehrman’s latest book, which ignores entirely what I actually say about this in the work in question (which is not Proving History, but Not the Impossible Faith, chapter 9, pp. 247-57), where I explicitly articulate the view that the first Christians did not equate Jesus to God, but as an adopted son and agent of God, one who represents God, as all kings, high priests, and prophets of God did. That Ehrman didn’t pay attention to what I actually argue in the book, and McGrath just bought Ehrman’s mischaracterization of it, demonstrates why we can’t rely on Ehrman’s treatment of mythicist authors in general (and that’s not the only thing wrong with Ehrman’s book on the subject).
And yet McGrath overlooks the fact that Jesus was still more than that. He is a pre-existent celestial being with the assigned powers of God in our earliest texts: e.g. Philippians 2, 1 Corinthians 8:6 and 15:24-28. Which means he would have been in pagan parlance a god–what Jews typically called an archangel, a type of being pagans only understood as a deity. The notion of an “archangel” was a peculiarly Jewish way to have gods without calling them gods; the closest equivalent in pagan parlance was Lord, kyrios, the exact same thing Christians called Jesus, a word that was for pagans simply a synonym of “god” when referring to celestial powers, as Christians understood Jesus to be in our earliest recorded texts.
Thus, there were two systems of vocabulary in antiquity, and when translating from Jewish to pagan thought-concepts, Jesus would have been understood as a god from the beginning–just not the God, a hugely important distinction, even for pagans. Christians would not consider Jesus to be the God until well into the second century (at the earliest); but they already considered him a god from its earliest recorded time, if we use pagan but not Jewish vernacular.
For example, Paul slips into pagan vernacular when he calls Satan a god (2 Corinthians 4:4); obviously if Satan could be called a god, so could Jesus, who was his celestial superior. It just depended on whose vocabulary Paul chose to employ. For example, in contrast to the pagan, Jewish vernacular made a distinction between “deities” and “gods” (1 Corinthians 10:20: “the things which the Gentiles sacrifice, they sacrifice to demons [daimonia, "deities"], and not to God”), which made no sense in pagan vernacular, where these were synonyms. More exactly, the words Christians and Jews used for “demon” were two in number, daimôn and daimonion, the former meant “deity” and was applied to gods of all rank by pagans (it’s simply a synonym of theos), while the diminutive daimonion more often meant deities of lower rank (i.e. subordinate gods). Other than the fact that occasionally daimôn could be used in a related sense, distinguishing demigods from full gods (but both still gods in modern parlance), making this distinction between theos and daimôn is a Jewish invention, a way to continue having many gods while still appearing to be monotheistic. “Good” gods were angels; “bad” gods were demons. But they are all still “gods” by any pagan definition of the time (and to be honest, by our definition today, when speaking of gods generally instead of God specifically).
About one thing McGrath is right: understanding this context is crucial to understanding the origins of Christianity.