Joseph Atwill is a crank. His acolytes at Postflaviana are cranks.

They are all conspiracy theory nutcases who prefer wildly, implausibly complex interpretations of history that insist Shakespeare was a black Jewish woman, the writers of Jane the Virgin may have attended “quasi-Masonic meetings of the Ordo Templi” and are now sending us coded signals in their TV show, astrological signs may have alerted CEOs to go to a conference to avoid being killed on 9/11, the 9/11 attack was itself part of a global conspiracy of an ancient cabal of aristocratic families even more secretive than the Illuminati, John Lennon tried to warn us about this cabal by putting coded messages in Beatles lyrics, The Catcher in the Rye is a Freemason brainwashing vector developed in connection with the MK Ultra mind control program to undermine society by causing the sexual revolution, indeed hired anthropologists “were used by the CIA against the American people in the creation of the 1960’s counter culture,” and all of this ties back to the Flavians, because not only had “the Flavian Caesars, Vespasian and Titus, invented Christianity, more or less in the form we know it today,” but there was also “an early Roman and Herodian-controlled form of Christianity which was invented well before the time of the Flavians,” and Paul was its shill. And that’s all just “Straight-Talking Common Sense.” Holy balls.

Atwill, and his crank Flavian invention theory, I’ve already covered enough. He has no relevant skills, training, qualifications, or education. As far as I can tell, he can’t read Greek, doesn’t know how to consult manuscripts, doesn’t have any informed understanding of ancient politics and culture, and has no comprehension of how logic or historical method actually work. He just fabricates elaborate conspiracy theories from random tea leaf patterns, that just as often don’t even exist. There is no logically valid connection between the evidence he adduces, and his theories being likely…when he even adduces evidence at all; often enough, he and his cronies confuse random speculation with evidence. It’s tinfoil hat. All the way down.

Atwill also lies a lot. Recently, for example, he attempted to spread the lie that I once refused to debate him, and for some bizarre reason he made up. In actual fact, I assented to a moderated debate online with a fair format. He refused. As soon as a balanced format was proposed, he dropped out. Atwill fabricated stories about me after that that have no relation to anything that actually happened. His acolytes remain duped, even after being shown complete copies of the email thread in question. Because they are crazy.

So this is the context you should be aware of when I survey their 2015 attempt to critique my book On the Historicity of Jesus on their kooky website in Richard Carrier: Meet Barnum & Bailey. Because after all these years, a patron has now funded the writing of a response.

Sound Bayesian Reasoning

Bayesian reasoning. If you’re not caught up, you can get a refresher from my past writing on the point. It’s important. Because once we know the difference between sound and crank applications of Bayesian reasoning, we can see how cranks fly off their witch’s broom handle right into stardust and faerieland. Crucial to sound Bayesian reasoning is (1) not leaving evidence out, (2) deriving priors from actual documented background knowledge, (3) not leaving out expected consequences of a theory that should be observed when estimating likelihoods, and (4) including reasonable margins of error as a hedge against all our cognitive biases. It also requires having a sound grasp of probability theory. And one of the biggest fails cranks and tinfoil hatters fall victim to, is not understanding how common coincidences are, and thus having no idea how to control for that possibility. They see design in everything; when in fact most stuff, even seemingly patterned stuff, is just random noise.

The Jerry Russell Review

I’ll leave aside the question of whether Jerry Russell is an actual person or just a sock puppet for Joe Atwill himself. It hardly matters. What’s more important is that right off the bat, Russell identifies admitting our uncertainty as “weakness.” This is the difference between cranks and real historians. Real, honest historians admit we must have wide margins of error, when we have such a paucity of evidence and all of it as problematic and compromised, as we have for the first three centuries of Christianity. And they will therefore correctly account for that. Instead, Russell declares that my book’s conclusion that the probability Jesus existed given what little evidence we have lies “between 1 in 3 and 1 in 12500” is “amusingly wide” and entails “the weakness of Carrier’s approach.” That’s crank 101. “We can be far more certain than that, even though the evidence is shit” is precisely the foundational logic of all crankery, everywhere.

Russell then gives us almost immediately after that another marker of crankery 101: the conflation of objective methodology with political and rhetorical goals. The exact opposite of sound and honest history, which starts with admitting you cannot have an objective methodology, if you cannot separate it from desires. That’s the very point of constructing a valid professional methodology in history: to not define conclusions, or select methods, based on what we want to be true for political or social or emotional reasons. And yet behold: Russell says “Carrier further undermined the importance of his own effort” by admitting the evidence Jesus didn’t exist is far too poor to use such a thesis as an argument against the truth of Christianity, and advising that critics of Christianity stick to simply demonstrating what is far more easier to prove: that its miraculous basis cannot be true. Which one can more effectively do by granting historicity in the fashion of most mainstream scholars.

Note the significance of this: simply because we can’t use my conclusion to defeat Christianity, therefore my methods are questionable. I’ve “undermined the importance” of my effort by conceding uncertainty. The uncertainty actually objectively entailed by the state of the evidence. Russell thus thinks it’s far more important to defeat Christianity than to tell the truth; therefore any method that gets results too uncertain to be useful for that political aim, must be discarded. Not because it isn’t a valid method or because there is any other method better able to ascertain the truth; but because the truth is a liability to our political goals. Therefore objective methods must go. This is the summation of a crank’s very modus operandi.

Not Reading the Book

In OHJ I compare two hypotheses: minimal historicity, allowing for the widest possible range of historicity theories (while effectively not even bothering with supernaturalist ones even though they are logically included), and a minimalist version of what elsewhere I call “the Doherty thesis.” And then address the theories thus excluded (pp. 53-55). Failing at logic, Russell thinks it’s “astute” to notice “both” theories could be true because they “are by no means mutually exclusive.” Far from being astute, that’s obtuse.

Minimal historicity is excluded by several of the enumerated elements of minimal mythicism. A Jesus who communicated with his followers “only” through dreams and visions, clearly cannot have been “an actual man at some point named Jesus [who] acquired followers in life who continued as an identifiable movement after his death.” These are literally mutually exclusive propositions. A Jesus who was believed to have been incarnated and died “in a supernatural realm” cannot have been “the same Jesus who was claimed by some of his followers to have been executed by the Jewish or Roman authorities.” Indeed, even in the most rudimentary sense, a Jesus who had only been incarnate in the heavens cannot also have been incarnate on earth. And a Jesus who died in the sky, cannot also have died on earth. And a Jesus who was only later “placed … on earth” cannot also have been a Jesus who had already been on earth. These are all, again, mutually exclusive propositions. And for a critic not to notice that can only be explained by either laziness or stupidity.

Confusing Speculation with Evidence

Russell claims my argument (in Ch. 11 of OHJ) that Paul appears to only ever refer to a celestial Jesus, never an earthly one, “completely fails once we realize that Paul might be describing only one out of two or more pre-Christian sects of his time, including historicizing as well as non-historicizing sects.” Notice he just assumed this as a fact, and then inserted this thus-fabricated fact as evidence against my thesis. This is how cranks operate. It is of course in the most basic sense a possibiliter fallacy (see Proving History, pp. 26-29), which even serious historians are frequently prone to—as shown by Fischer with numerous examples in his superb Historians’ Fallacies (pp. 53-56; he there calls it “the fallacy of the possible proof”).

Bayesian methodology forces you to confront the question of how likely is any such assumption, before using it in the logical structure of any argument. Cranks don’t ever confront that question. Anything possible, is automatically all but 100% certain. As long as it supports whatever weird thing they want to argue. Instead, we have to look at the evidence, and assess the likelihood ratio: Is there any evidence of this being the case? And is the evidence we do have even congruent with this being the case? Is it even 50/50 that it’s the case?

Of course the answer to all three questions is no. There is no evidence of there being those two versions of Christianity in Paul’s day, one with and one without an earthly Jesus. There were certainly many competing sects, about which we aren’t told details. But there is no evidence of that particular divergence existing then, anywhere. To the contrary, the evidence that is in Paul (and all other potentially early sources, e.g. 1 Clement, 1 Peter, Hebrews: OHJ pp. 308-15, 529-31, 538-52, respectively) is wholly incongruent with there having been the earthly creed alongside Paul’s celestial creed. Precisely for the reasons I explain in OHJ (e.g. Chs. 11.1, 11.2, 11.4, 11.6, 11.8): had that been the case, Paul would constantly be confronted with it and have to address it; and Paul could not have argued his gospel was congruent with the first apostles’ to so many congregations who well knew those apostles and their preaching (and they did; that’s precisely why Paul claims he was in alignment with them, and confesses only one incongruence—regarding the requirements of conversion—that he argues they allowed him: see Element 20, Ch. 4).

In effect, Russell wants to argue for historicity (ironically for an Atwillian) by conceding that all the evidence of Paul confirms mythicism, “but” there “must” have been evidence of historicity from some other people in Paul’s day, even though we have no evidence of that ever having been the case. This wild speculation, contrary to all evidence and based on no evidence, then becomes Russell’s “evidence for” historicity. See how that works? That’s what making a tinfoil hat looks like.

Standard Crank Tactics

I’ll pass by Russell’s amusing typo in saying I think “Paul was writing during the 5th decade of the first century” (the 50s A.D. is the 6th decade, FYI). He wants to double down on adding speculations to his “evidence” by asserting we should conclude even Paul didn’t exist and that his Epistles were written a century later. That’s simply not probable, in any usable degree. It only makes the non-existence of Jesus even more unlikely, by requiring an additional already-improbable supposition. This is a common mistake made in probability reasoning, but especially common among cranks: adding even more improbable suppositions to a theory, and thinking that this increases the probability of that theory. No. It logically necessarily reduces that probability. It’s basically a conjunction fallacy gone amuck.

I’ll also not bother much with Russell’s totally crank suggestion that all the messianic movements Josephus documented were versions of Christianity. There is exactly zero evidence of that, and no plausibility to it whatever (see OHJ, Ch. 6.5 & Element 4, Ch. 4). Russell seems to think any messianic movement should be considered “Christian,” which is an equivocation fallacy: we do not mean by “Christian” just any “messianism,” but specifically the belief in a resurrected savior whose death atoned for all sins, a belief present in no other messianic sect attested. But it’s just like a crank to go off on pattern hunting to a level of lunacy, and rest a bizarre conclusion on fallacious tricks with words. Even more typical of crankery is reasoning by bizarre non sequiturs that experts actually agree with you when in fact they do not. Russell does this when he says “Carrier’s belief” is that Christianity was like “a singular kernel [that] exploded through a process of conversion driven by evangelism and managed by a tightly controlled hierarchical leadership,” and “yet as Carrier himself admits, nothing could be farther from the truth,” since “we know … the chaotic nature of early Christianity.” This is all false.

I do believe Christianity launched from a singular origin: because that’s what our only eyewitness to its formative years tells us (1 Corinthians 15; Galatians 1). And there is zero evidence of it being otherwise. But I have never said it was “managed by a tightly controlled hierarchical leadership.” To the contrary, I explain in OHJ that that same witness confirms the leadership could only barely control anything, and the movement fractured and spiraled into numerous competing sects (OHJ, Element 21, with Elements 10, 15, and 20, in Ch. 4). It is also a total non sequitur to argue from that, that therefore Christianity didn’t start exactly as that same witness says who attests to its fracture; and outright bizarre to conclude that therefore “I admit” it didn’t. (It’s doubly bizarre of someone who thinks these Epistles are 2nd century forgeries to keep citing them as evidence of what was going on in the early first century…incoherence being another common feature of conspiracy theory cranks.)

Non sequiturs are also present in Russell’s strange move in arguing “surely the near-silence of history regarding that exploding Church is … far more damning to the view that Paul’s epistles reflect the genuine historical state of the Christian movement.” I struggle to discern any logic at all in this statement. Why would the obscurity of a tiny fringe movement, resulting in next to no document survival (and on Russell’s crank view, no document survival at all, as he thinks all the letters of Paul are forged), argue against it being too obscure and fringe to have documents survive? The idiocy in this reasoning is perplexing to me.

The reason we have next to no documents about the early church, and indeed no third party references to it either, is because it was too small and inconsequential for anyone to notice it in any records we now have; and the much later church that effectively acquired total document control chose not to preserve any of its first century of documentation (apart from some very few edited letters it deemed safe). No other explanation is possible. If there were many parallel versions of Christianity in the way Russell maintains, the silence of the record is just as peculiar. Unless you adopt exactly the same explanation for it. This evidence is therefore equally likely on either theory. It therefore cannot argue for Russell’s theory. That’s how sound Bayesian reasoning works. But not, evidently, how crank reasoning works.

Ignorance

Cranks are also ignorant. Being amateurs who are totally convinced of their every thought, it rarely occurs to them that they have to do research and check facts and actually understand the historical periods and sources they are rambling on about. Russell gives us a golden example when he complains that my finding parallels in Mark to Roman passion plays for Romulus comes from Plutarch who wrote after Mark. But he clearly didn’t look down the page at my footnotes, which extensively cite sources confirming Plutarch is only summarizing facts and sources well known before Christianity, like Cicero and Livy and Ovid. Indeed, Plutarch is describing annual passion plays that his narrative implies had been celebrated for centuries, as confirmed by pre-Christian sources. Does Russell imagine Rome completely invented these Romulus passion plays after Mark wrote his Gospel in (at its earliest) the 70s A.D.? Based on what wild speculation contrary to all evidence and plausibility?

Russell then bizarrely insists euhemerization “is actually defined as the process by which a human person is converted into a myth.” Um. No. Euhemerus did not take a historical Uranus and Zeus and mythologize them. He took a mythical Uranus and Zeus and historicized them. That’s why that is called “Euhemerization.” That Russell doesn’t know this shows he didn’t read my book’s discussion of Euhemerism, and doesn’t even know why it’s called Euhemerism—why it is named after Euhemerus, what Euhemerus did that the term references (see Euhemerization and my Brief Note). Total ignorance. Yet he burns several paragraphs in total confidence he is right about this. Yet more crankery gold.

It’s even more gold that Russell doesn’t get the irony when he then tries to argue for the historicity of Zeus and Uranus by saying “it is impossible to say for sure whether these tales are purely legendary, or whether there is some real-life hero,” which means it’s impossible to say they were ever historical. Just like Jesus. Cranks have a hard time grasping the significance of having absolutely no evidence for something. “But maybe there was a real Zeus at some point” is just another possibiliter fallacy run amuck.

It’s all the worse that in many cases we do have evidence they began as celestial or mythical beings, and were never conceived historically until far later. But Russell would actually have to study and check facts to know that. He’d also have to pay attention. Because I actually allow for a relatively high frequency of such heroes having been historical. 1 in 3, which is slightly below the 50/50 total ignorance would dictate only because of the evidence we have that so many were indeed mythical.

Cranks Don’t Like Peer Review Because They Can Never Get Anything Past It

Equally ignorant is Russell’s bizarre assertion that “the standard academic peer review process is seldom if ever applied to books, but only to journal articles.” That’s wildly false. And I cannot fathom how he ever came to believe this. It’s false even as a general statement of academic publishing. But it’s especially false in my particular case. As I wrote in 2013 (emphasis added so you don’t miss it):

I sought four peer review reports from major professors of New Testament or Early Christianity, and two have returned their reports, approving with revisions, and those revisions have been made. Since two peers is the standard number for academic publications, we can proceed. And Sheffield’s own peer reviewers have approved the text. Two others missed the assigned deadline, but I’m still hoping to get their reports and I’ll do my best to meet any revisions they require as well.

It’s important to note that clarification: Sheffield-Phoenix selected its own peer reviewers to vet my book, as they do all academic treatises they publish. That’s the entire point of an academic press. This was after I also submitted peer review reports from multiple prominent professors of Biblical studies I had used to pre-vet my manuscript, to ensure it would pass any peer review a publisher engaged. It’s also not uncommon for academic presses to ask the submitter of a manuscript to supply a list of suitable peer reviewers. But whether Sheffield-Phoenix relied on any of the peer reviewers I selected, I won’t have been told.

In history, peer review of course does not establish that what is published is true or correct, or even without flaws or mistakes. What it does do, is verify that the work meets the standards of the field, in its use of evidence and arguing to a conclusion, that it isn’t ignorant of basic facts an expert in the subject should know, and isn’t crank amateur wishwash academics can safely ignore. The point of it, in other words, is to signal to other scholars that this work is worth their time reading, considering, and criticizing. The process that should then follow is a widening awareness and discussion of the work, so the consensus can be reviewed against the new, vetted challenge to it. Just as happened with Thomas Thompson’s challenge to the consensus on the historicity of Moses in the 1970s. Which after a decade or two of resistance and debate, ended up changing that consensus. The process is slow (and is being betrayed currently). But Russell apparently has no knowledge of any of this.

A few years ago some nutter tried making the same challenge to my publisher. Here is the response of the General Manager of Sheffield-Phoenix (sent 11 January 2016):

Many thanks for your email, we are happy to clarify on this for you, whatever may be the case in the USA, in the UK we are familiar with two kinds of publishers (at least): academic and trade. All academic publishers operate on the same basis, of making publishing decisions on the merits of a proposal as assessed by experts in the field (‘peer review’). This is true for university publishers like Cambridge University Press and non-university publishers like Bloomsbury Academic, Routledge, Macmillan, Blackwell. There is no difference between the groups in the reviewing process. Nor is there any difference in prestige between university presses and independent academic publishers (e.g. some authors may like to be published by Oxford, but no one can say it is a ‘better’ publisher than Macmillan, for example).

Sheffield Phoenix Press (and before it Sheffield Academic Press) are independent academic publishers. A publishing house can only call itself ‘University of X Press’ if the university agrees, and perhaps if it ‘owns’ it in some sense. It would be a mistake to think that what a ‘university’ press publishes is somehow approved of or authenticated by the university. Publishers of university presses would be affronted if their university attempted to influence their publishing decisions.

For example, even Yale University Press is independent of Yale University (it’s “financially and operationally autonomous,” hence the President of Yale cannot tell the press what to publish, just as neither could the University of Sheffield). This is normal for university presses. In Sheffield Phoenix’s case, its editors were all Sheffield University faculty, and it was housed on the Sheffield campus. A couple years after I published with them the University disbanded and disbursed its Biblical Studies Department, but its succeeding Center still collaborates with the Press, and Sheffield-Phoenix has continued as an independent academic press in the same manner as Bloomsbury, Routledge, Macmillan, and Blackwell. In America, the prestigious Society of Biblical Literature continues as its distributor. But even when I published with them, peer reviewers may or may not have been Sheffield professors or emeriti; no academic press limits its peer reviewers to their associated school.

And what the manager of Sheffield Phoenix said about UK presses is also true in the US: there are academic and trade publishers; and academic publishers all use peer review for most of what they produce. Anyone who doesn’t know this is just ignorant.

Sucking at Math

Cranks also suck at math. To be fair, so do most folks. But cranks are particularly vulnerable to innumeracy.

For example, Russell concludes that some evidence I adduce, like peculiar features I find in Acts, isn’t evidence against historicity because he can think of “possible” ways historicity could produce that evidence. This is a basic math fail. It is, sure, yet another possibiliter fallacy. But more importantly, it betrays a complete lack of understanding of what probability even is. The whole point of assigning a probability is to account for possibilities like he alleges. The question is not “are there possible ways historicity or mythicism could cause this evidence,” since the answer will almost always be yes. The question is with what frequency? In other words, how often will one or the other theory produce that evidence? In a more colloquial sense: How weird is the evidence?

And also, of course, we also must ask, what’s the prior probability of your alternative? In other words, not just is it possible; but how likely is it even? You can’t say 100%, as that entails we have already conclusively confirmed your supposition is true, so we don’t even need to verify it! And in point of fact, we have no such confirmed knowledge that mere suppositions are true. That’s what distinguishes them as unproven suppositions. That’s the difference between a speculation and evidence.

This is the problem with critics like Russell never explaining what probabilities they would assign to such evidence other than mine. To the contrary, Russell basically just says he agrees with my assignments. He offers no others. Nor any argument for them being different than mine. But this is why understanding Bayesian reasoning is so important. Since it is completely useless to claim to refute any conclusion by just asserting “but maybe it was something else.” We need to sit down and ask the relevant question instead, for any hypothesis h we are testing or questioning: Is this exactly what we’d expect on h? Or is it a little weird on h? Or even extremely weird on h?

As expected, since Russell demonstrates repeatedly that he has not actually read my book, he didn’t notice I have a whole section on how one might question or challenge my estimates, and the example I employ is indeed Acts! Check it out: On the Historicity of Jesus, Ch. 12.2 (cf. pp. 603-05). Here is the key section (emphasis added):

[I]n fact Acts contains very strange oddities on the assumption that Luke was making the whole thing up. The notion that Luke would not put the family of Jesus, or a plausible historical plot, or plainly historicizing trial speeches into his fabrication is simply not that likely. Such strange omissions really make sense only if Luke was altering and embellishing some actual outline of the early church’s first trials and internal struggles. Only then might he not realize that his outline lacked any mention of the family of Jesus having anything to do with the church, that the trial speeches lacked any mention of a historical Jesus, and that the plot made no sense unless Jesus was only known as a revealed being even to the Jewish and Roman elite. Certainly, we must admit that these oddities are not what we’d expect from such a text. They cannot therefore have a 100% probability. My estimates reflect that fact. So must yours. Quite simply, the strange content of Acts must have some probability below 100%.

And yet even if we adopted such an extreme tactic of claiming (against all common sense) that Acts is exactly as we should expect it on historicity, we still don’t find that Jesus’ existence is likely.

And I demonstrate both points there mathematically. I highlighted certain things here to make a point: Russell only mentions “the trial speeches lacked any mention of a historical Jesus,” as if I didn’t also adduce many other oddities in Acts, and tabulate all of them in their mathematical effect. Omitting the best arguments for a conclusion, “refuting” the weakest one instead, and claiming to have refuted them all, is a common cognitive error humans are prone to. We could call it a straw man fallacy, but it’s more pernicious, as it shows the critic is literally blanking on the best arguments for a conclusion, and not noticing they have no argument to answer them; they instead jump on the weakest argument and assume they’ve won. This is irrational. But standard for a crank.

More important, though, is the Bayesian point. Let’s assume Russell had a valid argument regarding the trial speeches. He asserts “maybe” Luke used a Pauline source that contained a mythical-celestial Jesus, and had no sources from the parallel sect Russell invented earlier that was based on a historical Jesus. This is the same fallacy I already noted above: confusing unevidenced speculations as evidence that changes a probability. That’s not how speculation works. Russell’s alternative being possible is already accounted for in my calculations. So how can he use them to arrive at a different calculation?

And indeed, look at my estimates here: the odds I assign that the trial speeches would look like they do if Jesus existed is 9 in 10 (OHJ, p. 386; while that same evidence is 100% expected on the contrary thesis). In other words, roughly a 90% chance they’d look like that. That’s veritably inconsequential as evidence against historicity. But Russell doesn’t grasp math and thus doesn’t notice I already assigned a very high likelihood to this evidence on historicity, and thus weigh it extremely weakly as evidence against historicity. Exactly as he says I should! I do put a lower bound on this evidence of 50/50, but by definition, being the lower bound, I’m assuming a lower probability of alternative explanations there. I allow a higher probability of them in my upper bound. That’s the point of the upper bound.

What objection does Russell have to my upper bound then? Does he think it should be 91/100 instead of 9/10? 95/100? 99/100? Why? Based on what? Saying “possibly it’s this other thing” has no effect here, because I am already saying “there is a roughly 90% chance it’s this other thing” (or something else compatible with historicity). So I have already included his “possibly it’s this other thing” objection. And not just included it, but gave it an absurdly high probability—based on no evidence whatever. How much more generous can we honestly be? Russell can adduce absolutely no evidence that the probability of his alternative is higher than 90% (or even that it is 90%). So what’s his criticism here? There is none I can discern.

Conclusion

Russell claims it’s a “glaring flaw” that my book’s wide margins of error and thus confirmed uncertainty in its conclusion is “pathetically vague,” perfectly illustrating the mind of a crank. Admitting uncertainty is a horror. Which is precisely why cranks like him are always so certain of everything. Objective methodology is a bane to them, because it compels us to admit there are things we don’t know and can’t be certain about.

Otherwise, Russell’s claim that my conclusion in OHJ is “demonstrably ridiculous” is based on no arguments whatever. He completely gets wrong what’s in the book, strings together a bunch of obvious, even bizarre fallacies and non sequiturs, and repeatedly demonstrates his ignorance of history and sources, while deploying nearly every tactic in the crank’s handbook. And in the end, he never even challenges any of the actual evidence I used to adduce probabilities from, nor argues for any alteration of those probabilities!

Face. Palm.