Yesterday I started this two part series explaining what I am doing and why (so please catch up there), and went through an extended example from R.G. Price. Today I will finish with the second example, now from David Oliver Smith (channeling Robert Eisenman). And I then wrap the series below with a conclusion explaining what we can and should learn from both. I will sometimes refer back to the examples and discussion from Price in my discussion of Smith, so you will need to read Part 1 to get it all.

D.O. Smith and Overtrusting Cranks

My example from Smith is from The Bible Tells Me So (2022), chapter three. There Smith argues that the Dead Sea Scrolls are Christian texts that refer to the Apostle Paul as the Spouter of Lies. Which warrants initial doubt owing to the fallacy of Suspicious Coincidence (better known as Base Rate Neglect): it would be weirdly lucky that, of the dozens of sects and hundreds of communities in Judaism of the time, the only first-century sectarian documents we find a stash of “just happen” to be from the same sect and even people we want them to be. That’s initially so unlikely you really need very good evidence of it. Otherwise it’s too improbable to assume.

Yet Smith has no actual evidence for this, much less “very good” evidence. Because right away a common mistake across his whole chapter’s argument is:

  • Evidence already expected to exist cannot be evidence for your theory.

I’ll just give one example, but there are many across the chapter. His table showing common phrases between Paul and the Scrolls (and that they often cite and riff on scripture) contains nothing remarkable (83–87); with another (91–94) that merely jumbles common messianic ideas with commonalities of exegesis and ethics and problems entirely generic to any sect relative to (or indeed even prior and ancestral to) Christianity. Some of the claims in these tables, however, are actually false. And I will be focusing on a key selection of those below. But don’t forget the other half of the methodological lesson: for a fact to be evidence for a theory, that fact must be unexpected on any other explanation; and to be good evidence, it must be very unexpected. Nothing here qualifies. Everything he lists would be worth listing alongside good examples. But there are no good examples. And that’s Smith’s single most pervasive mistake here.

And by this I mean none of the other stuff on these tables is peculiar enough to only be there on Smith’s specific radical thesis. They are all expected for any person of the era of any similar sectarian bent. Not even the same sect is necessary; any ideological cousin or even predecessor would use the same vocabulary and ideas. To get an actual demonstration of awareness (which is only half Smith’s theory) requires a lot more than that (again, see my discussion of mimesis criteria in Proving History at the end of chapter five); and to go the remaining mile and get the more amazing conclusion—that these are the very same people talking about the very same things—you need even better evidence than that. And Smith doesn’t have that. Instead he fills this chapter with outrageously overconfident assertions of “evidence” that a reader unawares won’t know doesn’t really exist or isn’t peculiar enough to carry the required weight.

But the false facts are worse. So I’ll now turn to those.

I am really disappointed in this, because unlike his usual practice, I think Smith was dangerously lazy in his research and thinking here. I’ll start with the most fatal mistake: dating the dead sea scrolls. Very inexplicably Smith declares these “were written between the 2nd century BCE and the 2nd century CE.” Every expert knows they had to have been written before they were deposited in the caves and abandoned, which occurred during the Jewish War and thus before the year 70. So none can have been written even in the later 1st much less 2nd century. Carbon dates that make later dates possible are relatively rare and thus consistent with expected error. And you cannot build a hypothesis on what is already expected by error. Yes, some serious scholars have made that mistake, but it’s still a mistake. Which is another reason amateurs need to be more cautious: if you can’t tell when an expert is making a mistake, you need to divert more time and resources to making sure they aren’t.

I am assuming Smith did not mean a “literal” between, as in, excluding the second centuries BCE and CE, because that’s false on the low end (many were indeed written in the second century BCE, not “all” of them after). Either way, maybe Smith is only guilty of catastrophically confused or tendentious wording here (he certainly didn’t get any of this from his cited sources). But although that’s its own problem, the real problem is worse—because unlike that, this cannot be corrected and his argument survive: Smith derives his strange conclusion almost entirely from the Habakkuk pesher, by simply assuming without argument that it dates to the late first century, and even indeed refers to the Ananias who served as high priest in 62 AD, executing James in Josephus—a James whom Smith says is the Teacher of Righteousness referred to in the scrolls, and Ananias the “the Wicked Priest” (see instead this mainstream summary of the relevant cast of characters).

Smith’s entire chapter’s argument depends on this. The problem is that Pesher Habakkuk has been carbon dated to the century before that. The probability of it dating to the mid-60s A.D. or any year A.D. is too small to credit. Accordingly, all mainstream scholars believe the Wicked Priest belongs to a historical narrative in the community’s past that could have occurred in the first, second, or even third century BC, and has nothing to do with the Roman era at all.

Unfortunately, Smith relies overmuch here on fringe expert Robert Eisenman. But that guy’s work is rife with wildly implausible literary pareidolia and a bizarre immunity to contrary evidence, to the extent that, in my opinion, he is probably insane. Which may have gotten worse over time, similarly to Jerry Vardaman (an analogous case I document in Hitler Homer). Eisenman’s ideas are so rambling and crazy he was literally shouted off the stage at an academic conference once by several sober colleagues in the audience—and I know this because I was there. No one should be taking anything this guy says without stern triple-checking skepticism. And here Smith’s failure at that kicks the legs out from under his whole theory: Smith has gullibly just believed Eisenman’s crank efforts to deny or obfuscate the carbon date of pHab (without even mentioning that has happened), and then bought into a whole crazy theory built on top of Eisenman’s own red string board—for Eisenman’s the one who first argued this is all about Christians and Paul. Everyone else laughs this off, well they should.

Not knowing your source is a crank, and not checking whether anything they are saying is true or even arguable, is a common amateur mistake. This may be because amateurs are confused by cranks with actual credentials, falsely believing a relevant PhD and cv or even prestige (just as with Vardaman) immunizes them from being full of shit. Nope. Sorry. You have to check first. Yet Smith didn’t even read his own academic source (Doudna, 455–56), who warns that Eisenman’s redating claims have not held up and are probably false. For pHab the 95% interval maxes at 2 AD, which means the probability is below 2.5% that it came later, and a great deal below that for “later by sixty years.” The 68% interval at the most recent and reliable testing tells us the most probable period of transcription is 88–2 BC (Doudna, 474). Paleographic cross-dating (by handwriting style) gets the same result: last half of the first century BC, not AD.

It really is worth noting whom Smith is gullibly trusting here. Eisenman teamed up with the infamous amateur Joseph Atwill (yes, that Joseph Atwill) to try and contest the carbon dating results for this text (because the evidence, as it actually is, refuted him) and to insist instead that there was a decent probability that this pesher did date to the early first century—albeit to less than a one in six chance, an actual mathematical consequence of his argument he kind of avoids telling anyone, because he wants to rely on the possibiliter fallacy that any nonzero chance he is right entails he is right. But he’s not. Even with a 95% confidence level you can expect 1 in 20 results to be outside a given range by error alone, so finding one that is, across hundreds tested, is expected, and therefore cannot support a belief in a rare erroneous date. But cranks loathe probability theory. It is the enemy of all tinfoil. Which is why amateurs should not let themselves be snookered by their bullshit. They need to understand probability theory.

All of which was patiently explained to Eisenman by an actual scientist (painfully pointed out again by Doudna). Which means this is very much like young earth and flat earth and indeed shroud of Turin enthusiasts insisting the science that refutes them “must be wrong,” so then scientists have to patiently explain to them why they don’t know what the fuck they are talking about and can they please go away. Smith didn’t get the memo. He fell for Eisenman’s tinfoil hat. But even if Eisenman were right, that would still entail the probability of this Smith-Eisenman theory about Paul is true will be less than 17%. Because the premise it depends on (the date of pHab) cannot be more likely than that even granting Eisenman’s argument about the carbon dating. So any theory relying on that premise must be less likely still. This is, again, a basic reality of conjunction, just as I explained yesterday for Price. And since Eisenman’s bollocks about carbon dating isn’t even correct, the relevant probability is even less than his imagined one-in-six chance (surely below even a 2% chance).

Which is the second lesson here:

  • When your entire argument relies on a crucial premise, you need to make sure that premise is probable, and establish it before continuing to build on it.

“My conclusion follows if we trust a bunch of cranks refuted by real scientists” is methodology failure 101. Do not do this. And when you catch someone doing this, count it as strike one. They get only two. Third strike and you spend no more time on it. Shelve the book and move on to some other author or subject. Because while “maybe” the improbable thing is true, you have no warrant to believe it is.

So what could Smith do to recover from this? Not much. To reverse the consequences of such a low probability premise you need evidence that’s, say, a hundred times more likely on the unexpected date than any other explanation of that evidence, and that simply isn’t what Smith or Eisenman have. Their explanations are all less likely than mainstream ones, not more—much less a hundred times more. While Smith could try to rewrite his entire book with the high chronology model of Christian origins, which places the letters of Paul in the 50s BC and the origin of Christianity in the 70s BC (see How Do We Know the Apostle Paul Wrote His Epistles in the 50s A.D.? along with the scholarship in Historicity §8.1 and Paradigm 57–60), abandoning his belief that pHab refers to Ananias, that still reduces the probability of his thesis to below the probability already attending that model. So, for example, if Smith’s theory is (let’s say) “as likely as” any alternative high chronology model, and high chronology is (let’s say) 10% likely, then Smith’s theory cannot be more than 5% likely. It’s hard to make any progress from there. And as I explained for Price above, once you start stacking these improbable “total rewrites of existing evidence” (or any other “giant leaps of logic”) in your premises, that “less than 5%” quickly becomes “so far below 1% as to vanish under rounding error.” You’re pretty much done at that point.

And that’s what happens here. Because Smith goes on to assert a string of falsehoods as facts, and I can only assume he is getting this string of baloney from Eisenman and just believing it all for some reason. For example, Smith says pHab describes the Spouter of Lies “as creating churches based upon blood” (57). No such passage exists in pHab. There are no churches. The Spouter of Lies is never described as creating such a thing, nor “upon blood” in any relevant sense. Somehow this is supposed to convolutedly refer to Paul preaching a blood atonement in Christ, and therefore preaching that “his congregations should ignore the Law of Moses.” But this doesn’t exist there either. None of those things show up or connect in the text as Eisenman’s crazed mind imagines.

Smith cites 1QpHab 10:9–10 here, which doesn’t say any of that. I’ll use a translation of the commentary here free to access but I checked several (including Lim’s, below) to confirm it’s not being tendentious. This scroll is commenting on the condemnation of military violence in Habakkuk 2:12–13 (“woe to him who builds a city with bloodshed and establishes a town by injustice!” and very specifically it says a city, not a synagogue or congregation, and clearly is referring by blood to violence), which the commentary quotes (pHab 10:6–8), followed by its secret interpretation (pHab 10:9–13):

The interpretation of the matter concerns the spreader of lies who deceived many, building a worthless city by blood and forming a community by lies for his own glory, making many toil at useless labor, teaching them in shameful falsehood. Their toil will be for nothing. Because of this they will undergo firey punishments because they blasphemed and reviled God’s chosen ones.

In preceding and following sections it is clear the commentator has in mind actual (not figurative) military oppression (pHab 9, it’s about “the priests of Jerusalem” who “will gather unrighteous riches from the plunder of the people”; 12, “murder in the city and violence in the land” means the actual city of “Jerusalem” and “the cities of Judah,” etc.)—spilling blood by violence, and for greed and power, and in aid of adorning actual cities, not some convoluted ritual role for blood, nor some figure for churches. There is nothing here that connects to Paul or Christianity at all. Yes, Eisenman piles gigantic stacks of epicycles to “extract” a convoluted different meaning here, but that’s all made up. The evidence as it is does not contain any of this. So we cannot claim it’s here.

This is like Price “changing” the evidence refuting him into evidence supporting him by rewriting all the evidence with convoluted theorizing. But once again that goes in the opposite direction: the improbability of his every conjecture stacks against the conclusion, not in favor of it. If we need, say, five conjectures to reinterpret all the words and senses and contexts in this entry of pHab, and each is even an admirably 30% likely (relative to plainer explanations), not only is Smith’s then-resulting conclusion less than 30% likely (which means: not likely, and therefore we should not believe it), but it’s less than 0.30^5 = 0.30 × 0.30 × 0.30 × 0.30 × 0.30 = 0.00243 or one fifth of one percent likely. Which is so unlikely it vanishes to zero under rounding error (Bayesians might note that this can be avoided with an argument from relative likelihoods, but that is not an available escape here). So why is Smith telling us this dubious conclusion is a fact? Why does he not mention that this is an extremely tendentious and dubious understanding of the text rejected by almost every expert alive? Does he not know that? Did he just believe Eisenman? Sure, “every expert alive” can be wrong. But you need to admit to your readers that’s what you are claiming. And then you’d better have some kickass proof. Yet assertions are not even “a” proof, much less as would kick any ass.

Smith’s whole chapter is like that, a litany of factually false claims or unproven claims asserted as facts. I clocked one to ten errors like this on every page. And that’s just not excusable. You can’t just make a bunch of wild statements as if they are established facts and not even argue for them. That’s quintessentially bad method. It’s most pernicious because the reader is being told these are facts. They thus have no idea which sentences are actually asserting completely dubious epicyclic interpretations. Price at least was clear which he was doing. Smith was not. And that’s bad.

Readers won’t know what’s left out either. For example, Smith’s bizarrely convoluted narrative seems to flatly contradict Paul’s eyewitness repeatedly, without explaining why that’s okay or even that it’s happening. For example, Paul says he was never known in Judea until after he converted, and was an avid harasser of the church before he converted. This was almost a credally accepted fact. But Smith says Paul started a member of the community and then rebelled against it. That simply doesn’t track. It wildly deviates from all available evidence. Which makes Smith’s theory even more improbable. Smith also depends on many other dubious facts, such as that James was called the “Just” at any time prior to late second century mythology, and “therefore” James is the Dead Sea Scrolls’ “Teacher of Righteousness,” when that is the least likely evidence-based identification in the literature (see Harkins and Chilton and Painter and Davies).

But as just another representative example of this bad reasoning, Smith is duped by Eisenman’s claim that pHab refers to Paul preaching that the Gentiles need not adopt Torah. Note that Paul never preaches that Jews should abandon Torah—the first that ever shows up as an idea is Hebrews, and as Paul never has to defend himself from being associated with such a teaching, we can be sure it did not exist when he last wrote. But regardless, neither teaching shows up in the Dead Sea Scrolls. Contrary to Smith, it’s not even in the Damascus Document, which only describes violators of the community’s own special interpretation of the Law of Moses and what to do about them. It never mentions anyone preaching the abandonment of Torah, nor even preaching rejection of their interpretation. It only refers to breakers of the law. That’s it. Smith cites CD 1:14–17 as if it does, but that describes “scoffers” who mocked or broke the law in the time of Nebuchadnezzar, hundreds of years before, and not some current Jewish leader arguing believers could abandon Torah and still be saved or Gentiles could join up without it. Indeed CD describes those people as long dead and no longer an active concern. The DSS fragments of the CD are too damaged to contradict this. So there is no factual reality for Smith to refer to here. But his readers won’t know that. That’s bad.

So when Smith goes on (gullibly trusting Eisenman) and quotes pHab 5:11–12 as referring to the Man of Lies as one “who rejected the Torah in the midst of all their assembly” (a thing Paul never actually argued), the reader won’t know that “the Man of the Lie that had rejected The Law in the midst of all their company” refers to that community’s specific interpretation of the law, and to the Man not heeding it when arriving there. There is no reference to this man preaching that anyone could be saved without following Torah, or even that anyone should “reject, spurn, despise” Torah rather than their interpretation of it. It is just saying he refused to follow their law on his visit. And they are using a common phrase for anyone who breaks the law (e.g. Isaiah 5:24, Jeremiah 6:19 and 14:19, Amos 2:4). There is no description here of a Torah-free gospel being taught. Nor is any connection made to blood here (at all, much less as permitting abandoning Torah). These things can only be invented by unevidenced epicycles. Which is changing the evidence; not building on it. For a proper methodological treatment of this section see Timothy Lim, The Earliest Commentary on the Prophecy of Habakkuk (80–84). Eisenman’s exegesis is rank crankery by comparison.

And that’s the merest sample. There is more. The stack of all the improbabilities across Smith’s chapter weighs a ton and a spool. I can’t honestly justify fisking the scores of errors that plague Smith here. He’s well past strike three. Book gets shelved. I’m moving on. But as with Price, I will end with how Smith, too, could be on to something: a far more credible case can be made that these are texts from the same sect precedent to Peter (not Paul) snaking it in a new direction, even using the history and ideas it built up. So we could be looking at the texts of the very sect that Christianity was spawned out of, from the generation or two prior. But even that conclusion is unprovable; just tantalizingly possible, but at least closer to 50% rather than zero. And it doesn’t get any of the other conclusions Smith wants to argue in his book. So he’d have to throw most of that away and start over.

Conclusion

Amateurs have a high rate of failure. That’s why training and experience matter. Getting a relevant PhD literally reduces your failure rate, by skilling you in the relevant background knowledge that can catch more errors and develop better contextual estimates of priors and likelihoods, and by burning away mistaken modes of reasoning through thousands of hours of professional advice and criticism. Which includes training you to avoid handwavy arguments and rely on well-tested and careful methods of fact-checking, wording of points, and citing evidence and sources—and stating when confidence must be low, rather than replacing that caution with overconfidence. One of the most important outputs of this is supposed to be teaching you to tell when you are misusing a method you saw someone else use but didn’t understand why it was valid in their case and thus might not be in yours.

The rate of bad arguments (and even outright crankery) from PhDs tells us the vetting process is not so reliable that you can just “trust” all this has happened and therefore an expert “can’t” be making any of the mistakes they were supposed to have been taught not to. You do still have to check. And checking means more than just fact-checking (did they correctly describe provable facts, and, crucially, did they clearly demarcate facts from inferences or conjectures) but also logic-checking. And logic-checking requires understanding logic, and especially in empirical sciences like history, probability logic. Why do methods work when they do? Why do they fail when they do? The answer is always in how they are moving from one probability to another. And when you are falling victim to literary and cognitive pareidolia, this means being able to identify when that’s happening to you—and thus know when you are building a genuinely sound argument from and to a probability, and when you are not.

In R.G. Price’s case, he relied on too many improbable theories, a whole stack of wild theories, to get the one wild theory at the end to work—and didn’t realize that reduces rather than increases the probability he is right. I suspect he mistook valid backgrounding, and valid methods of stacking coincidences to make them unlikely rather than expected, as warrant for rampant epicycling, and ignoring multiple comparison and conjunction fallacies. He also didn’t correctly apply knowledge of ancient Greek and literary composition theory, as an expert would always know to do.

In D.O. Smith’s case, he overtrusted what he should have known to be cranks, and by not acting more skeptically and carefully, got snowed into someone else’s red string board, on a premise he didn’t realize was scientifically too improbable to carry the argument. Which reduced the probability of his cognitively elated idea about Paul and Qumran below the probability of that dubious premise, which was already pretty low. He then failed to logic-check anything his crank sources argued, and gullibly ended up just repeating their illogical bullshit as if it were a body of established facts. More competence would have skilled Smith in spotting, on his own, the fallacies here that would have made him suspicious of his sources’ reliability, and thus begun a cascade of logic- and fact-checking that would have unraveled the whole conspiracy theory he was being seduced by.

We need to all be on our guard against these mistakes because we are all prone to make them in every aspect of our lives, in deciding what to believe, in deciding who to trust and distrust, in deciding what’s happening in the world politically and economically and culturally. It’s not just about an obscure method applied to obscure topics in an obscure field of study. The mistakes we are seeing here are common across all human thought—and that means across everything you yourself think about. So you need to learn these lessons. That they are being taught here with fringe biblical studies stuff is incidental to the lesson. It’s just the example from which you can learn, to adapt the same principles to things more momentous and less obscure.

For more on probabilistic logic see my category tags critical thinking, historical method, and Bayes’ Theorem.

§

All comments go to moderation except for Patrons etc. See Comments & Moderation Policy.

Share this:

Discover more from Richard Carrier Blogs

Subscribe to get the latest posts sent to your email.