Proving History: Bayes's Theorem and the Quest for the Historical Jesus, Richard Carrier 2012:

# ch2

Hence, the burden of proof clearly falls on anyone who would challenge an existing consensus, despite repeated attempts to deny this. For example, in the matter of whether Jesus actually existed as a historical person, historicists have already met the burden of evidence to produce a consensus of qualified experts. So the deniers of historicity must overcome that burden with their own. Attempts to argue that the current consensus has been improperly generated have some merit (e.g., many historicists do make assertions out of proportion to the evidence, or simply cite “the consensus” without checking how that consensus was actually generated). But such arguments are still inadequate, since there is certainly prima facie evidence for a historical Jesus (hence, historicity need not be asserted dogmatically, even when it is). Moreover, such a claim (that there is an improperly generated consensus) must still meet its own burden of evidence. In fact, any claim that the consensus is actually wrong (and not merely unfounded) requires meeting an even greater burden of evidence (in defense of some alternative theory).

Axiom 8: A conclusion is only as certain as its weakest premise.

It's essential to watch for the weakest link in any argument, because very often a single weak link will render all resulting conclusions just as weak—or, by their accumulation, even weaker. This frequently happens in historical reasoning when qualifiers are snuck in without accounting for their logical consequences. For example, in any argument, analyzed formally, if there is a premise of the form “maybe x,” then any conclusion depending on that premise cannot be any more certain than “maybe.” This follows for any qualifying language (like “probably,” “possibly,” or “perhaps”). For example:
MINOR PREMISE:
Alexander might have wanted to assassinate his father Philip.
MAJOR PREMISE:
If Alexander wanted to assassinate his father, then he probably arranged his assassination.
CONCLUSION:
Therefore Alexander probably arranged Philip's assassination.
This is a fallacious conclusion, since the minor premise does not say “probably” but “might,” which makes it a weaker premise than the major premise requires. But the conclusion cannot be more certain than the weakest premise. Therefore, the only valid conclusion one could produce here would be “Alexander might have arranged the assassination of Philip,” which is such a trivial conclusion as to be useless (since it's true of practically anyone of the time, and predictive of nothing).
This may seem an obvious point to make, but ignoring it is a frequent error among historians, who often make bold assertions from more hesitant premises, forgetting that somewhere along the line of their argument one of their key supporting points was more speculative than their conclusion would suggest. And it only takes one weak premise to render all resulting conclusions equally weak, no matter how strong every other premise may be or how many other strong premises there are. Yet I've seen both sides of an argument take ambiguous evidence like this and derive unambiguous conclusions from it, making this a remarkably common error.

Nevertheless, this axiom cannot justify relying on gratuitously bad scholarship. Those who should get little or no mention are scholars who do not employ an adequate method of citation and referencing and consequently make many dubious or false claims or claims incapable of confirmation. Their work is of no use to laypeople (who can't assess which claims are credible or dubious) and of little use to experts (who have to redo all their research anyway before trusting what they say, which negates the point of reading them). Some scholars straddle the line, having insightful things to say, yet I have to fact-check anything they said before relying on it myself—the punishment for which is that I don't cite them if I had to do the work. I just cite the evidence instead.

10. This was an actual debate, which has been conclusively resolved from extensive examination of the evidence (comprising a good example of the complexity of generalizations and the evidence required to establish them): everyone in antiquity read silently when they wanted to, read aloud as performance and entertainment more frequently than we do, and just as we read aloud sometimes even to ourselves, so did they. See William Johnson, “Toward a Sociology of Reading in Classical Antiquity,” American Journal of Philology 121, no. 4 (Winter 2000): 593–627; A. K. Gavrilov, “Reading Techniques in Classical Antiquity,” Classical Quarterly 47 (1997): 56–73; M. F. Burnyeat, “Postscript on Silent Reading,” Classical Quarterly 47 (1997): 74–76; and Bernard Knox, “Silent Reading in Antiquity,” Greek, Roman, and Byzantine Studies 9 (1968): 421–35...32. This happens a lot in history, an outcome many scientists will find unfamiliar. Imagine a drug study for treating the common cold in which all five of the first patients in the cohort immediately die as soon as they receive the treatment; the study would be canceled at once, no further collection of data required. Though it's statistically possible those deaths were a fluke, the odds are already too small to credit that hypothesis, even with a sample of only five. Analogously, once we find five diverse and independent cases of silent reading in the extant evidence from antiquity, we don't have to bother collecting more data to conclude ‘ancient readers could and sometimes did read silently’ (see n. 10 for chap. 2, pp. 298–99).
Shared publicly