Testing (or not) Historical Sources for Reliability

Continuing from the previous post. . . .

Fallibility of eyewitness accounts

Eyewitness accounts are not necessarily more reliable than other sources. Timothy Good compiled 100 eyewitness accounts of the assassination of President Lincoln and its immediate aftermath in We Saw Lincoln Shot: One Hundred Eyewitness Accounts. David Henige comments in Historical Evidence and Argument (2005):df

Reading these reminds us of the omnipresent Rashomon effect, and also that a secondary account that collects and evaluates a number of primary sources might actually be preferred to these, even when it paraphrases them, as long as it does this well, and as long as it allows access to all the evidence. (2005: 48 — Formatting and bolding mine in all quotations)

We have all heard of the studies that demonstrate the depressing unreliability of memories of events witnessed and experienced. Henige cites several articles addressing many of these studies and I attempted to follow up a few to flesh out details. One common theme is the way false memories can be implanted as a byproduct of others asking a witness questions that introduce the possibility of details that were not originally seen (e.g. Wells and Olson).

Here are a few pertinent sections from Toward a Psychology of Memory Accuracy by Goldsmith, Koriat and Pansky:

• Although thinking about a perceived event after it has happened helps maintain its visual details, thinking about imagined events also increases their vividness, and may therefore result in impaired reality monitoring for these events (Suengas & Johnson 1988). Goff & Roediger (1998) found that the more times subjects imagined an unperformed action, the more likely they were to recollect having performed it. . . . .
• The fact that people know at one time that a certain piece of information was imagined, dreamt, or fictional does not prevent them from later attributing it to reality (Durso & Johnson 1980, Finke et al 1988, Johnson et al 1984). . . . ;
• In comparing the results for an immediate test with those for a test given two days later, the proportion of accurate recall declined over time, whereas false recall actually tended to increase (McDermott 1996).

Nor does the research support the belief that false memories are necessarily the product of trauma and psychological repression:

Many cognitive psychologists, however, doubt these assertions (Lindsay 1998, Loftus et al 1994), pointing instead to evidence suggesting that false memories may arise from normal reconstructive memory processes.

Henige’s conclusion:

We can hardly re-enact the life experiences of eyewitnesses from the past to judge their capacity with respect to memory. The alternative is to conduct large-scale and repeated experiments that test various kinds of memory. As noted, hundreds of these have been carried out and in general the results have not been encouraging for any historians who might wish to believe eyewitnesses implicitly.

Testis unus, testis nullus, One witness is no witness

In ancient history scholars can find themselves depending more often than not single sources for what they know. One would expect this difficulty to make historians more cautious about how they interpret and rely on this solitary pieces of data for various arguments but unfortunately the opposite is found to be the case far too often.

There is a natural tendency to treat unique evidence with kid gloves.²²  (2005: 49)

Henige’s footnote no. 22 brings us to a biblical scholar as a negative example:

22 Or even attempt to turn it to advantage, as R.N. Whybray does when he writes: “[t]o regard as useless for the historian’s purposes the only account of a nation’s history written by its own nationals is, to say the least, extraordinary.” Whybray, “What Do We Know,” 72.

Naturally an “only find” does deserve preservation. No-one disputes its importance. However,

that fact by itself should persuade the historian to apply every form of internal criticism possible. (2005: 49)

Take Josephus. Many historians (Henige remarks) place a very high value or give a greater credibility to Josephus simply because he is the only source for much of what we (think we) know about first century Palestine.

Another example Henige uses to showcase the point are the scores of scholarly publications that have appeared attempting to explain what was meant by the Star of Bethlehem. All of these authors generally accept that there really was some kind of astronomical phenomenon even though we have only one source for the tale — the Gospel of Matthew. Many scholars have sought to unravel what this phenomenon rather than asking why we have no other source (not even in any other gospel) making any mention of it. The real mystery, Henige points out, is that no other Gospel mentions it even though their authors must have associated with the same circles:

Written at about the same time and from within a similar milieu, Luke’s silence weighs heavily against accepting Matthew’s version, and suggests that this unexpectedly comprehensive disagreement is the real mystery, and that all studies devoted to defining and dating the Star of Bethlehem are largely futile. (2005: 50)

Appeals to earlier sources

When Josephus about events in the history of Tyre to show how they supported the Biblical account of Solomon’s building of the Temple he took pains to assure readers that he was making nothing up. So in Against Apion we read:

I will now therefore pass from these records, and come to those that belong to the Phoenicians, and concern our nation: and shall produce attestations to what I have said out of them. There are then records among the Tyrians, that take in the history of many years: and these are public writings, and are kept with great exactness; and include accounts of the facts done among them, and such as concern their transactions with other nations also: those I mean which were worth remembering. Therein it was recorded, that the temple was built by King Solomon at Jerusalem, one hundred forty three years, and eight months, before the Tyrians built Carthage: and in their annals the building of our temple is related. . . .

And again

Now, that this may not depend on my bare word, I will produce for a witness Dius, one that is believed to have written the Phoenician History after an accurate manner.

And again

And now I shall add Menander the Ephesian, as an additional witness. This Menander wrote the Acts that were done both by the Greeks and Barbarians, under every one of the Tyrian kings, and had taken much pains to learn their history out of their own records. Now when he was writing about those kings that had reigned at Tyre, he came to Hirom, and says thus: “Upon the death of Abibalus, his son Hirom took the kingdom; he lived fifty-three years, and reigned thirty-four.

(Dius is otherwise unknown; Menander is also largely unknown)

These assurances by Josephus of his reliability have impressed a number modern historians. They along with Josephus’s abundance of details pleased H. J. Katzenstein so much in The History of Tyre (1983) that he wrote:

The reliability of the document is uncontested by scholars, and the fact that the numbers which pertain to the life-spans of the kings or their regnal years are not rounded, underscores the veracity of the list and its contents. (1983: 118)

The mere mention of the Tyrian archives by Josephus likewise impressed William Barnes sufficiently for him to write in Studies in the Chronology of the Divided Monarchy of Israel (1991):

Few would dispute the basic authenticity of Josephus’ tradition, especially in light of his own appeal to the ‘public records’ to corroborate his polemic against the enemies of the Jews and his attempts to glorify the Jewish people. (1991: 32)

Yet notice (as Henige certainly does) that Barnes’s affirmation itself contains reasons to question Josephus’s claim. The context is polemical and Josephus is driven by an agenda “to glorify the Jewish people.” Clearly Josephus’s opponents were not as impressed by his claims as are many modern historians.

Barnes’s very language seems to make a case against him — a case strengthened by thousands of examples of false referencing throughout history. (2005: 52)

Recall Henige’s warning that the fact that we have only a single source for any event

should by itself should persuade the historian to apply every form of internal criticism possible.

And against historians who naively embrace at face value Josephus’s assurances that he can be trusted to clear the polemical air with the true facts of the matter

Josephus’s testimony is surprisingly, even suspiciously, rich. He provided not only the lengths of reign of several rulers, but their ages at death as well. . . . According to Josephus, Ithobaal succeeded to the throne by assassinating his predecessor. By this testimony Ithobaal was a wonderfully precocious young man: assassin at sixteen, but father at nine! And if Pygmalion was Metten’s son (implied though not stated), then Metten was a father at eleven. (2005: 51)

(Henige does not even get to the point of discussing the failure of archaeologists to uncover any indications of a Solomon led kingdom undertaking the construction of a great temple with or without the aid of the king of Tyre.)

Discrepancies like these are rarely enough to undermine a willingness to trust a sole source. Variant readings of the lengths of reigns in the manuscript tradition are not interpreted as evidence that early copyists noticed and attempted to rectify the flaws in Josephus’s original account but are seized upon as lucky finds when they can be used to show that Josephus’s narrative happily fits the Bible’s chronology.

Not even contradictory archaeological finds have been able to seriously tarnish confidence in Josephus.

When archaeologists uncovered an Assyrian inscription that introduced a new king of Tyre, Ba’limanzer, who was not found in Josephus’s list supposedly drawn from the Tyrian state archives. To maintain the credibility of Josephus’s account some historians interpreted Ba’limanzer as an alternative for the earlier king Balezor in Josephus’s list; others (e.g. W.F. Albright) interpreted the Assyrian evidence as a name that had been accidentally omitted by Josephus or his sources. Those who chose the former, of course, were also obliged to make adjustments to the length of his reign and different historians found different ways to do this (e.g. Katzenstein, Lipinski, Penuela).

Attempts to rehabilitate Josephus by bouncing Balezor around like a pin-ball (866-849? 859-853? 856-850? 856-830? 855-841? 848-830? 847-841? 846-841?) do nothing for the credibility of either Josephus or modern historiography. (2005: 53)

Giovanni Garbini (History and Ideology in Ancient Israel, 1988) also discusses various discrepancies in Josephus’s account and concludes:

The chronology of the kings of Tyre which is displayed here with a mass of detail is aimed simply at demonstrating that ‘in these [i.e. in the Annals of Tyre] it is written that in Jerusalem the temple was built by King Solomon 143 years and 8 months before Tyre founded Carthage’ (Contra Apionem 1,108), thus relying, evidently rightly, on the gullibility of readers who would have found it quite natural that what happened in Jerusalem was recorded in Tyre, not to mention calculating the year with a reference to future events.

Since the year cited with such exactness in Contra Apionem does not correspond to that given in the Jewish Antiquities, it seems clear that there was virtually nothing in the Annals of Tyre about Jerusalem and its temple.

Flavius Josephus, harmonizing the information in the Bible with that of the other ancient sources in his own way and without any concern to avoid contradictions, used precisely the same procedure as that followed by modern scholars who accept the biblical data without discussion, adapting everything else to them;

Josephus was not a historian but simply the first Alttestamentler – and it is not for nothing that his Jewish history, like the modern histories of Israel, is only a paraphrase of the biblical text. (1988: 23f)

Josephus was evidently not overly worried by any possibility that readers would dash off to consult the Tyrian archives for themselves — a scenario a number of modern critics appeal to in order to justify naive readings of ancient texts. Garbini doubted that such archives would contain details about the building of Solomon’s temple but Henige asks a more radical question:

How did these records survive as late as the third century BCE (Menander‘s floruit) anyway [or even to Josephus’s own time]?

After all, after the supposed events linking the king of Tyre with Solomon Tyre was besieged repeatedly: by the Assyrians, by the Egyptians, by the Chaldeans, and was finally razed by Alexander the Great in 332 BCE. Massacres and deportations followed the wholesale destruction of the city.

It is unreasonable to expect the Tyrian “archives” to have survived, and even less justifiable to profess it without discussion. (2005: 53)

We looked at the evidence for ancient historians more generally fabricating sources for their claims (especially their more fabulous ones) in the previous post. Much more could be and has been written on this tendency among our ancient writers and I don’t think David Henige will mind if I conclude this section with one passage from Katherine Stott’s Why Did They Write This Way? (2008):

In examining source citations in the [Hebrew Bible], an interesting book to consider is Esther, which also contains references to written documents. One well-known example is a reference to “the annals of the kings of Media and Persia” (Esth 10:2), regarded by many as reminiscent of similar citations in the book of Kings. However, it has long been doubted by biblical scholars that this book actually existed. The reference is widely regarded as nothing more than a fictional imitation of the references in Kings, designed to bolster the credibility of the book of Esther. As A. Berlin has observed, 

the Annals are the functional equivalent of the pea at the end of Hans Christian Anderson’s story “The Princess and the Pea.” That story closes with the words: “So the Prince took her for his wife, for now he knew that he had a true princess; and the pea was put in the museum, and it is there now, unless someone has carried it off. Look you, this is a true story.”

(2008: 72)

One wonders if the scholarly scepticism towards the Annals of the Kings of Media and Persia would be mitigated if they were otherwise introduced to confirm more critical chronologies of the Bible.

Independence of sources and diminishing probabilities 

We understand why independence is “a cardinal attribute” of sources. Anyone who reads or watches crime mystery stories knows how anyone who comes forward confessing to a murder is put to the test to see if he can repeat details of the crime that only the real perpetrator would know. If they can repeat nothing more than what is already in the public domain they can be dismissed. Police interview suspects separately to test each one for damaging discrepancies or unprompted consistencies. But of course suspicion is raised if the consistencies are too word-for-word perfect since in that case we appear to have collusion instead of corroboration.

Establishing the independence of sources is critical.

Thus it is that historians strive heroically to demonstrate — or just assert — the independent status of sources on which they depend. (2005: 54)

Once again Henige looks at the biblical scholars:

Pondering the historicity of the Acts of the Apostles, Joseph Fitzmyer writes that “[t]here are a number of incidents that Luke has recounted that find confirmation elsewhere.” It turns out that “elsewhere” is seldom farther away than the epistles of Paul. For many, this will seem too few degrees of separation. (2005: 54)

We dare not mention scholars like Bart Ehrman (Did Jesus Exist? pp. 74-78) who have insisted that the canonical gospels stand as four independent accounts of Jesus (yes, at least two of them used Mark but Ehrman argues they are still independent insofar as they add material not in Mark), and that the noncanonical Gospels of Thomas and Peter and the Papyrus Egerton 2 bring our total of “independent witnesses” (sic!) for the life of Jesus to seven.

Some biblical scholars string together multiple points from various sources, each one of which contains only a limited probability, and erroneously conclude that such an exercise produces “an argument of cumulative weight”. (This method lies at the heart of Maurice Casey‘s and James Crossley‘s attempts to prove that the Gospel of Mark was originally composed in the late 30s or early 40s CE.) The truth is that such a method only reduces, not increases, the probability of the conclusion. This is something Richard Carrier has made clear in his discussions of Bayes’ theorem but the point is made just as succinctly by Henige himself:

The load-bearing capacity of an argument based on a string of unconfirmed sources is dangerously modest. The probability that such an argument is correct founders on the axiom that probabilities multiplied are probabilities diminished. Since no single testimony can have a probability of 1.0, each time these are associated with (i.e., multiplied by) another one, the overall argument is increasingly in danger. Even three sources that the historian deems 75 percent correct amount, when used in concatenation, to a probability of .75³ or about 4 in 10. Reduce the individual probabilities to 60 percent and the result is barely 20 percent. And so on. (2005: 55)

Henige covers other questions that arise from the historian’s use of sources but I’ve addressed those points that are or ought to be in the forefront of our minds when critically evaluating historical reconstructions and arguments concerning Christian origins.

Granting sources Christian charity and civil liberties

One aspect I have not raised here is the problem of not knowing what our sources do not tell us. Authors necessarily take many details for granted when writing for an audience. To cover every detail would obviously be unnecessary and tedious. The problem arises when we approach from a distance of millennia documents produced by a lost culture. It is all too easy to fill in many gaps with anachronisms and we see this often enough when scholars too easily and with minimal justification introduce third and fourth century and even later rabbinic teachings as backdrops to the gospels.

Some biblical scholars (e.g. Richard Bauckham) even argue that “Christian charity” ought to be extended towards their sources: since they are the words of fellow Christians we should follow the Christian ethic and believe them unless extraordinary reasons warn us not to. Apparently this problem is not confined to theologians because Henige adds

In the end, the historian’s attitude about the relationship between his work and his sources comes down to whether he grants the latter civil liberties or not. Is evidence to be treated as true until proved false? Or the reverse?

And what will constitute proof? And, short of proof, what degree of probability is sufficient to build on?

Should greater efforts be made to prove or disprove evidence and hypothesis? Should the weight of the readily available evidence suffice, or should the search for new evidence continue regardless? What are the consequences of being wrong? But if historians are the jury, they are also the prosecution and the defense, or at least have an obligation to take on these roles as part of their investigation of the past, unless certain issues become so debated that the courtroom scenario is replicated. (2005: 56f)

 

---

Giovanni G. 1988, History and Ideology in Ancient Israel, Crossroad, New York.

Henige, D. 2005, Historical Evidence and Argument, University of Wisconsin Press, Madison.

Koriat, A, Goldsmith, M & Pansky, A 2000, ‘Toward a Psychology of Memory Accuracy‘, Annual Review of Psychology, vol. 51, pp. 481-537.

Stott, K. 2008, Why Did They Write This Way?, T&T Clark, New York.

---

---

---

If you enjoyed this post, please consider donating to Vridar. Thanks!

---