Understanding Historical Sources: Primary, Secondary and Questions of Authenticity

.

.

In my recent post Comparing the sources for Caesar and Jesus I referred to Historical Evidence and Argument (2005) by the historian David Henige. It contains an excellent chapter on the problems historians face with various kinds of source materials. It’s the sort of work not a few theologians who regard themselves as historians yet who have had little formal training in history beyond their field of biblical studies would do well to read. As for the rest of us, it can help clarify our understanding of the sources that lie behind the stories and arguments we read about the origins of Christianity.

Sources are commonly said to fall into two types. (Henige discusses more than two but I focus here on the main ones.)

1. Primary sources

Confusion sometimes arises depending on whether the historian is referring to “absolute” or “relative” primary sources.

The latter approach [i.e. primary in the relative sense] allows considerably more latitude, perhaps too much, in that whichever sources we have that are — apparently — closest to the events we are interested in are duly termed “primary,” even though they might be separated by centuries from these events. By this way of thinking, historians would always have access to something called “primary” because each historian can define the term idiosyncratically. (Henige 2005: 43)

What is meant by primary in the “absolute” sense?

Leopold von Ranke, and before him John Lingard, held a more stringent view; only a source that was at least “contemporary” can justly be considered primary.1 This sounds reasonable and would help provide consistency . . . (pp. 43-44)

The footnote is to the following:

Ranke needs no introduction; for Lingard see Jones, John Lingard.

If a name needs no introduction we can assume that every historian worth her salt knows the person and their contributions to their discipline. Long term readers of this blog and of my own exchanges with two New Testament professors have experienced the pain and dismay of learning that among scholars of biblical studies who insist they are “historians” there are those who have not heard of Ranke.

These biblical historians rely upon the relative meaning of primary sources. They have no choice. No “absolute” primary sources exist for the study of Christian origins.

So what are they missing out on? What sorts of primary evidence lies (either directly or indirectly) at the base of most other historians’ works? Let’s continue with my take on Henige’s discussion.

Even the idea of a “contemporary” source can be a fuzzy one. Both a high-brow national broadsheet and a local rag can publish a story on the same day. The two stories may be given very different emphases and be told in very different detail. What were the respective reporters’ sources? Were they eyewitnesses? Were they equally well placed to observe? How observant were they? How much did they recall and how much additional colour did their memories soon create?

Henige tells us of three levels of (absolute) primary evidence: participant, eyewitness, hearsay.

The first of these, the participant, sounds like the most reliable, but in actual fact it might be the least. Take the journal of Columbus. That would appear to be a first-hand account but scholars have been troubled by the fact that it’s not in Columbus’s own handwriting. Our earliest copy has been penned by Bartolomé de las Casas (two generations after Columbus). That inevitably raises questions about how much we really know about Columbus’s own thoughts:

but the journal is riddled with inconsistencies and incongruities, suggesting a massive early rewrite by Columbus for his own purposes. It can also be shown that, as a transcriber, Las Casas wielded an attitude as well as a pen. (2005: 44-45)

2. Secondary sources 

Henige here is speaking of sources closest to the events they describe but are not contemporaneous. Secondary sources are necessarily dependent upon other sources that may be lost to us. (Compare “relative” primary sources above.) Even if we accept Richard Bauckham’s case that the canonical gospels are directly dependent upon eyewitnesses they are still secondary sources in our context here.

The problem with secondary sources is that we don’t always know what sources (if any) that their authors used, or if we do know, we don’t always know how creative they have been with their use of them. Henige cites three examples to illustrate:

1. The Atlantic slave trade. Historian Philip Curtin decided to check the many secondary sources about the Atlantic slave trade that routinely cited a figure of around 50 million Africans being transported from their homelands. The 50 million figure had become virtually “canonized”: it was part of “common knowledge”. He learned that all secondary literature on which these numbers relied in fact ultimately derived it from a single mid-nineteenth century journalist’s account. Further investigation led him to conclude that the figure was an exaggeration by a factor of around five. (Curtin, Atlantic Slave Trade, 3-13, and passim.)

In another publication Henige quotes the following from the above work:

2. Smallpox epidemic in Hispaniola in 1507. “Belief in a smallpox epidemic in Hispaniola in 1507 also gathered speed and credibility by a chain of reciprocal, if often unannounced, dependence.” (Henige, “When did Smallpox Reach the New World?” — The full text of this chapter is available online via the link.)

Henige’s chapter on smallpox stands as an instructive contrast against many works of biblical history by too many New Testament scholars. Firstly, Henige reminds us that an argument from silence is not necessarily a non-starter:

The argument from silence can never be completely foolproof, but in this instance it must be regarding as convincing. If nothing else, it hardly bears imagining that Las Casas (who was in Hispaniola in 1507) would have failed to mention an epidemic of smallpox in that year had it occurred; after all, he was not entirely averse to mentioning other catastrophes which had not occurred.

Finally, I might mention one further early account which adds more silence, yet more muscle, to the argument that there were no smallpox epidemics in the New World prior to 1518. (1986: 18f)

Henige understands the temptation historians face with the secondary evidence so cogently offers clear answers and explanations. At the same time he recognizes the circularity of such solutions when they are based entirely on secondary sources.

I began by pointing out that the date of 1507 had recurred with increasing regularity and that it would prove a valuable and much-needed argument for the proponents of large numbers of Indians in Hispaniola and in other parts of the New World since, by their model, it would have provided an early and excellent mechanism for rapid depopulation and, better yet, a mechanism which would provide no opportunities for disconfirmation.

Although the proponents of large populations have not yet pushed this particular argument very strenuously, the present discussion may have some value in precluding at least this part of the very circular, yet somehow persuasive, line of reasoning that characterizes that cause.

. . . . It is easy enough to assert that the 1518-1519 epidemic was simply the first recorded one there [i.e. the first one to be officially documented] . . . but to do so merely begs the question in a particularly desperate way.

While it is certainly neither possible nor expedient to argue that no newly-imported infectious diseases raged outside the purview of European observation early in the sixteenth century, this must remain (probably for all time) no more than a not unreasonable assumption, entirely devoid of the evidence required to permit serious arguments based on it. (1986: 21)

3. The bison population. When speaking of the bison population in North America for a long time historians “blindly followed an early estimate of 75 million or more” but research has shown the the number should be less than half of that figure. (Isenberg, Destruction of the Bison, 23-30).

Only three examples, but a clear pattern that shows

• the need to test the apparent independence of our sources,
• and reminds us that, while we might begin with the source nearest to the events described, we should certainly never end with it.

(Henige 2005: 45 — my formatting and bolding in all quotations)

The above line of reasoning would pull the rug from beneath the feet of the well-guarded traditional models of Christian origins. That does not mean similar models might justifiably replace the current ones, but the key word here is “justifiably”.

Distinguishing among sources

One of the historian’s most formidable tasks is distinguishing among sources. It is tempting to treat a source as primary if it says what we want it to. It might well be primary or it might be a dependent source masquerading as primary. (2005: 45)

Pompeii illustrates the point. The true primary source is the excavated site itself. Eyewitness reports are necessarily relegated to a lesser status. Over time material is removed from the excavated site and reconstructions are erected there. So over time the excavated site of Pompeii can become less primary than accounts of it before the renovations began.

Sources can thus change over time the way medieval churches evolved through successive generations to be extended and reshaped to become quite different from their original plan.

To bring this back to our sources for earliest Christianity, in the same way texts (as well as oral accounts) can be regularly modified over time as they serve the “changing collective memory and self-identity of a community”. Critical scholars of the Bible have generally recognized that our canonical Gospels have emerged through such a process although it appears that a growing number of theologians are now questioning the extent to which this view holds weight.

A source can be either primary or secondary depending on the use to which we put it. A secondary source can become a primary source when it is used to study the beliefs and culture of the time and society in which it was composed.

The Lure of the Eyewitness

Historians instinctively warm to evidence that purports to be eyewitness. How much more primary can a source be? 

Historians from Herodotus to the present have taken care to assure their readers that at least some part of their testimony was ocular. Some scholars regard Herodotus’s visit to Egypt as fictional, but this is hardly Herodotus’s fault. More than thirty times he assured his readers that he saw or heard what he wrote. (2005: 46).

Speaking of Herodotus (our renowned “father of history”) scholars have come to increasingly doubt his assurances that he relied upon eyewitness sources. One of these is Detlev Fehling, author of Herodotus and His ‘Sources’ (1971, English trans 1989). I hope to post one day details from this book along with related subsequent scholarship but for now I’ll briefly note the following:

I begin with a fairly straightforward question. Up till now scholars could believe they had a rough idea of Herodotus’ travels. But now that nearly all his statements based on avowed personal inspection (autopsy) have turned out, contrary to the natural assumption, to be pure fiction and the cornerstones of the prevailing opinion have thus been wrenched away, we are left in complete uncertainty, at any rate as to what can be stated positively. . . . (1989: 240f)

Indeed, there survive today royal inscriptions carved in rock that Herodotus reported on in order to give authenticity to his account. Even though we know these monuments were there for anyone to check for themselves the claims Herodotus makes about them are simply false:

In 2.106 Herodotus mentions two rock reliefs in Asia Minor, which are still in existence; and what he says about them is obviously wrong on several points.

• The most important point is that he states quite incorrectly that the inscription runs across the shoulders.
• Furthermore, he describes the script, again quite incorrectly, as Egyptian hieroglyphics (in fact it is an example of the so-called Hittite hieroglyphics), a misstatement that cannot be explained away as a simple error, since to anyone who has seen the former once or twice they are completely unmistakable.
• To this second point we must then add a further misstatement obviously intended to fit in with it, that the figure is wearing a mixture of Egyptian and Ethiopian armour. Not that there is any point in asking what type of equipment Herodotus could have identified as Ethiopian, as commentators do. Herodotus certainly was not worried. He simply gives a general description suitable for Sesostris of Egypt, who, according to 2.110.1, was the only Egyptian king to rule over Ethiopia . . .  (1989: 135)

It’s not just Herodotus, either. Fehling cites hosts of other ancient authors who falsely claimed to be relying upon eyewitnesses (either those they spoke with or themselves) and other “authentic” testimony. Here is a very small sample:

Diodorus Siculus 1.10ff.: the account of Egypt, much of which is Herodotus in varied and extended form, is given in indirect speech as coming from the Egyptians themselves. 

Dio Chrysostom 11.37f.: an Egyptian priest tells the true story of the Trojan War, drawing on temple archives and (lost) stelae.

Scriptores Historiae Augustae 19.4.1: certain it is that Maximinus ate forty pounds of meat a day; indeed according to Iunius Cordus it was as much as sixty.

Ctesias is said by Photius bibl.72 to have presented most of what he recounted as directly experienced by him or else told him by the Persians. —idem ap. Aelian nat. an.4.21 (= frg.45d): he has seen the fabulous Indian beast called the martichoras in the Persian court.

Apion FGrHist 616 F 5 (= Aulus Gellius 5.14; cf. Aelian nat. an.7.48): claim to have personally witnessed the lion recognising Androcles.

Pausanias 9.28.2: Pausanias heard from a Phoenician how a snake killed a man with its breath.

Aelian nat. an.5.47: claim to have witnessed a blinded lizard’s sight being restored. The  passage  is  related  to Pliny  h. nat. 29.129f.,

Seneca benef. 2.19.1: claim to have witnessed something like what Apion claimed to have witnessed (see above).

Aulus Gellius frequently disguises the fruits of his reading in the form of a conversation with friends etc.

Tacitus Germ.3: an altar built by Odysseus and other monuments with Greek inscriptions.

Ammianus Marcellinus 19.9.9: a tribune mentioned by name counts 30,000 Persian dead after the battle, his task being facilitated by the fact that the Roman bodies decomposed rapidly, whereas the Persian ones simply dried up.

And Richard Bauckham notwithstanding, the Gospels of Matthew and Mark don’t claim any eyewitness testimony while that of John attests eyewitness verification of some rather unusual and miraculous events. As for Luke, see posts on the prologue. Enough said.

Cornucopia of details to support the argument

A stern test of the historian’s integrity comes when he is faced with evidence that is particularly full and seems to offer aid and comfort to his argument. (2005: 47)

Example: Hernando de Soto’s expedition through the southeastern United States. Four volumes are known to exist; three appeared within a few years of the return of the survivors, the fourth sixty years later. The first three are largely matter-of-fact first person accounts; the fourth, longer than those three combined, is a rhetorically florid and detailed account purporting to be the remarkably detailed reminiscences of an unnamed participant. This fourth volume is rich in data about what the Indians and Spaniards “did, said and even thought”.

Even in its practical details, however, Garcilaso’s account taxes credulity on every page. Sizes of rooms and fields, distances that arrows bounced off helmets, the giveaway florid speechifying by all hands presumably remembered by his remarkable informant—these and scores of other clues warn us that Garcilaso [=this fourth volume] is not to be trusted very far. (2005: 47)

Despite all of these warning signs the Garcilaso volume is still the one most relied upon by modern students. Most depressing of all is the following observation:

True, there are now some feeble and formulaic cautions expressed by those haring after de Soto’s route. But once expressed, these are cast aside, and the hunters continue to use Garcilaso unflinchingly, just as if he had passed every test rather than failed them.

Compare similar observations I have posted here before:

Laziness is common among historians. When they find a continuous account of events for a certain period in an ‘ancient’ source, one that is not necessarily contemporaneous with the events , they readily adopt it. They limit their work to paraphrasing the source, or, if needed, to rationalisation. — Liverani, Myth and politics in ancient Near Eastern historiography, p.28.

There has been a very strong tendency to take the Biblical writing at its face value and a disinclination to entertain a hermeneutic of suspicion such as is a prerequisite for serious historical investigation. It is shocking to see how the narrative of the Nehemiah Memoir has in fact been lazily adopted as a historiographical structure in the writing of modern scholars, and how rarely the question of the probability of the statements of the Nehemiah Memoir have been raised. (Clines, What Does Eve Do to Help, p. 164)

These observations warn us that even among scholars who are without any faith allegiance we could expect the core outline of the Jesus narrative to be taken for granted as reflecting genuine historicity.

Another illustration Henige gives is Caesar’s Gallic War. Beautiful: an ancient text written by an eyewitness. But wait, . . .

[T]he earliest surviving manuscripts date from the ninth to twelfth centuries, and the two major manuscript traditions differ in numerous ways.

And so it goes for many works from antiquity that have been preserved by the monastic copyists.

Using these sources uncritically forces historians to assume that nothing happened to alter them in the millennium or so between their first writing and our first glimpse of the text. As a working assumption this is unavoidable, but it can seldom bear the onus of serious scrutiny.

Yet blood has been spilled over dogmatic beliefs in the authenticity of single words and phrases in the New Testament.

The fact is that the self-witness of a text cannot be the arbiter of how we read any text.

Several narratives stipulating participant status have recently come under fire as possible fabrications.

A work professing to be a pre-Marco Polo account of southern China [=Jacopo d’Ancona, City of Light] recently appeared, but the ‘editor’ of the work has steadfastly refused to reveal the whereabouts of the manuscript on which it purports to be based, daring critics to disbelieve him, and they have obliged. 

In another case, the editor/author has also declined to produce one of the manuscripts on which he claims the work is based. This work has been used heavily and been granted the weight of its claims. It was published by a university press and assertions about its accuracy and authenticity accompanied it, so why not just believe? [i.e. Sharlet, “Author’s Methods,” referring to Earp, I Married Wyatt Earp.] (2005: 48)

Such discussions among non-biblical historians ought to put an end to accusations by too many New Testament scholars that raising radical questions about sources is not hyper or selective scepticism: some of us are merely attempting to treat the biblical texts the same way any historian ought to approach any primary or secondary source.

There is much more to come yet. Henige even wades into instances of biblical scholarship himself. To be continued, etc…..

 

---

---

 

Fehling, D. 1989, Herodotus and His ‘Sources’, Francis Cairns, Leeds.

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

---

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

---