Peter Kirby

Your train of thought is interesting. To my mind sprung an analogy from my Computer Science work, that of pre-conditions and post-processing. Let me elaborate on what I mean (excuse the "automatic writing" style).

What I mean is that the so-called criteria (whatever criteria they are) may not be one-off functions of truth. They may have pre-conditions and post-processing. (They may also have post-conditions and pre-processing, but I've probably confused a large segment, so let me back up.)

These are analogies. All processing and condition-evaluation are assumed to be undertaken by fallible humans rather than fallible computers.

Pre-conditions: statements which must be true for the function to work.

Example: the text is written within X years of alleged event.
Example: the text belongs to Y genre or has the intent of Z.
Example: the text does not have many matches of the criteria for a high incidence of error.
Example: the text succeeds in meeting a minimum number of matches by the primary criteria on a first pass.

Post-processing: cleanup steps that need to be taken once the main work is done.

Example: two contradictory claims passed the first stage, so hand them off to an arbitration function.

Now, the post-condition is simple: the claims that pass should be true with a reasonable margin of error.

As to pre-processing: cleanup steps that need to be taken before the main work is done.

Example: a claim is scientifically impossible, so exclude.

Note that this pre-processing and post-processing could be performed on the text in a first, heuristic pass in evaluating the fourth example pre-condition, "the text succeeds in meeting a minimum number of matches by the primary criteria on a first pass."

Now, idle thought, and possibly confusing to the reader, I wonder whether there is a way to use "artificial learning" to evolve the best methodologies.

It is still a serious problem how to settle which facts to accept as facts for use as a benchmark for evaluating methodologies.

Note: If all this talk of claims and facticity makes you uncomfortable, we could transfer the methodological question to another domain, such as text criticism, and how to evaluate the criteria for determining the earliest text. That might turn down the temperature of future discussion a bit, and possibly even shed a little light.

So we have, using the domain of text criticism:

Pre-conditions: statements which must be true for the function to work.
Example: manuscript, or its examplar, was written within X years of early form.
Example: the manuscript belongs to F text family (?).
Example: the manuscript does not have many matches of the criteria for a high incidence of textual error.
Example: the manuscript succeeds in meeting a minimum number of matches by the primary criteria on a first pass.

Post-processing: cleanup steps that need to be taken once the main work is done.

Example: two contradictory readings passed the first stage, so hand them off to an arbitration function.

Now, the post-condition is simple: the readings that pass should be in the early form with a reasonable margin of error.

As to pre-processing: cleanup steps that need to be taken before the main work is done.

Example: a reading is misspelled, so correct spelling (?).

Please tell me if this analogy helps at all in the discussion of methodology.

kind thoughts,
Peter Kirby

krosero

I know nothing about Computer Science, and if your analogies are going to get more complicated than this, start again and walk me through it like I'm a 4-year-old. But let me see if I can paraphrase the quote above: the so-called criteria may not produce conclusions of true/false in one step; there are several steps, such as making clear the prior conditions for a criterion to be applied, then applying it, and finally processing any remaining issues. For instance, you can't just apply embarrassment like a one-off rule; you have to first set the condition that if the text was heard in its time as a fictional story, the embarrassment criterion cannot weigh so heavily, and may even have to not be applied; but if the text was heard as historical in its time, you apply the criterion more strongly. Then process the remaining issues in the same manner.

That's my best guess as to what you were saying. Please simplify or concretize if you can.

Maybe the analogy with "artifical learning" would apply some needed rigor. But I fear it would take much of the intuition and heart out of the work. I would question the whole implication that historiography is the same as the hard sciences, or that an objective truth exists which can be scientifically measured. How are you going to measure the number of times that a text meets a criterion? Well, you can come up with your own count of the number of times, but other scholars will score with different "hits", and will disagree with the ones you picked out. Why does that happen? Lack of rigor? Or the strength of personal and social presuppositions?

I've just started James Dunn's "Jesus Remembered" and I find very promising his argument that reading a text is a personal encounter, one in which the reader and his/her reactions to the text is as much on the table as the data of the text. If that stuff is on the table, there's a chance to limit subjectivity and actually get somewhere.

I would tend to think that people come to radically different conclusions about texts, in most cases invoking science and displaying a prodigous number of "hits" next to low number of "misses", precisely because their presuppositions are not out on the table. The personal encounter is not accounted for; everyone's measuring things precisely, and everyone's got an argument that sure sounds measurably strong, but no one agrees.

Still, I'd like to see your model tried; I'm just not sure I even get it yet.

Peter Kirby

That's it exactly!

That's a possible application of the framework.

Will concretize below with an example from another domain.

Let me explain in a nutshell what I understand by "artificial learning." It's often use to determine the weights given to evaluation function in telling who is winning a chess game. For example, the beginning chess player is told that a pawn is worth 1, a knight 3, a bishop 3, a rook 5, and a queen 9. A simple evaluation function would sum up the materiel on each side and say who's winning or whether it's a dead heat. This is useful for looking ahead several moves and seeing who would take the advantage with each possible maneouver. But how do we know that these are the best weights to use for the pieces? We could set it up against a second evaluation function in which the pawn is worth 1, a knight 2.5, a bishop 3, a rook 5, and a queen 8.5. Then we would have the two evaluation functions "duke it out," play a few games as black and white (and vice-versa) and see which tends to come out on top. The winner moves on (while the "better losers" might be kept around to test against future winners, in the hope of triangulating a new weighting system to resolve any rock-paper-scissors situation--this parenthetical comment can be ignored). After all that, the best of the best is selected as the weighting system to put in a "Deep Blue" (or "Chessmaster 9000").

In this case, it might be a good idea to artificially evolve the weighting system used in the arbitration function. (It's a pity neither Dom nor J.P. thought of this.) To get an idea of how this might be done, take a look at this blog entry on the weighting of various criteria used in HJ research:

Criteria for Authenticity by Loren Rossen

Of course this should be applicable (or be made applicable) to all varieties of hearsay-based (or, oral tradition sifting) history, but HJ research is perhaps the most well-developed field in such methodology. It would be interesting to see finally some cross-polination of disciplines going the other way, as J. Z. Smith once remarked in another context.

With that explained, I will continue with your note...

The refrain to get your biases out in the open is stated so often these days that one wonders whether it is presented as a panacea. It is not, of course. Any Tom, Dick or Harry can say, "I am a middle-class American white male academic with a bent towards...." without the slightest improvement in their work. I won't name names.

Let me take a concrete example from law. Law requires human judgment. Nonetheless, it is often broken down into criteria. There are criteria for all sorts of things, from the non-establishment clause to the principle of fair use. Let me quote an example of the latter:

Without criteria, all we could say is, "that's not fair!" With criteria, we break down the process of forming a judgment into several smaller judgments. These smaller judgments are made based on criteria, sometimes-fuzzy guidelines as to how to make the smaller judgments. Often, people will come to agree on the smaller judgments, when a kneejerk reaction towards the whole would have produced disagreement.

Well, sometimes.

kind thoughts,
Peter Kirby

Toto

After thinking about it, I think that no criteria should be used unless it can be validated with hard scientific evidence - archeology or other sorts of forensic evidence. Otherwise admit that you are just doing literary criticism and have no hope of deciding anything about the underlying history.

These criteria seem to be confined to NT research, for a reason: NT scholars are the only group who feel that they have to find some real history in documents full of legendary tales.

I notice that Loren Rossen rates the criteria of embarrassment as a 3 on a scale of 4 in usefulness. What could this possibly be based on? Has anyone done a study of embarrassing facts in history and legend, and correlated embarrassment with truthfulness? Has anyone even defined what is embarrassing?

Peter Kirby

Please give an example of what such validation would look like.

It's a little early for any said sort of confession by anyone in this thread. Only you and krosero have responded, and none of the suggestions have really been tried out by us.

This impression is because NT scholars, or a subset of them, are more methodologically conscious than their counterparts studying other traditions, who largely (to use Meier's phrase of Vermes) "muddle through." Further, a proper search could reveal that HJ research criteria, or some of them, have analogues in secular history. Finally, there would be nothing wrong with the theology-related scholars being trailblazers, so long as those in other fields can attempt to apply similar techniques.

That is the point of this thread. (Loren's figures are attributions based on reading various scholars.)

Now, an unresolved question. Whence comes our store of pre-methodological historical 'facts'?

kind thoughts,
Peter Kirby

Huon

Sometimes i consider that history is a huge jigsaw puzzle, with millions pieces (is pieces the right word ?). If a group of pieces can be reasonably assembled together (they fit with the other pieces), then, they could make a part of the jigsaw puzzle. Until a new piece comes under the look, and could replace one of the previous pieces, already tentatively assembled. Now, this is embarrassing !

Huon

What is the methodology for writing the (possible) history of a character who is regarded (simultaneously) as a man and an emanation of God ? Obviously, the two aspects cannot be dealt with in the same manner.

When it comes to JC, all the historians have the same written data, from the same (more or less) antique authors. And yet, we have images of perhaps ten different JC. One criterion is to dismiss the descriptions which are too "modern" to be plausible. JC was not, even remotely, a sort of bolshevik !

Toto

For example, archeology fails to validate the Conquest. So we assume it didn't happen.

Forensic evidence validates that Philip of Macedon was buried in a certain grave in northern Greece. So we assume that he was a real person and a king and stories about him might be true.

Did Midas exist? King Arthur? Who knows? and who cares??

Have you done the proper search? Have you found any examples of other fields adopting theological methodologies? I am not aware of any.

Are NT methodologies any better than "muddling through?" I sincerely doubt it.

I suspect that many of our "facts" are incorrect, and are the result of our own mythologies, and I think that it is dangerous to base anything on alleged historical "fact" in religion or politics.

I mean, there are people who believe that George Washington was a Christian and wrote a prayer book and prayed on his knees during the Revolutionary war. We can trace how those myths developed.

But what if someone in the year 3000 is trying to write the history of America in 2000, and the only surviving documents are the collected works of Ann Coulter and some Republican campaign material?

There are people killing each other over national boundaries or other alleged facts derived from historical documents. How crazy is this? History should be for recreational purposes only.

Huon

Is this history ? Why, or why not ? If it is history,could it have been told from the point of view of Herod, and from the point of view of John the Baptist ?

Flavius Josephus, Jewish antiquities 18.109-119 :

Now some of the Jews thought that the destruction of Herod's army came from God as a just punishment of what Herod had done against John, who was called the Baptist. For Herod had killed this good man, who had commanded the Jews to exercise virtue, righteousness towards one another and piety towards God. For only thus, in John's opinion, would the baptism he administered be acceptable to God, namely, if they used it to obtain not pardon for some sins but rather the cleansing of their bodies, inasmuch as it was taken for granted that their souls had already been purified by justice.
Now many people came in crowds to him, for they were greatly moved by his words. Herod, who feared that the great influence John had over the masses might put them into his power and enable him to raise a rebellion (for they seemed ready to do anything he should advise), thought it best to put him to death. In this way, he might prevent any mischief John might cause, and not bring himself into difficulties by sparing a man who might make him repent of it when it would be too late.
Accordingly John was sent as a prisoner, out of Herod's suspicious temper, to Macherus, the castle I already mentioned, and was put to death. Now the Jews thought that the destruction of his army (by Aretas, the king of the Arabian city Petra) was sent as a punishment upon Herod, and a mark of God's displeasure with him.

Ben C Smith

The second part of this statement might be problematic, but the first part resonates with me. Why does it seem that NT research is full of discussion of criteria, while other historical inquiries are not? Perhaps I am just not reading the right kind of outside histories, or enough of them. But when I pick up R. Syme, The Roman Revolution, a classic piece of historical writing, I read no discussion of which criteria the historian found most or least useful in deciding whether or not Augustus really said such-and-such en route to being named dux and princeps. Syme just dives in and presents his hypothesis, and (apparently) hopes that it will come out more cogent and cohesive than its competitors.

Likewise those Patristic scholars that have come up recently on the Gaius thread. They are all (Hill, Brent, Gwynn, and the rest) attempting to reconstruct something of the historical Gaius of Rome (from evidence, I might add, far more fragmentary than we have for Jesus of Nazareth) and yet none of them ever seems to think of laying out his criteria for historicity. Each just presents as plausible a picture as he can based on the available evidence.

I find it interesting that Peter Kirby did not mention criteria in the OP (he spoke of methodology), yet the ensuing discussion (inevitably for Jesus research?) shifted to criteria as soon as the second post. I am not saying that this is a bad thing, nor a good thing, but am rather noting that it seems to differ from the usual historical procedure (again, perhaps I am just not well-read enough to have found discussions of criteria cropping up all over historical Alexander research or such).

So far as the actual historical arguments that I have read go, I find that the concept of historical analogy seems to crop up quite a bit, whether explicitly or implicitly. For example, Allen Brent appears to argue that certain Hippolytan quotations relevant to the question of Gaius of Rome are legendary, at least in part by analogy with other Hippolytan quotations which are (already) better established as legendary. Historians interested in the broad sweep of things often argue that this kind of (political, social, religious) revolution is often countered by that kind of (political, social, religious) reaction, and then try to explain the data in the relevant extant texts according to such a model. That is argument by analogy.

As for the second part of your statement, Toto (NT scholars are the only group who feel that they have to find some real history in documents full of legendary tales), is that not begging the question at hand? What use would you have for a methodology that sifts historicity from ahistoricity if you have already decided in favor of ahistoricity? If it is merely a matter of saying that document X is clearly legendary here and there, so it must also be legendary everywhere, such an assumption would rule out the vast majority of ancient texts as witnesses for history. (Lucian, How to Write History 60-61, presupposes that the historian will include myths in his history, and merely gives advice on how to present them responsibly.) Perhaps that is your point. Perhaps you do not think that it is possible to arrive at very much actual history from the ancient period.

Interesting topic.

Ben.