More on ur-source theories vs. undesigned coincidences
(Originally published at What's Wrong With the World. Link to original post at 'permalink' below.)
Several years ago, after receiving a question by e-mail, I wrote this post about attempted "ur-source" hypotheses as an alternative explanation of undesigned coincidences. If you're interested in this question, I strongly urge you to read that post (if you haven't already). This post is meant to be a supplement to it, not a replacement.
In the last few days I've been writing quite a bit more about this type of objection to undesigned coincidences because I received some questions about it again. I wrote up so much material in response that I've decided to post some of it here for others who might find it useful.
As used in this post, an "ur-source hypothesis" intended to explain away an undesigned coincidence would go something like this: The accounts we have in our actual Gospels are separated pieces of what was a larger, earlier, oral tradition. They fit together in an explanatory way because the earlier "tradition story" fit together in a plausible way, though it may or may not have been true. It could be seen to be probably true only by some other argument--for example, because it was in existence early. Of course, we do not have this earlier oral tradition and are only conjecturing about its content and appearance. Then (goes the theory), our gospel authors chose to report different parts of that earlier oral ur-source, which produced the appearance of a coincidence between the gospel accounts that we have.
I urge you (again) to read the earlier post, which makes several very important points in response to this. These include, for example, the implausibility of a copier's leaving something highly surprising and inexplicable unexplained by copying only a fragment of an earlier tradition. Also included in the earlier post is the fact that any critic like Bart Ehrman who implies that the unique material in the Gospel of John was invented much later than the other Gospels cannot consistently, simultaneously, hold that the unique Johannine material existed in a stable oral tradition, combined with material now found in Mark and the other synoptic Gospels, before the Synoptic Gospels themselves were even published.
I am not going to deal in this post with a different version of an "earlier source" hypothesis according to which only a portion of what we now find as a UC was already "known to the community" and a later fictionalizer simply added fictional aspects to the story that fit in with other facts already present "in the tradition." That type of theory is already extensively dealt with in Hidden in Plain View itself, since in the book I assume for the sake of the argument that Mark was available to Matthew and Luke, Matthew was also available to Luke, and all three of the Synoptics were available to John. I then discuss, repeatedly, the implausibility of a later fictionalizer's subtly connecting his fictions with material already available in the earlier written Gospels. The very same considerations I give there apply to already available oral versions of the stories.
Concerning "ur-source" hypotheses according to which both sides of a UC were present in some oral ur-source and just got separated in the Gospels we have, I have decided from my more recent correspondence that it is very important to emphasize this: The mere fact that many scholars think that there were some kind of earlier oral traditions in existence prior to the writing of the Gospels does not mean that such earlier "oral traditions" took a form that would make an ur-source hypothesis a good alternative to an undesigned coincidence between the accounts we have. Some kind of earlier oral traditions could mean many different things. Such a phrase, for example, could merely mean that the eyewitnesses themselves talked about their experiences. That is not the same thing as a compendium including both/all parts of an undesigned coincidence. Even if some witness's version of his story were repeated by others fairly faithfully and (say) Luke got hold of it at a couple of removes and put it into his Gospel, it would still be just one side of the story and hence could easily participate in a UC with something told in John. It wouldn't in that case be an "ur-source" from which Luke broke off a piece.
An idea that is popular (partly popularized by Richard Bauckham) is that there were stable, oft-told story versions of the stories in the Gospels (particularly the Synoptic Gospels) and that these were presided over by "tradents" (e.g., the apostles) to make sure that their retelling did not stray from the truth of what actually happened.
Bauckham is using such a theory as an alternative to the much more liberal ideas of form criticism, of which he is an opponent. The more liberal alternative is that the retellings of the stories varied without control and that what has made it into our Gospels is something many times removed from what actually happened and greatly morphed. In other words, Bauckham is arguing against the "telephone game" idea such as one hears from Bart Ehrman.
But it's important to remember that, if a Gospel author either was an eyewitness himself or talked to eyewitnesses, he would not need to rely on such formal, oft-told versions of a story. Bauckham expressly denies Matthean authorship, so he apparently considers the "traditional story" theory to be particularly relevant both to Matthew and to Luke. He thinks that Mark shows strongly the witness perspective of Peter and that John (though not John the son of Zebedee) was a disciple of Jesus and an eyewitness of many things, so the "tradent-certified story version" theory seems most relevant, in Bauckham's view, to Luke and Matthew. If Matthew was an eyewitness (as I believe he was and as patristic evidence supports), the theory becomes even less relevant.
Even for Luke, it's important to remember several things: An oft-told, preserved story version could have been much like the one-sided story we have in our Gospels, not at all like a larger ur-source. (This is similar to what I said above about what could have been the case even if Luke got his version of a story at a couple of removes.) And if the author was able to talk to a witness himself, such a formal version, told by a non-eyewitness, is an unnecessary theory. Since we know that Luke probably did have such opportunities, although it is not impossible that he got some stories at multiple removes, we should not assume that his Gospel is heavily dependent upon formal earlier "oral traditions" as opposed to informal interviews. And we certainly have no reason whatsoever to think that Luke's Gospel is heavily dependent upon formal oral traditions that looked very unlike what he records, off of which he has broken bits and pieces.
The fact that a theory of tradent-certified oral traditions is an improvement over an ever-morphing telephone-game tradition, finding its way ultimately into our Gospels, does not mean that we should settle on the idea that our Gospels were constructed out of such oral traditions as opposed to being known in more natural ways.
Of course, the theory that the Gospels were written by witnesses or in direct consultation with eyewitnesses is "on the table" only because we know on independent grounds that this is historically quite possible. If we knew independently that they were written several hundred years later, they would have to be based upon some kind of earlier, preserved materials, whether oral or written, though that wouldn't (even so) mean that these earlier materials took the form of umbrella sources containing all the material, subsequently dispersed into multiple documents. (Even in the case of Old Testament stories in Kings and Chronicles, where the books themselves may have been compiled long after the fact, we rightly do not assume that what lies behind somewhat different versions of a story in Kings and Chronicles was some ur-source that somehow combined both.)
Any hypothesis that a Gospel was constructed out of much earlier oral traditions is even more implausible, if possible, for the Gospel of John, where the plethora of unique material (which has so often been used to argue against the Gospel's historicity) is actually evidence that the Gospel was written by a witness who wished to supplement what was already known. At least, that is the case once we see John's unique material confirmed over and over again, as it is by undesigned coincidences and external confirmations.
It is possible that the idea that the Gospels (including John) were constructed out of earlier oral traditions is thought to be independently supported because of some term usages by New Testament scholars. "Moderate" NT scholars have a "complimentary" way of using a phrase like "based on earlier tradition" that absolutely must not be allowed to confuse us. When, for example, such a scholar comes upon a case where he believes that a pericope satisfies one or more of the famous "criteria of authenticity," he may magnanimously confer upon that pericope the accolade of "based on earlier tradition." This does not mean that there is some evidence there in the contents that the story was based on earlier tradition as opposed to being based more directly or more informally on reality. It does not mean that the pericope shows that it was based on oral tradition rather than being written directly by an eyewitness (the author of the gospel) or rather than being based on an interview with an eyewitness. On the contrary. The evidence of historicity is ipso facto evidence that supports more simple and direct connections with reality, at least in cases where this is otherwise possible, as it is for the Gospels. It is simply that the scholar does not want to say that. He prefers, thinking of himself as being somehow "cautious" and "careful," to speak of being "based on tradition." The alternative he has in mind, the more "liberal" alternative he is rejecting, is that the pericope was mostly made up out of whole cloth. That is why the phrase "based on earlier tradition" is considered a compliment. But let us not become confused into thinking that this means that there is independent evidence that strongly supports a vast network of earlier, formal, oral tradition, recited regularly by non-eyewitnesses, predating that particular Gospel, out of which the Gospel was constructed.
D.A. Carson explicitly notes that C.H. Dodd uses a phrase like "based on tradition" in a sense indistinguishable from "historical." (Carson, “Historical Tradition in the Fourth Gospel: After Dodd, What?", p. 106) In other words, Dodd uses this phrase as a compliment to a passage in the Gospel of John, to single that passage out as something that he thinks may actually be historical. This doesn't mean that Dodd has objective evidence for an oral tradition predating John as opposed to John's having witnessed the event or having spoken to an eyewitness. Far from it. It's just that he prefers "based on tradition" to "based on reality."
It may be some improvement over, "The Johannine community made up this incident" to say, "This incident is based on earlier tradition." But when the "tradition" hypothesis begins to be set up as a competitor to "based on reality," we must speak up and point out that such a competitor has never been established at all. It would be ironic indeed if evidence supporting historicity (e.g., criteriological evidence that leads a scholar to say that a story in some Gospel is "based on earlier tradition") were ultimately taken to undermine other evidence for historicity (such as undesigned coincidences) because the bias of scholars makes it appear that "based on earlier tradition" is the most that we can ever say.
One almost wonders if scholars have the faulty idea that postulating some source or tradition rather than reality is always "simpler" than postulating reality. No doubt such a confusion could arise from the mere pressure of scholarly trends. Saying that John or Matthew "knew what happened" is taken to sound naive and incautious. So saying instead that they appear to be "based on earlier tradition" is a use of scholar-ese that signals that you are not one of those hasty fundamentalists and are taking the more "careful" and hence (allegedly) the more justified position. But this is not actually true. As I pointed out in the earlier post, never talking about reality but always about traditions is a history-dissolver. It means we literally never get around to talking about reality! It is also a violation of simplicity to hypothesize an extra layer of "tradition" in between what we actually have and the events as they occurred when this is unnecessary, when the author could have known of the events in some more direct way. After all, reality is out there. So the intervening "tradition" is just an extra entity. And it becomes all the more complex as one has to attribute to the "tradition" a set of highly specific characteristics in order to use it as a replacement for reality in our model.
But someone might ask, "Is there not evidence in Paul's epistles of the existence of earlier oral traditions?" Yes, there is some, but by no means in a sense that would support an "ur-source" hypothesis for undesigned coincidences. The existence of something like the "creed" in I Cor. 15 indicates only (at most) that early believers were catechized using traditional formulas. It does not mean that any of the Gospels themselves were composed by putting together formal, oral traditions as opposed to being based upon reality more directly. (It's worth noting here that no such "creed" is incorporated into the text of any of the Gospels.)
At this point, readers may be wondering about the two-source hypothesis and Markan priority. But the two-source hypothesis concerning the synoptic gospels does not at all mean that the synoptic Gospels, nor even Matthew and Luke, were based upon a large network of pre-Gospel oral traditions. Mark itself is neither a hypothetical entity nor an oral entity. It is a document that we possess and can evaluate on its own merits. If Matthew and Luke are based on it in part literarily, it does not follow that Matthew and Luke are, for their other information, based upon formal "oral traditions."
In any event, Matthew may have been an eyewitness himself. Luke had opportunity to interview eyewitnesses. And, as I point out repeatedly in Hidden in Plain View, the evidence of undesigned coincidences cuts across the "synoptic problem" and the two-source hypothesis, because Matthew and Luke show evidence of independent access to the truth at multiple points, even in some places where their narratives bear similarity to that of Mark. There is no good reason to infer that, since Luke and Matthew may have been partly based literarily on Mark, they are otherwise composed by putting together formal oral traditions as opposed to witness testimony--their own or that of others. And it would be even more implausible to think that they were composed by putting together mere fragments of such oral traditions.
When I went back to Hidden in Plain View recently, I was especially struck by how many of the undesigned coincidences discussed in it bring together information from more than one incident--more than one "pericope," as they are called. For example, the UC concerning Jesus' trial and the claim that he threatened to destroy the Temple brings together that trial scene and a completely different context--his words in John when he cleanses the Temple in John 2: "Destroy this Temple, and in three days I will raise it up." The explanation of John the Baptist's words, "I saw and bare record that this was the Son of God" in John 1:32-34 brings together that scene, in which John the Baptist is later talking about the baptism of Jesus, and the actual story of Jesus' baptism as told in the Synoptic Gospels. The UC concerning Peter's boast that he will never deny Jesus and Jesus' later words to him, "Do you love me more than these?" in John 21 also brings together different scenes and incidents. And there are many more.
I had brought this point up briefly in the earlier post, but I hadn't realized just how many of the UCs have this feature. It comes up again and again. This is especially relevant for any attempted ur-source hypothesis concerning "oral traditions," because any common "oral tradition story" idea will concern specific stories as told in standardized form. Thus a parable might (allegedly) have circulated in such a "tradent-cerified" form, or the story of a healing, and so forth. Hence, the idea of such earlier oral stories does nothing to explain away the occurrence of UCs that cut across two or more pericopes in the Gospels. There is no reason to think that a "tradition story" of Jesus' words as told in John when he cleansed the Temple (if there even was such a thing, which there may well not have been) would dovetail with the "tradition story" of the accusation against him at his Sanhedrin trial unless both were true. In other words, in that case the same UC argument that applies to the Gospel accounts would apply to separate "tradition stories." If a "tradition story" of a single pericope is a conjectural entity (as it is), a single "tradition story" that just happened to contain two completely different pericopes concerning different parts of Jesus' life is truly wildly conjectural and is not even what is usually meant by the concept of an "oral tradition."
This is not to say that an "ur-source" theory works at all well even for a single incident, such as the feeding of the five thousand. For all the reasons given here and in the earlier post, a conjecture of a composite "tradition story" of the feeding containing all the information needed for the UCs surrounding that incident, which later fragmented into the Gospels we currently have, is highly implausible and unjustified. But the "ur-source" hypothesis is, if possible, epistemically even worse if one were to try to apply it to UCs that cut across multiple incidents.
One more topic has occurred to me in this most recent exchange on ur-sources, which I didn't discuss in the earlier post: Apparent discrepancies.
Take, for the sake of concreteness, the feeding of the five thousand. I can think immediately of two apparent discrepancies concerning the feeding which have puzzled scholars for a long time: 1)The apparent difficulty concerning "pros Bethsaidan" in Mark and the statement in Luke that the feeding took place near Bethsaida. I discuss that here. 2) The apparent discrepancy between Mark 6:45 and John 6:15-16 concerning whether the disciples went away by boat before or after Jesus went up into the mountain by himself. Nor are these the only apparent discrepancies in the feeding of the five thousand.
If there is one thing that apparent discrepancies are good for, it is opposing various causal dependence theories. And an ur-source theory is a causal dependence theory. (This is an area of my professional specialty in probability theory.)
If there were ever (implausibly enough) some compendium version of the feeding of the five thousand story that contained all of the details now found dispersed in various gospels, or all of the details used in a given undesigned coincidence, we would expect that compendium version to be at least coherent in the sense of resolving such apparent discrepancies. Jesus would either appear to dismiss the crowd and go up into the mountain before the disciples leave in their boat or after, not both. Such a version of the story (being just a single version) would be expected to be somewhat clearer about what direction they rowed after leaving the place of the feeding so as not to appear (even prima facie) to contradict the statement that the event occurred near Bethsaida. A well-honed official story version has the opportunity to smooth out these rough edges or to avoid producing them in the first place. Indeed, part of the job of creating one version of a story that will be told over and over again would presumably be making choices about which direction to go if one's underlying evidence had some apparent discrepancies in it already.
But separate witness testimonies, as cold-case detective J. Warner Wallace has repeatedly pointed out, often contain small apparent discrepancies concerning ancillary parts of the story. Hence, the apparent discrepancies between the feeding accounts are yet another characteristic that makes them look like independent witness testimonies in the form that we now have them, not like broken-off parts of a pre-existing ur-narrative from which they were copied.
It cannot be stressed too strongly: The fact that NT scholars have an unfortunate tendency to construct sources or traditions at the drop of a hat and to endow them with all sorts of specific characteristics on the basis of no independent evidence does not mean that doing so is reasonable or constitutes an epistemically problematic challenge to UCs. Undesigned coincidences tend to support a simpler model, and if that simpler model is regarded as "radical" when compared with the more convoluted approach that one might expect "critical scholarship" to take, we need to let the data speak for themselves. Hobbling the conclusions that can be drawn from UCs by pointing out that a scholar or a skeptic could make up a hypothetical "ur-source tradition" and gerrymander it so that it included everything found in the Gospels' separate stories is not a good scholarly practice.
Nor does such a practice deal with the inconvenient fact that, when all is said and done, the Tradition mountain in the middle has, now that the Gospels exist, magically turned back (in the accounts we have) into varied accounts with the properties that we find normal witness testimony to have--puzzling incompleteness, casualness, inclusion of vivid and unnecessary details, fitting-together with other partial accounts, and occasional apparent discrepancies. That is quite a coincidence if these are merely broken-off fragments of some ur-tradition rather than being independent testimony to reality.
As I said in the earlier post, contemporary NT scholars appear to be largely unaware of or uninterested in undesigned coincidences, though one does see occasional positive references to them (whether by that name or not) in the works of more conservative scholars. Both Leon Morris and Craig Blomberg discuss them in the context of supporting the historicity of John's Gospel.
The fact that there is no contemporary, back-and-forth literature to speak of on undesigned coincidences means that the NT scholarly guild has not yet taken much of a whack at explaining them away. This is just as well from my perspective. I tend to think it would be a weariness of the mind and flesh to have to read the convoluted theories that critical scholars, motivated to discredit UCs as independent evidence for the Gospels' historicity, might try to come up with. But so far, we must conjecture about what they "would say."
Since ad hocness, construction of phantom sources on a case-by-case basis, and rampant violation of simplicity are pretty common faults of the discipline, it is possible that some sort of ur-source hypotheses will end up being the favored "go-to" if and when more "mainstream" critical scholars start trying to dismiss undesigned coincidences. The fact that the question comes up from time to time among my correspondents may indicate that the wind will blow in that direction. That will be particularly interesting to see insofar as it involves the Gospel of John: Will we then see a scholar who has talked expansively of the inventive proclivities of the late-1st-century "Johannine community" suddenly start solemnly telling us that John's story of Jesus' trial before Pilate was part of a larger, oral ur-source that existed before Luke and that John's version of the feeding of the five thousand is a broken-off piece of a larger, oral ur-source that was standardized in the Christian community prior to Mark? I suppose anything is possible, and NT critics are, in practice, great believers in the maxim that consistency is the hobgoblin of little minds. I would like to think that, if that happens, there will be plenty of others besides me around to point out, once again, that the Emperor is unclad. If needed, these posts should enable us to make that point in some detail.
Post a Comment