Tuesday, December 12, 2017

Garrow's Flaw

Over on Bart Ehrman's blog, there has been some fun relating to the Synoptic Problem. A commenter promised $1,000 to Prof. Ehrman's blog's charities if he engaged with Alan Garrow's work on the Synoptic Problem. Since I have read and thought about Dr Garrow's work, I thought I'd offer to write a post on it, drawing attention to what I see as one of the flaws in the argument. Prof. Ehrman has posted those comments today, and I am cross-posting them here.

I've enjoyed the social media reaction to this, with comments also on Alan Garrow's blog, The Logos Academic Blog, and no doubt elsewhere.

Please note that I do this to encourage people to visit and subscribe to my friend and colleague's blog, with the hope that others will support these great causes, including one local to us, The Urban Ministries of Durham.


In a recent comment on Bart Ehrman's blog, "Evan" suggested that Alan Garrow's arguments are so compelling that he effectively "proves beyond any doubt that Matthew used both Mark and Luke". He says that it is "virtually impossible to believe in the Q theory once you’ve seen this data". The same commenter goes on to offer $1,000 for the charities on the blog in return for an assessment of Garrow's case, asking in particular for "holes in his arguments." As a self-confessed synoptic nerd, and as one who has spent some time with Garrow's work, I thought I would offer this critique by way of response. Since "Evan" is particularly interested in holes in Garrow's case, I thought I would focus on one particular flaw that places a major question mark over this model.

I should point out that while Bart and I are on different sides on the Synoptic Problem, he a staunch supporter of the Two-Source Theory and I an advocate of the Farrer Theory, we are both agreed that Matthew did not know Luke, which is what is under discussion here. I have enjoyed reading Bart's recent blog entries (and the multiple comments!) on these issues, and I hope to find a moment to respond, even if just to provide an excerpt from something I have written. But back to the immediately pressing issue.

Alan Garrow has made his case in a series of videos on his blog, Streeter's 'Other' Synoptic Solution: The Matthew Conflator Hypothesis and An Extant Instance of Q, and it is these videos that "Evan" is referring to. Garrow's more detailed, scholarly argument is laid out in two articles by the same names recently published in the premier journal in the field, New Testament Studies. The fact that one of Garrow's articles is entitled "An Extant Instance of Q" illustrates that it can hardly be the case that "Alan Garrow has compiled an extremely compelling argument that Q never existed" (so "Evan", on Bart's blog). Garrow is actually arguing that Matthew and Luke did use Q, and that we can see exactly how they used Q, because Q has not been lost. Q is, in fact, the Didache! The Didache is a fascinating early Christian work, first published in 1883, but Garrow is the first -- to my knowledge -- to identify it with the hypothetical Q.

Garrow's synoptic model works with Marcan Priority (something shared by Bart and me) but he adds a couple of further elements: (1) Matthew and Luke both know and use Q, which is now unveiled as the Didache; and (2) Matthew also knows Luke's Gospel. In his videos and articles, he works with the analogy of a multiple vehicle car crash. All the data, he says, need to be explained. He then pays special attention to a key element in the car crash -- the variation in rates of verbatim (word for word) agreement between Matthew and Luke in the double tradition (i.e. in passages found only in Matthew and Luke). He points out, quite correctly, that sometimes Matthew and Luke have very high verbatim agreement with one another, and sometimes they have rather low verbatim agreement with one another. This variation in verbatim agreement, he says, demands an explanation.

Garrow argues that his model provides a good explanation of both the high verbatim and the low verbatim passages. High verbatim passages are the result of Matthew directly copying from Luke. They are places where Matthew has just Luke in front of him. Here, Matthew is copying Luke "without distraction." Low verbatim passages are the result of Matthew conflating Luke with Q (=the Didache), i.e. places where Matthew does not agree as much with Luke because he is distracted by one of Luke's sources, Q (=the Didache). As he expresses it, "High DT [double tradition] passages are best explained by Matthew’s copying of Luke without interference from any other entity.”

Garrow is right that the spectrum of agreement in Matthew's and Luke's double tradition requires an explanation. Anyone studying the Synoptic Problem should certainly make sure that they have a good account of why Matthew and Luke sometimes agree very closely and why they sometimes provide the same material in very different words. But is Garrow's diagnosis correct? I don't think so. The difficulty is that it is contradicted by his own model. Several of the passages with very high verbatim agreement in Matthew and Luke are passages where Matthew, on Garrow's own theory, is also copying from Mark, passages like John's messianic preaching (Matt. 3.12 // Luke 3.17), the Beelzebub Controversy (Matt. 12.22-30 // Luke 11.14-23), and the Sign of Jonah (Matt. 12.38-42 // Luke 11.16, 29-32). In passages like these, Matthew and Luke can be remarkably close in wording, and yet these are passages where there are also parallels in Mark.

In other words, Garrow has constructed a model where Matthew is supposed to be agreeing very closely with Luke when there is no distraction, in places where only Luke has the passage in question. And he is supposed to agree with Luke much less when he is distracted by another source, the Didache (or Q). But this is sometimes manifestly not the case. We can test Garrow's thesis by asking how Matthew behaves when the evangelist is copying from both Luke and Mark. And since these passages (usually called "Mark Q Overlap passages") feature a lot of very high verbatim agreement between Matthew and Luke, it is clear that "distraction" has nothing to do with it.

Garrow summarizes his own argument by saying that, "[W]hen Matthew copies Luke without distraction he produces High DT passages. When, however, Matthew knows differing versions of the same event he conflates them – resulting in a Low DT passage." But on Garrow's own thesis, Matthew is quite capable of producing high verbatim agreement when he "knows differing versions of the same event." It may be worth adding that there are plenty of low verbatim passages in Matthew and Luke where there is no parallel in Mark or the Didache, i.e. Matthew is perfectly capable of producing a low verbatim passage on his own with just one source. Whatever we might make of the wisdom of comparing one's synoptic model to a car crash, it has to be said that high verbatim agreement is simply not diagnostic of an author working from only one source, just as low verbatim agreement is not diagnostic of an author working from more than one.


Ken Olson said...

I would add to your point that Garrow has not made clear how or why Matthew decides which of his sources he will follow and which will "distract" him when he finds that his different sources contain differing versions of the same story or saying. On the traditional two source theory, there is a limited number of Mark-Q overlap passages and Matthew conflates them (while on the Farrer theory Luke follows Matthew instead of Luke in these cases). On Garrow's theory there is a much larger number of passage found in both Mark and Q. In some of these Mark-Luke overlaps (i.e., the traditionally identified Mark-Q overlap passages) and especially the Beelzebul pericope, Matthew carefully conflates the two versions of Mark and Luke, while in most cases he follows Mark with a few touches of Luke (the traditional "Minor Agreements'). It seems like the mere existence of a parallel source does not determine whether Garrow's Matthew will be "distracted" by it or not, or by how much. A further level of explanation is needed on how Matthew makes the decision about which sources to use and how much to use them.

Unknown said...

Thanks, Mark!

And if - as you draw out from Garrow's proposal - Matthew were looking only at Luke when copying a passage designated a "Mark-Q Overlap", he would be siding with Luke against his own predilection for the Gospel of Mark!

Mark Goodacre said...

Excellent point, Ken. Thanks.

Mark Goodacre said...

Thanks, Andrew. To some extent, that is how the major agreements between Matthew and Luke are explained on the Farrer theory. Normally speaking, Luke follows Mark as his main source with supplementary material coming in from Matthew, but on several occasions, he follows Matthew, with supplementary material coming in from Mark. It's actually a spectrum.

Mark Goodacre said...

In addition to your point, Ken, it's worth remembering that Matthew produces very high verbatim pericopes in the Triple Tradition in places where he ought to have been "distracted" by Luke. In my fuller response, if I ever get around to writing it, I will draw this point out more fully, but the value of stressing the so-called Mark-Q overlap passages in this context is that these are among the very passages listed by Kirk as supposed examples of Matthew copying Luke without distraction.

Unknown said...

I am more inclined to accept a spectrum for Luke than for Matthew. As you (may) know, I am firmly convinced Matthew is a "Markan Christian" (for want of a better term!).

Jens Knudsen said...

What about the other part of the proposal? The Didache as the origin of what was believed to be sayings material ('Q')? That idea should work equally well with the Farrer hypothesis. Personally I find it hard to see how a non-sayings document like the Didache would result from sayings believed to come from 'The Lord'. It makes more sense that Luke (or Matthew, if Harrow is right) put those words in Jesus' mouth the same way Mark did with Paul's.

Jens Knudsen said...

I don't know why I show up as "Unknown". Sorry.

I'm reading my way through your Rewalking the "Way of the Lord" and I see you pointed out the similarity with the opening of the Didache. Isn't that a point in favour of Luke preceding Matthew, if this theme is especially dear to Luke? It would also turn the aesthetic argument on its head. Luke didn't retain any order from Q, he made it himself. Matthew according to the Q advocates found that unsatisfactory and reordered it to the 'beautiful' discourses.

- Sili