Monday, May 24, 2010

James Edwards on McIver and Carroll

I have recently finished working through James R. Edwards, The Hebrew Gospel and the Development of the Synoptic Tradition (Grand Rapids: Eerdmans, 2009). I am reviewing the book for CBQ, but I have quite a lot to say about it and there are several things I won't have space for in the review, so I will let them overflow here in the blog.

In no particular order, the first of these is a discussion of Robert K. McIver and Marie Carroll, "Experiments to Develop Criteria for Determining the Existence of Written Sources, and Their Potential Implications for the Synoptic Problem", Journal of Biblical Literature 121/4 (Winter, 2002): 667-687. Edwards brings up McIver and Carroll's work in order to corroborate both Marcan Priority and the notion that "Matthew concludes and consummates the Synoptic tradition" (251), though it is not quite clear what he means by this given that elsewhere "double tradition" is retained as a source for Matthew's and Luke's non-Marcan material (Chapter 7; cf. the diagram on p. 262 in which Matthew and Luke are independent of one another).

Edwards reviews one of the key conclusions of McIver's and Carroll's study, quoting with approval their comment that "Any sequence of exactly the same 16 or more words that is not an aphorism, poetry or words to a song is almost certain to have been copied from a written document" (681; Edwards, 250; italics original). Edwards then draws attention to the 23 parallel passages listed by McIver and Carroll as including 16 words or more in verbatim agreement. He notes that three of the pairs are Mark // Luke, nine are Matthew // Mark and eleven are Matthew // Luke. He then writes:
Of a total of 23 pairs shared by Luke and Mark, or Luke and Matthew, or Mark and Matthew, Matthew shares 20 of them. Matthew, in other words, is almost twice as likely to share material in common with Mark, or in common with Luke, as either Luke or Mark is to share material in common with the other. This statistic suggests that Matthew is either the source of material for Mark and Luke; or, conversely, that Matthew is the recipient of material from either Mark or Luke. (251).
The deduction from the data is puzzling. McIver and Carroll's data clearly show that there is literary dependence of some kind in these passages, but the data tell us nothing in themselves about the direction of dependence. Edwards goes on:
The theory of Markan priority supplies solid and repeated evidence for the assumption that Matthew has been the recipient of Markan material (251).
I am sympathetic with this view, but it cannot be deduced from McIver and Carroll's data. What the data demonstrate is copying of some kind, but there is no indication of direction of dependence.  If there is one thing that might be deduced from their data with respect to Marcan Priority it is that Matthew and Luke are unlikely to have used Mark independently of one another. Two of the eleven Matthew // Luke pairings are actually in triple tradition, Matt. 8.1-4 // Luke 5.12-16 and Matt. 16.21-28 // Luke 9.21-27, i.e. they feature Minor Agreements that generate extended conjoined sequences of verbatim agreement, which suggests that if Matthew and Luke used Mark, there is also a direct link of some kind between Matthew and Luke.

Incidentally, McIver and Carroll appear unaware of this themselves. They list one of these passages, Matt. 8.1-4 // Luke 5.12-16 (Leper) among their pared down list of nine passages that "almost certainly contain copied material" without noticing that this runs contrary to the Two-Source Theory that elsewhere they apparently assume (683, "perhaps Q?"; see further John C. Poirier, "Memory, Written Sources, and the Synoptic Problem: A Response to Robert K. McIver and Marie Carroll", JBL 123/2 (Summer, 2004):  315-322 (316-7).)

But back to Edwards.  His discussion of McIver and Carroll concludes:
The foregoing arguments, and especially the list of examples on pages 249-50 above, argue the same, that Matthew is the recipient of Lukan material. The work of McIver and Carroll seems to supply further statistical evidence that Matthew concludes and consummates the Synoptic tradition. (251).
The "list of examples" on pp. 249-50 of the book consists of suggestions that Matthew is a later Gospel than Luke, itself something I may return to in due course.  But the conclusion here, that McIver and Carroll's work supplies statistical evidence in favour of Matthean posteriority, is not self-evident.  It is simply asserted.

There are several minor issues that are also worth mentioning: (1)  Edwards also speaks of "46 passages" (250), but the figure should be 45 -- he has counted Matt. 16.21-28 twice; (2) McIver and Carroll do not weight these 45 passages equally.  They eliminate 14 pairs because they are aphorisms and distinctive sayings and they put stress on only 9. (3) McIver and Carroll are only dealing with conjoined sequences of words; this work needs to be supplemented with other Synoptic data; (4) The points made in Poirier's perceptive critique (citation above) need to be taken seriously rather than ignored.


Stephen C. Carlson said...

This looks like Edwards is making some kind of Middle Term argument for Matthew (but just with pairs of long verbatim strings). Of course, as Butler, the middle term phenomenon works for all three positions: first, second, and third, not just first (or, as sometimes done, first and third).

Mark Goodacre said...

Yes, I wonder if it is something like that, Stephen. It's impossible to tell, though, because Edwards does not make clear what it is about the data that leads him to this view. He says, ". . . . conversely, that Matthew is the recipient of material from either Mark or Luke" and that appropriates to the view he holds, but his "either . . . or" here further confuses me.

Stephen C. Carlson said...

I can understand the confusion because he's using "either ... or" twice in the same sentence, but with different meanings. The first one is disjunctive (exclusive possibilities), while the second is conjunctive (non-exclusive possibilities). I would say that using the word "either" is unusual for expressing the conjunctive meaning.

After years of studying logic, I didn't realize that "and" and "or" are actually ambiguous in English until I took a patent contracts class. This issue about the ambiguity of "or" actually shows up in Greek exegesis too, e.g. Gal 1:10, which I was just looking at.

Mark Goodacre said...

Yes, I think you have parsed that well. The subsequent sentences confirm that the either / or is intended as conjunctive. Unfortunately, he does not specify what it is about the data that leads him towards the view that Matthew is either first or third, though I think your guess is a good one.

Stephen C. Carlson said...

I've read someone somewhere saying that the middle term phenomenon means that the middle term is either first or third. I can't remember who offhand (McKnight?).

Mark Goodacre said...

I've read that too. Sanders and Davies pretty much say it in Studying the Synoptic Gospels.

Stephen C. Carlson said...

That must be it.

Jim Deardorff said...

Regarding the "and/or" ambiguity, there is also the hypothesis that (Hebraic/Aramaic) Matthew came first, as per the early church fathers, and Greek Matthew last. It easily explains the verbal agreement, with Matthew's later translater having inserted a few pro-gentile passages and a bit of reverential upgrading.

Somehow Butler overlooked this possibility, in believing that Hebraic Matthew was translated into Greek before Mark and Luke were written.

Stephen C. Carlson said...

Hi Jim. Butler discusses that possibility (thought w/out considering Luke) on pp. 159-161 of his Originality of St. Matthew (1951).

Jim Deardorff said...

Thanks, Stephen, for reminding me of that. But the failure by Voste and Butler to include Luke as also written before Hebraic Matthew was translated into Greek, leaves the strong verbal agreement of the double tradition unexplained by their Matthean priority hypothesis. I.e., If Luke copied the double-tradition material from a Greek Matthew, carefully replicating long strings of consecutive words, why would he not have done so even more fastidiously with Markan parallels?

I believe Butler's arguments for rejecting Greek Matthew as being written subsequent to Mark can be reasonably explained even allowing for priority of a Semitic Matthew.

Judy Redman said...

Mark, I agree that McIver and Carroll's work cannot be used as a tool to determine anything other than that there might have been copying. The direction of any copying requires another methodology and different data. It seems as though Edwards is trying to give the stats more weight than they are able to carry.

Jim Deardorff said...

That there had been copying I think has to be accepted as a given. It's the anomalous verbal-agreement aspect of this copying that should attract our attention. Why are there a significant number of very lengthy word strings in sequential order, (between parallel passages of Matthew/Mark and of Matthew/Luke), for word sequences of length 17 to 31 words? Upon examining their distribution of shorter word sequences, one finds that they decay to essentially zero, as expected for an exponential or geometric distribution, for word strings of length up to 16 or 17. So how could so many anomalously long word strings exist unless purposely replicated by a redactor?

The hypothesis that Semitic Matthew was the first gospel allows that this redactor had been the translator of that initial gospel into Greek, after Mark and Luke had been written.