Wednesday, May 26, 2010

A Flaw in McIver and Carroll's Article about detecting the existence of written sources in the Gospels

I've recently found myself thinking again about 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, partly because it is briefly discussed in James R. Edwards's recent book (James Edwards on McIver and Carroll), partly because it came up in Judy Redman's recent post on Speeches of Jesus (3) -- human memory experiments, with which I have been interacting over there on Judy's blog.

I have always been uncomfortable with McIver and Carroll's article, in part for reasons explained in John Poirier, "Memory, Written Sources and the Synoptic Problem: A Response to Robert K. McIver and Marie Carroll", JBL 123/2 (2004): 315–322.  But today I was able to locate the source of my discomfort  and it relates to an important flaw in the way that McIver and Carroll apply the results of their experimental data with students to the texts of the Synoptic Gospels.  The mistake that they make is simple but important: their experimental data is in English whereas the Synoptic Gospels are in Greek.

Now one may well say at this point, "So what?" and "Well, of course their experimental data is in English! Contemporary undergraduate students do not speak fluent first century Koine Greek!"  But here is the problem:  McIver and Carroll make no adjustment for the differences between the English student experiments and the Greek texts of the Gospels.  Their criterion for the determining the presence of a copying from a retained written source in English is 16 words in conjoined sequence (later adjusted in a subsequent article to 18).  This criterion for English material is then transferred without analysis or comment to the Greek Gospels.

This is important because it takes many more words to say something in contemporary English than it takes to say the same thing in Koine Greek. Therefore, 16 (or 18) words in English correlates to fewer words in Greek, and so the 16 (or 18) word criterion is much too high for Greek.  The point can be illustrated by looking at English translations of McIver and Carroll's list of passages in that feature 16 or more words in conjoined sequence.  Notice the kinds of difference between expression in Greek and expression of the same thing in English:

  • The 18 Greek words in conjoined sequence in Matt. 8.2-3 // Luke 5.12-13 (miscounted by McIver and Carroll as 17) are 26 words in the RSV.
  • The 31 Greek words in Matt. 10.21-22 // Mark 13.12-13 becomes 49 words in English translation. 
  • The 24 words in Matt. 3.9-10 // Luke 3.8-9 becomes 29 words in English translation.
  • Matt. 6.24 // Luke 16.13 goes from 26 Greek words in conjoined sequence to 33 in English translation.
  • Matt. 7.7 // Luke 11.9 goes from 24 Greek words in conjoined sequence to 42 in English translation.
  • Matt. 8.9-10 // Luke 7.8-9 goes from 25 conjoined Greek words (miscounted by McIver and Carroll as 21) to 37 in English translation.

In other words, attention to this important point about the difference between English and Greek would have resulted in a vastly greater number of passages that satisfy McIver and Carroll's 16 (or 18) word criterion. They have derived a criterion from work with one language (English) and have applied it to texts in another language (Greek), in spite of the fact that it takes many more words to say the same thing in English than it does in Greek.

And there is a further, related difficulty that I will deal with in my next post.


Jim Deardorff said...

Wouldn't McIver & Carroll be able to sidestep this particular objection by saying that "Of course, for Koine Greek the criterion of 16 should be reduced to 9 or 10"?

Mark Goodacre said...

Yes, but as soon as the criterion is reduced to 9 or 10, they no longer have just 9 passages in which copying has definitely occurred but dozens.

Judy Redman said...

this may depend on whether the limiting factor for memory is words or the ideas they express. If it's just words, then you possibly can just say 18 words in English = 18 words in Greek as long as you are working with fluent speakers of each language. If content also contributes, then the fact that it takes lots more words to say the same thing in English is a significant issue.

Mark Goodacre said...

Well, it is something we can test, isn't it? We can compare English translations with the Greek to get a feel for the how important the difference might be. In every case that I have looked at so far, the English takes more words than the Greek to say the same thing, sometimes massively more, e.g. in Matt. 7.7 // Luke 11.19, from 24 words to 42, 75% longer!

Unknown said...

it takes many more words to say something in contemporary English than it takes to say the same thing in Koine Greek.

Hmmm... would it be a similar problem to transpose the authentic sayings of Jesus from Hebrew/Aramaic to Koine Greek?