3. Major Agreements between Matthew and Luke; Minor Agreements between Mark and Luke
Christopher Tuckett (Review of The Case Against Q, NovT 46 (2004): 401-403) acknowledges my points concerning Mark-Q overlap passages summarized in the previous post and comments:
In these passages [viz. Mark-Q overlaps], one can indeed refer to Luke’s use of Matthew’s additions to Mark, and/or to extensive non-trivial Matthew-Luke agreements. However, any non-Q theory has to explain Luke’s apparently almost pathological refusal in some of these texts to use any Markan material at all (e.g. the Beelzebul controversy, or the Mustard Seed). As Gerald Downing argued many years ago, Luke’s procedure on the Farrer-Goulder-Goodacre model appears totally at odds with his procedure elsewhere (where, according to Goodacre and others, Luke knows Mark far better than Matthew and uses Mark in preference to Matthew). In these passages, Luke must have studiously avoided all the points where Matthew and Mark agree and reproduced only Matthew’s additions to Mark. (402).The description of the data here is inaccurate. It is not the case that Luke lacks Marcan material in these passages. Both of the specific examples given by Tuckett, the Beelzebub pericope (Matt. 12.25-32 // Mark 3.23-30 // Luke 11.17-23, 12.10) and the Mustard seed (Matt. 13.31-2 // Mark 4.30-32 // Luke 13.18-19), feature several triple agreements, as well as minor agreements between Mark and Luke. On Sanders's count (E. P. Sanders, "Mark-Q Overlaps and the Synoptic Problem", NTS (1973): 453-65, 458), the Beelzebub Controversy features 31 triple agreements, 35 Matthew-Mark agreements, 5 Mark-Luke agreements and 65 Matthew-Luke agreements. Similarly, the Mustard Seed, on Sanders's count, features 14 triple agreements, 11 Matthew-Mark agreements, 6 Mark-Luke agreements and 11 Matthew-Luke agreements. There is no "almost pathological refusal" to include Marcan material here. It is true, of course, that there is a substantial degree of agreement between Matthew and Luke against Mark in these passages, and it is this that forces Q theorists to view these as "Mark-Q overlap"; there are far too many Matthew-Luke agreements for these to be the result of independent redaction.
Let us be clear about how the situation is explained on the Farrer Theory. It is quite straightforward. On occasions like this, where Matthew is the middle term among the Synoptics, Luke is working with Matthew as his primary source and not Mark. The usual triple tradition situation, where there are major agreements with Mark and minor agreements with Matthew, is reversed and, instead, there are major agreements with Matthew and minor agreements with Mark. If Luke is working with both Matthew and Mark, it is not surprising that on occasions Luke turns to Matthew as his primary source, even in triple tradition material. It is interesting to see how often this happens where Matthew has a fuller account than Mark, in the John the Baptist material, the Temptations, Beelzebub, the Mission discourse.
The article to which Tuckett refers, by F. Gerald Downing, has now been answered persuasively by Ken Olson, "Unpicking on the Farrer Theory" in Mark Goodacre and Nicholas Perrin (eds.), Questioning Q (London: SPCK, 2004): 127-50.