Minor Agreements Against Mark
This post has been a while in the making and there's still more to do. Following several debates on the synoptic problem earlier this year (found here: Matthew 3:16 , Synoptic Problem split from Matthew 3:16, and Three Arguments for Lukan-Matthean Dependence) I decided to take a more systematic look at the Matthew-Luke minor agreements against Mark. I used Neirynck's, The Minor Agreements of Matthew and Luke against Mark as a fairly comprehensive list of the agreements. Neirynck offers a very useful classification of agreements at the end of the book, but I had to go beyond it since my goal was to estimate the significance of the agreements.
This first attempt at this task entailed classifying the agreements in four broad categories. The first category is of the agreements in omission (where Matthew and Luke agree in leaving out some text found in the Mark in the triple tradition). The next three categories are of the positive agreements (where Matthew and Luke agree in a text differing from Mark in the triple tradition). The three agreements are: 1) literal agreements, where the agreement entails the exact same word and spelling, although, not always in the same order, 2) root agreements, where Matthew and Luke agree in using the same word, but have different forms, 3) sense agreements, which includes agreements in any of the following: tense, person, number, meaning, or any other grammatical or syntactic agreement that doesn't fall into the earlier categories. The only agreements I chose not to include are agreements in order, since they are very few and well discussed at the end of Neirynck's book.
Below are my counts of the words in each of my categories. There might be slight adjustments in these in the future since reading through 140 pages of agreements can get rather dull and I may have made errors somewhere. The "instances" counts are according to how Neirynck separates them. These instance counts are, admittedly, somewhat arbitrary.
Omission: 2687 words / 553 instances / against 412 Markan verses (4.86 words per instance, 70.9% of all agreed words)
Literal: 721 words / 347 instances / against 254 Markan verses (2.08 words per instance, 19% of all agreed words)
Root: 208 words / 154 instances / against 127 Markan verses (1.35 words per instance, 5.5% of all agreed words)
Sense: 176 words / 109 instances / against 94 Markan verses ( 1.61 words per instance, 4.6% of all agreed words)
It is important to note that the instances and verses in the above counts are not mutually exclusive. Often, there would be a combination of agreements in omission, literal, root and sense against a particular Markan verse, although, in most cases where there were multiple agreements against a Markan verse, they'd consist of a couple literal agreements and a few negative ones.
For reasons already cited in earlier debates, I do not take the agreements in omission very seriously. Instead, in this post, I will focus on the positive agreements, and specifically the literal agreements, since they are the most numerous positive agreements and comprise the most significant hurdle for the 2-Source Hypothesis.
At first glance, the agreements seem quite numerous and, at least in number, not very "minor." It should be noted that the minor agreements are similarly problematic for competing documentary hypotheses that posit a Mark redacted from Matthew and Luke since one would have to account for why the Markan redactor would so often depart from his sources when they both agree. When, however, examining the literal agreements more closely, the brute force of the numbers begins to lose its mystique. The first observation is that each instance of an agreement on average involves only 2 words. Indeed, the majority of the literal agreements is only in one word per instance (198/347 of the instances), and when we examine later what these agreements are, it'll become clear why many of them are irrelevant. 280/347 of the instances of literal agreements are only in 1 or 2 words per instance. 35 instances involve 3 words. 16 instances involve 4 words. There are only 16 instances of literal agreements that involve 5 or more words.
Why so many instances of literal agreements involving only one word? When examining what the literal agreements are, we find that the largest class involves conjunctions and particles. There are 93 instances where: 1) Matthew and Luke use DE for Mark's KAI (4 cases of the reverse), 2) Mark and Luke use the same conjunction or particle where Mark has asyndeton (1 reverse), 3) Matthew and Luke agree in using a participle where Mark uses KAI + finite verb, 4) a number of other cases of agreements concerning 'INA, 'OTI, ALLA, GAR, etc.
The next largest group is that of avoiding the historic present or other tense agreements (which result in a literal agreement) with 53 such cases. Next we have 30 cases of Matthew and Luke agreeing in making explicit a subject which Mark leaves implicit. This may not seem like an obvious correction for independent redactors to make, but when one looks at the actual agreements, one finds that almost all of these agreements are simply in adding an article, 'O or 'OI where Mark avoids it, and the next largest number of these agreements consisting of making IHSOUS explicit. There are 8 cases of agreement in using the participle of LEGW instead of a finite form or another speaking verb, and 5 cases where Matt and Luke avoid the awkward HRXATO + infinitive. Thus, this next group of agreements involving slight adjustments to grammar adds up to 96 instances.
Thus, we can easily see why insignificant literal agreements in Matt and Luke involving 1 word comprise 57% of the instances of literal agreements.
Now, agreeing that such changes ought to be obvious and occur both to Matt and Luke regularly in the same place is convincing on its own to an extent, but it helps to look at some specific probabilities. For my example I have chosen the DE/KAI interchanges since the potential variations from which to choose by any redactor are fairly low: in most cases where Mark uses KAI in the triple tradition, Matthew and Luke choose between KAI and DE. Out of the 209 instances in the triple tradition where Mark uses KAI (if my count based on Neirynck's tables is correct), Matthew changed 111 to DE and Luke changed 128 to DE. In other words, Matthew chose DE over KAI, 53% of the time and Luke 61% of the time. Thus, by random chance, Matthew and Luke ought to coincide in changing KAI to DE 32% of the time. In fact, there are 48/209 times where they both agree in using DE against Mark's KAI, or 23%, which is well within probability. If anything, there should be more agreements in DE against KAI.