A Statistical Approach to the Synoptic Problem: Part 1—Triple Tradition

This article aims to contribute to the body of empirical data—particularly in the matter of the thousands of words involved in the “minor agreements” between the Gospels of Matthew and Luke against the Gospel of Mark—which must be accounted for by any viable theory that attempts to describe the interrelationships between the Synoptic Gospels, the so-called Synoptic Problem. It should be clear why the Synoptic Problem has become such a battleground:^[18] Whatever solution is adopted will have tremendous influence on Gospel scholarship, all the way from textual criticism to the attempts to summarize the theologies of the respective Gospels.^[19] With stakes that are so high, it is important to analyze the available data on an objective empirical basis. To that end, I have developed and adapted several new methods of quantifying and testing synoptic theories. The method I will discuss here is to evaluate all the options of linear dependence between three authors. Only six theoretically possible options exist, and the question to be asked is whether any of the options can stand up to objective empirical analysis or whether they all fail the test.

Six Theoretical Options of Linear Dependence

Theoretically speaking, it is possible that no literary dependence among the Synoptic Gospels exists at all.^[20] On the other hand, all kinds of complicated schemes of interdependence are theoretically possible. Even in the case of linear dependence, the simplest of the hypothetically possible schemes of interdependence, it is theoretically possible that only two of the gospel writers had some literary relationship and the third was completely independent of the other two. The method I will discuss here is intended to test out the viability of the various schemes of linear dependence in which all three authors are involved, of which there are six theoretically possible relationships:

• Matthew→Luke→Mark
• Mark→Luke→Matthew
• Luke→Matthew→Mark
• Mark→Matthew→Luke
• Luke→Mark→Matthew
• Matthew→Mark→Luke

• ^[1] For bibliography on options that have been argued in the past, see Thomas R. W. Longstaff, The Synoptic Problem, A Bibliography 1716-1988, (Macon, Ga.: Mercer, 1988). For a recent and extremely detailed argument against any interdependence see the German scholar Eta Linnemann, Eta, Is There a Synoptic Problem? (Grand Rapids: Baker, 1992). ↩
• ^[2] Robert M. Morgenthaler, Statistische Synopse (Zürich and Stuttgart: Gotthelf, 1971). ↩
• ^[3] I am operating on the assumption that in spite of Morgenthaler's bias toward Markan Priority, the kind of facts with which I am dealing, i.e., verbal identities, will prove able to override the possible infelicitous pericope divisions that may be there. In any case, there are more divisions than in the more widely used Aland Synopsis, and on the level of verbal identities this provides a higher level of precision in the counting and comparisons. ↩
• ^[4] Ibid., 66-68. ↩
• ^[5] Ibid., 33-65. ↩
• ^[6] The reason for going to the summary chart first is that it is easier at a glance to pick out the 90 TT parallels. ↩
• ^[7] Ibid., 33-65, and the summary chart on 66-68. ↩
• ^[8] Note that Mark is not the shortest Gospel, but the longest Gospel in parallel passages. ↩
• ^[9] These statistics had to be totaled from the counts of the individual pericopae according to the list of the TT pericopae of each Gospel listed on that summary chart albeit in Markan order. ↩
• ^[10] In order to count minor agreements color coding was applied to the parallel identical and continuous words, IFS, in the horizontal line synopses of both Swanson and Neirynck. See Reuben Swanson, The Horizontal Line Synopsis of the Gospels, Greek edition, Vol. I, The Gospel of Matthew (Dillsboro: Western North Carolina Press, 1982); Frans Neirynck, T. Hanson and F. Van Segbroeck, The Minor Agreements of Matthew and Luke Against Mark, with a Cumulative List (Leuven: University Press, 1974). ↩
• ^[11] Literally thousands of words are involved in these CAs where Matthew and Luke agree contrary to Mark. The designation of these agreements as “minor” stems from Markan priorists who attempted to minimize their significance in contrast to the seemingly more impressive major agreements between Mark and the other two Gospels, especially in the matter of pericope order. ↩
• ^[12] In the Triple Tradition contexts considered here the total word counts are as follows:
• 1. Mark is the longest gospel with 7889 words.
• 2. Matthew uses 6682 words to tell the same stories.
• 3. Luke uses 6482 words. (These counts are not found in Morgenthaler, but can be totaled a bit laboriously from the three pages of his “Gesamtresultat,” ibid., 66-68. The counts of the IFS are available in the same way.) ↩
• ^[13] If one starts with the IFS shared between the first and the third writer and then subtracts all of those IFS not found in the middle writer, then what remains are only those IFS words that are shared by all three writers. An equivalent way would be to start with the IFS of any other pair (the first and the second or the second and the third) and subtract the IFS not present in the third, and the same result would be obtained, i.e., the IFS words shared by all three Gospels. ↩
• ^[14] These statistics force one to eliminate either the working hypothesis of simple linear literary dependence or the assumption of independence between the first and third Gospel, because obviously the third Gospel shares more material with the first than could have come through the second. ↩
• ^[15] There are actually 2,356 words in Mark which are omitted in the parallel accounts of Matthew and Luke. This count is easily obtained by counting the words that are underlined with a broken line in the Markan text as it appears in Frans Neirynck, The Minor Agreements in a Horizontal-Line Synopsis, (Leuven: University Press, 1991). I did not include in this count those words that were only partly underlined (to indicate where Mark had a different form of the same word). So the number of common omissions is even slightly higher than what I have counted. One can easily see the relatively longer text of Mark in TT by looking at the red colored section of the graphs in the comparative graphs found in the appendix. ↩
• ^[16] These counts are based on the horizontal line synopsis of the Gospel of Mark and all of its parallels in Matthew and Luke published by Frans Neirynck, The Minor Agreements of Matthew and Luke Against Mark (with a Cumulative List) (Leuven: University Press, 1974). I colored, categorized and counted every Minor Agreement. ↩
• ^[17] Only Lockton comes close with his count of 750, which means he probably included the identical words involved in inversions, in which case our counts are perfectly identical. See William Lockton, Certain Alleged Gospel Sources (London: 1927). Cf. J. Schmid, Matthaeus und Lukas, 181 n. 2. My source for this Lockton count is Neirynck, Minor Agreements, 43. Neirynck also quotes the counts of others: De Solages: 393; Abbot: 446; Hawkins: 240; Burton: 275; Streeter: some 220 cases; Schmid: about 250: McLoughlin: about 342. See ibid., 37. Neirynck’s chapter on “The Study of the Minor Agreements,” (ibid., 11-50), is full of such massive detail and bibliography. That is why I can concentrate on the research of the specific scenario of linear dependence, since detailed surveys and bibliographies of the Synoptic Problem in general and the minor agreements in particular already exist. ↩
• ^[18] This has become all the more clear from the analyses of Stoldt and Meijboom. They wrote independently, and a century apart from one another, yet they agree amazingly on the role of David F. Strauss as scaring people into Markan Priority. See Hans-Martin Stoldt, History and Criticism of the Marcan Hypothesis (Macon, Georgia: Mercer University Press, 1980), 227-235; H. A. Meijboom, History and Critique of the Origin of the Marcan Hypothesis1835-1866, (trans. John J. Kiwiet; Macon, Georgia: Mercer, 1993), 9-11. ↩
• ^[19] Presently all critical editions of the Greek New Testament text which are in common use are products of text critical scholars who accept the theory of Markan Priority. An example of this, noted by David Flusser, is the insertion of the notion of ascension into the text of Luke 24:51 from Mark 16:19 on the biased judgment call that certain papyri, which include the ascension phrase, outweigh all the other texts, which omit the phrase. The old Nestle text did not have this bias at this point; it did not include the phrase about ascension in its text of Luke. This Markan priority bias also artificially creates a conflict between this supposed ascension in Luke 24 and the fuller account in Acts 1 which takes place after forty days. Reuben Swanson has published a new critical edition based on an existing historical document, Codex Vaticanus, with all the variants presented in the apparatus. See Reuben Swanson, New Testament Greek Manuscripts: Variant Readings Arranged in Horizontal Lines Against Codex Vaticanus (4 vols.; Sheffield: Sheffield Academic Press, 1995). Of course, no text will be free from the bias of its editor(s), but there is no good reason why theorists who question Markan Priority should be forced to work from a text already biased in that direction. ↩
• ^[20] See Eta Linnemann, Is there a Synoptic Problem? (Grand Rapids: Baker, 1992). ↩