The Priority of St. Matthew’s Gospel

Daniel J. Castellano


Part V: Linguistic and Stylistic Evidence
   5.1 Identifying Marcan Style
   5.2 Marcan Redactional Syntax
   5.3 Marcan Vocabulary
   Footnotes to Part V

Part V: Linguistic and Stylistic Evidence

5.1 Identifying Marcan Style

The degree of verbatim similarity between Matthew and Mark could more easily be explained by shared oral and written sources than by the supposition that one evangelist had access to the other’s Gospel. If the claim that Matthew used the Gospel of Mark as a written source is to have any weight, it should be shown that (1) he sometimes replicates writing style found in Mark; (2) this style is truly idiosyncratic, not common usage; and (3) this style really comes from Mark as editor or author, rather than his sources. All three of these statements must hold in order to indicate direct or indirect literary dependence on Mark.

The third claim is the most difficult to establish, since it requires source critical judgments about the Gospel of Mark. Since we are using these judgments to evaluate the Synoptic two-source hypothesis, there is a danger of circularity. Any judgments distinguishing “source” and “redactional” material ought to abstract from the theory of Marcan priority we are trying to test.

The difficulty of distinguishing source and redactional material was already felt by Bultmann, who recognized no way of distinguishing between Mark’s “editing” of oral tradition and editing of written sources. Today we appreciate that it is improper to speak of “editing” oral tradition, since this has no fixed textual form. Still, since our only versions of the original oral and written tradition are the completed Gospels, the difficulty of distinguishing the two forms remains, especially as texts were mediated orally and oral traditions may have been partly written.

Nonetheless, form critics from Bultmann onward have attempted to distinguish Mark’s editorial material from his sources, finding the former mainly in the introductions and transitions between pericopes, as well as in parenthetic comments. The redactional verses identified by various form critics have been listed by E.J. Pryke in Redactional Style in the Marcan Gospel (1978).[1] By our count, we find that 426 Marcan verses contain redactional material, or 64% of the total. These are distributed throughout the Gospel, generally with only one or two verses of “source” in between. Many redactional pieces, however, are partial verses, so if we count each of these as a half verse, the total becomes 350, or 53% of the total. If Mark’s sources were purely literary, this would be a ridiculously complicated editing project.

On the contrary, granting that most of his sources were oral, Mark could have composed most of his Gospel in a natural style, drawing on oral tradition from memory or recitation, while occasionally inserting some textual material. This would be consistent with what historical testimony tells us, so that it is improper to think of Mark as a mere compiler or redactor. Still, even oral traditions have their fixed phrases for sayings or key statements, so if we are looking for distinctively Marcan style, it is safest to steer clear of all source material.

Pryke (op. cit.) applied a linguistic method that is independent of form critical judgments, relying on statistical evidence of Marcan vocabulary and syntax. Following earlier linguistic studies that had identified distinctively Marcan syntactical features, Pryke tested whether these features were more frequent in what form critics had identified as redactional as opposed to source material. Further, he invoked the presence of distinctively Marcan vocabulary and other factors to determine which instances of syntactical features found in “source” ought to be reclassified as “redactional.” These reclassifications have not been generally followed by later critics, and Pryke himself admits that his arguments for such are of uneven weight.

Since we will use linguistic evidence to test Marcan priority (which Pryke took for granted), it is important for us to identify only that redactional material which was definitely composed by Mark himself, to see if this is copied by Matthew. Some of the syntactical features and vocabulary discussed by Pryke are not unique to Mark, but may also have been used by his oral and written sources. Much of what Pryke calls “redactional” may come from a textual source (“redacting” oral tradition) rather than Mark himself. Thus we will invoke our earlier distinction between “oral” and “textual” pericopes to help clarify which instances of syntactical features are unequivocally Marcan. If Marcan style were copied by Matthew in an “oral” pericope, this would be stronger evidence of dependence on Mark, since there would be no question of copying from a shared written source.

We will examine the more significant Marcan syntactical features, most of which are discussed by Pryke, namely: parenthetical clauses, ευθυς, explanatory γαρ, παλιν, historical present (not in Pryke), impersonals, αρχομαι plus infinitive, λεγω ‘οτι, genitive absolute, two or participles with the main verb. Afterward, we look at some vocabulary characterized as Marcan either because of their frequent occurrence or uniqueness to the Gospel of Mark.

In general, we will find only equivocal or non-existent evidence that any Marcan style has been copied by Matthew. Given the generally high degree of verbatim similarity between the two Gospels, this finding is astonishing. If Matthew used Mark, somehow he had a keen sense of distinguishing source from redactional material. Yet if he chose only to use Mark’s sources, there is no need to invoke the supposition of access to Mark, especially if the similarity of order, as we have seen, can be otherwise accounted.

[Top of page]

5.2 Marcan Redactional Syntax

5.2.1 Parenthetic Clauses
5.2.2 ευθυς
5.2.3 Explanatory γαρ
5.2.4 Connective παλιν
5.2.5 Historical Present
5.2.6 αρχωμαι plus Infinitive
5.2.7 λεγω οτι
5.2.8 Impersonals
5.2.9 Genitive Absolute
5.2.10 Double Participles

5.2.1 Parenthetic Clauses

As mentioned previously, the use of parenthetic clauses is more characteristic of writing than speech, as the latter tends to integrate short idea units sequentially. Extended parentheses require long, complex sentences to be integrated in a logical structure. The main structure of the sentence must be resumed at the end of the parenthesis, which is easier to do in writing than in speech. Any extended parenthesis in Mark breaking the flow of a sentence would be incompatible with oral tradition, and so may be regarded as evidence of redaction, either by the evangelist himself or some written source.

There are twenty-nine parenthetic clauses in Mark, eight of which merely introduce the translation of a term with ‘ο εστιν (“that/this is”), and do not truly interrupt the flow of the sentence. Of the remaining twenty-one parenthetic clauses, Pryke considers thirteen to be clearly redactional, three of which may also be “traditional or community sayings.” Another three are partly redactional and partly source or traditional, while the remaining five are redactional in only part of the verse. [Pryke, op. cit., p.32.] Thus, in Pryke“s estimation, at least eighteen and possibly all twenty-one parentheses are redactional.

Only three of the twenty-one Marcan parentheses are also found in Matthew. These are: Mk. 2:10c (Mt. 9:6), Mk. 7:6-7 (Mt. 15:7-9), Mk. 13:14b (Mt. 24:15). None of these three are clearly attributable to Mark rather than a written source used by him.

The first of the three, “he saith to the paralytic,” [Mk. 2:10c, Mt. 9:6] clarifies that this part of the speech is directed to the sick man rather than the scribes and Pharisees addressed earlier. The use of the present tense here is indicative of oral tradition (so the more literary Luke uses the past tense of the synonym ειπεν), and such a dialogue clarification hardly qualifies as the sort of subordinate clause structure that is specific to written tradition. Thus the parenthesis could just as easily come from an oral source, or a written source using oral modes of composition.

The second of the three [Mk. 7:6-7, Mt. 15:7-9] is not really a parenthesis, but a quotation of Isaiah by Jesus, which is essential to the dialogue. Thus it is attributable to the oral source, though any writer would naturally assimilate the language to the Septuagint.

The last of the three, “he that readeth let him understand,” appears in the Synoptic apocalypse, which we have seen to belong to a common textual tradition. There is no reason to believe Matthew copied this parenthesis (omitted by Luke) from Mark unless we suppose Mark to be the author of this apocalyptic discourse. Yet we find that absolutely none of the distinctively Marcan syntactic features are found in the Synoptic apocalypse, reinforcing our impression that all three Evangelists had recourse here to a common written tradition.

Some other Marcan parentheses are found only imperfectly in Matthew, with a difference in grammar or meaning. In Mk. 1:1-4, Mt. 3:1-3, we again find a quote from Isaiah, rather than an interpolated subordinate clause, so this is hardly incompatible with source usage.

Mk. 6:14b-16 has the parenthetic clause: “for his name was made manifest.” In Matthew 14:1, however, this is not interpolated in a sentence, but starts a passage: “At the time Herod the Tetrarch heard the fame of Jesus…”

Three more parentheses are not really digressions, and might easily exist even in an oral source. Mk. 11:31-32, Mt. 21:25-26 is not really a parenthesis at all, but describes at the right place what the scribes and priests were thinking prior to their reply. Mk. 16:8c, Mt. 28:8 are straightforward explanatory clauses, not really out of place or interrupting the sentence. The same is true of Mk. 4:31b, Mt. 13:32, since explaining the significance of the mustard seed is essential to the teaching, so we may expect to find it even in source.

It is at least plausible that the Marcan parenthesis in Mk. 12:12b (“For they knew that he spoke this parable to them”) is copied by Matthew [Mt. 21:45]. This explanation is placed in a more logical order by Matthew, right after the parables, so he might have cleaned up Mark’s more garbled order. Yet it is also conceivable that both used a common source with (a) parables and (b) the desire to lay hands on Jesus, and independently produced the same explanation. The verbal similarities in the explanatory verses, however, suggest that this explanation was in the common source, and Matthew does a better job of cleaning it up. (Luke 20:19 uses the same words as Mark.)

The parenthesis, “For the woman was a Gentile, a Syrophoenician born,” [Mk. 7:26a, Mt. 15:22] is a clear interpolation in Mark, yet it is not the sort of extended subordinate clause characteristic of writing. It is common in oral compositions to bring in details that more logically ought to have been mentioned earlier, so this could easily belong to source material. Matthew has this in a cleaner order, introducing the narrative, “And behold a woman of Canaan...” If one Gospel were literarily dependent on the other, we should say Matthew improved Mark, but the verbal similarity is not so strong as to require such direct dependence. (Luke has no parallel pericope.)

There is no strong evidence that Matthew copied from Mark any of the twenty-one parenthetic clauses that are not introducing translations. We now consider the eight mini-parentheses that simply introduce translations of terms. Seven of these use the phrase ‘ο εστιν an apparent Latinism derived from id est or hoc est. This Marcan peculiarity is consistent with traditional testimony that the evangelist wrote for the Roman Church.

The seven instances of ‘ο εστιν introducing translations of Aramaic terms are Mk 3:17, 5:41b; 7:11-12, 7:34; 12:42b, 15:16b; 15:34, which are all redactional according to Pryke. Yet five of the first six instances (excluding Mk. 12:42b, which has no Matthaean parallel) fall within pericopes we have identified as having only mild verbal similarity among the Synoptics, consistent with mediation by shared oral tradition. None of these six have ‘ο εστιν replicated in Matthew or Luke, confirming that the phrase is a Marcan interpolation, but disconfirming any evidence of Matthaean copying of Mark. The seventh instance, Mk. 15:34, pertains to the Aramaic words spoken by Christ on the Cross, a pericope we have classified as reflecting a shared textual tradition. The term εστιν is found in Matthew 27:46, but as part of the grammatically correct Greek phrase το&upsilonτ εστιν. This hardly constitutes strong evidence of copying from Mark, especially if we accept that this Aramaic utterance of Christ was contained in the source. Then it would only be natural for any Greek writer to offer a translation of this phrase. Matthew’s translation could well be independent of Mark’s, as seems indicated by their verbal differences in Greek.

The eighth parenthesis introducing a translation is Mk. 5:9c, which does not include the phrase ‘ο εστιν This is not an overtly editorial insertion, as it is contained in the quotation: “My name is Legion, for we are many.” The latter clause explaining the Latin name might well have existed in source, either to explain the term to a Jewish Christian audience, or, assuming the term was known, to explain why it was used as a name. At any rate, neither the name nor its translation is reproduced in Matthew.

Although the phrase ‘ο εστιν is almost certainly a Marcan characteristic, we find that it is never replicated in Matthew (or Luke). Either Matthew had no need of Mark for these passages, or else he deliberately omitted the redactional portions.

[Top of page]

5.2.2 ευθυς as Connective

The use of ευθυς (“then; immediately”) as a connective or introduction to a pericope is thought to be an example of Marcan syntax, since this is far more frequent in the Second Gospel. There are at least 40 such usages (43 by Pryke’s count); in Mark. Matthew uses some version of the expression at least 15 times (17 by some counts) in its parallels to Mark, but only once elsewhere. Since 59.6% of Matthaean words have Marcan parallels (by the count in Robert H. Stein’s The Synoptic Problem: An Introduction), it is argued that the stronger correlation in the usage of ευθυς is evidence that Matthew copied the expression from Mark.

Yet there are two credible alternative explanations of this correlation. First, Matthew and Mark might both take the expression from a common source. Second, the Matthaean-Marcan double tradition is mostly narrative, so a Jewish writer would be more likely to invoke the expression in these sections, especially if ευθυς is a translated Hebraism meaning “at that time.”

I have examined all forty-three of Pryke’s Marcan instances of ευθυς, and compared them with Matthaean parallels in Farmer’s Synopticon, updating readings with the 28th edition of Nestlé-Aland when needed. Fifteen possible Matthaean parallels were found, which we discuss below.

Pryke distinguishes instances of ευθυς and και ευθυς, treating each expression separately. Beginning with ευθυς, he finds 7 Marcan uses in redactional texts, and 9 uses in source texts, 3 of which Pryke would reclassify as redactional. Only one Marcan redactional use out of seven (or ten if we count those reclassified by Pryke) is replicated in Matthew. This is the instance of ευθειας in Mark 1:3a, which is a quotation from Isaiah. Clearly, this text is “redactional” only as contrasted with oral tradition, not as proving a uniquely Marcan contribution, since this is just a quotation from the Septuagint. Even Luke, who shuns the Hebraic usage of ευθυς includes the term. Since it is uncontroversial that all three Evangelists had access to the Septuagint, there is no proof of dependence here.

Three out of nine usages from Marcan source (or three out of six following Pryke) are found in Matthew. Instances of ευθυς in Mark 4:16 and 4:17 are also in Matthew. ευθυς in Mark 14:45 is seen in Matthew as ευθεως, which we will find to be a preferred Matthaean form of the term.

In short, Matthaean replication of Marcan ευθυς is strong in source, but non-existent in redaction. This does not support Marcan priority, and gives counter-evidence to the notion that ευθυς is a uniquely Marcan expression. Since it is found in sources which come from oral tradition, it is plausible that this expression would be common among Hellenized Jewish Christian narrators.

For και ευθυς, there is at least some evidence of Matthaean replication of Marcan redactional use. Pryke counts 16 instances in redactional texts, and 11 uses in source, 4 of which are possibly redactional. We find only 4 out of 16 (or possibly 20) redactional Marcan uses also in Matthew, three of these only imperfectly similar.

Mark 1:20 (“And forthwith he called them…”) has και ευθυς, but the Matthaean parallel is οι δε ευθεως instead. (The conjunction και appears only much in the verse.) This pericope on the call of the disciples (Mk. 1:16-20, Mt. 4:18-22), has strong verbal similarity (66% of Marcan words in Matthew) indicating some literary relationship. The question is whether Matthew or Mark’s version is more primitive. We will find that Matthew generally prefers the spelling ευθεως over ευθυς, while Mark uses only the latter. This difference in usage also occurs in source texts, including a rare instance where Matthew and Luke agree against Mark. (Mt. 8:3, Lk. 5:13, Mk. 1:42). While there is nothing strongly probative either way, it is at least plausible that Matthew’s form better preserves a shared written source.

και ευθυς in Mark 6:45 (“And immediately he obliged his disciples...”) is paralleled by και ευθεως in Mt. 14:22. (This expression is considered a variant in NA25, but is the primary reading in NA28.) Since context does not demand this expression, the parallel is not likely to be coincidental. There is some literary relationship, but this could be a shared written source, as in the previous case.

In Mark 10:52 and Matthew 20:34, however, there need not be a literary relationship at all. In both cases (και ευθυς in Mk. and και ευθεως in Mt.), the expression is called for by the context, to emphasize the immediate effect of Jesus’ healing power. We have seen that this pericope has only an oral degree of similarity (25% of Marcan words in Matthew).

Mark 11:2 is the only redactional instance of και ευθυς that is replicated exactly in Matthew. While this pericope generally shows only an oral degree of similarity, the second verse quotes the words of Jesus, and we find that important quotations can be preserved verbatim even in oral tradition. The words “and immediately at your coming thither”... are functionally essential to the quotation, which gives instructions on where to find a colt, so the expression may well have been in oral tradition. This would account for why even Matthew uses the form ευθυς here, not daring to alter the conventional recollection of Jesus’ words.

We may further note that a redactional use in Mark 14:43 is not replicated in Matthew although there is otherwise strong verbal similarity for this pericope.

Ultimately, there are only two instances of και ευθυς (Mark 1:20 and Mark 6:45) that are indicative of some literary relationship between Gospels. This relationship need not be one of Matthaean dependence on Mark, but could be the result of a shared written tradition. The expression is not idiosyncratic to Mark, as is proved by its comparably frequent appearance in source text (which is 40%-50% of Mark by form critical standards).

Of the 11 Marcan instances of και ευθυς in source, seven of these are at least partly replicated in Matthew. This is a much higher rate of replication than for the redactional instances, which argues against Matthaean dependence on Mark, for on that assumption, Matthew should had little means of distinguishing source and redactional usage.

As in the redactional instances, Matthew’s wording and spelling is more diverse than that of Mark. For the six Marcan instances of και ευθυς in source (Mk. 1:10, 1:18, 1:42, 4:5, 11:3, 14:72), Matthew has (1) simply ευθυς (not in same place, as words are all in different order) in Mt. 3:16; (2) οι δε ευθεως in Mt. 4:20; και ευθεως in (3) Mt. 8:3 and (4) Mt. 13:5; and και ευθυς in (5) Mt. 21:2 and (6) Mt. 26:75. Where Mark has ‘ο δε ευθυς (Mk. 6:50), Matthew has ευθυς δε (Mt. 14:27).

Luke generally eschews the Hebraism και ευθυς, but in the one place where he retains it (Lk. 5:13), he agrees with Matthew’s version (Mt. 8:3) against Mark (1:42). There is no question that this instance belongs to source, since it is essential to the narrative that the leper was immediately cleaned. Here Luke seems to indicate that Matthew has more accurately preserved the usage of source in his preference for ευθεως. He could not have gathered this from Mark, so he must have had independent access to the source material. Once this is granted, dependence on Mark would be superfluous.

[Top of page]

5.2.3 Explanatory γαρ

The term γαρ (“for”) is frequently used to introduce an explanatory phrases in Mark. By Pryke’s count, 33 out of 66 such uses are redactional or editorial. (Some counts have 34.) Matthew, by contrast, is said to use γαρ editorially only 11 times out of 123 total uses. Ten of these are parallel to Mark, from which it is urged this is evidence of dependence on Mark on statistical grounds.

First, it is critical to this analysis to determine which uses of γαρ are truly editorial, since, based on overall usage, γαρ is no more Marcan than Matthaean. Second, granting that a use is editorial, it might be plausible for both Matthew and Mark to independently make such an insertion from time to time, since both frequently use this construction. The fact that Matthew’s usage mostly appears in pericopes with Marcan parallels might simply be a consequence of the fact that explanatory phrases are much more useful in narratives (shared with Mark) than in sayings material (shared with Luke). In other words, there may be other features distinguishing parallel and non-parallel material, such as genre, which could account for the statistical disparity.

Pryke, following M.E. Thrall and others, explains Mark’s redactional use of γαρ with his free writing style, not planning long sentences in advance. In consequence of this style, details are written as they come to mind, sometimes out of logical order. Yet we have seen that this style is characteristic of oral modes of composition, so we might just as well find it in source as in redaction. Further, not all of these interpolations need to be explained by an author’s forgetfulness, as there are sound storytelling reasons to delay mention of a detail.

If we survey the 33 “redactional” uses of γαρ in Mark, we find that they are replicated in Matthew only in pericopes where we have found a shared “textual” source. When the degree of similarity is only “oral,” (e.g., Mk. 3:10), the term is not found in Matthew. This implies that Matthew uses explanatory γαρ phrases only when found in written tradition, not on his own. Mark, by contrast, does this both on his own and when citing written tradition. This is consistent with general findings that Matthew is a more literary writer, whereas Mark composes his text orally.

The 33 redactional uses of γαρ in Mark are: 1:16,22,38, 2:15, 3:10,21, 5:8,28, 6:14,31,48,52, 7:3, 9:6(x2),31,34,41, 10:22,45, 11:13,18(x2),32, 12:12,23, 13:11, 14:2,40, 15:10, 16:4,8(x2). Fourteen of these appear in double tradition pericopes that have a “textual” degree of similarity: 1:16,22, 2:15, 6:48,52, 9:6(x2), 10:22,45, 11:32, 12:12,23, 13:11, 14:40. Eight of these fourteen are found in Matthew. Another fourteen Marcan uses are found in pericopes with only an “oral” degree of similarity: 3:10, 5:8,28, 6:14, 7:3, 9:31,34,41, 11:13, 14:2, 15:10, 16:4,8(x2). Only two of these fourteen are also in Matthew. The remaining five uses are in pericopes with no Matthaean parallel: 1:38, 3:21, 6:31, 11:18(x2).

The similarities in “textual” pericopes would prove Marcan priority only if we could eliminate the possibility that Matthew and Mark are using a common written tradition, and show that the explanatory γαρ is a distinctively Marcan characteristic. Yet other writers such as Herodotus use similar construction, so we cannot discount the possibility that Matthew and Mark found these phrases in written sources.

More promising evidence of Marcan priority is in the two instances replicated by Matthew in “oral” pericopes. Since Matthew generally does not prefer this construction, he is hardly likely to introduce it unless he found it in his source, and it may strain credulity to suppose that both evangelists independently inserted the same explanatory remark.

On closer inspection, these two instances do not clearly support dependence on Mark. The Matthaean analogue of Mark 5:28 (“For she said...”) differs in its surrounding construction, so the similarity could be coincidental. Further, even the Marcan version need not be a redactional lapse, as the interpolation would make sense within oral storytelling. After all, there is better dramatic effect if the woman first touches Jesus’ garment before we are told what her reason was. Action before explanation makes for more compelling narrative.

Likewise, Mark 15:10 and its Matthaean parallel (“For he knew that for envy...”) use a rational construction that could easily come from source without appeal to a Marcan lapse. It would not make narrative sense to give this explanation prior to Pilate’s question. There is no need to regard this instance as a Marcan idiosyncrasy, especially as Matthew is not strongly averse to explanatory γα&rho phrases when found in source.

In any case, a mere two parallels is hardly probative of dependence on Mark, given Matthew’s tolerance of this construction. Luke, who mostly shuns this construction, even when it is in source, has as many parallels in the Matthaean-Marcan “oral” pericopes. The Lucan pericope parallel to Mark 5:8 has a “textual” degree of similarity, so its γαρ phrase could be based on Mark. The Lucan parallel to Mk. 14:2 (“for/but they feared the people”), by contrast, is not an irregular usage, so it could well be coincidental.

Pryke uses a “linguistic method” to argue that many of the explanatory γαρ phrases in source should be reclassified as redactional. His arguments are not all equally convincing. A stronger case for redaction exists where γαρ introduces a Biblical citation (Mk. 7:10, 10:27, 13:8,19). Looking closely at these, we find that the γαρ phrase in Mk. 7:10 is quoting Jesus, “For Moses says…,” so this could well be in source. A clearer case of redaction is in Mark 10:27, where γαρ is inserted in the Scriptural text, yet this is not replicated in Matthew. The phrase “For nation shall rise...” in Mark 13:8 is part of the Synoptic apocalyptic discourse, which seems to have had an early written tradition, so it could come from source. Lastly, Mark 13:19 “For in those days...” is not really a Scriptural citation, so it is not necessarily redactional.

In sum, we find at best equivocal evidence of Matthaean dependence on Mark, and some evidence positively opposed to such dependence. Whenever there is an unequivocally Marcan instance of the explanatory phrase, it is not replicated by Matthew. If Matthew did have access to Mark, we once again find that he had an uncanny intuition about which Marcan uses were from source rather than redaction.

[Top of page]

5.2.4 Connective παλιν

The term παλιν (“again”) is used in Mark as a redundant connective, likely a translated Hebraism. Pryke counts 17 redactional uses: Mk. 2:1,13, 3:1,20; 4:1, 5:21, 7:14,31, 8:1,13, 10:1(x2),10,24,32, 11:27, 14:40. Only four of these are in pericopes we have identified as showing an “oral” degree of similarity between Matthew and Mark: Mk. 5:21, 7:14, 7:31, 8:13.

Of the seventeen Marcan uses, only two have a parallel use in another Gospel, in both cases Matthew. Neither instance has the same adjoining words in Matthew as in Mark, making it plausible that the similarity is coincidence.

In the first instance, we find: “Jesus again answering, says to them…” (Mk. 10:24) versus “And again I say to you…” (Mt. 19:24) This is hardly a clear case of copying Marcan usage.

In the second instance, we have: “he found them again asleep” (Mk. 14:40) versus “…and he cometh again and findeth them sleeping.” (Mt. 26:43) Again, the difference in wording implies that this need not be a case of Matthew copying Mark. Further, the use of “again” here is not redundant, but eminently appropriate, as this is the second time the apostles are found sleeping. Lastly, the term παλιν is found repeatedly in the Passion narrative, so this usage could be characteristic of a shared source.

If anything, the Marcan παλιν is evidence against Matthaean dependence on Mark, since Matthew never replicates its usage as a redundant connective. On the assumption that Matthew used Mark as a literary source, it is curious that Matthew would systematically delete this term, especially since he is not averse to a pleonastic παλιν. (cf Mt. 4:7,8, 5:33, 13:44,45,47, 18:19, 19:24)

[Top of page]

5.2.5 Historical Present

One stylistic trait not discussed by Pryke is the use of the historical present, i.e., speaking of the past in the present tense, as if the historical events were happening now. This is not a Marcan peculiarity, but can be found in many ancient authors, notably the early historians Herodotus and Thucydides. In Mark, however, the frequency of the historical present is especially pronounced compared with the other Synoptics. This usage is more typical of oral histories, and may reflect Mark’s less literary mode of composition.

By the count of J.C. Hawkins, there are 151 instances of historical present in Mark, 78 in Matthew, and 9 in Luke. Five of the Lucan instances occur in parables rather than narration, so they are attributable to source, and we may say Luke hardly uses the historical present at all. Marcan priorists argue that Luke removed all the historical presents from his Marcan source, while Matthew did this only some of the time, while in other instances leaving the Marcan version unaltered. This argument is weak, since the fairly frequent usage of the historical present by Matthew implies he had no strong objection to this usage. In other contexts, irregularity over uniformity is an argument for priority. Further, Matthew need not have relied on Mark for the historical present, since such usage would have been present in any oral source. Lastly, we find that the Gospel of John also has a high frequency of this usage, with over 160 instances. Matthew’s usage could be similarly independent.

[Top of page]

5.2.6 αρχωμαι plus Infinitive

Mark frequently uses αρχωμαι (“begins to”) as an auxiliary preceding the infinitive verb. In most cases, this is a purely redundant stylistic choice, adding nothing to the meaning. According to Pryke, all 26 instances of this usage should be considered redactional, including 6 in verses assigned to source.

Matthew has 12 instances of αρχωμαι preceding an infinitive. As shown by J.W. Hunkin, however, αρχωμαι is never altogether redundant (pleonastic) as in Mark. [J.W. Hunkin "'Pleonastic' archomai in the New Testament," JTS 25 (1924): 391-93, 395.] Thus this usage cannot be invoked as evidence of copying a peculiar Marcan style.

[Top of page]

5.2.7 λεγω οτι

The expression λεγω ‘οτι (“saying [that]”) occurs frequently in Marcan redactional text. This form is used to introduce direct speech, which Mark strongly prefers over indirect speech, again reflecting an oral rather than literary mode of composition. The ostensibly redactional instances are:

Mk. 1:15,40, 2:12,17, 3:11,21,22,28, 4:21, 5:28,35, 6:14-16(x3), 7:20, 8:28, 9:1,11(?),12,30,40, 10:32, 11:23, 12:43, 13:6,30, 14:18,27(x2),57,58(x2) [Pryke, op. cit., p. 73. (Corrected 9:31,41 to 9:30,40.)]

Only 11 of these 32 instances have at least partial Matthaean parallels. On close inspection, it is far from clear that the shared instances are redactional, since they are mostly given in the first person form with Jesus as speaker. All four canonical Gospels give αμεν λεγω ‘υμιν (“Amen I say to you”) as a distinctive expression of Jesus, including the unique prefatory amen. Thus it is likely that most of these shared instances come from oral tradition.

In most of the parallels, Matthew omits the term ‘οτι found in Mark. This means there is no stylistic similarity other than the use of direct speech, which is hardly peculiar to Mark.

Mark 3:28 has Jesus say αμεν λεγω ‘οτι, but in Matthew 12:31 it is “Therefore I say…,” dropping the ‘οτι.

Mark 5:28 introduces the speech of the unclean woman: ελεγεν γαρ ‘οτι (“For she said (that)…”). Matthew has this without the ‘οτι. It is unremarkable that both evangelists choose direct speech here, and may expect the same from an oral source.

Mark 9:11 introduces the reply of Jesus to the Pharisees: “said to them…” Again Matthew omits ‘οτι.

The Matthaean parallel to Mark 9:12 (λεγω ‘υμιν ‘οτι) retains the ‘οτι but inserts a disjunctive δε: λεγω δε ‘υμιν ‘οτι (Mt. 17:12) The rest of the verse has only fragmentary verbal similarities, no more than two consecutive words. This phrase is in the middle of Jesus’s speech, so it could easily come from source. Unlike the other Marcan uses, here ‘οτι is grammatically demanded, so there is no evidence of stylistic borrowing.

The Matthaean parallels to Mark 9:30, 9:40, 11:23, and 13:6 all omit the adverb ‘οτι favored by Mark. In the case of Mk. 9:30, even the verb form is different, for Mark has ελεγεν. This means there is no distinctively Marcan style in Matthew here.

In the Jerusalem narrative, we find three instances of Jesus’ distinctive expression αμεν λεγω ‘υμιν ‘οτι (“Amen I say to you” that), which include the adverb ‘οτι. Unlike other Marcan instances of this term, these are consistently replicated in Synoptic parallels, consistent with our supposition that the Passion narrative comes from a shared written source. The account of the widow’s mite (Mk. 12:43) has no Matthaean parallel, but Luke 21:3 has λεγω ‘υμιν ‘οτι, dropping the amen. The full expression is found in Mark 13:30 and its parallels, and in the Matthaean parallel to Mark 14:18. In short, the style is identical only in instances likely attributable to source.

By contrast, in the two instances of this construction in Mark 14:27 (λεγει… ‘οτι and ‘οτι γεγραπται) lack the adverb ‘οτι in Matthew, who reproduces every other word in this verse but the leading και. On the assumption of dependence on Mark, Matthew shows an uncanny knack for omitting only redactional terms. It would be more parsimonious to suppose that Matthew had access to Mark’ source (or some version of it) and that Mark introduced his mode of diction here. Thus Matthew in this instance better reproduces the written source of the Passion narrative.

In short, this Marcan construction offers no evidence of stylistic copying by Matthew. On the contrary, Matthew’s pattern of usage and non-usage indicates that he was unaware of Mark’s redaction of their shared source material.

[Top of page]

5.2.8 Impersonals

C.H. Turner found Mark’s use of impersonal terms such as “they” or “others,” denoting unspecified groups of people, to be an important stylistic feature. In some cases is it used as a substitute for the passive voice, which is avoided in Aramaic and Hebrew. Such usage is not peculiar to Mark, however, and it is found frequently in the Septuagint, notwithstanding its generally literary style. Thus we might expect to find impersonals frequently used in any text that relies on traditions translated from Aramaic.

On the rare occasion when we have an impersonal usage that is part of a purely Marcan introduction, Mark 9:14 (Pryke, op. cit., p. 110), there is no parallel clause in Matthew. This is the case for six of the twenty-nine Marcan impersonals that occur in redactional verses:

Mk. 1:45 “they flocked to him” (clause not in Mt./Lk.)
Mk. 3:21 “For they said: He is become mad.” (pericope not in Mt./Lk.)
Mk. 5:35 “some come from the ruler of the synagogue’s house” (clause not in Mt.)
Mk. 6:15 “others said” (sentence not in Mt.)
Mk 9:14 “they saluted him” (clause not in Mt. or Lk.)
Mk 14:12 “when they sacrificed the pasch” (clause not in Mt.; Lk. 22:7 “it was necessary that the pasch should be killed”)

Of the remaining twenty-three Marcan impersonals, at least thirteen are definitely not replicated in their Matthaean parallels.

Mk. 1:21 “they entered Capharanum”; Mt. 8:5 “he entered Capharnaum”
Mk. 1:29 “they came into the house of Simon”; Mt. 8:14 “when Jesus was come into Peter’s house”
Mk. 5:1 “they came over the strait”; Mt 8:26 “when he was come on the other side”
Mk. 6:33 “And they saw them going away...”; Mt. 14:13 “the multitudes having heard of it”
*Mk. 6:54 “And when they were gone out of the ship, immediately they (presumably the crowds) knew him”; Mt. 14:33 “And they that were in the boat came and adored him” Mt. 14:35: “And when the men of that place [Genesar] had knowledge of him...”
Mk. 9:33 “But they held their peace”; Mt. 18:1 “the disciples came to Jesus”
*Mk. 10:13 “And they brought to him young children”; Mt. 19:13 “Then were little children presented to him” (no verbal similarity; no parallel clause in Luke)
Mk. 10:32 “And they were in the way going up to Jerusalem, and Jesus went before them”; Mt. 20:16 “And Jesus going up to Jerusalem, took the twelve disciples apart...”; Lk. 18:31 “Then Jesus took unto him the twelve”
*Mk. 11:12 “when they came out from Bethania”; Mt. 21:18 “returning into the city, he was hungry” (no parallel pericope in Luke)
Mk. 11:15 “And they came to Jerusalem”; Mt. 21:12 “And Jesus went into the temple”; Lk. 19:45 “And entering into the temple, he…”
Mk. 11:20 (x2) “And when they passed by in the morning they saw the fig tree dried up”; Mt. 21:20 “And the disciples seeing it wondered” (no parallel pericope in Lk.) Possible second instance in Mk. 11:19 variant: “whenever evening came they would go out” (verse not paralleled in Mt.)
Mk. 11:27 “And they come again to Jerusalem”; Mt. 21:23 “And when he was come into the temple”; Lk. 20:1 “as he was teaching the people in the temple”

Marcan priorists may argue that Matthew has here cleaned up Mark’s impersonals, but some instances have such thorough differences in wording and structure as to efface any dependence on Marcan redaction. For example, the Matthaean parallel to Mark 6:54 is split into two non-consecutive verses, with such different form and content as to suggest independent redaction. Likewise, the entire verses Mk. 10:13 and Mt. 19:13 have strong verbal dissimilarity. The change in structure of the Bethania pericope is also suggestive of independence. (Mk. 11:12; Mt. 21:18) Comparison with Luke in the last instance (Mk. 11:27) shows what a variety of constructions are available to a redactor.

If Matthew did feel a need to rewrite Marcan impersonals, it was not out of any special aversion to this construction, as he uses them frequently. Ten of the Marcan redactional impersonals have some apparent Matthaean replication, though several of these are verbally dissimilar. None of these constitute unequivocal evidence of dependence on Mark.

  1. Mk. 1:22 “they were astonished”; Mt. 7:28 “the people were in admiration”
  2. Mk. 1:32, Mt. 8:16 “they brought to him” (in both cases, antecedent is the apostles, since previous verse has Simon’s mother ministering to them)
  3. Mk. 6:55 “they began to carry about in beds”; cf. Mt. 14:35 “they [the men of that place]… brought to him all that were diseased”
  4. Mk. 9:29 “departing from thence, they passed through Galilee”; Mt 17:21 “And when they abode together in Galilee” [not in Lk.] (unclear if antecedent is the Twelve or the disciples)
  5. * Mk. 11:1 “And when they were drawing near to Jerusalem”; Mt. 21:1 “And when they drew nigh to Jerusalem”; Lk. 19:28 “he went before, going up to Jerusalem”; Lk. 19:29 “when he was come nigh to Bethpage” [triple tradition]
  6. Mk. 11:11? “having viewed everyone/thing around… he went out to Bethania”; Mt. 21:17 “And leaving them, he went out of the city into Bethania” (antecedent is priests/scribes) [no parallel in Luke]
  7. Mk. 12:13, Mt. 22:16 “And they sent to him”; Lk. 20:20 (in Mt. and Lk., the antecedent is in previous verse: “Pharisees” in Mt.; “priests and scribes” in Lk.)
  8. * Mk. 14:22 “And whilst they were eating”; Mt. 26:26 “And whilst they were at supper”; Lk. 22:14 “And when the hour was come, he sat down, and the twelve apostles with him”
  9. * Mk. 14:26 “And when they had said a hymn”; Mt. 26:30 “And a hymn being said, they went out”; Lk. 22:39 “And going out, he went… and his disciples also followed him.”
  10. * Mk. 14:32 “And they came to a farm called Gethsemani”; Mt. 26:36 “Then Jesus came with them into a country place which is called Gethsemani”; Lk. 22:39 “And when he was come to the place” (does not name Gethsemani)

In the first case, the use of an impersonal is unremarkable. In the second, there is clear antecedent in both Mark and Matthew. In the third, sixth and seventh, the Marcan impersonal has a clear antecedent in Matthew. Notably, in the seventh, there are different antecedents in Matthew and Luke. The fourth instance is not a highly ambiguous impersonal, since it refers to either the Twelve or the disciples in general, who are not always distinguished by any evangelist. Further, the wording is highly distinct.

This leaves the fifth, eighth, ninth, and tenth instances, which all occur in the triple tradition Passion narrative. These impersonals could come from a shared source, especially as none of them are examples of Marcan confusion. In all four cases, “they” are likely Jesus and the disciples. Luke removes the impersonal in all cases. Matthew retains an impersonal, but has different wording from Mark, especially in the last instance, where he partially removes the impersonal by naming Jesus. There is little evidence of verbal dependence on Mark.

Marcan impersonals provide no clear evidence of Matthaean dependence on Mark, which is noteworthy considering how many potential opportunities for such evidence they provide.

[Top of page]

5.2.9 Genitive Absolute

The Greek genitive absolute may be translated as “while [noun], [verb],” but in Greek there is no “while,” only the noun in the genitive form, coupled (not always immediately) with a verb participle (typically the present participle or “-ing” form in English, though past and future participles are also possible). The participle should not correspond to the subject or object of the sentence (i.e., it is in a subordinate clause). Many uses in Koine Greek ignore this rule, however, often having the participle agree with the subject of the main clause. While modern scholars have held that this is grammatically improper, or at least in bad style, since a genitive absolute construction is not necessary in such cases, it is hardly illogical, e.g.:“Looking around, he walked slowly.” Further, this “improper” usage was so common even among highly literate ancient authors that we should rather question whether this rule was even recognized in first century. What is essential to the construction is its function as a circumstantial participle, not an “absolute” distinction from the main clause.[2]

More blatantly ungrammatical is the double use of the genitive absolute in the same sentence. This occurs three times in Mark:

Mk. 6:21-22 “And when a convenient day was come… and when the daughter of the same Herodias had come in” (not in Mt. or Lk.)
Mk. 8:1 “when there was a great multitude… calling” (not in Mt.; no Lucan parallel)
Mk. 14:3 “having an alabaster box… and breaking the alabaster box” (not in Mt. or Lk.)

None of these ungrammatical doubles are replicated in the three Matthaean and two Lucan parallels. It may be argued that the other evangelists cleaned up Marcan grammar, but this supposition cannot constitute positive evidence of verbal dependence on Mark.

Mark has a strong preference for genitive absolutes, so it is useful to look at all redactional instances:

MarkProper?Genitive absolute used in Matthew?
1:32yesyes (8:16)
5:2no (dat.)no
5:21no (acc.)no
5:35yesno (but in Lk.)
6:54no (nom.)no (no parallel to “when they were gone out of the ship”)
8:1 no (nom.
9:9 [9:8]no (dat.)yes (17:9 Transfiguration)
9:28 [9:27]no (acc.)no (Mk. εισελθοντος αυτον “while he came indoors”; Mt.17:19 πρσελθοντες “[the disciples] having come”)
10:17no (acc.)no
10:46yesyes (Mk. ‘and as he went out of Jericho, with his disciples;” Mt. 20:29 “And when they went out from Jericho”)
11:27no (acc.)yes (dat.) (Mt. 21:23 completely different verb; also has additional verb “as he was teaching”)
13:1no (dat.)no
13:3no (acc.)yes (dat.)
14:3no (gen.)no

[Vulgate verse numbers in brackets. Case of matching referent noted parenthetically for Mark, and for Matthew where different from Mark.]

Ten of the 24 redactional uses of the genitive absolute in Mark have some parallel in Matthew; seven of these are in the Passion narrative (out of 12 in Mark). Five of ten “proper” uses in Mark, and only five of 21 improper uses, have a Matthaean parallel. Three of the five Matthaean proper uses and four of the five improper uses are in the Passion narrative. Evidently, Matthew had no problem retaining the improper use when he found it in a source, i.e., the common source of the Passion narrative.

The three improper uses marked with an asterisk may be considered proper use, insofar as the genitive absolute refers to Jesus and the disciples collectively, while the subject of the main clause is Jesus alone. This would agree with L.K. Fuller’s (2006) count of genitive absolute constructions in Mark. She finds that, overall, Mark uses improper genitive absolutes less frequently than Matthew (13/35 = 37.1% for Mark; 23/52 = 44.2% for Matthew), but for Matthew, most of these refer to main clause nouns in the dative case (17 of 23), while in Mark there are 2 nominative, 6 accusative, 4 dative, and 1 genitive.[3]

If we view accusative referents of genitive absolutes as being characteristic of Marcan style, while dative referents are characteristic of Matthew, we can refine our comparison of parallel uses. The six accusative instances in Mark have no Matthaean parallel in four cases, and in the other two cases (Mk. 11:27, 13:3) Matthew uses the dative, consistent with his practice even in passages without Marcan parallels. It might be argued that Matthew has “corrected” Mark to his own stylistic preference, yet there are three instances of an accusative referent in the Gospel of Matthew, showing the evangelist did not generally find this to require correction.

There are only three possible replications of an improper use that preserve referent case: Mark 9:9, 14:18, 14:22 and parallels. The first of these, in the Transfiguration narrative, uses the dative case, which is generally characteristic of Matthew and hardly evidence of copying Mark. Further, it is especially doubtful whether this is improper, since the ‘they’ in the genitive absolute includes Jesus, while the dative pronoun refers only to the disciples. Likewise, the latter two instances, from the Passion narrative (likely a shared written tradition rather than Marcan redaction), do not have a precise match in subject with the main clause. So there are no clear cases of Matthew copying an idiosyncratic Marcan improper use of the genitive absolute.

Again, when we have a truly distinctive Marcan stylistic feature, it is nowhere to be found in Matthew. The negative evidence against Matthew copying Mark accumulates. Matthew uses the genitive absolute more frequently in parts with Marcan parallels, as we should expect, since this construction is more apt to be used in narrative, where time is important. Yet he does not replicate most of the redactional uses in Mark, and where he does have a parallel genitive absolute, it is generally not in the same style as Mark, often having a different verb or referent. Matthew may have had some of the same written sources as Mark, but he apparently did not rely on Mark’s Gospel, as there is no reason for him to omit the grammatically correct usages, as he often does.

[Top of page]

5.2.10 Double Participles

The Gospel of Mark frequently uses two or more participles before or after the main verb. This is an unrefined style in Greek, and may be evidence of translation from Hebrew or Aramaic, which use run-on sentences with multiple indicative verbs. The translator dispenses with the conjunctions and makes some of the main verbs into participles. Whatever the reason for this construction, there are 27 instances of it in Mark, 11 of which are in clearly redactional verses. [Pryke, op. cit., p. 119]

Double Participle in MarkParallel Text in MatthewPericope
Mk. 1:14-15 preaching… and sayingMt. 4:17 began to preach and to sayoral
Mk. 1:39 And he was preaching… and casting outMt. 4:23 teaching… and preaching… and healingnone
Mk. 1:40 beseeching him, and kneeling downMt. 8:2 came and adored him, sayingtextual?
Mk. 3:31 standing without, sent unto him, calling himMt. 12:46 stood without, seeking to speak with him.textual?
Mk. 7:25 For a woman having heard of him, whose daughter had an unclean spirit, having come, fell at his feet.Mt. 15:22 And behold a woman… having come out, cried out, sayingoral
Mk. 8:13 And having left them, again having embarkedMt. 16:4 And having left them he went away.oral
Mk. 8:33 Who turning about and seeing his disciplesMt. 16:23 Who turning, said to Petertextual?
Mk. 10:17 a certain man running up and kneeling before himMt. 19:16 And behold one came and said to himtextual
Mk. 10:50 Who casting off his garment and leaping upMt. 20:29-34 [no parallel to verse]oral
Mk. 12:28 And there coming one of the scribes… and seeing that heMt. 22:35 And one of them, [a doctor of the law,] testing himnone
Mk. 15:46 And Joseph buying fine linen, and taking him downMt. 27:57 And Joseph taking the body, wrapped itoral
[Verbatim similarity defined in Sec. 4.5.1, using data in Appendix A.]

If only Matthew would at least once inadvertently retain Mark’s distinctive syntax, we could be persuaded that he was copying Mark, cleaning most of it up, but alas! he never fails to correct the grammar. Yet Matthew has no strong objection to multiple participles: he uses three in Mt. 4:23, which has no verbal similarity at all to the Marcan parallel. He also retains double participles from source materials, as in Mt. 12:48-49. So it hardly seems likely that he would systematically remove all redactional double participles from Mark, if Mark was his source.

Looking closely at the data, we find that in no case does Matthew retain a Marcan redactional double participle, though there is a parallel text in 10 of 11 cases, four of which are in pericopes with strong verbatim similarity. There are two cases (Mt. 4:23, 15:22) where Matthew uses his own multiple participle by sheer coincidence, employing different verbs. These are in pericopes with little or no verbatim similarity.

This “redactional” material shows no verbal sign of having been copied from one evangelist to another, though the substantive content of this material is frequently similar. The evidence suggests that the physical facts related by Matthew and Mark were part of a shared tradition (in some cases mediated by texts), yet the verbal particulars of how to express these facts were left to the discretion of each narrator, not fixed by tradition. This agrees with what we know of predominantly oral modes of composition, so that much of the “redactional” material really ought to be thought of as the compositional discretion allowed to all representations of controlled oral tradition. Accordingly, the distinction between “redactional” and “source” material breaks down.

The fact that we cannot neatly reverse engineer how the Gospels were composed suggests that their mode of composition was considerably more nuanced than a two-source hypothesis of literary redaction would imply. Instead of the anachronistic supposition that an evangelist had some written sources laid out in front of him from which to copy and paste, we find the stylistic evidence is much more indicative of non-verbatim usage of common written and oral sources, using mostly oral modes of composition.

[Top of page]

5.3 Marcan Vocabulary

It is difficult to identify distinctively Marcan word usage, since the evangelist had a relatively limited Greek vocabulary, though not as much as is commonly averred. He was far from barbarous, and had a reasonable command of Greek as a second language. This is shown, for example by his usage of 25 words (in 32 instances) found nowhere else in the New Testament. [Pryke, op. cit. pp. 28, 136-38]

Pryke identifies Mark’s characteristic vocabulary as words that appear uniquely in his Gospel, as well as those used at least five times in the Gospel, with at least half of the uses being redactional. For the purpose of synoptic comparison, we naturally can make no use of the uniquely Marcan terms. More promising, however, are the frequently used, predominantly redactional terms. On the assumption that Matthew used Mark as a source, we may expect to find distinctively Marcan vocabulary occasionally copied.

We lay down the following criteria for selecting Marcan redactional words (from the list in Pryke, pp. 136-38) for synoptic comparison:

  1. Some other common equivalent is available in Greek, so that word choice is stylistic rather than demanded by content.
  2. Frequent usage (10 or more times), for a decent sample size.
  3. High percentage of redactional use (at least 70% of total), so Mark clearly favors these terms more strongly than do common Gospel sources.
  4. Not a highly common word.
  5. Not a proper noun practically necessitated by Gospel content.
  6. Not a syntactic feature already discussed in last section.

After παλιν and ευθυς, the most frequently used redactional term in Mark is ακουειν, “to hear.” This fails our criterion against highly common words, and in general Mark uses it only slightly more frequently than Matthew. There are 47 total uses in Mark versus 71 in Matthew, a ratio of .662. This is close to the ratio of the lengths of the Gospels in words: 11,292/18,345 = .616.

If we restrict our scope to two third person indicative inflections, ακουειν and ακουει, we at least find some difference in usage, so we might test for verbatim copying. Synoptic parallels are noted in brackets.

Uses of ακουειν
Mt 24:6 [Mk. 13:7, Lk. 21:9 “when you shall hear” ακουσητε]
Mk 4:9 [Lk], 4:23, 4:33, (7:16), 7:37
Lk. 5:1, 5:15, 8:8[Mk], 14:35, 15:1, 21:38, 23:8
(Also 3x John, 3x Acts, 1x Romans, 1x Revelation)

Uses of ακουει
Mt. 7:24, Lk 10:16(x2)

In only one case (Mk. 4:9, Lk. 8:8) is a use of either of these two inflections replicated by another Synoptic Gospel. This lack of overlap suggests that usage was driven more by content or a shared source than by reliance on Mark as a textual source.

Forms of the verb ανιστεμι, “rise up,” appear in these instances. They are “redactional” except where marked (S) for “source.&rdquo.

Mt. 9:9 [triple trad., Mk. 2:14], 12:41 [Lk], 22:24 [triple trad., diff. inflection, Mk. 12:23(S)], 26:62 [Mk 14:57]
Mk 1:35, 2:14, 3:26(S), 5:42, 7:24, 8:31, 9:9, 9:10(S), 9:27(S), 9:31, 10:1, 10:34, 12:23(S), 12:25(S), 14:57, 14:60(S), [16:9(S)]
Lk. 27 times

Here is another example of how, paradoxically, Luke is closer to Mark in verbal form (relative to the amount of shared material), while Matthew has more shared content with Mark. Still, here at least Matthew shares two redactional instances of the term. One of them, the call of Levi/Matthew (Mk. 2:14, Mt. 9:9), belongs to a clearly textual shared tradition. The other, from the trial before the Sanhedrin (Mk 14:57, Mt 26:62), has only oral similarity. Here the usage proves no literary dependence, since any account of Jesus’ testimony would have to mention the Son of Man “rising up.”

Usage of αρχεσθαι, “to begin,” seems genuinely stylistic in many cases, as when saying, “he began speaking,” rather than “he spoke.” Uses in Mark and Matthew (noting Marcan parallels in the latter) are:

Mk. 1:45, 2:23, 4:1, 5:17, 5:20, 6:2, 6:7, 6:34, 6:55, 8:11, 8:31, 8:32, 10:28, 10:32, 10:41, 10:47, 11:15(S), 12:1, 13:5, 14:19, 14:33, 14:65, 14:69, 14:71, 15:8, 15:18; 25 redactional (R), 1 source (S)[4]
Mt. 4:17, 11:7, 11:20, 12:1 [Mk. 2:23], 14:30, 16:21 [Mk. 8:31], 16:22 [Mk. 8:32], 18:24, 20:8, 24:49, 26:22 [Mk. 14:19], 26:37 [Mk. 14:33], 26:74 [Mk. 14:71]; 13 uses

Mark clearly has a greater stylistic preference for the term than Matthew. The latter has only seven uses not paralleled in Mark, two of which are in pericopes with Marcan parallels (Mt. 4:17, 14:30). On the assumption that Matthew used Mark as his source, he retained six Marcan uses, suppressed 19 uses, and added two of his own. In non-Marcan material, he used the verb five times, at least one of which, Mt. 11:7, is a purely stylistic choice. Among the six uses shared with Mark, four of these, Mt. 16:21,22, 26:22,37, seem to be purely stylistic choices. The other two parallels (Mt. 12:1, Mk. 2:23; Mt. 26:74, Mk. 14:71) are not overtly stylistic, being determined by narrative content.

Looking at the six parallels closely, we find: (1) Mt. 12:1 says the disciples “began to pluck ears,” while Mk. 2:23 says they “began, as they went, plucking ears.” Matthew uses the infinitive while Mark uses a participle. (2) In the second prediction of the Passion, Mt. 16:21 and Mk. 8:31 use different infinitives: “Jesus began to show” and “began to teach,” the latter word being a Marcan preference. (3) Mt. 16:22 and Mk. 8:32 have the same phrase “began to rebuke” verbatim. (4) In the Passion narrative, Mk. 14:19 has “began to be sorrowful and to say,” whereas Mt. 26:22 reads: “And they were exceedingly sorrowful, and began every one of them to say…” (5) Matthew says Jesus began to be sorrowful and heavy, while Mark says he began to fear and be heavy. (6) Matthew and Mark agree that Peter began to curse and to swear, but this use is hardly idiosyncratic, being dictated by content, and the rest of the wording is dissimilar.

Four of the six Matthaean parallels involve stylistic choice, and only one of these has the verb phrase replicated exactly. Mt. 4:17, 11:7, 14:30 show that Matthew need not be following Mark in order to make this choice in usage. The stylistic parallels are sufficiently infrequent to be consistent with coincidence in most cases. In at least one instance (Mk. 16:22, Mk. 8:32), however, the similarity is likely due to shared written tradition or literary dependence.

The verb διδασχειν, “teach,” is favored by Mark, who uses it 17 times, as contrasted with only 14 in the much longer Gospel of Matthew. Parallels are marked with an asterisk.

Mk. 1:21,22*, 2:13, 4:1,2, 6:2*,6,30,34, 7:7(S)*, 8:31, 9:31, 10:1, 11:7(S), 12:14(S)*, 12:35, 14:49*; 14 R, 3 S[5]
Mt. 4:23, 5:2,19(x2), 7:29*, 9:35, 11:1, 13:54*, 15:9*, 21:23, 22:16*, 26:55*, 28:15,20

Matthew uses the same verb in two out of three Marcan source uses, and in only three of 14 Marcan redactional uses. The first two of the latter (Mk. 1:22, Mt. 7:29; Mk. 6:2, Mt. 13:54) are practically required by the subject matter (teaching in synagogue), and the third (Mk. 14:49, Mt. 13:54) is in direct speech by Jesus, so it could be source. In short, we find little evidence of Matthew copying Marcan style, even in some pericopes with strong verbatim similarity (Mk. 2:13-17, 4:1-9, 6:32-44, 8:31-33, 10:1). This suggests that their literary relationship involves having common sources, rather than direct dependence on each other.

δωδεκα, “twelve,” is another redactional term prevalent in Mark (15 redactional uses out of 17 total in Pryke). Unfortunately, this preference is not unique to Mark, and the term is found frequently throughout the New Testament, due to its significance in Jewish religion and culture.

A more promising term is εξερχεσθαι (or εξερχομαι), i.e., “ come forth/out.” The uses are:

Mk 1:25(S),26,28,29,35,38,45, 2:12,13, 3:6*,21, 4:3*, 5:2,8,13(S)*,14(S)*,30, 6:1,10(S)*,12,24(S),34(S)*,54, 7:29(S),30(S),31, 8:11,27, 9:25[24](S),26[25](S)*,29[28](S),30[29], 11:11*,12, 14:16(S),26,48,68(S), 16:8(S)*; 25 R, 14 S (Pryke: 25 R, 13 S)
Mt 2:6, 5:26, 8:28,32*,34*, 9:26,31,32, 10:11*,14, 11:7,8,9, 12:14*,43,44, 13:1,3*,49, 14:14*, 15:18,19,21,22, 17:18[17]*, 18:28, 20:1,3,5,6, 21:17*, 22:10, 24:1,26,27, 25:1,6,30,55,71,75, 27:32,53, 28:8*; 44 uses
[Vulg. in brackets]

Nine of 39 Marcan uses are found in Matthew. Again, we follow the classification of redactional and source material used by Pryke, though our understanding of oral tradition suggests that many of the “redactional” uses probably belong to source (e.g., Mk. 4:3, 5:8,30). Nevertheless, even on the form critics’ assumptions, only 3 of 25 (12%) Marcan redactional uses are in Matthew, versus 6 of 14 (43%) source uses.

Two especially clear instances of stylistic usage are Mk. 6:34 and 8:27. Matthew replicates only the first, which is in source, but not the second, redactional use. We do not count the pseudo-parallel of Mk. 5:2 and Mt. 8:28, since the verb has different subjects and objects: Jesus coming out of the boat, and demoniacs coming forth from sepulchres.

Mk. 4:3 opens the parable of the sower, “Behold the sower went out to sow,” so this was likely to be recalled in oral source. At any rate, the verb choice is indicated by content, not a stylistic insertion. Likewise, the usage in Mk. 11:11 (Mt. 21:17), “went out of the city into Bethania,” is demanded by the content. This pericope has only an oral degree of similarity, making copying especially unlikely. The only likely redactional replication is in Mk. 3:6 (Mt. 12:14), “And the Pharisees going out,“ which differ by a conjunction: και εξελθοντες ‘οι Φαρισαιοι in Mark; εξελθοντες δε ‘οι Φαρισαιοι in Matthew.

“Being called” is an important concept in the Gospel of Mark, as shown by fairly frequent use of the term προσκαλεισθαι “to call to oneself, to summon.” Mark uses the indicative προσκαλεισται only twice (Mk. 3:13, 6:7), and this is found nowhere else in the NT. More commonly, he and other NT authors use the participle προσκαλεσαμενος, “calling to himself.” These uses are:

Mt. 10:1, 15:10 [Mk. 7:14], 15:32 [Mk. 8:1], 18:2, 18:32, 20:25 [Mk. 10:42]; 6 uses
Mk. 3:23, 7:14 [Mt. 15:10] 8:1 [Mt. 15:32], 8:34, 10:42 [Mt. 20:25], 12:43, 15:44(S); 7 uses
Lk. 7:19, 15:26, 16:5; 3 uses
Acts 13:7, 23:17,18,23; 4 uses

The only other place in the NT where some form of the verb is used is James 5:14. Clearly this term is peculiar to the Synopticists, and especially favored by Mark. Oddly, we find that for once Matthew replicates a Marcan redactional term and Luke does not. In all three instances (Mk. 7:14, 8:1, 10:42), Luke does not have a parallel to the Marcan verse at all. Without Lucan parallels, we cannot use the argument that Mark is the common source of Matthew and Luke. Yet the double tradition agreements are exactly symmetrical: each has three of his six redactional uses in common with the other. This leaves no basis for establishing priority.

The verb φοβεισθαι, “to be afraid” is favored by Mark, notably in the peculiar ending at 16:8. The uses are:

Mk. 4:41, 5:15,33,36(S), 6:20,50(S)*, 9:32, 10:32, 11:18,32*, 12:12*, 16:8; 10 R, 2 S[6]
Mt. 1:20, 2:22, 10:26,28,31, 14:5,27*,30, 17:6,7, 21:26*,46*, 25:25, 27:54, 28:5,10; 16 uses

Two uses are especially indicative of an editorial choice: Mk. 9:32 “were afraid to ask him,”, and Mk. 10:32, the third prediction of the Passion. Unfortunately, neither of these is replicated in Matthew.

Mark uses the term more frequently in proportion to the length of his work. On the other hand, this only to be expected, since the additional sections of Matthew focus on the preaching of Christ, who generally did not mention fear except to counsel against it. Even the exhortation to fear God is made only in contrast to the counsel: “fear not those who destroy the body.” (Mt. 10:28)

We exclude the doubtful parallel between Mt. 6:20, saying Herod feared John and Mt. 14:5, which says he feared the multitude. Both the object and inflection of the verb differ.

This leaves only two possible redactional parallels. The use in Mk. 11:32 (Mt. 21:26) is in the words of Jesus, so this could be source. Lastly is Mk. 12:12 (Mt. 21:47), which says those who sought to lay hands on Jesus feared the people.

The evangelists had a variety of Messianic titles from which to choose, so Mark’s frequent use of ‘υιος του ανθρωπου, “Son of Man” might reflect a stylistic choice. This title, however, abounds in all four Gospels.

Mt. 8:20, 9:6*, 10:23, 11:19, 12:8*,32,40, 13:37,41, 16:13,27,28, 17:9*,12*,22*, 19:28, 20:18*,28* 24:27,30(x2)*,37,39,44, 25:31, 26:2,24(x2)**,45*,64*; 30 uses
Mk. 2:10(S)*,28*, 8:31,38(S), 9:9*,12*,31*, 10:33*,45(S)*, 13:26(S)*, 14:21(x2)**,41*,62(S)*; 14 uses; 9 R, 5 S[7]
Lk. 5:24, 6:5,22, 7:34, 9:22,26,44,58, 11:30, 12:8,10,40, 17:22,24,26,30, 18:8,31, 19:10, 21:27,36, 22:22,48,69, 24:7; 25 uses
Jn. 1:51, 3:13,14, 5:27, 6:27, 6:53,62, 8:28, 9:35, 12:23, 12:34(x2), 13:31; 13 uses

Even accounting for differences in length, this expression is actually more frequent in Matthew than in Mark. It is likewise abundant in Luke, and makes a fair showing even in the higher theology of the Gospel of St. John. The most probable source of this common use is none other than Jesus. That ‘Son of Man’ was Christ’s favorite title is consistent with what we know of his humility, his solidarity with mankind, and his reticence about revealing the fullness of his mission too soon (shown by repeated admonitions, often ignored, not to spread news of his miracles).

Matthew and Mark share eight “redactional” uses. (1) Mk. 2:28 (Mt. 12:8) “the Son of Man is lord of the Sabbath;” (2) Mk. 9:9 [Vulg. 9:8] (Mt. 17:9) and (3) In Mk 9:12 [9:11] (Mt. 17:12), from the Transfiguration; (4) Mk. 9:31 [9:30] (Mt. 17:22) from the second prediction of the Passion; (5) Mk. 10:33 (Mt. 20:18) from the third prediction of the Passion. (6-7) Mk. 14:21 (Mt. 26:24), Christ predicts his betrayal; (8) Mk. 14:41 (Mt. 26:45), Christ says the Son of Man is delivered to sinners. All of these instances are in the direct speech of Christ, except Mk. 9:9, which attributes this by indirect speech. They could easily all belong to source.

Indeed, even the non-paralleled “redactional” uses by Mark are in the direct (10:45) or indirect (8:31) speech of Christ, giving us further reason to doubt whether any of this usage is truly characteristic of Mark, rather than the shared Gospel tradition. The prevalence of this commonality even with the Gospel of John, and in pericopes with only oral degrees of verbal similarity, suggests that the Synoptic evangelists all had some sources in common.

As a further counter-example against copying from Mark, we note that Matthew does not copy the phrase “Son of Man” from Mark 8:31, even though this pericope generally shows strong verbal similarity. Matthew would have no reason to suppress this title, as he favors it even more than Mark. This suggests that Mark’s use may be redactional in this instance, while both Gospels rely on a shared written tradition.

Reviewing all the findings regarding vocabulary usage, the evidence for Matthew’s dependence on Mark is consistently weak or non-existent. Considering how much overlap these Gospels have in content and even in verbatim similarity, the negative evidence for use of Marcan redactional vocabulary is staggering. Not once do we find a dead giveaway of Matthew copying from Mark, as if he went to some extraordinary lengths to cover his tracks. Yet how would he be able to do such a thing, unless he had modern critical abilities to distinguish redactional material from source? If he had independent access to the source material, what need would he have of Mark, except perhaps to order the material?

Overall, a close examination of Marcan syntax and vocabulary shows that it is unlikely, or at least unnecessary, that Matthew should have relied on Mark as a source, though we might allow at this point that it influenced how he ordered his material. We might further salvage Marcan priority by supposing that Matthew at least depended on Mark for the Passion narrative, which certainly has strong verbatim similarity, including narrative or editorial material. In the absence of evidence of stylistic copying, however, we cannot know if this similarity is to be explained by attributing the original written Passion narrative to Mark, Matthew, or a shared source.

Regardless of whether we accept a more limited use of Mark (i.e. for the Passion narrative or for ordering material), we can scarcely avoid the impression that the main body of the Gospel of Matthew was written in ignorance of Mark’s work. We do not pretend to refute or disprove the possibility of some restricted literary dependence on Mark, but that is not our objective. We have only to show that such a hypothesis is unnecessary, and that the internal evidence in its favor is far too weak too overturn the external evidence against it.

Throughout our syntactic analysis, we have assumed the distinction between redactional and source material identified by form critics and compiled by Pryke. In the process, however, we have seen that the boundary between redaction and source is often difficult to discern, and perhaps undefinable. This is precisely what we should expect on the assumption that the Gospels were composed by predominantly oral modes of composition, using oral and written source traditions. Once this is admitted, the hypothesis of direct literary dependence is no longer necessary to explain the similarities and dissimilarities between Matthew and Mark.

[Top of page]

Continue to Part VI

Footnotes to Part V

[1] E.J. Pryke. Redactional Style in the Marcan Gospel (Cambridge: Cambridge Univ. Press, 1978).

[2] Lois K. Fuller. “The ‘Genitive Absolute” in New Testament/Hellenistic Greek: A Proposal for Clearer Understanding.” Journal of Greco-Roman Christianity and Judaism 3 (2006) 142-67.

[3] Ibid., p. 145.

[4] Pryke, op. cit., has 26 redactional and 27 total, erroneously counting two uses in Mk. 5:18-20, pp. 136, 141.

[5] Pryke, op. cit., p. 135 has 15 redactional out of 17 total, but only 14 in fact are in the redactional verses listed in pp. 139-48.

[6] Pryke, op. cit., p. 138 has 8 redactional out of 17 total, but 10 in fact are in the redactional verses listed in pp. 139-48.

[7] Pryke, op. cit., p. 137 has 8 redactional out of 15 total, but this undercounts the double instance in 14:21 (pp. 147, 172), so his adjusted count is 9 R, 6 S.

© 2015-16 Daniel J. Castellano. All rights reserved.