The Gospel of Matthew

| 11 →

Introduction

This commentary greatly differs from other modern commentaries on the Gospel of Matthew. The difference results from the particular methodological approach which has been adopted therein. Instead of explaining the Matthean Gospel in historical-critical terms as a result of redactional use of earlier sources or traditions, in narratological terms as a set of narrative-organizing devices, etc., this commentary aims at explaining it in a critical intertextual way as a result of a sequential hypertextual reworking of the Acts of the Apostles.

This methodological approach, unlike many others, does not originate from any particular literary theory. It rather reflects the recent discovery of the phenomenon of the sequential hypertextual reworking of earlier texts in numerous biblical writings. This phenomenon occurs in the writings of both the Old and the New Testament: Gen, Exod-Lev-Num, Deut, Sam-Kgs, Chr; Mt, Mk, Lk, Jn, Acts, Rom, Gal, Eph, 2 Thes, Hebr, 2 Pet, and Rev.¹ These writings, taken together and measured by their extent, constitute almost a half of the Christian Bible.

Accordingly, it is fully justified to perform a thorough analysis of the Matthean Gospel, taking this important literary discovery into consideration.

Matthew and Mark

Modern scholars generally agree that the Gospel of Matthew is a reworking of the Gospel of Mark. This opinion concerning the relationship between these two ← 11 | 12 → Gospels constitutes an important element of probably the earliest solution to the synoptic problem. It was first witnessed in the so-called ‘testimony of Papias’ (Eusebius, Hist. eccl. 3.39.15-16).

According to this patristic text, the Gospels of Mark and Matthew were based on oral traditions which were handed down by Peter and presumably also by other apostles. For this reason, the Gospel of Mark was for centuries regarded as a work which was based on Peter’s oral catecheses. Likewise, the Gospel of Matthew was by most ancient, medieval, and modern scholars regarded as preserving authentic sayings of Jesus, which were first recorded in Hebrew (or Aramaic), possibly in the form of a protogospel2 or a later lost sayings source,³ and thereafter translated into Greek.⁴

However, a close analysis of the composition of the so-called ‘testimony of Papias’ reveals that this text was not primarily concerned with the sources of the material which is contained in the Gospels of Mark and Matthew. Otherwise, it would refer to the origin of all four canonical Gospels, and not just two of them. In fact, the bipartite structure of this patristic text reveals that it was only aimed at explaining the differences between the Gospels of Mark and Matthew, as well as the striking features of the Matthean Gospel.

The author of the so-called ‘testimony of Papias’ rightly perceived the Gospel of Matthew as having two apparently contradictory features. On the one hand, this Gospel seems to be a result of literary enhancement and rhetorical improvement of the relatively short and simple Gospel of Mark. On the other hand, in difference to the Marcan Gospel, the Gospel of Matthew, with its particular wording and theological stance, seems to be a very ‘Hebrew’, so apparently primitive, Jewish Christian Gospel. Consequently, it is quite natural to ask whether the Matthean Gospel should be regarded as written after or before the Gospel of Mark. ← 12 | 13 →

The so-called ‘testimony of Papias’ presents an early Christian attempt to answer this difficult literary-theological question. According to this patristic text, the Marcan Gospel originated from a set of Peter’s oral catecheses and therefore, as the ‘testimony of Papias’ repeatedly stresses, it was not well organized in terms of a carefully composed literary work (οὐ… τάξει, οὐχ ὥσπερ σύνταξιν: Eusebius, Hist. eccl. 3.39.15).5 The patristic text further suggests that as a consequence of this fact, ‘so then (μὲν οὖν) Matthew arranged the [Lord’s] oracles […] in an orderly way’ (συνετάξατο: Eusebius, Hist. eccl. 3.39.16).⁶ Accordingly, the so-called ‘testimony of Papias’, through its correlated references to the Gospels of Mark and Matthew, explained the evident posteriority of the apparently ‘Hebrew’ Gospel of Matthew against the apparently ‘Gentile’ Gospel of Mark in terms of necessary literary improvement of the allegedly poorly organized Marcan Gospel.

In order to lend credence to this thesis, the author of the ‘testimony of Papias’ argued that the things which could be rearranged in the Marcan Gospel, without compromising the truth of them, were the Lord’s and Peter’s allegedly isolated oracles or discourses (λόγια: Eusebius, Hist. eccl. 3.39.15-16). However, the suggestion that the Gospels of Mark and Matthew had their origin in some orally transmitted discourses or oracles of the Lord evidently functioned in the ‘testimony of Papias’ as a secondary, in fact merely postulated element of the main argument concerning the literary-rhetorical reasons for rearranging the contents of the Gospel of Mark into the better organized Gospel of Matthew.⁷

In fact, the pattern of the relocations and modifications of the Marcan (and Lucan) material in the Gospel of Matthew is very complex. Hitherto given scholarly explanations of this fact in terms of, for example, Matthew’s composition of three sections with three miracle stories and words of Jesus in each of them in Mt 8–11⁸ or the use of two overlapping Marcan sequences in Mt 8–9⁹ are only ← 13 | 14 → partly satisfactory because they do not explain numerous other relocations and modifications of the Marcan (and Lucan) material in the Matthean Gospel (e.g. Mk 5:24 in Mt 8:1; Mk 10:46-52 in Mt 9:27-31; Mk 6:34 in Mt 9:36; Mk 13:9 in Mt 10:17; Mk 3:22.25 in Mt 10:25; Mk 9:41 in Mt 10:42; Mk 9:22 in Mt 17:15; Mk 11:22-23 in Mt 17:20; Mk 10:15 in Mt 18:3; Mk 12:34 in Mt 22:46; Mk 9:34-35 and Mk 10:43 in Mt 23:11). Consequently, the complex pattern of the Matthean relocations and modifications of the Marcan (and Lucan) material remains an oddity,10 unless the likewise complex pattern of the Matthean reworking of the Lucan bipartite work is taken into due consideration.

Matthew and Luke

In the twentieth century, the problem of the relationship between the Gospels of Matthew and Luke did not draw much attention in biblical scholarship. This fact mainly resulted from the popularity of the so-called Two-Source hypothesis, which assumed mutual literary independence of the Gospels of Matthew and Luke, in favour of the existence of a hypothetical ‘Q source’.¹¹ However, in the twenty-first century this widespread consensus concerning the relationship between the Gospels of Matthew and Luke is more and more seriously challenged. Scholars began to realize that the arguments against the so-called Two-Source hypothesis, especially those pointing to hundreds of Matthean-Lucan agreements against Mark, are in fact really strong.¹² Therefore, nowadays the so-called Two-Source hypothesis gradually loses the status of a widely agreed-upon scholarly axiom, and comes to be treated as merely one of the synoptic hypotheses.¹³

The studies on the issue of the relationship between the Gospels of Matthew and Luke, which had been published until the year 2009, were already analysed in my previous book on the synoptic problem.¹⁴ Therefore, only most recent contributions to this issue will here be presented. ← 14 | 15 →

Mark-Matthew-Luke

The hypothesis of the Mk-Mt-Lk order of literary dependence of the Synoptic Gospels has a relatively strong group of supporters, who were more or less directly influenced by the work of Michael D. Goulder.

Werner Kahl has argued that the so-called ‘minor agreements’ of Matthew and Luke against Mark, which are commonly regarded as an argument against the ‘Q source’ theory, are in fact not ‘minor’ because they are numerous and significant.¹⁵ Moreover, according to the German scholar, they often appear in clusters (especially in Mk 1:1-13 parr. and in Mk 14:32-16:8 parr.), so that their presence in Matthew and Luke should not be attributed to mere chance.¹⁶

Kahl’s argument for the Lucan dependence on Matthew, and not vice versa, is based on the philological observation that the Greek style of Luke is better than that of Matthew.¹⁷ However, his general reference to alleged Lucan corrections of Matthew in Mk 11:7 par. Mt 21:7 par. Lk 19:35¹⁸ is hardly convincing. Likewise, Kahl’s argument that Luke often locates his additions to Mark in a different context (especially in the ‘Travel Narrative’), whereas Matthew locates them in their Marcan contexts,¹⁹ cannot solve the problem of the direction of literary dependence, given the evidently differing treatment of the Marcan material by both later evangelists. On the other hand, Kahl’s argument that the Matthean introduction to the Sermon on the Mount (Mt 5:1-2), unlike the Lucan Sermon on the Plain, is not consistent with its context (Mt 4:18-22; 7:28-29)²⁰ is quite convincing, but this inconsistency can also be explained as resulting from Matthew’s reworking of the Acts of the Apostles.

Francis Watson has argued that the hypothesis of the Lucan dependence on Matthew, without the existence of the hypothetical ‘Q source’, should be called the ‘L/M (= Luke/Matthew) theory’.²¹ However, in his rejection of Q, the British scholar merely mentions the possibility of the reverse Matthean-Lukan dependence, namely the Matthean dependence on Luke, using the argument that, as he ← 15 | 16 → states, ‘[…] it is impossible to investigate every possibility at once’.22 As concerns the ‘primitivity’ and ‘secondariness’ between Matthew and Luke, Watson rightly argues that ‘each of the competing [synoptic] hypotheses can produce plausible examples of apparent “primitivity” or “secondariness” ’, so that the synoptic problem should be solved by ‘attending to the compositional procedure required’ for the postulated later Gospel to emerge from the earlier ones.²³

Watson rightly notes the thematic and lexical similarities between the Lucan prologue and the ‘testimony of Papias’, but he again uncritically accepts only one direction of possible dependence between them: from Papias’ view to Luke’s view.²⁴ Watson also suggests that since Papias and Irenaeus mention both Mark and Matthew, and Irenaeus adds to them Luke and John, the Gospels of Mark and Matthew were probably the earliest ones.²⁵

The British scholar notes numerous notable similarities between the Gospels of Matthew and Luke, which extend beyond the material traditionally attributed to the ‘Q source’: genealogies, annunciation stories, the introduction of John the Baptist before the Isaiah citation, the brood-of-vipers speech, etc.²⁶ However, even if Watson rejects the hypothesis of mutual independence of the Gospels of Matthew and Luke, he does not reject the hypothesis of mutual independence of the Gospels and the letters of Paul,²⁷ although both hypotheses are logically constructed in the same way: differences between the analysed texts are taken to prove their mutual independence, notwithstanding the existence of important similarities between them.

Watson also tries to explain the Matthean relocations of the Marcan material in the section Mt 3–13 by suggesting that Matthew used the material of Mk 1:1-6:6 in an ordered way, namely in two overlapping sequences (Mk 1:1-4:34; 4:35-6:6).²⁸ However, this analysis is rather superficial. It does not take into consideration the Matthean relocations of the material of Mk 1:39; 3:7-8 etc. to the narrative point after Mk 1:20 (in Mt 4:22-25), Mk 3:13 to the narrative point ← 16 | 17 → after Mk 1:20 (in Mt 5:1),29 Mk 10:46-52 to the narrative point after Mk 2:22 (in Mt 9:27-31), Mk 3:22.25 to the narrative point after Mk 2:22 (in Mt 10:25), etc.

John C. Poirier, in the introduction to the recent multi-author monograph on ‘Marcan priority without Q’,³⁰ clarifies its main title. He notes that although the label ‘Marcan priority without Q’ is welcomed by many supporters of the ‘Farrer hypothesis’ (Mk-Mt-Lk order without Q), it is quite ambiguous because it can also be applied to the opposite hypothesis, namely that of the Mk-Lk-Mt order of direct literary dependence, also without the hypothetical ‘Q source’ (the so-called ‘Matthean posteriority hypothesis’). Moreover, Poirier explains the factors which nowadays encourage more numerous scholars to adopt the Farrer hypothesis (and in fact, at least partly, also the Matthean posteriority hypothesis), rather than the Two-Source hypothesis: (a) increased awareness that the evangelists were creative writers and reshapers of earlier material, rather than strict compilers; (b) wider acceptance of the view that Luke wrote in response to other Gospels; (c) increased awareness of Luke’s literary ability; and (d) wider acceptance of a late date for Luke.³¹

In his article contained in this monograph, John C. Poirier rightly criticizes Delbert Burkett’s recent support of the Two-Source hypothesis by pointing to his outdated understanding of the evangelists as slavish copyists and compilers of earlier sources, and not as creative authors and composers of literary works.³² Poirier’s critique of Burkett mainly refers to the contested plausibility of various aspects of Luke’s reworking of Matthew. Such issues are notoriously difficult to solve in a convincing way. For example, the argument of the supporters of the Farrer hypothesis concerning Luke’s postulated ‘editorial fatigue’ in his reworking of Matthew in fact refers to the difference between Luke’s greater variatio locutionis and Matthew’s greater uniformity in style and vocabulary. Therefore, it can endlessly be debated whether it was Luke who introduced the variation into Matthew’s text or it was Matthew who uniformed the style of the Lucan material. However, the text-critical rule lectio difficilior potior faciliori generally favours Matthew’s corrective uniformization of the original Lucan variation. ← 17 | 18 →

Eric Eve analyses the Beelzebul controversy in the Gospels from the point of view of the Farrer hypothesis. He argues that Mt 12:22-37 can be regarded as an expansion of Mk 3:20-30, and Lk 11:14-23 as a reworking of Mt 12:22-37. Accordingly, he argues that the hypothetical ‘Q source’ can here be omitted.33 However, he offers no detailed arguments for Luke’s use of Matthew, and not vice versa. For example, the agreement in order between Lk 11:14-28 and Mt 12:22-50 is simply taken as an argument for Luke’s dependence on Matthew. The British scholar does not consider the reverse possibility of the understanding of Lk 11:27-28 as a Lucan original composition, and not as a reworking of Mt 12:46-50. In fact, a similar Lucan original composition, evidently not suggested by the text of Matthew, can be found in Lk 23:29 (μακαρια* + ἡ κοιλία ἡ + καὶ μαστοὶ οἵ).

Stephen C. Carlson analyses the non-aversion principle, which was used by some defenders of the Two-Source hypothesis to make the judgments concerning redactional traits of a given evangelist non-reversible. According to this principle, a given feature can only be identified as truly secondary if the other evangelist in a compared Gospel had no aversion to it; otherwise, the other evangelist could also have been secondary in his omission of this feature for his particular reasons. Carlson shows that the use of this apparently logical principle in Lk 20:47-21:4 par. Mk 12:40-44 in fact leads to erroneous results because Luke often changed Marcan vocabulary even if he was elsewhere not averse to it.³⁴ Therefore, it can be argued that the use of this principle cannot lead to conclusive (non-reversible) results because the evangelists could freely change the wording of earlier Gospels for some barely identifiable reasons. Consequently, merely linguistic considerations cannot conclusively solve the synoptic problem, especially if they are applied to isolated fragments of the Gospels.

Heather M. Gorman assesses the plausibility of the Farrer hypothesis against the background of ancient rhetorical tradition, as it is witnessed in the extant progymnasmata and rhetorical handbooks. In her opinion, the overall order of the Lucan Gospel, which includes, as she argues, the section mainly concerning Jesus’ deeds (Lk 4:14-9:50) and the section mainly concerning Jesus’ teaching (Lk 9:51-19:28), suits Quintilian’s advice that an encomium could include such sections.³⁵ One might debate whether Quintilian’s phrase operum (id est ← 18 | 19 → factorum dictorumque) contextus (Inst. 3.7.15 [sic]) in fact refers to such large sections and whether the Lucan Gospel really contains such two major parts, and consequently whether Luke’s Gospel is really well arranged (cf. Lk 1:3) in terms of ancient rhetoric. Likewise, Gorman’s argument that Luke considerably shortened and rearranged Matthew’s Sermon on the Mount for the sake of rhetorical brevity and clarity is not very persuasive, given Luke’s predilection for quite elaborate speeches elsewhere in his Gospel and Acts.

Mark Goodacre rightly argues that the instances of very high Mt-Lk verbal agreement in their ‘double tradition’, reaching uninterrupted strings of 24–27 identical words in the same order in Lk 3:8-9; 7:8-9; 10:21-22; 11:32; 16:13 par., in fact disprove the Two-Source hypothesis and favour the hypothesis of Mt-Lk direct literary dependence.36 However, his argument that the higher Mt-Lk verbal agreement in their ‘double tradition’ than in the ‘triple tradition’ is best explained by the Farrer hypothesis is, alas, unconvincing. The fact that, as Goodacre rightly notes, the highest verbal agreement can mainly be found in Mt-Lk and Mt-Mk pairs, and much more rarely in Lk-Mk pairs, in reality favours the Matthean posteriority hypothesis, according to which Matthew consistently quite faithfully copied from both Mk and Lk, and not the Farrer hypothesis, according to which Luke surprisingly oscillated between a relatively free mode of literary reworking (Lk-Mk) and a relatively faithful one (Lk-Mt). Such an oscillating pattern of Lucan literary reworking of earlier texts (e.g. Paul’s letters) is notably absent in Acts.

Ken Olson, similarly to Stephen C. Carlson, argues that Luke could have omitted the unparalleled Matthean expressions in the Lord’s Prayer because they repeat ideas which are expressed earlier in the text, and Luke generally avoided repetitions.³⁷ However, Olson fails to explain the fact that the idea of subjection to God’s will (Mt 6:10b) was evidently important for Luke at crucial points of his narrative (Lk 22:42; Acts 21:14), so its omission in the model prayer of the disciples (Lk 11:2-4) would be really surprising.

Andris Abakuks applies several models of statistical analysis to Matthew’s and Luke’s use of Mark. The use of a simple chi-square test, time series modelling using logistic regression, as well as using hidden Markov models, commonly reveal that in the so-called ‘triple tradition’ both the Farrer hypothesis and the Matthean posteriority hypothesis are more plausible than Matthew’s and Luke’s independent use of Mark, with no particular clue as to the superiority of the ← 19 | 20 → Farrer hypothesis over the Matthean posteriority hypothesis or vice versa. The use of hidden Markov models also suggests that Matthew’s or Luke’s rather loose, so maybe somehow correlated reworking of Mark can most likely be found in Mk 1:40-44; 2:8-12; 3:28-33; 6:37-44; 12:36-38 parr.38 Abakuks’s analyses would be even more persuasive if he used the NA²⁸ and not the NA²⁵ edition of the text of the Gospels.

Jeffrey Peterson notices some important Mt-Lk thematic and linguistic agreements against Mk in the conclusions of the Gospels (and in the birth stories). He also argues that they are best explained by the Farrer hypothesis and not by the Matthean posteriority hypothesis.³⁹ However, the arguments for the latter claim are rather weak. The Lucan repeated references to ‘the eleven’ (Lk 24:9.33; Acts 1:26; 2:14) are understandable after Judas’ betrayal (Lk 22:3.47-48), but in Matthew’s Gospel the phrase ‘the eleven’ appears only once, in a text which is thematically related to Luke (Mt 28:16 par. Lk 24:33.36). On the other hand, the allegedly Matthean verb προσκυνέω (Mt 28:17 par. Lk 24:52) was repeatedly used not only by Matthew, but also by Luke (3 times in his Gospel and 4 times in Acts). The scriptural justification of the mission to ‘all the nations’ is hardly more explicit in Luke (Lk 24:46-47) than in Matthew (Mt 28:18-19; cf. Dan 7:14 LXX), and in any case such a phenomenon cannot be taken as proving only one direction of reworking. Likewise, the Matthean idea of Jesus’ spiritual presence with his disciples (Mt 28:20; cf. 18:20) can be taken as a reworking of the Lucan idea of Jesus’ presence in the Spirit (Lk 24:49; Acts 1:4-5.8 etc.), and not vice versa. Notwithstanding these difficulties, Peterson’s idea that the relationship between Matthew and Acts should be analysed more carefully is certainly insightful.

David Landry argues for a late dating of Luke (c. AD 115–160) on the basis of Luke’s use of Mark, Matthew, John, Josephus, and Paul (with the Pastoral Epistles), and on the other hand on the basis of Marcion’s and Justin’s use of Luke.⁴⁰ Even if the direction of the dependence between Luke and John was in fact reverse, and Luke’s use of Matthew is something that Landry wants to prove, the other arguments are generally correct. On the other hand, Landry’s arguments for an earlier dating of Matthew (c. AD 80–90), which are intended to prove the Farrer ← 20 | 21 → hypothesis, are much weaker because they rely on the highly debatable dating of the letters of Ignatius of Antioch, which post-date Matthew, to c. AD 110.

In brief, the recent multi-author monograph concerning the Farrer hypothesis as opposed to the Two-Source hypothesis displays no significant interaction with the ‘mirror’ solution, namely the Matthean posteriority hypothesis. Moreover, as often happens in the discussion on the synoptic problem, the three Gospels are here generally only compared with one another, as though they were written in a literary vacuum. The Pauline and post-Pauline letters, the Acts of the Apostles, classical and Hellenistic literature, etc. are generally not taken into consideration as potential hypotexts for the Gospels. Therefore, much work in this field has still to be done.

Mark-Luke-Matthew

One of the advocates of the Two-Source hypothesis has interestingly noticed that ‘Matthew’s dependence on Luke has rarely been proposed […] Given the number of passages where the advocates of Q have suggested that Luke’s version is more original, this is perhaps surprising’.⁴¹ However, several scholars have recently argued for some variant of the hypothesis of the Lk-Mt order of literary dependence between these Gospels.

Paul N. Tarazi has argued that Matthew borrowed numerous fragments from the Gospel of Luke in order to illustrate various Pauline ideas. Thus, he closed the canon of the New Testament writings.⁴²

James R. Edwards in his work on the Hebrew proto-Gospel has argued that because the Lucan special material apparently contains more Semitisms than does the Gospel of Matthew, then the latter Gospel should be considered posterior to the Lucan one.⁴³ Moreover, according to the American scholar, in comparison to the Lucan Gospel the order and formulas of the Gospel of Matthew are more balanced and proportional, its Greek style is more clean and consistent, and its christology is more developed.⁴⁴

David L. Mealand has carried out a stylometric analysis of various fragments of the Matthean Gospel and has come to the conclusion that ‘M samples were ← 21 | 22 → distinct from those attributed to Mark and Q, and the latter from each other’.45 This result can be used against the hypothesis of the Mk-Mt-Lk order of direct literary dependence (in which both Q and M fragments are attributed to Matthew), but it favours not only the Two-Source hypothesis,⁴⁶ but also the hypothesis of the Mk-Lk-Mt order of direct literary dependence (in which Q is attributed to Luke, so it can be stylistically distinct from both Mark and M).

Robert K. MacEwen has recently published a monograph devoted to the ‘Matthean posteriority hypothesis’, which postulates Matthew’s use of both Mark and Luke (who in his turn had also used Mark) as a solution to the synoptic problem.⁴⁷

In the presentation of his own arguments in favour of the Matthean posteriority hypothesis, MacEwen convincingly argues that the Matthean posteriority hypothesis best explains the presence of the Aramaic word μαμωνᾶς once in Mt 6:24 and 3 times in Lk 16:9.11.13, which is surrounded by Lucan Sondergut (Lk 15:8-16:12; 16:14-15). Luke generally avoided Aramaisms, so Luke’s combination of sources (Q 16:13 with much Lucan Sondergut on the Two-Source hypothesis) or the reverse direction of borrowing (on the Farrer hypothesis) would be here rather implausible.⁴⁸ Likewise, MacEwen convincingly argues that the word εἰρήνη in Mt 10:13 par. is generally Lucan and non-Matthean, which favours the Matthean posteriority hypothesis.⁴⁹ Similarly, he persuasively argues that the Matthean saying concerning care for one sheep on the Sabbath (Mt 12:11), which is inserted into the Marcan story Mk 3:1-6, was borrowed from Lk 14:5, which is an integral part of the story Lk 14:1-6, and consequently the reverse direction of borrowing (on the Farrer hypothesis) or the existence of an isolated saying Q 14:5 (on the Two-Source hypothesis) would be highly implausible.⁵⁰

Describing the results of the statistical analysis of strings of verbal agreement which contain four or more words in the Synoptic Gospels, MacEwen rightly argues that on the Matthean posteriority hypothesis Matthew was rather consistent in his reworking of both Mark and Luke with roughly the same, high level of ← 22 | 23 → verbal agreement. However, on other synoptic hypotheses, the latest evangelist (or evangelists in the Two-Source hypothesis) would display surprisingly variegated level of agreement with his (or their) two main sources.51