Mark’s Gospel—Prior or Posterior? A Reappraisal of thePhenomenon of Order – by ArchBishop Uwe AE. Rosenkranz

Mark’s Gospel—
Prior or Posterior?

A Reappraisal of the
Phenomenon of Order

David J. Neville

Journal for the Study of the New Testament
Supplement Series 222


Copyright © 2002 Sheffield Academic Press

A Continuum imprint

Published by

Sheffield Academic Press Ltd

The Tower Building, 11 York Road, London SE1 7NX

370 Lexington Avenue, New York, NY 10017-6550

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage or retrieval system, without permission in writing from the publishers.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN 1-84127-265-5


Preface and Acknowledgments



Part I
Preliminary Issues for Analysing
the Phenomenon of Order

Chapter 1
Pericope Divisions and the Question of Parallels

Pericope Divisions

Paralleling Pericopes

Chapter 2
Towards an Arrangement of Parallel Pericopes

Describing the Phenomenon of Order

Issues Involved in Arranging Parallel Pericopes

Matthew-Mark Disagreements in Order on Three Different Pericope Arrangements

Different Pericope Arrangements in Relation to the Two-Gospel Hypothesis


Annotated Table of Parallel Pericopes

Part II
The Value of Formal Arguments from Order

Chapter 3
Two Formal Arguments from Order

The Refutation of Streeter’s Formal Argument from Order

The Neo-Griesbachian Objection to Streeter’s Inference from Order

Re-evaluating the Neo-Griesbachian Formal Argument from Order

Chapter 4
The Phenomenon of Correlation

Farmer’s and Riley’s Attempts to Verify the Phenomenon of Correlation

A Statistical Test of the Phenomenon of Correlation

Conclusion and Preview

Part III
Compositional Arguments from Order: A Reappraisal

Chapter 5
Compositional Conventions in the First Century ce

The ‘Oral Environment’ of the First Century ce

Writing Equipment and Compositional Habits in Antiquity

Compositional Conventions

Chapter 6
The Disagreements in Order in Luke 3:1–5:11

Luke 3:19–20, Herod’s Imprisonment of John the Baptist

Luke 4:16–30, Jesus at Nazareth

Luke 5:1–11, The Call of the First Disciples


Chapter 7
The Disagreements in Order in Matthew 4:23–9:35

Matthew 4:23–5:2, Matthew’s Introduction to the Sermon on the Mount

Excursus on Matthew 8:1–9:34

Matthew 8:1–4, Jesus Heals a Leper

Matthew 8:18–34, Jesus Calms a Storm and Exorcizes Two Demons

Matthew 9:18–26, Jesus Heals a Woman and Resuscitates a Girl


Chapter 8
Is Mark’s Gospel Posterior?
The Evidence of Pericope Order

Recent History of the Two-Gospel Hypothesis: An Overview

Mark 1:21–22, Mark’s Initial Transition from Matthew to Luke

Mark 3:7–12, Mark’s Second Major Transition

Mark 6:1–13, Mark’s Alternating Agreement with Matthew and Luke



Appendix 1
The Segmentation of the Synoptic Gospels

Appendix 2
Statistical Tests of the Phenomenon of Correlation


Index of Authors

Preface and Acknowledgments

This study is a sequel to my Arguments from Order in Synoptic Source Criticism: A History and Critique, and was conceived as part 2 of a two-part work. The work as a whole grew out of a project initiated by Richard K. Moore to work through a Greek synopsis using a colour-coding and underlining method designed (by Moore) to display not only similarity in wording and expression, but exact morphological agreement and difference in inflexion. This method also enabled those involved in the project to depict transpositions of words, phrases, sentences and even subsections within pericopes.2 On the basis of this research, we compiled a statistical record of various categories of verbal agreement and disagreement between the gospels.

Whereas my Arguments from Order in Synoptic Source Criticism surveyed ways in which scholars have appealed to the phenomenon of order (the pattern of agreement and disagreement in the order and arrangement of pericopes) in source-critical arguments, this study evaluates the significance of the phenomenon of order for resolving the question of the literary relations between the Gospels of Matthew, Mark and Luke, with special reference to the theories of Markan priority and Markan posteriority. Arguments from order, both formal and compositional, have been especially prominent in defence of both Johann Jakob Griesbach’s theory of Mark’s posteriority to Matthew’s and Luke’s Gospels and the theory of Markan priority first advocated by Gottlob Christian Storr. Moreover, a strongly defended alternative to the still-dominant Markan hypothesis is the two-gospel hypothesis, which is a renovation of Griesbach’s source theory. As a result, this study focuses on these two hypotheses. Arguments from order are certainly important for other source theories, for example, the Farrer—Goulder hypothesis that Matthew used Mark’s Gospel while Luke used both Mark’s and Matthew’s Gospels.5 In Chapter 6 I attend to Michael D. Goulder’s arguments in relation to Lk. 3:19–20, 4:16–30 and 5:1–11, but to have given more attention to the significance of the phenomenon of order for determining the relation between the Gospels of Matthew and Luke would have made an already lengthy study longer still.

After an introduction that (a) outlines various reasons for the continuing value of gospel source criticism, (b) surveys some alternative theories of synoptic relations and (c) reviews some recent appraisals of the significance of the phenomenon of order, Part I examines three methodological issues relevant to any analysis of the phenomenon of order: the delimitation of pericopes, the question of what constitutes parallelism between pericopes and the arrangement of synoptic parallels.

Part II: critiques formal arguments based on the phenomenon of order. Accepting that the argument for Markan priority based on the relative absence of agreement in order between Matthew and Luke against Mark is inconclusive, I demonstrate that the argument for Markan posteriority to Matthew and Luke based on the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke is also inconclusive. I also discuss the ‘phenomenon of correlation’ between agreement in order and agreement in wording, which some regard as evidence of Mark’s dependence on Matthew and Luke. Having devised a statistical test of the phenomenon of correlation, I argue that the results are inconclusive. In short, I argue that any formal argument from order is inconclusive.

Regarding my mathematical calculations in Part II, especially my statistical test of the phenomenon of correlation in Chapter 4, I acknowledge my limitations as a mathematician. My discussion lacks mathematical sophistication because it fails to demonstrate that the results are significant in a statistical sense, but perhaps someone with better mathematical credentials will be prompted to conduct a sound statistical analysis of my results or to devise a better test.

Part III: examines compositional arguments from order, that is, arguments that aim to provide plausible reasons for disagreements in order by comparing the relevant pericopes and parallels in their respective contexts and making source- and redaction-critical judgments about which synoptist is responsible for such disagreements. To avoid anachronistic judgments about how the gospels were written, Chapter 5 examines what can be known about compositional conventions in the first century of the Common Era (ce). The final three chapters examine select pericopes from the Gospels of Luke, Matthew and Mark, respectively, to ascertain whether the theory of Markan priority or the theory of Markan posteriority best accounts for the phenomenon of order. A concluding chapter sums up the results.

Since I completed the research for this book, a number of books relevant to various aspects of my work have appeared. None has led me to alter my conclusions, but some are particularly noteworthy. In connection with Chapter 5, see Harry Y. Gamble, Books and Readers in the Early Church: A History of Early Christian Texts (New Haven and London: Yale University Press, 1995), and Loveday Alexander, ‘Ancient Book Production and the Circulation of the Gospels’, in Richard Bauckham (ed.), The Gospels for All Christians: Rethinking the Gospel Audiences (Grand Rapids, MI: Eerdmans, 1998), pp. 71–111. With respect to Chapter 6, see Michael Prior, Jesus the Liberator: Nazareth Liberation Theology (Luke 4:16–30) (Sheffield: Sheffield Academic Press, 1995), especially Chapter 2 on the source question. Advocates of the two-gospel hypothesis have published Beyond the Q Impasse—Luke’s Use of Matthew, edited by Allan J. McNicol, with David L. Dungan and David B. Peabody (Valley Forge, PA: Trinity Press International, 1996). According to the research team of the International Institute for the Renewal of Gospel Studies, Beyond the Q Impasse—Luke’s Use of Matthew supersedes its series of reports on the ‘Narrative Outline of the Composition of Luke according to the Two Gospel Hypothesis’ in the Society of Biblical Literature Seminar Papers between 1992 and 1995. (This same research team intends to publish a redactional analysis of the Gospel of Mark based on the two-gospel hypothesis.) With respect to Chapter 7, readers should note Evert-Jan Vledder, Conflict in the Miracle Stories: A Socio-Exegetical Study of Matthew 8 and 9 (JSNTSup, 152; Sheffield: Sheffield Academic Press, 1997), which interprets the miracle stories in Matthew 8–9 through the sociological lens of conflict theory. In connection with Chapter 8, see David L. Dungan, A History of the Synoptic Problem: The Canon, the Text, the Composition, and the Interpretation of the Gospels (New York: Doubleday, 1999). Finally, see Peter M. Head, Christology and the Synoptic Problem: An Argument for Markan Priority (SNTSMS, 94; Cambridge: Cambridge University Press, 1997), which like this book not only evaluates an aspect of the synoptic problem with reference to the Markan hypothesis and the Griesbach (or two-gospel) hypothesis but also utilizes a comparative redaction-critical method. Head comes to firmer conclusions than I, hence the subtitle of his book, but his work is a good example of the kind of approach I advocated in my earlier book, namely, that in source-critical argumentation one should compare at least two possible explanations, assuming different source theories, and assess the relative probabilities of each.

It remains for me to acknowledge the support, guidance and encouragement of Richard Moore, who supervised my doctoral research. I also thank Ashley Lucas and Jessica Spratt of the Department of Modern Languages at Trinity Grammar School in Sydney for checking my translations of French and German materials.


AB    Anchor Bible

ABD    David Noel Freedman (ed.), The Anchor Bible Dictionary (New York: Doubleday, 1992)

ANRW    Hildegard Temporini and Wolfgang Haase (eds.), Aufstieg und Niedergang der römischen Welt: Geschichte und Kultur Roms in Spiegel der neueren Forschung (Berlin: W. de Gruyter, 1972–)

ATR    Anglican Theological Review

BAGD    Walter Bauer, William F. Arndt, F. William Gingrich and Frederick W. Danker, A Greek—English Lexicon of the New Testament and Other Early Christian Literature (Chicago: University of Chicago Press, 2nd edn, 1979)

BECNT    Baker Exegetical Commentary on the New Testament

BETL    Bibliotheca ephemeridum theologicarum lovaniensium

Bib    Biblica

BibRes    Biblical Research

BJS    Brown Judaic Studies

BSac    Bibliotheca Sacra

BTB    Biblical Theology Bulletin

CBQ    Catholic Biblical Quarterly

CBQMS    Catholic Biblical Quarterly, Monograph Series

CTM    Concordia Theological Monthly

CTQ    Concordia Theological Quarterly

CurTM    Currents in Theology and Mission

DR    Downside Review

EKKNT    Evangelisch-Katholischer Kommentar zum Neuen Testament

ETL    Ephemerides theologicae lovanienses

ExpTim    Expository Times

GH    Griesbach Hypothesis

HTKNT    Herders theologischer Kommentar zum Neuen Testament

HTR    Harvard Theological Review

ICC    International Critical Commentary

IDB    George Arthur Buttrick (ed.), The Interpreter’s Dictionary of the Bible (4 vols.; Nashville: Abingdon Press, 1962)

Int    Interpretation

JBL    Journal of Biblical Literature

JETS    Journal of the Evangelical Theological Society

JHS    Journal of Hellenic Studies

JSNT    Journal for the Study of the New Testament

JSNTSup    Journal for the Study of the New Testament, Supplement Series

JTS    Journal of Theological Studies

N-A26    Nestle, Aland et al. (eds.), Novum Testamentum Graece (26th edn)

NGS    New Gospel Studies

NovT    Novum Testamentum

NovTSup    Novum Testamentum, Supplements

NTG    New Testament Guides

NTS    New Testament Studies

P(ST)J    Perkins (School of Theology) Journal

RB    Revue biblique

RSB    Religious Studies Bulletin

RTR    Reformed Theological Review

SBLDS    SBL Dissertation Series

SBLMS    SBL Monograph Series

ScrB    Scripture Bulletin

SNTSMS    Society for New Testament Studies Monograph Series

SQE    Synopsis Quattuor Evangeliorum

TD    Theology Digest

TSK    Theologische Studien und Kritiken

TZ    Theologische Zeitschrift

USQR    Union Seminary Quarterly Review

WBC    Word Biblical Commentary

WUNT    Wissenschaftliche Untersuchungen zum Neuen Testament

ZNW    Zeitschrift für die neutestamentliche Wissenschaft

ZTK    Zeitschrift für Theologie und Kirche

2DH    Two-Document Hypothesis

2GH    Two-Gospel Hypothesis

2ST    Two-Source Theory


The Value of Gospel Source Criticism

Why do questions about the origins and interrelations of the gospels continue to demand attention? One often reads assertions about the relevance of gospel source criticism for other issues such as text criticism, exegesis, early church history, the quests of and for the historical Jesus, and New Testament Christology and theology. For example, as Adela Yarbro Collins has noted, ‘Most questions of interpretation are affected by the interpreter’s presupposition regarding which is the earliest Gospel’. Such claims are not always explained, however, and some readers probably fail to appreciate the full import of such statements.

First, source criticism is one way to become better acquainted with the gospels and, in doing so, to learn more about early Christian convictions concerning Jesus and his significance. In the closing paragraph of his paper for the Pittsburgh Festival on the Gospels (1970), Albert Outler, the Methodist church historian, alluded to the intrinsic value of gospel criticism:

I have spoken of this Festival as a celebration of the survival of interest in the Gospels after eighteen centuries, an act of recognition of the perennial relevance of Jesus for man’s hopes—’this Jesus, whom we crucified and whom God made to be both Kyrios and Soter‘ (cf. Acts 2:36). Is it visionary to hope that, before the week is out—or at least in later retrospect—this Festival may prove to have been the sign of an actual revival of that interest: the renewal of the ancient Christian conviction that in these four little tracts … there really is an unquenched power to illuminate the human mystery and to fortify man’s aspirations? For it is in the Gospels, as nowhere else, that the story of Jesus is told and it is in this Jesus, as in no one else, that men may discover ‘the power of God and the wisdom of God’ (1 Cor. 1:24). It is, therefore, to these four Gospels—as prisms that catch and refract the light of God’s self-witness everywhere—that we must return, ever and again, to learn each for himself of God’s reconciling love, incarnate in Jesus Christ and deathless in the world through his ever-present Spirit.

Anyone who shares Outler’s concern for what he described as a ‘metahistorical’ presupposition with respect to the gospels, namely, that sense of ‘the cruciality for modern man of four documents that tell us the story of Jesus—God’s revelation of his special love for man, his clue to man’s highest and best hopes’, needs no other reason or motivation for engaging in gospel source criticism.

The point is worth reiterating. Source criticism is important in its own right, but its by-product—greater familiarity with the gospels—is arguably more important still. As William O. Walker, Jr, once remarked,

It is important to solve the Synoptic Problem if it can be solved, but even if a solution to this particular problem remains elusive, past and continuing explorations of the relationships among the gospels will be well worth the time and effort devoted to them, for the ‘spin-off’ effects of such explorations are that the more we investigate gospel relationships, the more we learn about the gospels, and this is all to the good.

The importance of source criticism for textual criticism was disclosed by Bruce M. Metzger in his Textual Commentary on the Greek New Testament. He pointed out that when judging between variant readings in the manuscript tradition of the gospels, the editorial committee of the United Bible Societies’ Greek New Testament used the theory of Markan priority as one of its text-critical criteria.6 How significant this is in practice is difficult to determine, but it takes little genius to perceive that the uncritical application of such a criterion is capable of corrupting the texts of the gospels in significant ways, thereby undermining the source-critical enterprise. The circularity of arguing for Markan priority on the basis of texts constructed (at least partly) on the basis of Markan priority is obvious. No wonder, then, that source critics have been advised not to ignore textual variants. As J.K. Elliott has warned, ‘A reader who works with a synopsis and ignores the textual variations does so at his peril. The synoptic problem and textual criticism are inextricably linked’.8 To ascertain the influence of text critics’ source-critical presuppositions, David L. Dungan has even suggested utilizing various source theories to establish different critical texts of the gospels.

The relevance of source criticism for our understanding of the historical Jesus and New Testament Christology was underscored by Dieter Lührmann in ‘The Gospel of Mark and the Sayings Collection Q’. Lührmann acknowledged that the two-document hypothesis was one of the weapons devised by German critics in the nineteenth century to combat the views of David Friedrich Strauss. According to Lührmann, ‘This becomes very clear from the introduction that Heinrich Julius Holtzmann wrote to his 1863 publication on the Synoptic Gospels, in which the two-document hypothesis was presented in a comprehensive, definitive way’.11 Lührmann’s overview of research on the origins and interrelations of the gospels between 1863 and 1924 illustrates how critical the synoptic problem is for our understanding of Jesus: ‘So ever since Holtzmann, and largely in opposition to Strauss, to speak about “The Gospel of Mark and the Sayings Collection Q” has in fact meant nothing less than to raise the question of Christology’.

In this connection, it is noteworthy that for the Jesus Seminar, which was convened in 1985 so that members could work collaboratively towards a new understanding of the historical Jesus, the priority of Mark and Q constitute two of its ‘seven pillars of scholarly wisdom’ for ‘liberating’ the real Jesus from the theological portraits of Jesus in the four canonical gospels. Among other revealing statements, the following is particularly important:

The basic solution to the synoptic puzzle plays a fundamental role in historical evaluations made by members of the Jesus Seminar and other scholars. Mark is now understood to be the fundamental source for narrative information about Jesus. The priority of Mark has become a cornerstone of the modern scholarship of the gospels.

In a discussion of ‘The Theological Stakes in the Synoptic Problem’, John S. Kloppenborg noted that ‘the ultimate interest of source criticism is, presumably, to achieve an adequate textual basis for understanding (a) the literary and theological achievements of each gospel, (b) the place of each gospel in the history of primitive Christianity, and, eventually, (c) the history of primitive Christianity itself’. He recognized how influential a source theory can be both for reconstructing the theology (or theologies) of the gospel writers and for the general acceptance of any account of Christian origins, including developments in early Christian thought about Christology, soteriology and ecclesiology. After first evaluating Mark’s historical and theological emphases based on the assumption of Markan posteriority, then making some observations on the usefulness of theological tendencies for determining priority and posteriority in the gospel tradition, Kloppenborg identified some ‘theological stakes’ in the synoptic problem: ‘One, for example, has to do with whether reflection on the role of the Torah was at the heart of the first attempt to codify the Jesus traditions. On the GH, clearly it was. On the 2DH, by contrast, Jesus’ posture towards the Torah is not a programmatic concern either for Mark or for Q’.

However, the primary theological dispute caused by recent developments in synoptic source criticism has to do with the centrality or otherwise of Jesus’ death interpreted as having redemptive significance. As Kloppenborg observed,

In the history of the synoptic debate, the key theological objection to the 2DH has to do not with Mark, but with Q and its lack of a salvific interpretation of Jesus’ death. On the GH, the passion of Jesus and its saving effects are present from the very start. On the 2DH, at least one of the principal documents of formative Christianity did not, apparently, feel it necessary to locate Jesus’ saving significance there.17

In an earlier study, William R. Farmer had already noted that the absence of any ‘salvific interpretation of Jesus’ death’ is a significant theological implication of the two-source theory. In ‘The Church’s Stake in the Question of “Q” ‘, he responded to James M. Robinson’s remark that ‘even if it has survived only incompletely, Q is surely the most important Christian text we have’, a claim Robinson considered incontestable ‘in spite of the recognition of the ongoing debate about the Synoptic Problem, and with all due respect to the canonical books of the New Testament’.19 According to Farmer, Robinson’s views about Q are based largely on speculation, with a minimum of literary and historical support. He also found Robinson’s speculations about the Christology of Q anomalous, especially because Robinson conceded that members of the so-called ‘Q-community’ were familiar with the kerygma that Jesus ‘died for them’. In that case, Farmer asked, ‘Where in the history of the Church, or the Jesus movement, does this anomalous christology belong?’ He also asked whether this historical anomaly does not undermine the validity of the Q-hypothesis itself: ‘Must we not grant that reasonable conclusions which follow from the premise of “Q” are important for judging the validity of that premise? Is it not a meaningful test of the “Q” hypothesis that it produces results that the historian finds anomalous?’

In a later study, Farmer raised the question of the canonical relevance of Robinson’s predilection for Q:

James Robinson has organized a group of scholars to decide upon a text for the hypothetical document Q. The motive is to fix in publishable form an authorized text of what Robinson terms ‘the most important Christian document [sic] we have’. Once again we have the question of canon being raised: for, once Q has been reconstructed and translated some will insist that copies of Q should be put in our churches. For Q, assuming its existence, gets us back behind our canonical Gospels, closer to Jesus, the font of Christian authority. This reconstructed Q, assuming there ever was such a document, makes no reference to the redemptive consequences of the death of Jesus Christ. On the assumption that there once was such a document in the apostolic church, Robinson is rendering an important theological service in fixing its text. On the other hand, if such a document never existed (certainly a decided possibility in the minds of most scholars, and a strong probability in the minds of others), in ‘reconstructing the unreconstructible’ Robinson is leading scholars on what would appear to be a wild goose chase of enormous and far-reaching canonical consequences for all Church leaders (and their number is legion) who continue to believe in Q.

Farmer’s negative appraisal of Robinson’s efforts to reconstruct Q is counterbalanced by Arland D. Jacobson and others in pursuit of the wording, order and theology of Q. Despite disagreement between Farmer and Jacobson on the value of Q-research, they largely concur on why so much energy is currently devoted to Q. Jacobson no doubt spoke for many when he wrote:

Why enter this wilderness called Q? Because many of us suspect that deep in its heart may lie the secret of that fateful time when Judaism gave painful birth to Christianity. Q was, as far as we can tell, the earliest gospel, if by ‘gospel’ we mean a theologically-shaped presentation of Jesus tradition. And then there is Jesus of Nazareth. To the believer he is, as exalted Lord, close at hand; but as a historical figure he remains as mysterious and puzzling as ever. Rightly or wrongly, Western civilization has long believed that the nature of a thing can best be understood from its origins. Therefore, if Q is the first gospel, then it may hold clues both about the origins of Christianity and about Christianity’s central figure, Jesus.

Clearly, gospel source criticism has significant historical and theological ramifications, particularly (but not only) in relation to Q, which has been granted a new lease on life in recent decades.

Alternative Theories of Synoptic Relations

In the scholarly literature the two-source theory remains the regnant hypothesis. However, alongside the two-gospel hypothesis, there are various alternative theories of synoptic relations competing for the allegiance of critics prepared to commit themselves. Among them are the theory of Markan priority without Q, theories involving modifications to the Markan hypothesis by recourse to Ur-Markus or Deutero-Markus, the multi-stage theory, the theory of Lukan priority and the oral tradition hypothesis.

The ‘Farrer Model’

Numerous critics hold to the theory of Markan priority, yet have problems with the theory of Q as it is generally conceived. Arthur J. Bellinzoni, Jr, named this position the ‘Farrer model’, after Austin M. Farrer.26 An overview of this theory, providing useful historical context, was provided by Edward C. Hobbs in ‘A Quarter-Century Without “Q” ‘. Hobbs mentioned various British defenders of Farrer’s theory, including John Drury28 and Michael D. Goulder. More than anyone else, Goulder has shouldered the mantle of demonstrating Q’s dispensability. For many, however, his reliance on the concept of midrash to explain Matthew’s rewriting of Mark’s Gospel and Luke’s rewriting of Mark’s and Matthew’s Gospels makes the composition of the secondary gospels less credible than their dependence on a hypothetical Q. Although unconvinced that one need not posit sources for the sayings material in the gospels, E.P. Sanders and Margaret Davies nevertheless accept Goulder’s theory that Luke borrowed from both Mark and Matthew.30 Given the scholarly energy expended on Q-research in recent decades, their critical comments on Q are of considerable interest:

Historically most scholars have been conscious that ‘Q’ is a scholarly convention which explains the Matthew-Luke double tradition, and they have deliberately remained vague about whether or not it was one document, a loose assemblage of passages, or simply a convenient name for oral or ‘floating’ traditions. For many decades the effort to reconstruct Q, like the effort to reconstruct Proto-Mark, was abandoned. Now a few scholars are again attempting to define Q as a document: it really existed, it directly reflects the theology of a community, and one can even make a concordance of it. This work is mostly of curiosity value, since it shows how far a hypothesis can be pushed despite its lack of fundamental support.

Modifications to the Markan Hypothesis

While those who defend the ‘Fairer model’ of synoptic relations have difficulty with the Q-element of the two-source theory, others have difficulty with the priority of Mark. This has not caused them to abandon the Markan hypothesis but to modify it by appealing to earlier or later recensions or editions of Mark’s Gospel than appears in the New Testament. Theories involving earlier or later recensions of Mark’s Gospel—Ur-Markus or Deutero-Markus—are proposed by those who recognize that certain features in Mark’s Gospel are difficult to explain on the theory of Markan priority. An Ur-Markus theory contends that Matthew and Luke used a version of Mark’s Gospel earlier than canonical Mark, whereas a Deutero-Markus theory maintains that Matthew and Luke borrowed from a later edition (or editions) of Mark’s Gospel than canonical Mark.

Helmut Koester defended an Ur-Markus hypothesis at the Colloquy on New Testament Studies (1980). In his ‘History and Development of Mark’s Gospel (From Mark to Secret Mark and “Canonical” Mark)’, Koester identified evidence that to his mind suggested a number of different recensions of Mark’s Gospel before it reached its canonical form. Koester presupposed Markan priority, but his starting point was that ‘the text of the Gospel of Mark, as it is preserved in all ancient manuscripts, cannot have been identical with the text used by Matthew and Luke’. Influenced by non-canonical materials, especially the recently-discovered fragments of a Secret Gospel of Mark mentioned by Clement of Alexandria, Koester surmised that Mark’s Gospel went through the following editions: a preliminary Proto-Mark, used by Luke; an expanded Proto-Mark, used by Matthew; a thorough revision of Proto-Mark, now known as the Gospel of Matthew; a different revision of Proto-Mark, now known as the Gospel of Luke; a conservative revision of Proto-Mark that resulted in the Secret Gospel of Mark; a different edition of Secret Mark used by the Carpocratian sect, mentioned by Clement of Alexandria; a revised version of Secret Mark, now known as the canonical Gospel of Mark; and finally, later revisions of canonical Mark influenced by the Gospels of Matthew and Luke as illustrated by Mk 16:9–20. In response to Koester, David B. Peabody remarked, ‘Ur-Marcus is a valuable hypothesis for an advocate of Markan priority because one can make such appeals to the text of the lost Ur-Marcus to explain those places within the triple tradition where Matthew and Luke agree against canonical Mark, as Koester does here’.35

In his more recent discussion of the history and development of Ancient Christian Gospels, Koester proposed a similar reconstruction of the history of the Gospel of Mark: the earliest version, used by Luke; an expanded version, amplified by the ‘Bethsaida section’ (Mk 6:45–8:26), used by Matthew; a further edition containing redactional features that are not paralleled in the Gospels of Matthew and Luke and ‘closely related’ to the Secret Gospel of Mark; canonical Mark; and a later expansion of canonical Mark.

The principal advocate of Deutero-Markus is Albert Fuchs. For Fuchs, the verbal agreements between Matthew and Luke against Mark in the triple tradition are the major stumbling block to the priority of canonical Mark. Perhaps Frans Neirynck has summarized his position best:

The minor agreements are … the central theme of the synoptic studies of Albert Fuchs. In his view, to cope with the evidence of the minor agreements is only possible by postulating a Deutero-Markus, not simply a recension of Mark … but a real second redaction of the gospel including the rewriting of the text and the insertion of new material. The combination of Q with Mark is the work of Deutero-Mark, and some portions of Q (traditionally assigned to Q) are in fact expansions to the text of Mark by the deutero-Markan redactor.

Multi-Stage Theories

Most critics recognize that certain data in the first three gospels are not susceptible of satisfactory explanation and that any source theory must be content to explain most of, not all, the data. For example, those who hold to the Markan hypothesis know that some verbal agreements between Matthew and Luke against Mark are anomalous on their theory. Referring to these ‘minor agreements’ of Matthew and Luke against Mark, Neirynck has stated, ‘Some of these difficulties are easily answered, others are not and constitute the “unexplained remainder” which Markan priorists seem to tolerate without irritation’.

However, a growing number of critics refuses to accept simple theories to explain complex phenomena and instead advocates a more complex picture of gospel origins. This group includes Léon Vaganay, Xavier Léon-Dufour, Antonio Gaboury, Rainer Riesner and Philippe Rolland, but its most articulate spokesperson is Marie-Émile Boismard.

Boismard first presented his source theory in 1966. In 1972, he and Pierre Benoit published the second volume of their collaborative work, Synopse des quatre évangiles en français avec parallèles des apocryphes
et des Pères. Written largely by Boismard, volume II of this synopsis is a source-critical commentary on every pericope in the synoptic gospels and includes a lengthy introduction in which he elaborated and defended his theory. Expositions and evaluations of Boismard’s complex theory generally focus on this important work.41

Boismard’s later study, ‘The Two-Source Theory at an Impasse’, illustrated his source-critical approach. Focusing on the summaries preceding the accounts of the multiplication of the loaves in Mk 6:31–34, Mt. 14:13–14 and Lk. 9:10b–11, he proposed that two primitive sources, A and B, represent the earliest stages of tradition. Mark’s Gospel incorporated both of these primitive sources. One of these sources, document A, was also used in producing a first edition of Matthew’s Gospel. Luke’s Gospel was based primarily on the first edition of Matthew’s Gospel with supplementary dependence on the Gospel of Mark. A second edition of Matthew’s Gospel was dependent on Mark’s Gospel.

Although other critics have also proposed a multi-stage hypothesis, there is no single multi-stage theory. In 1984, at an important conference in Jerusalem, Boismard opened his overview of the multi-stage theory by making this very point:

Contrairement aux deux groupes concurrents qui rassemblent …, le groupe C [multi-stage team] n’est pas homogène; il englobe des théories qui ne se rejoignent qu’en partie. Ces théories, certes, sont unies par un principe fondamental, mais qui peut être appliqué selon des modalités différentes. Il vaudrait done mieux parler de ‘théories (au pluriel) des Niveaux Multiples’.

This absence of agreement between advocates of a multi-stage theory on how best to explain the complex relations between the gospels is likely to be the Achilles heel of this ‘family of theories’.

Despite differing on specifics, advocates of a multi-stage theory concur on one point: the relations between the canonical gospels are indirect rather than direct; that is, they are indirectly related by virtue of their dependence on common, non-extant sources. As Boismard explained,

Les diverses théories des Niveaux Multiples ont en commun le principe fondamental suivant, par lequel elles se distinguent de la théorie des Deux Évangiles et de la théorie des Deux Sources: même lorsqu’il s’agit des matériaux communs aux trois Synoptiques, les rapports entre ces évangiles doivent s’expliquer, non par dépendance directe, mais en faisant appel à des sources hypothétiques plus anciennes dont ils dépendent.

He contrasted this view with the more simple two-source and two-gospel hypotheses, in which points of contact between the Gospels of Matthew and Mark are explained by direct dependence one way or the other. However, in his view,

les contacts litteraires entre Mt. et Me devraient s’expliquer … par dépendance envers une ou plusieurs sources hypothétiques qui pourraient être, soit un proto-Matthieu, soit un proto-Marc, soit des documents plus anonymes. Le problème synoptique est complexe; il ne peut être résolu que par une solution complexe.

Boismard maintained that his complex solution to the synoptic problem reconciles to some extent the contradictory features of the two-source and two-gospel hypotheses. He then identified the main difficulties for the two-source theory: verbal agreements between Matthew and Luke against Mark; Mark’s duplicate expressions in triple-tradition material; Matthew’s and Luke’s shared omissions of Mark’s unique materials, for example, Mk 4:26–29, 7:32–34 and 8:22–26; and Luke’s omission of Mk 6:45–8:21. The two-gospel hypothesis fared no better under Boismard’s gaze. In his view, its primary difficulties (‘faiblesses congénitales’) are: the presence of ‘doublets’ in the Matthew-Luke double tradition on one hand and in any one of the three following sets of data on the other—the triple tradition, the Matthew-Mark double tradition and the Mark-Luke double tradition; the amount of material omitted by Mark if he borrowed from the Gospels of Matthew and Luke; and Luke’s failure to incorporate features of Matthew’s text that should have appealed to him, for example, Matthew’s tendency to highlight the miraculous in his account of the resurrection of Jesus. In short, according to Boismard, ‘La théorie des Deux Évangiles, en donnant la priorité absolue à l’évangile de Mt, ferme les yeux sur les trés nombreux cas où cet évangile apparaît secondaire par rapport aux deux autres Synoptiques’.

Finally, Boismard noted instances in which Matthew’s Gospel seems to have preserved the most primitive text and others where Matthew’s text appears secondary to its parallels in Mark’s and Luke’s Gospels. As a result, he concluded his discussion of the essential features of his multiple-stage hypothesis by saying, ‘Nos conclusions précédentes sont encore confirmées: pour expliquer les rapports entre les Synoptiques, il est nécessaire de faire appel à des documents plus anciens, sous peine de se heurter à des difficultés insurmontables’.

Lukan Priority

In the history of synoptic source criticism, few critics have advocated Lukan priority. Robert Lisle Lindsey is one who defended Luke’s priority, although his theory has affinities with Boismard’s multi-stage theory because he posited a number of earlier sources from which the synoptists borrowed. Although he originally proposed ‘A Modified Two-Document Theory of Synoptic Dependence and Interdependence’, his theory changed somewhat at least partly as a result of his translation of the Gospel of Mark into Hebrew. Further refinements were subsequently made, as may be seen from his introduction to A Comparative Greek Concordance of the Synoptic Gospels52 and his lecture series entitled The Jesus Sources: Understanding the Gospels. The following summary is based on these later presentations: (1) the earliest source was a Hebrew narrative; (2) this Hebrew text was then translated into Greek; (3) this Greek translation of the original Hebrew source was subsequently rearranged, with the result that the original order of the narrative was lost; (4) an effort was later made to reconstruct the original order by excerpting material from the rearranged narrative, the result being a shorter narrative described by Lindsey as the ‘first reconstruction’; (5) Luke used this ‘first reconstruction’ as a guide for organizing his Gospel, but also incorporated a great deal of material from its source (source 3 above), especially in his central section; (6) Luke must also have written Acts before the appearance of the Gospels of Mark and Matthew, because Lindsey postulated Markan dependence on Acts; (7) Mark used the rearranged source (source 3) but mainly borrowed from Luke in those sections in which Luke was most influenced by the ‘first reconstruction’ (source 4), with supplementary dependence on Acts; finally, (8) Matthew used Mark for his narrative outline, amplifying his Gospel with material from the rearranged source (source 3). In short, all three synoptists borrowed from the rearranged source (source 3); only the first synoptist, Luke, knew and used the ‘first reconstruction’ (source 4); Mark borrowed from Luke; Matthew borrowed from Mark, but was unfamiliar with Luke.54

Oral Tradition

In 1977, at the Colloquy on the Relationships among the Gospels, Albert Bates Lord, George Kennedy and Lou H. Silberman stressed the probable influence of oral tradition in the composition of the gospels. For Lord, an expert in the study of oral traditional literature, numerous features in the synoptic gospels suggest that relations between them are better explained when they are understood as ‘three interrelated oral versions of the material’ than if they are conceived as the result of literary dependence. By exploring, first, the presence within the gospels of what he described as ‘oral traditional narrative mythic patterns’ (for example, stories of miraculous births and of death and resurrection) and, second, parallel sequences of common episodes within the gospels, Lord’s analysis led him to conclude that the gospels are at least ‘closely related to oral traditional literature’.56

Lord’s emphasis on the probable role of oral tradition in the composition of the gospels was reinforced by both Kennedy, a classicist, and Silberman, a professor of Jewish Literature and Thought. As Silberman observed,

A century ago, scholars assumed unquestioningly that a literary work had its sources in literary works (for, after all, were not these scholars themselves ransacking literary works to fabricate new literary works?). And even now, when we have come to affirm that behind some or many of the literary works we deal with there is an oral tradition, we still manipulate such traditions as though they too were ‘literary’ works. We have not come to terms with a fecund world of ideas … We still march along the straight black line of the Gutenberg galaxy.

Alongside the importance and broad reliability of oral transmission, Kennedy emphasized that at the time when the gospels were written, note-taking was a common intermediate stage between oral transmission of material and the final composition of a work. He also stressed the importance of memory and the broad conception of ‘translation’ as understood in antiquity. In his view,

The experience of classicists seems to suggest that memory of oral teaching, especially if the teaching was heard repeatedly, could be retained with considerable integrity over an extended period of time, even though oral teaching was often converted into running notes by students and these notes were sometimes checked with the original speaker … After oral transmission and note-taking, a third stage would be the publication of a systematic or more literary work. The gospels are not themselves notes on preaching; they are systematic works, with the material appropriately organized. When a work was translated from one language into another, existing traditions in the second language often exercised influence on the form and style of the work, and considerable freedom of rearrangement or restatement was possible even if not inevitable.

Although Charles H. Talbert and Joseph A. Fitzmyer maintained that relations between the synoptic gospels are essentially, if not exclusively, literary, the role of oral tradition in the composition of the gospels has continued to receive attention. In 1989 and 1990, conferences in Dublin and Milan reaffirmed the influence of oral tradition on the composition of the gospels by examining issues related to the oral transmission of ‘Jesus-material’ (whether sayings of Jesus or stories about Jesus), not only in the period before but also subsequent to the writing of the gospels. In a review of the proceedings from these two conferences,60 J.K. Elliott made the sagacious recommendation that if—as the participants in these conferences maintained—oral transmission of Jesus-material continued well beyond the time of the composition of the gospels, ‘then it behoves investigators into the question of the synoptic problem to avoid too rigorous adherence to merely literary solutions’.

By contrast, in Studying the Synoptic Gospels, E.P. Sanders and Margaret Davies assert that ‘the relationship among the gospels is literary. It is the result of either direct copying from one to the other, or of common dependence on the same source or sources’. This was the shared judgment of the majority of critics for most of the last century. Although it is difficult to verify a synoptist’s use of oral tradition, a theory that allows for the continuing influence of oral tradition more accurately represents the historical situation than one that explains the relations between the gospels on strictly literary terms. Responding to the work of recent critics who have emphasized the role of oral transmission in the composition of the gospels, Elliott advised:

We should perhaps take the message from these and other publications that when the second and third Gospel writers put pen to papyrus they had at their disposal four types of material: a) stories and sayings ignored by or unknown to the first evangelist—some or most of these coming to them from the oral tradition; b) stories and sayings that happened to have been used by the first evangelist but which the later writers received from the oral tradition independently of the written account; c) stories and sayings that reached the later evangelists from the oral tradition that had developed from the first evangelist’s written account, as well as d) stories and sayings that were indeed taken directly from their predecessors’ written versions.

The view that the similarities and differences between the gospels are best explained by common reliance on oral tradition goes back to Johann Gottfried von Herder in the late 1790s and Johann Carl Ludwig Gieseler in the early 1800s. This view and its corollary, the relative independence of the gospels, were defended by Brooke Foss Westcott in his Introduction to the Study of the Gospels, which went through numerous reprintings in the latter half of the nineteenth century. For Westcott, the gospels were the result of ‘the successive remoulding of the oral Gospel according to the peculiar requirements of different classes of hearers’. Though influential for a time,66 his views were eclipsed as a result of the influence of William Sanday and others.

In recent decades, however, the oral tradition hypothesis has made something of a comeback. In 1978, the classicist John M. Rist accepted Luke’s dependence on Mark’s Gospel, but argued for the independence of Matthew and Mark. In 1986, Bo Reicke argued that the synoptic gospels are largely the result of independent efforts to present oral traditions in literary form. Although he allowed for personal contact between Mark and Luke in Caesarea between 58 and 60 ce, nevertheless ‘the relative parallelism between the Synoptic Gospels is fundamentally due to common traditions of the early church’. In 1991, John Wenham allowed for some interdependence between the synoptic gospels but maintained that oral tradition was responsible for most of their similarities and differences. He also defended the traditional Augustinian hypothesis that Matthew’s Gospel was written first, Mark’s second and Luke’s third, with each at least familiar with the work of his predecessor(s). In short, his is an ‘oral theory with some measure of successive dependence’.69

To acknowledge the important role of oral tradition in the composition of the gospels does not imply the absence of literary relations between them. Jacobson gave four reasons for thinking that the relations between the synoptics are primarily, even if not exclusively, literary, and Wenham offered three reasons for accepting at least ‘an important measure of interdependence’.71 They concurred on one point, namely, that the high level of correspondence in order of pericopes is difficult to explain solely by recourse to common use of oral traditions.

Recent Appraisals of the Significance of the Phenomenon of Order

The Symposium de Interrelatione Evangeliorum (Jerusalem, 1984) concluded with participants identifying four points on which they agreed and a much larger number of issues on which further research was required. The first issue on the published list of items requiring further research was ‘the phenomenon of order among pericopes’. Arguments based on the phenomenon of order have played a prominent role in the history of synoptic research, particularly in defence of the theory of Markan priority and of Griesbach’s theory of Markan posteriority.74

At the Jerusalem Symposium, both the two-source and two-gospel teams emphasized the importance of the phenomenon of order. In both his overview of ‘The Two-Source Hypothesis’ and his discussion of ‘Matthew 4:23–5:2 and the Matthean Composition of 4:23–11:1’, Frans Neirynck devoted much of his attention to the phenomenon of order. Indeed, his decision to concentrate on Mt. 4:23–11:1 was influenced primarily by his appreciation of the significance of the phenomenon of order: ‘The phenomenon of order, and particularly the relative order of Mt. 4:23–13:58/Mk 1:21–6:13, is a major issue in the discussion of the synoptic problem. For that reason, our analysis will concentrate on the crucial passage of Mt. 4:23–5:2 and the problem of dislocations in Mt. 4:23–11:1′. More recently, in his article on the ‘Synoptic Problem’ for The New Jerome Biblical Commentary, Neirynck claimed that ‘the argument from order, as understood since K. Lachmann (1835), constitutes the main reason for positing Marcan priority’.

In response to Neirynck’s paper at the Jerusalem Symposium, David Dungan disputed his procedure for analysing the phenomenon of order but agreed on its importance: ‘We agree with Neirynck that the argument from order of pericopes is the fundamental starting point’. In Farmer’s overview of the two-gospel hypothesis, he asserted, ‘The first step in proposing any solution to the Synoptic Problem is the recognition of the literary fact that Matthew, Mark and Luke all three agree significantly with one another to varying degrees in content and order of episodes’.80

Whereas Neirynck held that an argument from order was ‘the main reason for positing Markan priority’, another member of the two-gospel team, Allan J. McNicol, maintained that ‘the internal evidence of order of pericopes’ is one of ‘two foundational pillars’ upon which the two-gospel hypothesis rests. He reiterated this point in a response to Christopher M. Tuckett’s analysis and appraisal of arguments put forward by neo-Griesbachian critics:

The observation from order of pericopes is the foundational starting point for the Neo-Griesbachians. Their various studies on the linguistic and literary characteristics of the synoptic writers, as well as their more recent extensive studies on the provenance of the gospels in the development of early Christianity, all have their essential point of departure with this observation about the Synoptic Problem.

Given the significance attached to the phenomenon of order by advocates of both the two-source and two-gospel hypotheses, G.M. Styler’s comment is apposite: ‘Anyone brought up at the feet of Streeter will be surprised to discover that the principal, and strongest, argument in favour of Griesbach is the argument from order’. If the phenomenon of order could be shown to favour one of these theories over the other, it would represent considerable progress in the quest for a solution to the synoptic problem.

One particular difficulty, however, is confusion over different types of arguments based on the phenomenon of order. Neirynck and Tuckett have demonstrated that two quite different arguments from order have been used to promote Markan priority. Since Butler’s exposure of the so-called ‘Lachmann fallacy’, Markan priorists have dispensed with Streeter’s inference based on the absence of agreement in order between Matthew and Luke against Mark in the triple tradition. The other argument from order for Mark’s priority attempts to demonstrate that Matthew’s and Luke’s alleged transpositions of Markan pericopes are explicable and consistent with their redactional tendencies; in short, their transpositions of Markan materials cohere with what can be surmised about their redactional objectives on other grounds. Tuckett in particular has explained the importance of this ‘criterion of coherence’, and McNicol has endorsed its application.86

In response to McNicol’s critique of Tuckett, William O. Walker, Jr, stated that ‘on balance, the phenomenon of the order of pericopes appears to be more nearly compatible with the Griesbach Hypothesis than with the Two-Source Hypothesis’. However, during discussion at the colloquy at which McNicol and Walker presented their respective papers, he confessed,

to the extent that I have done the kind of careful textual analysis that both [McNicol] and Tuckett have done, most of the time … I have found … that the individual texts apparently point toward Markan priority, but not always. And so I’m in the very ambiguous situation that if I follow the criterion of coherence I seem to be pointed toward Markan priority but if I follow the phenomena of order I seem to be pointed in the other direction.

What Walker seems not to have noticed is that the criterion of coherence is applicable to the phenomenon of order in the same way that it is applicable to other literary data. Furthermore, the criterion of coherence and the phenomenon of order are not comparable ‘entities’. The phenomenon of order is a literary datum, whereas the criterion of coherence is a means of evaluating the source-critical significance of this datum (as well as other data). This illustrates how easily discussion about the value of the phenomenon of order can be thwarted by confusion over different types of arguments from order.

In any case, McNicol and Walker agreed on the need for further research on the phenomenon of order. In ‘A Statement of Closure: Items for Further Research’, they stated:

we believe that this discussion has brought to center stage the crucial importance of arguments based on the order of pericopes and for the order within pericopes. Specifically, we need to determine, if possible, whether such an analytic argument can be sustained that may be called foundational for a source theory. At present, the Neo-Griesbachians claim to have such a foundational argument. It needs to be examined rigorously. Tuckett, in principle, denies the validity of such an argument. The implications of his position—especially for Redaction Criticism—need to be explored.

However, not everyone agrees on the importance of the phenomenon of order. Persuaded by B.C. Butler’s demonstration that B.H. Streeter’s inference from order was inconclusive, there are those who regard the phenomenon of order as compatible with numerous source theories and therefore unhelpful for determining the relations between the gospels. For example, according to Pierson Parker,

The evidence from order is compatible with Griesbach’s theory, that canonical Mark drew from canonical Matthew and canonical Luke. But the sequences are equally compatible with other hypotheses, for example, that Mark was the source, which Matthew and Luke usually followed but departed from at their pleasure; or that Matthew was the source, Mark followed and altered it, and Luke often preferred Mark; or that Mark drew upon a Proto-Matthew (as I argued in 1953); or upon a Proto-Matthew and a Proto-Luke. The evidence from order is not decisive. It needs to be supplemented by other considerations.

Similarly, in discussing what might constitute corroboration of a source theory, Jacobson noted that one difficulty is the ambiguity of various literary phenomena that comprise the synoptic problem, which can be explained in a number of ways. ‘A good example of this’, in Jacobson’s view, ‘is the so-called phenomenon of order’.

Currently there seems to be no firm consensus either about the importance of the phenomenon of order or about the probity of arguments based upon it. This study aims to clarify the significance (or otherwise) of the phenomenon of order for resolving the synoptic problem, and to identify which of the arguments used in the past remain useful and which should be abandoned in the quest for a better understanding of the origins and interrelations of the synoptic gospels.

Part I

Preliminary Issues for Analysing
the Phenomenon of Order

Chapter 1

Pericope Divisions and the Question of Parallels

An analysis of the phenomenon of order is complicated by the nature of the disagreements in the order of material shared between the synoptic gospels. The gospel writers did not always arrange their materials in the same way. For example, Mt. 12:31–32 and Mk 3:28–29 (on blasphemy against the Holy Spirit) are clearly tethered to their respective accounts of the dispute between Jesus and the Pharisees (or scribes) over whether Jesus exorcized demons by the power of the chief demon (Mt. 12:22–37 and Mk 3:20–30), but Luke disconnected Jesus’ warning about blasphemy against the Holy Spirit from his account of Jesus’ response to the charge that he used the chief of demons to drive out demons (Lk. 11:14–26) and included it in an entirely different context (Lk. 12:10).

The same point is further illustrated by comparing this same material in its broader Matthaean and Markan contexts. Mark seems to have composed Mk 3:20–35 as a unified section because of the apparent connection between Mk 3:21 and 3:31–35, whereas the only connection Matthew created between Mt. 12:46–50 and what comes before it is the reference to the crowds in Mt. 12:46, which presumably refers to the crowds mentioned in Mt. 12:23. Even if Mt. 12:22–50 and Mk 3:20–35 were intended to be unified sections, Matthew’s division of the material is different from Mark’s by his inclusion of Mt. 12:33–45, which is either entirely absent from Mark’s Gospel or is duplicated in Matthew’s Gospel at the point where these two gospels share the same relative order (cf. Mt. 12:38–42 with Mt. 16:1–4// Mk 8:11–13). One is tempted not to subdivide Mk 3:20–35, but Mt. 12:22–50 breaks naturally into three subsections: 12:22–37, 38–45, 46–50.

A second difficulty for analysing the phenomenon of order arises from different accounts of what appear to be the same events. Luke’s Gospel demonstrates this problem most clearly. For example, Lk. 5:1–11 describes the call of Simon Peter, James and John, each of whom left everything to follow Jesus, but the setting of this event, its relative sequence and its vocabulary differ substantially from the similar accounts of the call of Simon, Andrew, James and John in Mt. 4:18–22 and Mk 1:16–20. The same is true of Lk. 4:16–30 (cf. Mt. 13:53–58 // Mk 6:1–6a) and Lk. 7:36–50 (cf. Mt. 26:6–13 // Mk 14:3–9).

A third difficulty is created by doublets, that is, pericopes within one gospel that appear to be duplicate accounts of the same incident. Matthew’s Gospel illustrates this difficulty most clearly. Mt. 9:27–31 appears to be a duplicate account of Mt. 20:29–34 (cf. Mk 10:46–52 // Lk. 18:35–43), and Mt. 9:32–34 appears to be a partial doublet of Mt. 12:22–30 (cf. Mk 3:22–27; Lk. 11:14–23). Matthew’s compositional and stylistic tendencies create similar problems. For example, Mt. 12:15–16 echoes Mt. 4:24–25, which makes it difficult to know which of these pericopes is the ‘contextual parallel’ of Mk 3:7–12.

Pericope Divisions

Determining the boundaries of literary units within each synoptic gospel is significant because a comparison of order sequences in two or more documents requires the division of those documents into comparable sections. With respect to the construction of a synopsis, Bernard Orchard noted, ‘The problem of the demarcation of the limits of each pericope unit … can vitally affect the layout’. Elsewhere he underscored the importance of pericope divisions by observing that ‘much of the argument about sequence depends on where the limits are drawn between one unit and the next’.4 Therefore the criteria for dividing a document into smaller sections for the purpose of comparison require careful consideration.

It is generally accepted that the synoptic gospels were written in the first century of the Common Era. Literary conventions such as headings and paragraph divisions were not employed by first-century writers. As a result, ancient witnesses to the gospel autographs do not provide this type of formal criteria for making divisions between pericopes or sections. But is this significant? Bo Reicke was aware of this issue but seemed relatively unconcerned by it. In The Roots of the Synoptic Gospels, he opened his chapter on the distribution of the material in the gospels with this observation:

Each of the four New Testament Gospels may be divided into sections called pericopes, of which a considerable number represent thematic and structural analogies in one or more of the other Gospels. A division of the material into such units is found, for instance, in the synopsis of Kurt Aland or that of Heinrich Greeven. Aland counts 367 pericopes in all four Gospels, and Greeven gives a number of 275 to the Synoptic Gospels. In both cases the boundary remains inevitably subjective and, above all, is uncertain among short sections. But the divisions at least give an approximate idea of the circumference and distribution of the material.

On the other hand, E.P. Sanders has shown how variation in pericope size can reveal or conceal certain data. For example, he noted that pericopes in Tischendorf’ s Synopsis were generally longer than those in Huck’s. As a result, ‘some instances in which neither Matthew nor Luke supports Mark’s order were overlooked because they were not full pericopes in Tischendorf’ s synopsis’. The important point, then, is that decisions about the division of the gospel texts into comparable sections or pericopes potentially affect one’s perception of the phenomenon of order.

The significance of pericope divisions has been addressed most thoroughly by David L. Dungan in two studies on synopsis construction. However, his earlier remarks on this issue appeared to minimize its importance. In his survey of different types of synopses designed to facilitate various research agendas, Dungan referred to W.A. Stevens and E.D. Burton’s Harmony of the Gospels for Historical Study (1893), in which the editors explicitly focused on pericope parallelism rather than word-for-word arrangement of the texts. Responding to their decision to subdivide the gospel narratives according to the original writers’ subdivisions, Dungan remarked:

This seems clear enough. Almost anyone could probably go through each of the Gospels and mark these divisions in the narratives where the original author was passing from one ‘meaning unit’ to the next. Modern editions of the New Testament all contain these handy paragraphs, usually with convenient headings, cross-references, etc.

Dungan listed ‘the three basic problems of Synopsis construction’ as: a) how to divide the material into pericopes; b) how to decide what are genuine parallels; and c) how to arrange the parallels throughout the synopsis. With respect to pericope divisions, he drew attention to some of the more obvious structural differences between the gospels, then observed: ‘The central fact is that the Gospels often vary in the amount of material they have within those meaning units which are apparently about the same event’.10 He then invited his reader to compare the total number of pericopes in five different synopses to gain some appreciation of the divergence of opinion on how the gospel narratives should be subdivided. Finally, he summed up by saying:

there seems to be no generally agreed upon method for making such pericope divisions. I suggest that one never really confronts the problem as long as one simply considers each Gospel in isolation. But as soon as the synopsis editor begins trying to match together parallel pericopes, the problems begin. Should the pericopes involving all three (or four) Gospels be made according to the editor’s subjective judgment as to the stages of Jesus’ ministry? Should he seek for the original evangelists’ own divisions? Should he slice the material into form-critical units?

In his earlier study, Dungan was primarily concerned with criteria for subdividing the gospel texts into their component pericopes. Although he recognized the difficulty of determining the limits of pericopes, this issue seemed to be overshadowed in his mind by the problem of paralleling pericopes from different gospels.

This situation was reversed in Dungan’s subsequent discussion, ‘Synopses of the Future’. He structured the first part of this study around three issues, each related to the question of whether a synopsis can be a neutral tool: the text of a synopsis, the arrangement of pericopes, and the division of the gospel narratives into pericopes. He pointed out that the second and third issues are closely related because the arrangement of pericopes in a synopsis depends to some extent on the prior division of each gospel into its constituent pericopes. After considering whether a synopsis can be neutral in terms of text and arrangement of pericopes,13 Dungan underscored the importance of pericope divisions and in doing so revealed a change of emphasis on his part:

Few realize that the question of the division into pericopes is a third area fraught with complications and paradoxes … Touching on this in my previous essay, I did not bring out sufficiently there the specific way in which a task as seemingly innocent as dividing the text into pericopes has an enormous impact on how one visually perceives the Synoptic Problem.

Dungan then emphasized the significance of Sanders’s point that one’s perception of the pattern of agreement and disagreement in the order of pericopes between the synoptic gospels is influenced by how the gospel narratives are subdivided and arranged in a synopsis or table of synoptic parallels: ‘the way one perceived the evidence bearing on the Synoptic Problem, in particular the traditional argument from order of pericopes, varied according to whether one used a synopsis having large or one having small pericopes’. As he explained, an argument based on the arrangement of pericopes in the first three gospels was fundamental to Griesbach’s theory of Mark’s dependence on the Gospels of Matthew and Luke, but the evidence to which Griesbach appealed was eventually advanced by others as evidence for the priority of Mark’s Gospel. Alongside other potential explanations for this reversal in scholarly consensus, Dungan suggested that the demise of Griesbach’s hypothesis occurred partly as a result of the production of synopses, like Huck’s, designed to illustrate other source theories:

What if Griesbach’s statement regarding the order of pericopes, namely, the alternating support he observed between Mark and the other two Synoptic Gospels, became meaningless because scholars began to use other synopses in which (a) the Gospel narratives were divided differently, and (b) the common order of parallel pericopes was differently arranged? Wouldn’t it be inevitable that the phenomenon Griesbach thought he saw would simply disappear? Of course! In point of fact, his ‘discovery’ was literally buried beneath an avalanche of alternative synoptic charts, diagrams and synopses.

To illustrate his point, Dungan reproduced Griesbach’s chart in his ‘Commentatio’ and drew attention to the large size of the pericopes, which frequently contain whole incidents or even series of incidents. He then invited his reader to see whether the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke on Griesbach’s chart is detectable using Huck’s Synopsis. It is ‘virtually impossible’, according to Dungan, partly because the pericope divisions are so small and partly because of the markedly different arrangement of pericopes:

Huck’s arrangement makes it virtually impossible to understand Mark’s use of his two Vorlagen if one assumes Griesbach’s theory of the order of composition. But equally as important, Huck’s much smaller pericope divisions completely obscure the methodical process Griesbach thought he detected in Mark’s artful combination of his two source documents, destroying completely all traces of Mark’s simple, step-wise procedure.

Whether the smaller size of Huck’s pericopes assisted in the demise of Griesbach’s theory is open to question because by grouping together smaller pericopes in the same order in two gospels, one is able to perceive the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke using Huck-Greeven or Aland. But Sanders and Dungan are surely right that there is some relation between the synopsis one uses and one’s perception of the pattern of agreement and disagreement in order between the synoptic gospels. For example, Aland’s Synopsis reveals 11 cases in which parallel pericopes in the Gospels of Matthew and Mark do not share the same order; Huck-Greeven’s shows only 5 disagreements in the order of shared material; and Orchard’s displays 18 differences in narrative order between these two gospels. Dungan’s advice is therefore apposite: ‘We must get in the habit of citing the synopsis we are using whenever we make any claims with respect to the evidence regarding the order of pericopes—or any other aspect of the Synoptic Problem. Different synopses exhibit different evidence’.

On the issue of potential criteria for subdividing the gospel texts into pericopes, Dungan maintained that by comparison with Griesbach’s and Tischendorf’s synopses, many recent synopses were influenced by the Formgeschichte method of gospel criticism:

One may recall … how exhilarating winds of change began blowing through the halls of Gospel research during the post-World War I days. A new approach calling itself Formgeschichte riveted the reader’s attention on the pre-redactional period, when the ‘Jesus material’ was allegedly being handed on (or invented) in the earliest post-Easter communities. Eventually, synopses appeared to facilitate this new brand of biblical scholarship, and they bore little resemblance to earlier synopses.

Dungan showed how W.E. Bundy’s Syllabus and Synopsis of the First Three Gospels (1949) was designed on form-critical assumptions and surmised that other twentieth-century synopses had been influenced in the same way. Whether or not he was right about Huck’s and Aland’s synopses, the question he raised about appropriate principles or criteria for segmenting the gospels is crucial.

Did Formgeschichte significantly influence synopsis editors? Frans Neirynck has found little evidence to support such a view. He admitted that ‘the synopses of Tischendorf, Huck and Aland show a gradual decomposition of the sayings complexes’, but in his view ‘this has to do with the study of the Synoptic parallels and is not directly related to form criticism’. In any case, whether or not synopsis construction has been influenced by Formgeschichte, this method of gospel criticism has certainly influenced the way critics view the gospel materials. From one vantage point, Formgeschichte is a promising avenue of research for determining pericope divisions. After all, if the gospel materials were originally transmitted as brief, self-contained units and were subsequently arranged together by the gospel redactors, then form-critical analysis of the gospels should provide an appropriate vehicle for deciphering the boundaries of the original units.

It is well known that in describing the pre-literary transmission of the gospel materials, Martin Dibelius and Rudolf Bultmann classified individual pericopes into various forms. Dibelius classified the gospel material under five headings—Paradigmen (Illustrations), Novellen (Tales), Legende (Legends), Mythen (Myths), and Paränese (Hortatory Sayings)—whereas Bultmann employed the categories Apophthegmata and ‘Dominical Sayings’ for discourse material, and Wundergeschichten (Miracle Stories) and Geschichtserzählung und Legende (Historical Narratives and Legends) for narrative material. The question raised by these two classifications, however, is whether the particular forms that Dibelius and Bultmann unearthed from the gospel texts bear any relation to what first-century writers understood to be literary units. Some critics have charged that Dibelius and Bultmann did not pay sufficient attention to the literary and rhetorical forms of the ancient Near East at the time when the gospels were written. As early as 1944, R.O.P. Taylor observed, ‘It seems strange that, in all the discussions about the Form-criticism of the Gospels, no appeal or reference has been made to the careful studies of literary form, which were made by writers of the first centuries of our era’.24 More recently, James R. Butts—whose doctoral research focused on Hellenistic compositional exercises, especially The Progymnasmata of Theon—made this indictment of the form-critical work of Dibelius and Bultmann:

Both R. Bultmann and M. Dibelius were aware of the chreia as a literary unit of antiquity when they did their groundbreaking analyses of the forms found in the Synoptic Gospels. But neither of them made use of it in those investigations. Indeed, the failure to use fully the literary categories of antiquity is a major weakness in the scholarly attempt … to identify and understand the various forms which the material of the Synoptic Gospels has taken.

Although Bultmann failed to explain why he considered the chreia not to be useful for his form-critical analysis, in the second edition of Die Formgeschichte des Evangeliums, Dibelius at least discussed the similarities between his ‘Paradigms’ (as well as Bultmann’s Apophthegmata) and the chreia, but to his mind the differences outweighed the similarities.

In 1976, Orchard suggested that five basic literary forms were used to transmit information in the Hellenistic world of the first century:

Now the principal literary forms used by the Greek Rhetors as the means of communicating the Greek tradition were five in number: the gnome (the maxim or pregnant saying), the parabole (‘an effort to put an idea before the mind’s eye’), the diegesis (or narrative), the apomnemoneuma (or recollection, the natural artless form in which an incident would be remembered), and the chreia (or pithy anecdote) … These are the real literary units of the Synoptic Gospels, and the other categories normally used by modern scholars and critics, e.g., ‘the miracle-story’, etc., are not based upon these classical divisions, but on theologically motivated considerations.

It might be argued that some of Dibelius’s and Bultmann’s forms approximate these classical distinctions, but the crucial point is that neither made a concerted effort to relate each of his forms to classical literary categories. From a methodological perspective, it would seem sound to try to divide gospel pericopes on the basis of classical literary categories rather than on subjective impressions of what constitutes a literary unit. As George Wesley Buchanan observed with reference to the chreia:

The advantage of beginning with the chreia … is that it is a definite literary form that was known and defined in antiquity. It was not conjectured in the twentieth century. It is well defined so it is not necessary to confuse it with a reminiscence, a maxim, or an extended narrative. It was used before, during, and after New Testament times, not only by Greek philosophers, church fathers, rabbis, and rhetoricians, but also by the gospel writers themselves and their predecessors.

In recent years, more and more scholars have heeded R.O.P. Taylor’s appeal to make use of classical literary instruction in the interpretation of the gospels. In 1962, for example, William R. Farmer appealed to the Hellenistic use of chreias in his analysis of various Lukan pericopes, as well as Mt. 18:21–35. In Jesus: The King and his Kingdom, Buchanan made extensive use of the chreia as a basic form for evaluating the historical reliability of the teaching of Jesus in the gospels. In 1989, Burton L. Mack and Vernon K. Robbins published a book in which they explored the usefulness of applying classical rhetorical theory to the interpretation and exegesis of various synoptic texts.31 Likewise, in Studying the Synoptic Gospels, E.P. Sanders and Margaret Davies included a chapter on chreiai.

Alongside the usefulness of recent analyses of the relevance of classical rhetorical theory for subdividing the gospel narratives into their basic building blocks, Robert W. Funk’s Poetics of Biblical Narrative is also helpful. In this book, Funk discussed formal criteria for determining the beginning and end of narrative units in the gospels. He distinguished three levels of narrative discourse from the most fundamental to the most complex: actions, events and sequences. Actions are related in ‘narrative statements’. For example, Mk 2:14 consists of five separate narrative statements. Actions grouped together are events, which are related as ‘narrative segments’.35 Events are arranged in sequences. ‘Narrative statements … are grouped into segments, and segments are arranged in sequences’. Using more traditional terminology, one might equate actions with clauses or sentences, events with pericopes, and sequences with sections or blocks.

According to Funk, basic devices called ‘focalizers’ begin a narrative segment using the following three elements—a change in interaction between characters, a change in locale or a change in time: ‘When two or more participants are brought together in a common time and place, the narrative may be said to be brought into focus, or focalized’. This means that a new segment begins any time a character is either introduced for the first time or drawn from the background into the foreground of the narrative to play a significant role. For example, Mt. 3:13 begins a new segment of Matthew’s narrative because Jesus reappears after being absent since Mt. 2:23. A new segment is also begun by a place change, as at Mk 1:21, or a time change, as at Lk. 9:37.

These same elements—participants, place changes, time changes, as well as certain types of actions—also serve to end or ‘defocalize’ narrative segments:

The defocalizing process is … the reverse of the focalizing of a narrative segment: when focalizing, one brings together a specific number of participants in a particular time and place in order to perform a series of actions; when defocalizing, one disperses these same elements, so to speak, so that the focused scene is now defocused.

Funk also noted certain ‘terminal functions’ that serve to end or defocalize a segment, for example, death, fear, awe, praise of God, conversion and faith, prayer and fasting, reports, ‘conversation stoppers’, as well as the narrator’s own commentary and recapitulation. Using Funk’s analysis of the most usual marks of narrative segmentation in concert with recent rhetorical criticism of the gospels, it is possible to subdivide each gospel into its constituent pericopes.40

Paralleling Pericopes

Once one has decided how and where to subdivide the gospel texts, the next step is to arrange parallel texts to facilitate comparison between them. Here the critic is confronted with serious problems:

The problems begin as soon as the synopsis editor tries to match the paragraphs from three (or four) Gospel narratives on a single page. What is a lengthy story in one Gospel appears as a mere fragment in the others, or as part of another story altogether. A striking saying at the end of a healing in one Gospel will appear, stripped of its context, in another as part of a chain of sayings. Two Gospels will proceed almost word for word throughout an entire story, while the other(s) will give a radically divergent account, in an entirely different context.

One of the problems Dungan raised is the significant question of what constitutes true parallelism between pericopes. How is one to discriminate between parallel and non-parallel passages? Are there guidelines for such decisions?

In many cases this is not difficult. Totally dissimilar pericopes pose no problem. Nor do pericopes that relate the same information in similar or nearly identical terminology, such as Mt. 4:18–22 and Mk 1:16–20 or Mk 12:41–44 and Lk. 21:1–4. Problems arise with passages in different gospels apparently reporting the same incident but in markedly different language and circumstances. For example, Lk. 5:1–11 reports the call of Jesus’ first disciples, but unlike Mt. 4:18–22 and Mk 1:16–20 it occurs within the context of a miraculous catch of fish and is related in almost totally dissimilar language. The same can be said for Lk. 4:16–30, which appears to report the same incident as Mt. 13:53–58 and Mk 6:1–6a. Other examples in which Luke’s Gospel differs from the other two include Lk. 4:14–15, 7:36–50, and 23:39–43. In addition, Luke’s wording differs so much from Matthew’s in the following cases that there is considerable doubt whether they relate the same material: Lk. 14:15–24 (the parable of the great feast); Lk. 15:1–7 (the parable of the lost sheep); and Lk. 19:11–27 (the parable of the pounds). Of the three synoptists, Luke was clearly the most independent in style and linguistic usage, but in these passages the differences in wording are accompanied by strikingly different circumstantial details that raise the question of independent sources or traditions.

While it is Luke’s distinctive features that highlight the problem of parallelism, it is the recurrence of similarities between the Gospels of Matthew and Mark that raises a similar problem. In other words, while a number of Lukan pericopes provoke the question whether they are parallel to Matthew’s and Mark’s obviously parallel pericopes, some of Matthew’s pericopes provoke the question of which of two Matthaean pericopes is the real parallel to Mark’s pericope. For example, there are a number of Matthaean doublets, such as Mt. 9:27–31 and Mt. 20:29–34, each of which is conceivably parallel to Mk 10:46–52. But which of these two pericopes is the true parallel to Mk 10:46–52? Perhaps the most crucial instance of this type of difficulty is the choice that must be made between Mt. 4:23–25 and Mt. 12:15–16 as the true parallel to Mk 3:7–12. Aland decided upon Mt. 4:23–25, while Huck-Greeven chose Mt. 12:15–16 as the true parallel to Mk 3:7–12.

These data evoke two questions: What constitutes a true parallel of pericopes? And further, what guidelines or principles help to discriminate between parallel and non-parallel pericopes?

Dungan’ s discussion of this issue intimated that percentage of ‘verbatim correspondence’ constitutes the fundamental criterion for deciding whether pericopes are genuine parallels: ‘Should the [synopsis] editor print only passages next to each other where there is 90% verbatim correspondence? What about 75%? 50%? Who decides where the cutoff point is?’ In this connection, it is worth reflecting on remarks made by Lou H. Silberman in the Seminar on Judaic Studies and the Gospels at the 1977 Colloquy on the Relationships among the Gospels. In his summary of the discussion generated by Silberman’s paper, Joseph A. Fitzmyer remarked:

Silberman raised the question, What constitutes a ‘parallel’? Noting a tendency to regard verbal agreement as the primary mark of parallelism, he argued that a true parallel does not exist unless two passages have similar functions; in the absence of such congruity of function, mere verbal agreement is not significant. It appeared, however, that members of the seminar did not completely agree at this point.

The disagreement Fitzmyer noted with respect to Silberman’s suggestion that two passages should function similarly before being regarded as parallels indicates that for some critics verbal correspondence is the most significant criterion of parallelism. However, Silberman’s comment about the importance of similarity of function should not be ignored. His remarks, as recorded by Fitzmyer, suggest that the two criteria of verbal correspondence and function are complementary when considering whether two passages are parallels.

If narrative function is important, as Silberman argued, what of pericopes such as Lk. 5:1–11 that function within Luke’s narrative in much the same way that Mt. 4:18–22 and Mk 1:16–20 function in their respective narratives? Does the relative absence of verbal correspondence between Lk. 5:1–11 and Mt. 4:18–22//Mk 1:16–20 discount it as a parallel to these pericopes?

In an important study, ‘Sequential Parallelism in the Synoptic Gospels’, Joseph B. Tyson observed that ‘the principles for judging parallelism are by no means clear’. He therefore suggested five general principles for determining parallelism between pericopes:

1.    Narrative material should have the same basic ‘line of action’ and should involve the same basic characters. However, parallelism should not be discounted if different ‘group designations’ are involved, provided the groups have the same function in the two narratives.

2.    Agreement in wording between the two blocks of narrative should be substantial or, if this is absent, there should be precise agreement of ‘unusual or technical expressions’.

3.    Conversely, where two blocks of narrative contain ‘exact verbal agreement of several words’, but the action and characters are not the same, the pericopes are not to be judged as parallels.

4.    Sayings are to be treated individually.

5.    However, pericopes containing a string of sayings should be counted as parallels if the majority of the sayings occur in both pericopes, if they have a high degree of verbal agreement, and if they have the same meaning. Order variations within the sets of sayings does not discount parallelism.

Tyson did not expect that with these guidelines critics would agree in every instance where parallelism between pericopes is disputed. However, his guidelines do provide a valuable starting point for discussing cases in which parallelism is uncertain. His second principle—that there should be substantial verbal agreement between narrative pericopes or at least agreement in the use of ‘unusual or technical expressions’—is probably justified if one is primarily concerned with historical or source issues relating to pericope parallelism, such as whether different types of transmission or variant traditions are required to account for different versions of the same event. For example, some critics argue that the verbal and situational differences between Lk. 5:1–11 and Mt. 4:18–22 // Mk 1:16–20 indicate another source for Luke’s account of the call of the first disciples. However, from a compositional perspective, it is probable that Luke was familiar with the version of the call story in Matthew’s and Mark’s narratives. If so, Luke either replaced Matthew’s and Mark’s version of the call story with an account from a different source or simply retold it in his own terms by combining it with a miracle story. Either way, he substituted another call story for Matthew’s and Mark’s version and in the process gave his account a new context, that is, after rather than before Jesus’ activity in Capernaum. Using Tyson’s principle, Lk. 5:1–11 is not parallel to Mt. 4:18–22 or Mk 1:16–20, despite a similar ‘line of action’, the same basic characters and agreement in the use of three common words in Lk. 5:11: ἀφέντες, ἠκολούθησαν, and αὑτῷ. In an analysis of the phenomenon of order, however, Lk. 5:1–11 must be considered in relation to Mt. 4:18–22 and Mk 1:16–20 because it clearly functions the same way in Luke’s narrative as Matthew’s and Mark’s call stories function in their narratives. Even if a variant tradition stands behind Lk. 5:1–11, Luke clearly changed the relative sequence of his call story. Similarly, Lk. 3:19–20, 4:16–30, 7:36–50 and 22:24–30 also function as parallels of Matthaean and Markan pericopes that are obvious verbal and contextual parallels.

Each of these two issues—the question of how pericope divisions are determined and that of determining parallelism between pericopes—is fundamental for ascertaining the degree of agreement and disagreement in narrative order between the synoptic gospels. Unless these basic issues are resolved, the task of comparing the relative orders of events in the synoptic gospels is less securely grounded, thereby making any argument based on the phenomenon of order less conclusive than it might be. To reiterate, comparing the relative orders of pericopes is fundamental to any argument based on the phenomenon of order. Before making any comparison, however, the critic must resolve the preliminary questions of pericope divisions and pericope parallelism or accept the resolutions of one who has already made these decisions—the synopsis-maker.

Chapter 2

Towards an Arrangement of Parallel Pericopes

The issues of pericope divisions and parallelism are certainly significant, but they are eclipsed in importance by the problem of arranging parallel and functionally parallel pericopes in a table to depict the pattern of agreement and disagreement in order between the gospels. It would seem a relatively simple task to construct a table showing where all three gospels share the same order of arrangement, where two gospels agree in order against the third, and where all three gospels disagree in their arrangement of parallel materials. But it is a complicated and intricate task.

Related to the problem of arranging parallel pericopes in a table or chart is the difficulty of describing the various agreements and disagreements in order between any two and between all three synoptic gospels. Indeed, an awareness of the difficulty of describing the pattern of agreement and disagreement in order between the gospels helps one to comprehend the complications of establishing a working scheme of parallelization. For this reason, this chapter examines the difficulty of describing the phenomenon of order before addressing various problems associated with constructing a table of synoptic parallels as a basis for analysing, or building arguments based upon, the phenomenon of order.

Describing the Phenomenon of Order

From time to time, critics have drawn attention to the difficulty of describing the phenomenon of order either accurately or in a neutral manner. According to E.P. Sanders,

Of all the synoptic phenomena, the phenomenon of order is perhaps the most difficult to state in general terms which can claim to accuracy. That Matthew, Mark and Luke have approximately the same arrangement of material is obvious, but the attempt to state just what the interrelations of the orders of the gospels are is agonizing.

Similarly, in ‘Sequential Parallelism in the Synoptic Gospels’, Joseph B. Tyson noted that ‘what we should require for sequential parallelism is not obvious. Should we look for parallel passages which follow one another without interruption, or should we look for pericopes which come in the same relative order but are scattered and assimilated into non-parallel sections?’ He decided to focus primarily on strict sequences of pericopes, in other words, ‘those cases where two or more parallel pericopes follow one another without interruption in two or three gospels’.3 Put negatively, he chose not to regard parallel pericopes as sequential parallels, or parallels in order, if two or more pericopes in strict sequence in one gospel are interrupted by a unique pericope or series of pericopes in another gospel. He was certainly aware of cases in which unique pericopes interrupted shared sequences of material, not to mention cases of inversion in which the strict sequence of pericopes in one gospel is simply reversed in another. Although he noted these phenomena, however, he chose not to incorporate instances of relative sequence into his analysis of sequential parallelism.

Christopher M. Tuckett criticized Tyson’s procedure for two reasons: first, it minimizes what can be regarded as occurring in the same relative order in two or more gospels; and second, it fails to discriminate between parallel pericopes in strict sequence in the same relative position in two gospels and parallel pericopes in strict sequence in completely different contexts. He illustrated his first objection with the following hypothetical order of incidents in documents X and Y:

















Here d is in the same relative position in X and Y, but because of the rigidity of Tyson’s method it is not counted as an agreement in order. Tuckett noted that this is an actual occurrence in the case of the parable of the mustard seed (Mt. 13:31–32 // Mk 4:30–31). An illustration of his second objection to Tyson’s method is the case of the parables of the mustard seed and the leaven, which follow each other in strict sequence in Mt. 13:31–33 and Lk. 13:18–21 but occur in markedly different contexts.

Tyson’s attempt to demonstrate what degree of sequential parallelism exists in the synoptic gospels was intended to provide a neutral description of the phenomenon of order: ‘The need for a neutral description of the order of material in the Synoptics should be clear to all who are interested in the Synoptic problem. The purpose of this study is to provide a means of describing the phenomenon of order’. Tyson’s intention was honourable. All too often, students of the New Testament have been introduced to the phenomenon of order under the guise of a source theory presupposed in the description of the phenomenon itself. Both Tyson and Tuckett criticized B.H. Streeter’s description of the phenomenon of order because it presupposed the source theory it was presented to support. Streeter wrote, ‘The order of incidents in Mark is clearly the more original; for wherever Matthew departs from Mark’s order Luke supports Mark, and whenever Luke departs from Mark, Matthew agrees with Mark’.6 Tuckett, who advocates Streeter’s ‘fundamental solution’, the two-source theory, admitted that Streeter’s description of the phenomenon of order was ‘clearly formulated under the presupposition of Markan Priority … and can hardly be said to be an unbiased statement of the facts’. However, while conceding that there are less biased ways of describing this phenomenon, he doubted whether the variations in order between the gospels are capable of neutral description.

Tuckett’s reason for doubting whether order differences between gospels can be described in neutral terms is that it is not always possible to determine which pericope is ‘out of order’ when two or more documents are compared. In other words, there are a number of ways of expressing which of a number of pericopes is or are out of order in relation to other pericopes. As he explained,

A disagreement is, at one level, a failure to agree, i.e. if one writer has ab and the other has ba. But can one be more precise beyond making the negative statement that there is a failure to agree? Suppose X and Y have four pericopes abed in the order X: abed, Y: acbd. Clearly there is failure to agree in order. Further, most would assume that a and d are in the same order in the two texts. But which element, or elements, is, or are, out of order? There are at least three ways of illustrating the parallels diagramatically:












b c



c b







According to (i) b is out of order; according to (ii) c is out of order; according to (iii) both b and c are out of order. At the purely formal level there appears to be no way of claiming in absolute terms which of these three is the most preferable way of describing the pattern of agreement and disagreement in order between X and Y. The way in which the parallels are initially set up inevitably affects in a significant way the description of the differences in order. A different scheme of parallelization produces a quite different set of non-parallels.

David L. Dungan has proposed that this inability to give a precise and unambiguous description of order differences between the gospels be dubbed ‘Tuckett’s Dilemma’. He has also given his own explanation for this difficulty: ‘The reason for this curious fact is because the original order must be known, in order to say which pericopes are “out of order” in the disarranged series’. In short, one must presuppose which order of arrangement is original to identify those pericopes or sections that have been transposed.

Dungan has given considerable attention to the implications of ‘Tuckett’s Dilemma’ in his discussion of the various ways that the gospels can be arranged in a synopsis. One of the principal concerns of his study on ‘Synopses of the Future’ was to dispel what he regarded as a naïve confidence in the ability to parallel pericopes in a neutral way. To begin with, he noted how dividing the gospel texts into comparable sections and arranging the parallels are intertwined:

In practice, the synopsis editor divides up the Gospel narratives with one eye constantly on divisions he has already made in the parallel Gospels. The process of dividing up the material and the process of arranging the primary parallels take place simultaneously and influence each other continuously as the synopsis editor moves back and forth, shaping and arranging the parallels in his synopsis.

Although Dungan did not define precisely what he meant by ‘primary parallels’, his point is clear enough. There is no self-evident scheme of arrangement that emerges from the synoptic data. Or, as Dungan put it, ‘There is, unfortunately, no Ariadne’s thread to guide the synopsis editor through the labyrinth of Synoptic interrelationships’, no objective criterion for choosing between ‘equally valid alternative configurations of primary parallel pericopes’. To clarify the issue, he referred to three basic options for positioning Matthew’s Sermon on the Mount relative to Mark’s narrative sequence. For Dungan,

The great problem is: which chain of pericopes will [the synopsis editor] create, running throughout the length of the Gospels? What series of pericopes will he match up, in the full knowledge that if he puts them one way, he will have to forego other possible points of common linkage? This is no longer the elementary question of keeping each Gospel’s order intact throughout the synopsis. This is a very different question: what common order or chain of pericopes among all three (or four) Gospels will he decide upon? A lot hangs on this decision … There are a number of ultrasensitive questions which hang in the balance as the Synopsis editor works his way through the Gospel narratives, arbitrarily deciding which pericopes to put next to each other in some sort of continuous chain. At the very least, he will be laying down what is widely regarded as the fundamental evidence for any source hypothesis.

As a result, he suggested that one should ‘prepare a preliminary redactional analysis of each Gospel independently before beginning the task of dividing up the narratives. Then these redactional studies would be used as a guide to determine which paragraphs to place next to each other’.

Although I have not adopted Dungan’s suggested procedure, my procedure is similar in some respects. In what follows, I examine what kind of redaction is involved on three different arrangements of parallel pericopes in relation to two source theories: Markan priority and the two-gospel hypothesis of Markan posteriority. Specifically, after discussing three issues that affect how one arranges parallel pericopes in a synopsis or table, I examine what is involved at the compositional level on three schemes of parallelization—with particular reference to disagreements in order—by comparing Matthew’s rearrangements of Mark’s materials, assuming Markan priority. I then repeat the procedure focusing on Mark’s transitions between Matthew and Luke, assuming the two-gospel hypothesis of Markan posteriority. My arrangement of parallel pericopes is informed by these ‘redactional soundings’.

Issues Involved in Arranging Parallel Pericopes

John Bernard Orchard is one of the few synopsis editors to have discussed some major decisions that must be made when constructing a synopsis. In his view, there are three fundamental issues whose resolution determines the shape of a synopsis: first, which gospel should be the ‘mean’, or, more simply, the column order in which the gospels should be placed; second, where Matthew’s Sermon on the Mount should be placed relative to Mark’s and Luke’s narrative outlines; and third, how to arrange the parables and commissioning discourses in each of the synoptic gospels. While column order and the arrangement of the parables and commissioning discourses are certainly issues about which decisions must be made, they are not as critical as the position of Matthew’s Sermon on the Mount relative to the other two synoptics. After dealing briefly with Orchard’s less critical issues, much of the remainder of this chapter is concerned with this decisive issue for devising a scheme of parallelization.

Most synopses, synoptic charts, or tables of synoptic parallels are constructed in three vertical columns with Matthew’s text in the left-hand column, Mark’s in the middle, and Luke’s in the right-hand column. In Orchard’s synopsis, however, Matthew’s Gospel is displayed in the left-hand column, Luke’s in the middle, and Mark’s in the right-hand column. His reason for doing so was to illustrate his own source theory: ‘the Griesbach Hypothesis holds that in the order of reality Lk. knew Mt. and that Mk knew both Lk. and Mt; and hence anyone wishing to construct a synopsis to illustrate it would of course apply Lk. to Mt. and then apply Mk to the product to give Mt.—Lk.—Mk’.

In a review of Orchard’s synopsis, Tuckett opined that the decision to arrange the vertical columns in the order Matthew-Luke-Mark-John was ‘unimportant’ and, indeed, ‘relatively trivial’. Frans Neirynck echoed Tuckett’s judgment:

In recent synopses it has become a sacrosanct principle to print the text of each gospel in its original sequence … With such synopses at our disposal, the question of which gospel is placed in the first column no longer has the importance it once had when the gospel texts in the second and third columns were rearranged in accordance with the first.

He also noted that to remove Mark’s text from the middle column was disadvantageous for the Griesbach hypothesis because it was then more awkward to see Mark’s alleged conflation of Matthew’s and Luke’s texts, not to mention Mark’s alternating agreement between Matthew’s and Luke’s orders.

In response to his critics, Orchard reiterated that ‘the Mt.—Lk.—Mk synopsis is necessary to illustrate the Two-Gospel Hypothesis’. He also conceded:

However, since the peculiar ‘zigzag’ effect can be seen more clearly by placing Mk between Mt. and Lk … there is also a good case to be made out for the inclusion of the synopsis with the Mt.—Mk—Lk. order to illustrate this fact of the problem … Therefore in my opinion both forms of the vertical-column synopsis play a part in illustrating the Two-Gospel Hypothesis. Above all, my own Synopsis is essential to illustrate properly the indebtedness of Lk. to Mt. and of Mk to both Mt. and Lk., according to the chronological development of the synoptic tradition. On the other hand, the Aland and Huck-Greeven synopses should theoretically be able to illustrate better the manner in which Mk has used both Mt. and Lk

This seems reasonable. Orchard’s column layout is useful for exploring the potential relations between the Gospels of Matthew and Luke and the ‘development of tradition’ within the synoptic gospels on the assumption of his source theory. On the other hand, a middle position for the Gospel of Mark affords the critic a better vantage point for examining Mark’s relation to the other two synoptics.

While advocates of the two-gospel hypothesis contend that Luke was dependent on Matthew’s Gospel, the viability of the theory rests primarily on the plausibility of Mark’s dependence on both Matthew’s and Luke’s Gospels. Only after critics are convinced that Mark’s Gospel is a compilation from the Gospels of Matthew and Luke, rather than the source of those Gospels, will they rethink the relation of Luke’s Gospel to Matthew’s Gospel. Because my purpose is to weigh the relative strengths and weaknesses of the Markan hypothesis and the two-gospel hypothesis, my first decision with respect to an arrangement of parallel pericopes is to place Mark’s Gospel between the Gospels of Matthew and Luke.

Orchard’s second issue was the relative order of the parables and commissioning discourses. The problem is that Matthew’s commissioning discourse occurs before his parable collection, whereas Mark and Luke included their parable collection before recounting Jesus’ instructions for the mission of his twelve apostles. Contrary to Huck’s and Aland’s decision to arrange pericopes so that all three parable collections occur in the same relative order, Orchard chose to place the three commissioning discourses in parallel order. He conceded that both alternatives are possible, but he maintained that to parallel the commissioning discourses more accurately reflects the compositional scenario implied by his source theory:

According to the New Griesbach Hypothesis, there is a question of principle involved here. This theory implies that Lk. has for a long time been following the general order of Mt, while at the same time transferring systematically a number of units (Lk. 4:38–41; 5:12–6:11) into a new order of his own with which Mk concurs. However, having got down to Mt. 12:21, Lk. breaks off at 8:3, and here Mk picks up Mt. at the point where Lk. left off, viz. at Mt. 12:22–30 = Mk 3:22–30, and goes on until Lk. Joins up with him again at Mk 4:1f. = Lk. 8:4f. In other words anyone constructing a synopsis on the Griesbach Hypothesis will follow the lead of Mk here when Lk. leaves off.

The practical effect of this decision is that although only Mark’s and Luke’s parable collections occur in parallel order, all three commissioning discourses occur in parallel order, as do the respective versions of the stilling of the storm, the Gadarene/Gerasene demoniac(s), and Jairus’s daughter and the haemorrhaging woman (Mt. 8:23–9:26 // Mk 4:35–5:43 // Lk. 8:22–56).

The effect of this decision is more negative than positive. In his review of Orchard’s synopsis, Tuckett noted two disadvantages for Orchard’s source theory: first, the extent of Luke’s rearrangement of Matthew’s materials is greater than on Huck’s schema; and second, Orchard’s synopsis actually obscures Mark’s alternating agreement with the Gospels of Matthew and Luke. Given that Orchard’s aim was to provide support for the two-gospel hypothesis, this decision seems to have backfired. I see no advantage in following Orchard on this point.

The chief difficulties for constructing a working scheme of parallelization occur in one block of synoptic material—Mt. 4:23–13:52; Mk 1:21–6:13; Lk. 4:31–9:6. It is not simply that Matthew’s order of events in this block differs so much from Mark’s and Luke’s more similar orders. A further complication is that within this same block, Matthew duplicates certain materials that appear only once in Mark’s and Luke’s Gospels. This sometimes creates the dilemma of not knowing which of Matthew’s duplicate passages is contextually parallel to Mark’s and Luke’s accounts of the same episodes. No such problem occurs after Mt. 14:1, Mk 6:14, and Lk. 9:7, despite some instances in which there is more than one way of displaying the similarities and differences in narrative order.

The key difficulty is the position of Matthew’s Sermon on the Mount relative to Mark’s and Luke’s order of events. One can place Matthew’s Sermon on the Mount in three different Markan and Lukan contexts: after Mk 3:19 and Lk. 6:19 (with Aland, Boismard-Lamouille, and Orchard); after Mk 1:39 and Lk. 4:44 (with Huck-Greeven); or after Mk 1:20 and Lk. 4:31 (with Neirynck and the Dutch Synopsis van de eerste drie evangeliën edited by A. Denaux and M. Vervenne).

Two factors seem to have influenced Aland’s decision to place Matthew’s Sermon after Mk 3:19 and Lk. 6:19. The first is the conceptual and verbal correspondence between Mt. 4:23, Mk 1:39, and Lk. 4:44, each of which is a summary of a preaching tour throughout Galilaean synagogues. The second is the strong verbal agreement between Mt. 4:25 and Mk 3:7–8, each of which lists a number of regions from which crowds followed Jesus. The verbal parallels are striking in each case, but by section 50 in Aland’s Synopsis Mark and Luke have related a full day’s activity in Capernaum (Mk 1:21–38; Lk. 4:31–43) as well as a series of healings and controversies that culminate in either a plot to destroy Jesus (Mk 3:6) or a discussion about what to do with him (Lk. 6:11), whereas Matthew has only summarized Jesus’ activity in Galilee in three verses, Mt. 4:23–25.

Huck’s decision to place Matthew’s Sermon after Mk 1:39 and Lk. 4:44 was also influenced by the verbal and conceptual similarities between Mt. 4:23, Mk 1:39, and Lk. 4:44. Nevertheless, he regarded Mt. 12:15–16 as the contextual parallel of Mk 3:7–12, even though this Markan pericope also has strong verbal connections with Mt. 4:24–25. Another factor may also have influenced Huck and his revisers to place Matthew’s Sermon after Mk 1:39. Orchard has suggested that Huck chose to place Matthew’s Sermon where he did at least partly because this arrangement resulted in ‘the maximum number of parallel verses in natural order in the Triple Tradition between Mt. 3:1–13:58 and its parallels’.

Against Huck, Orchard maintained that Matthew’s Sermon should be placed after Mk 3:19 and Lk. 6:19 because of the ‘close literary relationship between Mt. 4:24–5:2 and Lk. 6:12–20a and Mk 3:7–19’. Of course, Orchard constructed his synopsis to demonstrate the feasibility of Griesbach’s hypothesis:

For the Griesbach Hypothesis holds that in the order of reality Lk. knew Mt. and that Mk knew both Lk. and Mt; and hence anyone wishing to construct a synopsis to illustrate it would of course apply Lk. to Mt. and then apply Mk to the product in order to give Mt—Lk—Mk. And this order of application leads to a concentration on the parallels in Mt. 4:24–5:2 = Lk. 6:12–20a = Mk 3:7–19 as providing the correct location for the Great Sermons, to the exclusion of the other possibilities in the Markan and Lukan schemata.

Orchard emphasized the parallels between Mt. 4:24–5:2 and Lk. 6:12–20 at least partly because these pericopes introduce, respectively, Matthew’s Sermon on the Mount and Luke’s Sermon on a Level Place, which agree closely in the relative sequence of shared materials. While Luke’s sermon is considerably shorter than Matthew’s, with the single exception of the so-called Golden Rule the material common to these sermons follows the same relative order. In any case, Orchard made it clear that his decision about the appropriate position for Matthew’s Sermon on the Mount relative to Mark’s and Luke’s Gospels was consciously influenced by his preferred source theory.

In Neirynck’s proposal, a primary datum is the close verbal parallelism between Mt. 7:28b–29 and Mk 1:22. One should not overlook this, but verbal correspondence is an ambiguous indicator of parallelism in order. Whatever one decides about the relative position of Matthew’s Sermon, there will be close verbal parallels that are not in the same order. For example, on Neirynck’s arrangement of pericopes, Mt. 4:23 and Mk 1:39 are not contextual parallels, nor are Mt. 4:24–25 and Mk 3:7–12, even though both pairs of verses are reasonably close in content and wording. Matthew 7:28b–29 and Mk 1:22 are summary statements describing public reaction to Jesus’ teaching. In both gospels it follows the first occasion on which Jesus is said to have taught, although they are not the same occasions. Therefore this description of the crowd’s reaction to Jesus’ teaching was clearly not tethered to any particular occasion. Although Luke’s Gospel shares Mark’s order of incidents at Mk 1:21–28 and contains a similar description of the people’s reaction to Jesus’ teaching in the Capernaum synagogue (Lk. 4:32), this is not the first time that Luke recorded public reaction to Jesus’ teaching. In Lk. 4:15 Jesus is said to have taught in their [Galilaean] synagogues, being praised by all, and in Lk. 4:22 the initially favourable response to Jesus’ teaching in the Nazareth synagogue is recorded in words that convey the same impression as Mt. 7:28b, Mk 1:22a, and Lk. 4:32a. Luke 4:32 is therefore the third occasion in which Luke recorded a favourable reception to Jesus’ teaching. On this evidence one could argue that Mt. 7:28b–29, Mk 1:22 and Lk. 4:32 are best described as ‘stock descriptions’ of the way that people responded to Jesus’ teaching, so were used whenever it was considered appropriate.

More important for Neirynck is his understanding of Matthew’s redaction based on the theory of Markan priority:

The Sermon of Jesus’ teaching (cf. διδάσκων in 4:23 and ἐδίδασκεν in 5:2) is located in between Mk 1:20 and 1:22, as the Matthean parallel to the first mention of Jesus’ teaching in the gospel of Mark: ἐδίδασκεν (1:21). For the solemn introduction of the Sermon, Matthew’s motifs are borrowed from the Markan summaries in 1:39 (cf. 6:6b) and 3:7–8:13, but also from 1:28 (Mt. 4:24a) and 1:32–34 (Mt. 4:24b). Matthew extends Mark’s Day at Capernaum to a First Day of Jesus’ Teaching and Healing Ministry (4:23–8, 17).

As with Orchard, Neirynck’s arrangement of parallel pericopes is largely the result of his preferred source theory.

Were Matthew dependent on the Gospel of Mark, as Neirynck affirms, it would not have been difficult to omit Mk 1:21–28 and reserve Mark’s description of the crowd’s response in Mk 1:22 for the conclusion to his Sermon. But if Matthew did what Neirynck claims he did, he had to comb through Mk 1:22–3:13 for the introduction to his Sermon in Mt. 4:24–25. This seems a complicated procedure for writing a brief paragraph. Even if Mk 1:21 suggested a suitable position for Matthew’s Sermon, might it not have been easier to skip over Markan material until he found a suitable passage to rewrite for his own purposes, insert his Sermon, recall Mk 1:22 at its conclusion, and subsequently incorporate material that he had passed over temporarily? On this understanding of Matthew’s redaction of Mark’s materials, one might be inclined to place Matthew’s Sermon on the Mount where Aland and others place it—after Mk 3:19.

Clearly, when considering the most appropriate location for Matthew’s Sermon relative to Mark’s and Luke’s Gospels, one can appeal to various criteria: verbal parallelism, which is ambiguous; apreferred source theory, which preempts one’s source-critical investigations; redactional considerations, which are often based on a source theory and therefore also preempt one’s source-critical investigations; and the concern to parallel as much common material in the same sequence as possible. Another criterion is what might best be called ‘narrative development’. This consideration seems to be the basis for Neirynck’s remark, ‘It cannot be denied that the content of Mt. 4:25 is parallel with Mk 3:7–8, and Mt. 4:23 with Mk 1:39, but the parallel of order is with Mk 1:21′. This is incontrovertible in terms of the strict sequence of Matthew’s and Mark’s incidents; Matthew’s Sermon occurs at the same relative juncture as Mark’s and Luke’s respective accounts of events in the Capernaum synagogue.

One further consideration is that if one works backward from Mt. 14:1 and Mk 6:14—instead of forward from the beginning of the gospels—one finds that Mt. 12:15–16 most naturally parallels Mk 3:7–12 in relative order. This suggests that either Mk 1:21 or Mk 1:39 is the more likely position for Matthew’s Sermon relative to Mark’s narrative outline.

In any case, Neirynck’s contention that both the Griesbach theory of Markan posteriority and the Markan hypothesis are ‘well served’ by placing Matthew’s Sermon opposite Mk 1:21 seems to have persuaded at least some neo-Griesbachians. In ‘Narrative Outline of the Markan Composition According to the Two Gospel Hypothesis’, William R. Farmer, David L. Dungan, David B. Peabody, Allan J. McNicol and Philip L. Shuler state:

One major revelation came in our discovery of the precise and detailed way Mark skirted Matthew’s Sermon on the Mount at Mk 1:21, and yet put details from Matt. 4:23–25 in Mk 1:39 and again in 3:7–12. This has led successive generations of synopsis editors to place Matthew’s Sermon on the Mount next to all three locations … But which is correct? With the key of the Two Gospel Hypothesis, we believe we can see where it should properly be located according to Mark’s own thinking, namely, after 1:20.

Despite this apparent agreement between proponents of competing source theories, the point remains that there is no objective criterion for settling this vexed issue. The practical implications of this situation are twofold. First, different arrangements of parallel pericopes result in substantially different patterns of agreement and disagreement in narrative order, including differences in the number of disagreements in order between the Gospels of Matthew and Mark. Second, different schemes of parallelization result in different pericopes or sections being classified as ‘disagreements’ in order. For example, the three accounts of the healing of the leper in Mt. 8:2–4, Mk 1:40–45 and Lk. 5:12–16 occur in the same relative order in Huck-Greeven’s Synopsis, whereas Matthew’s version of this event occurs earlier than Mark’s and Luke’s accounts on Neirynck’s configuration of parallels and considerably later in Aland’s and Orchard’s synopses. Therefore, if one employs a compositional argument from order, one must explain different sets of transpositions depending on which configuration of pericopes one adopts. The data simply do not yield to the critic’s effort to display them in a neutral and unambiguous way.

Matthew-Mark Disagreements in Order on
Three Different Pericope Arrangements

In this section I examine both the number of disagreements in order between Matthew’s and Mark’s Gospels and the nature of Matthew’s transpositions on the assumption of Markan priority according to three arrangements of pericopes—Neirynck’s, Huck-Greeven’s and Aland’s.


Beginning with Neirynck’s scheme of parallelization, by placing Matthew’s Sermon on the Mount between Mk 1:20 and Mk 1:22 relative to Mark’s narrative outline, the following pericopes and transitional or summary statements disagree in relative order: (1) Mt. 4:23 // Mk 1:39; (2) Mt. 4:24–25 // Mk 3:7–12; (3) Mt. 8:2–4 // Mk 1:40–45; (4) Mt. 8:18, 23–34 // Mk 4:35–5:20; (5) Mt. 9:18–26 // Mk 5:21–43; (6) Mt. 9:35 // Mk 6:6b; (7) Mt. 10:2–4 // Mk 3:16–19; (8) Mt. 10:1, 5–15 // Mk 6:7–13; (9) Mt. 10:17–22 // Mk 13:9–12.

On this configuration of pericopes, all of Matthew’s transpositions were ‘anticipations’. In each case, Matthew reached forward in Mark’s narrative to transpose pericopes or summaries into an earlier context in his own narrative relative to Mark’s outline. His relocations of Markan materials were always from a later context than the point he had reached in Mark’s order of incidents.

Although Matthew inserted his Sermon on the Mount between Mk 1:20 and Mk 1:22, he used two later Markan pericopes as a ‘bracket’ or ‘frame’ for his Sermon, namely, Mk 1:39 (= Mt. 4:23) and Mk 1:40–45 (= Mt. 8:2–4). One could almost describe this as a transposition of a single section of Markan material, Mk 1:39–45, albeit involving the insertion of a large block of material between Mk 1:39 and Mk 1:40–45.

It would not be difficult to imagine Matthew recalling Mk 1:22 at the conclusion of his Sermon, but other details are not so easily explained, particularly Mt. 4:24–25. According to Neirynck, Matthew drafted the introduction to his Sermon by combining fragments from Mk 3:7–8 and 13, Mk 1:28 (cf. Mt. 4:24a), and Mk 1:32 and 34 (cf. Mt. 4:24b).

As for Mt. 8:18–34, Matthew must have skipped over nearly three chapters before incorporating Mk 4:35–5:20 into a much earlier context, indeed, prior to his parallel to Mk 2:1–12, the forgiving and healing of a paralytic.

After reaching so far forward in Mark’s narrative for Mt. 8:18–34, Matthew must then have resumed Mark’s earlier order for some 17 verses (Mt. 9:1–17 // Mk 2:1–22). Between Mk 2:22 and 2:23, however, he interposed two-and-a-half chapters that bear no resemblance to Mark’s sequence of incidents, even though he borrowed some of this material from Mark’s later chapters. Matthew began this section with the third episode in the Markan block from which he had begun borrowing at Mt. 8:18, namely, Mk 5:21–43. One could almost say that Matthew interposed Mk 2:1–22 between Mk 5:20 and Mk 5:21 because at this point he happened to be borrowing as much from Mk 4:35–6:13 as from Mk 2:1–3:6.

Matthew’s doublets in Mt. 9:27–34 are difficult to assess. Each would have to be considered a transposition were it not that its ‘narrative twin’ occurs in the same relative order as its Markan parallel (cf. Mt. 9:27–31 and Mt. 20:29–34 with Mk 10:46–52; cf. Mt. 9:32–34 and Mt. 12:22–24 with Mk 3:22).

Matthew began Mt. 9:35–38 by expanding the brief summary in Mk 6:6b, but he also made use of Mk 6:34 (cf. Mt. 9:36), half of which he copied more closely later at Mt. 14:14. His expansion of Mk 6:6b is similar to his expansion of Mk 1:39 at Mt. 4:23–25.

Matthew 10:1 continued Matthew’s use of this later Markan block by borrowing from Mk 6:7. Matthew 10:2–4 is a transposition within a transposition because not only did Matthew bring forward Mark’s commissioning of the Twelve into an earlier context (i.e. before his incorporation of the culmination to Mark’s controversy section in Mk 2:1–3:6, the question of Jesus’ true family and the dispute over the source of Jesus’ miraculous power in Mk 3:20–35, and Mark’s parable discourse), but he also transposed the names of the Twelve from a different context from that from which he had been drawing. Indeed, Mt. 10:2–4 is the only transposition of Markan material that disrupts Matthew’s pattern of transposing Markan materials in their relative Markan order. Matthew then reverted to Mk 6:7–13 when writing Mt. 10:5–14.

Of particular interest is that three of Matthew’s transpositions were taken from a single Markan block, Mk 4:35–6:13. Within this block only one incident reappears in its original context in Matthew’s Gospel (i.e. shortly before the narration of Herod’s consternation about John the Baptist and the Baptist’s execution): Jesus’ rejection at Nazareth in Mk 6:1–6. This block of Markan material is isolable because it immediately follows Mark’s parable section and concludes with Mark’s account of the mission of the Twelve. Matthew’s literary plan meant that he was leading up to his own account of the mission of the Twelve in Matthew 10. In addition, assuming Markan priority, Matthew was faithful to Mark’s order from Mt. 12:1 onwards, although in Matthew 12–13 he had to omit what he had already brought forward into his Gospel from Mk 3:13–6:13.

Matthew’s final transposition was the most far-reaching. While describing in Mt. 10:17–22 what treatment Jesus’ disciples could expect in their immediate mission while Jesus was still present, Matthew borrowed from Mk 13:9–12, a Markan section that describes what Jesus’ disciples could expect during their mission after Jesus was no longer with them.


Turning to Huck-Greeven’s configuration of pericopes, by placing Matthew’s Sermon on the Mount after Mk 1:39, the following pericopes and transitional or summary statements disagree in relative order: (1) Mt. 4:24–25//Mk 3:7–12; (2) Mt. 7:28–29//Mk 1:22; (3) Mt. 8:14–17//Mk 1:29–34; (4) Mt. 8:18–34//Mk 4:35–5:20; (5) Mt. 9:18–26//Mk 5:21–43; (6) Mt. 9:35//Mk 6:6b; (7) Mt. 10:2–4 //Mk 3:16–19; (8) Mt. 10:1, 5–15//Mk 6:7–13; (9) Mt. 10:17–22//Mk 13:9–12.

In terms of disagreements in order between Matthew and Mark, the only differences between Huck-Greeven’s arrangement of pericopes and Neirynck’s are Mt. 7:28–29//Mk 1:22 and Mt. 8:14–17//Mk 1:29–34. On Huck-Greeven’s arrangement, these transpositions are not ‘anticipations’ of Markan materials, but rather ‘recollections’ of previously omitted materials.

The omission of most of Mk 1:21–38 between two closely paralleled pericopes, Mt. 4:18–22 (=Mk 1:16–20) and Mt. 4:23 (=Mk 1:39), suggests that Matthew may have bypassed this Markan section while searching for an appropriate Markan location for his Sermon. If so, he anticipated Mk 3:7–12 in his expansion of Mk 1:39 to form an appropriate introduction to his Sermon on the Mount. He then recalled Mk 1:22 as a suitable conclusion to the Sermon. Following the Sermon he included the pericope that follows the Markan summary he had expanded to lead into his Sermon, namely, Mk 1:40–45.

If Matthew noticed the thematic unity of Mk 2:1–3:6, he may have decided to begin interfering with Mark’s sequence at this point. He inserted a non-Markan pericope, Mt. 8:5–13, then recalled Mk 1:29–34 because these verses illustrate Jesus’ wonder-working ability, which is what Matthew is largely concerned with in Mt. 8:2–9:34. On this arrangement of pericopes, Matthew retrieved suitable Markan materials from a previously omitted section before transposing Markan materials from a later section, which, with one understandable exception, are all incorporated in their relative Markan sequence.


Turning finally to Aland’s arrangement of parallel pericopes, by placing Matthew’s Sermon on the Mount after Mk 3:19, the following pericopes and transitional or summary statements disagree in order: (1) Mt. 7:28–29//Mk 1:22; (2) Mt. 8:1–4//Mk 1:40–45; (3) Mt. 8:14–17//Mk 1:29–34; (4) Mt. 8:23–34//Mk 4:35–5:20; (5) Mt. 9:1–17//Mk 2:1–22; (6) Mt. 9:18–26//Mk 5:21–43; (7) Mt. 9:35//Mk 6:6b; (8) Mk 10:1, 5–15//Mk 6:6b–13; (9) Mt. 10:2–4//Mk 3:16–19; (10) Mt. 10:17–22//Mk 13:9–12; (11) Mt. 12:1–14//Mk 2:23–3:6; (12) Mt. 12:15–16//Mk 3:7–12.

Assuming Markan priority, this arrangement of pericopes implies a significantly different compositional procedure on Matthew’s part. To begin with, he initially borrowed from Mark’s Gospel only once between Mk 1:20 and Mk 3:7, that is, at Mt. 4:23 (=Mk 1:39). This means that Matthew initially bypassed a large Markan block as he searched for an appropriate introduction to his Sermon. In turn, this implies that of his twelve transpositions, six were ‘recollections’ of previously omitted materials while five were ‘anticipations’ of later Markan materials. The materials in Matthew 8–12 that have Markan parallels suggest that Matthew quarried from his Markan source both backwards and forwards with no apparent pattern in either direction.

Matthew’s first three transpositions were ‘recollections’ but they were not incorporated in their original relative order. Matthew reverted to Mk 1:22 at Mt. 7:28–29, then further forward to Mk 1:40–45 for Mt. 8:1–4, then back again to Mk 1:29–34 for Mt. 8:14–17. It is true that all three of these passages were culled from a relatively brief Markan section, Mk 1:21–45, but the pattern of dependence is somewhat cumbersome.

After his first three transpositions, Matthew anticipated Mk 4:35–5:20, reverted immediately to Mk 2:1–22, then immediately turned forward to Mk 5:21–43. In short, Mt. 8:23–9:26 is a transposition within a transposition, or a ‘recollection’ sandwiched into a larger ‘anticipation’. Once again, Matthew’s compositional procedure is complicated. Nevertheless, to this point he had quarried two distinct Markan sections, Mk 1:21–3:6 and Mk 4:35–5:43, which suggests that his procedure, even if somewhat awkward, was not haphazard.

Matthew’s compositional procedure in Mt. 9:35–11:1 (and even to the beginning of Matthew 12) is no different on this schema from on Neirynck’s or Huck-Greeven’s. However, Mt. 12:1–16 must be considered another ‘recollection’. Matthew reverted to Mk 2:23–3:12 and in the process duplicated part of Mk 3:7–12.

This particular scheme of parallelization undoubtedly implies a more complicated redactional procedure for Matthew, particularly in the composition of Mt. 8:1–12:21, because he continually had to look backwards and forwards in his Markan source to create a coherent and discernably well-planned opening to his account of Jesus’ public activity.

Different Pericope Arrangements in Relation to the Two-Gospel Hypothesis

On the two-gospel hypothesis, one almost never has to explain a Markan transposition because Mark’s order invariably agrees with that of at least one of the other two gospels. The question, then, is not why Mark chose to transpose one of Matthew’s or Luke’s pericopes, but why he switched from following Matthew’s order to following Luke’s order, and vice versa.

When constructing his Synopsis, Orchard placed the sermons in Mt. 5:1–7:29 and Lk. 6:20–49 in parallel order: ‘When the moment arrived to add Mk to the Mt—Lk. complex, the option was taken to place the two Great Sermons between Mk 3:19 and 3:20’. Working solely with the Gospels of Matthew and Luke, it is reasonable, albeit not imperative, to place Matthew’s and Luke’s sermons in parallel order because of the close correspondence in the order of shared materials (apart from the so-called Golden Rule). Once this decision is made, however, one has no choice about how to match Mark’s Gospel against the two gospels already arranged in relation to one another. If Lk. 6:12–20a is placed alongside Mt. 4:24–5:2, Mk 3:7–19 will automatically take its place alongside Mt. 4:24–5:2 because of Mark’s and Luke’s closely concurrent orders in their early chapters.

Arranging all three gospels on the same page is an academic construct that introduces a false constraint into the historical situation in which the gospels were written. In fact, the second gospel writer’s decisions when using a source did not necessarily impinge on the third writer’s use of the same source. More precisely, Luke’s decision about where to insert his sermon, based on Matthew’s Sermon, need not have influenced Mark at all. Indeed, Mark’s switch to Lk. 4:31 at Mk 1:21 after his close dependence on Mt. 4:18–22 suggests that he deliberately avoided Matthew’s Sermon at Mk 1:20 and only incorporated elements from Mt. 4:23 at Mk 1:39 because he came across a similar summary at Lk. 4:44. One need not imagine that Mark perceived Matthew’s and Luke’s sermons as parallel in order. Indeed, the opposite is more likely because Luke delayed his sermon until well into his narrative, whereas Matthew placed his Sermon on the Mount at the beginning of Jesus’ ministry. On the two-gospel hypothesis, it seems more plausible that Mark followed Matthew’s Gospel fairly closely until Mk 1:20, but then decided to follow Luke’s order at this point because of Matthew’s Sermon. As a result, the position of Matthew’s Sermon relative to Mark’s outline would be after Mk 1:20, even if one wanted to argue that Mark glanced back at Mt. 4:23 when he came across Lk. 4:44. According to this scenario, Mark simply followed Luke’s order until he came to Luke’s sermon, which because of its similarity to Matthew’s Sermon was deemed an appropriate point to resume following Matthew’s order. Consequently, either Neirynck’s or Huck-Greeven’s arrangement of pericopes seems more suitable than Aland’s or Orchard’s for displaying Mark’s alternating agreement between Matthew’s and Luke’s orders.

On the two-gospel hypothesis, Matthew’s Sermon on the Mount should probably be placed after Mk 1:20 because one needs an explanation for Mark’s adherence to Luke’s order rather than Matthew’s within Mk 1:21–3:6. Statistics reveal that Mark’s vocabulary is generally closer to Matthew’s wording than to Luke’s. Assuming that Mark borrowed from the Gospels of Matthew and Luke, one would expect on the basis of this statistical evidence that Mark preferred Matthew’s account to Luke’s and would, where it suited his purpose, follow Matthew’s order of events. Placing Matthew’s Sermon on the Mount opposite Mk 1:21 provides a reason for Mark’s transition from following Mt. 4:18–22 at Mk 1:16–20 to following the sequence of Lk. 4:31–6:11 in Mk 1:21–3:6. Supporting this scenario is that Mark parallels Luke’s order until just before the Sermon on a Level Place in Lk. 6:20–49.

If one places Matthew’s Sermon after Mk 1:39 or Mk 3:7–12, it is difficult to explain why Mark abandoned Matthew’s order at Mk 1:21 unless one argues that Mark noted the similarities between Mt. 5:1–7:29 and Lk. 6:20–49, matched Matthew’s outline to Luke’s with the sermons paralleled in order, resolved to follow Luke’s order until the sermon, then began to follow Matthew’s order after Matthew’s parallel to the incident immediately preceding Lk. 6:12–19, namely, Mt. 12:9–14 (cf. Lk. 6:6–11). Noting that Mt. 12:15–16 is similar to Mt. 4:24–25, Mark wrote Mk 3:7–12 using details from Mt. 4:24–25, Mt. 12:15–16, and Lk. 6:16–19. Mark 3:13–19 was written using the motif of going up the mountain and electing the Twelve. He then drafted his own transition, Mk 3:20–21, to lead into the incident about Jesus’ alleged collusion with the chief of demons in Mt. 12:22–30, after which he was once again in step with Matthew’s narrative sequence.

On Huck-Greeven’s arrangement of pericopes, one must suppose that Mark abandoned Matthew’s order at Mk 1:21 and chose to follow Luke’s sequence until Lk. 6:11, even though Matthew’s order occasionally agrees with Luke’s order in this section, for example, Mt. 4:23, 8:1–4, 9:1–17 and 12:1–14. In short, at the compositional level, Huck-Greeven’s arrangement implies nothing different about Mark’s redactional procedure from Neirynck’s. In either case, Mark used the introductions to the respective sermons as the point at which to discontinue following the order of one gospel and switch to the order of the other. Coincidentally, the point at which Mark first deserted Matthew’s order, Mt. 4:23–25, is similar to the point immediately preceding the pericope that Mark used when he began following Matthew’s order once again, namely, Mt. 12:15–16.


While there is no single, decisive consideration that forces one to choose one position over the other two as the most appropriate point for Matthew’s Sermon on the Mount relative to Mark’s and Luke’s narrative outlines, the following considerations have led me to place it after Mk 1:20 and Lk. 4:30. The first is the absolute sequence of material in each gospel. Mark and Luke each relate a considerable number of the same incidents in the same order at the beginning of Jesus’ public ministry. On the other hand, Matthew has hardly begun his narration of Jesus’ public ministry before he records Jesus’ Sermon on the Mount. Comparatively speaking, Matthew’s Sermon on the Mount occurs where Mark relates Jesus’ activity in the Capernaum synagogue and where Luke records Jesus’ teaching and rejection in Nazareth. As for the transitional summary at Mt. 4:23, Mk 1:39 and Lk. 4:44, Matthew used it to lead into his Sermon on the Mount, whereas Mark and Luke used it to sum up the preceding events in Capernaum.

A second consideration is that by working backward from the point after which Matthew and Mark have no further discrepancies in order, namely, Mt. 14:1 and Mk 6:14, one finds that Mk 3:7–12 is the contextual parallel of Mt. 12:15–16. This suggests that Mt. 4:24–25 is not contextually parallel to Mk 3:7–12, thereby discounting Mk 3:7–12 as the most appropriate Markan location for Matthew’s Sermon on the Mount.

A third consideration is that Matthew’s redactional procedure on the Markan hypothesis is simpler and more straightforward if his Sermon is slotted into the Markan framework after either Mk 1:20 or Mk 1:39 than after Mk 3:7–12. Similarly, on the two-gospel hypothesis, Mark’s alternating agreement with Matthew’s and Luke’s orders is more explicable if Matthew’s Sermon is placed after Mk 1:20.

Finally, the agreement between Neirynck and the two-gospel team that both the Markan hypothesis and Griesbach’s theory of Markan posteriority are ‘well served’ by placing Matthew’s Sermon after Mk 1:20 is a convenient reason for adopting Neirynck’s proposal for the relative position of the Sermon on the Mount in my arrangement of parallel pericopes.

Annotated Table of Parallel Pericopes




Pericope Heading




John the Baptist.

(cf. 14:3–4)

(cf. 6:17–18)


Herod Imprisons John




Jesus is Baptized

(cf. 1:1–17)


[Luke’s Genealogy]




Jesus is Tested




Jesus Returns to Galilee

(cf. 13:53–58)

(cf. 6:1–6)


Rejection at Nazareth



(cf. 5:1–11)

First Disciples Called



Capernaum Synagogue


(cf. 1:39)

(cf. 4:44)

Matthaean Summary


(cf. 6:20–49)

[Sermon on the Mount]


(cf. 1:22)

(cf. 7:1)

Transitional Summary


(cf. 1:40–45)

(cf. 5:12–16)

Jesus Heals a Leper


(cf. 7:2–10)

[Centurion in Capernaum]




Simon’s Mother-in-law




Many Healed at Evening

(cf. 4:23)



Jesus Preaches in Galilee

(cf. 4:18–22)

(cf. 1:16–20)


Miraculous Catch of Fish

(cf. 8:2–4)



Jesus Heals a Leper

8:18, [19–22]

(cf. 4:35)

(cf. 8:22)

[On Following Jesus]


(cf. 4:35–41)

(cf. 8:22–25)

Jesus Stills a Storm


(cf. 5:1–20)

(cf. 8:26–39)

Two Demoniacs




Jesus Heals a Paralytic




Jesus Eats with Sinners




Question about Fasting


(cf. 5:21–43)

(cf. 8:40–56)

Woman Healed, Girl Raised


[Two Blind Men]


[A Dumb Demoniac]


Matthaean Summary


(cf. 3:16–19)

(cf. 6:14–16)

Names of the Twelve


(cf. 6:6–13)

(cf. 9:1–6)

Jesus Instructs the Twelve



(cf. 21:12–19)

Fate of the Disciples




[Largely ‘M’ and Q]




Plucking Grain on Sabbath




Jesus Heals on Sabbath



(cf. 6:17–19)

Jesus Heals Many (Mt./Mk)

(cf. 10:2–4)



Jesus Chooses Twelve

(cf. 12:15–16)

(cf. 3:7–12)


Jesus Heals Many (Lk.)



(cf. 11:14–23)

Jesus and Beelzebul



[Mainly Q-material]



(cf. 8:19–21)

Jesus’ True Family

(cf. 5:1–7:27)


[Luke’s Sermon]

(cf. 8:5–13)


[Centurion at Capernaum]


[Widow’s Son at Nain]

(cf. 11:2–19)


[John and Jesus]

(cf. 26:6–13)

(cf. 14:3–9)


A Woman Anoints Jesus


[Lukan Summary]




Introduction to Parables




Parable of the Sower




Purpose of Parables




Parable of Sower Explained

(cf. 5:15; 10:26)



Parable of the Lamp

(cf. 7:2; 13:12)



Measure for Measure



[Two Seed Parables]



(cf. 13:18–19)

The Mustard Seed


(cf. 13:20–21)






[Parables cont.]

(cf. 12:46–50)

(cf. 3:31–35)


Jesus’ True Family

(cf. 8:18, 23–27)



Jesus Stills a Storm

(cf. 8:28–34)



Gerasene Demoniac

(cf. 9:18–26)



Woman Healed, Girl Raised



(cf. 4:16–30)

Jesus Rejected at Nazareth

(cf. 10:1–15)



Mission of the Twelve




Jesus, Herod and John




Jesus Feeds 5000



Jesus Walks on the Sea



Jesus Heals Many



Unwashed Hands



On What Defiles



A Gentile Woman


7:31–37; 8:1–10

Healings; 4000 Fed



Demand for a Sign



Leaven of Pharisees


[Blind Man]




Peter’s Confession




1st Passion Prediction




On Following Jesus







Descent from Mountain




A Possessed Boy




2nd Passion Prediction


[Temple Tax]




True Greatness



The Strange Exorcist





[‘Little Ones’ cont.]


[On Forgiveness]




Jesus Leaves Galilee


[Luke’s Central Section]



On Divorce




Jesus Blesses Children




On Riches


[Labourers in Vineyard]




3rd Passion Prediction



(cf. 22:24–27)

Zebedee’s Sons




Jesus Heals Blindness




[Parable of the Pounds]




Jesus Rides into Jerusalem


[Lament for Jerusalem]


(cf. 11:15–19)

(cf. 19:45–48)

Jesus Clears Temple (Mt.)



Jesus Curses Fig Tree

(cf. 21:12–17)



Jesus Clears Temple



Withered Fig Tree




Jesus’ Authority


[Parable of Two Sons]




Parable of the Tenants


[Wedding Feast]




Taxes to Caesar?




Sadducees’ Question



(cf. 10:25–28)

Greatest Commandment




Is Messiah David’s Son?




Warning against Scribes


(cf. 11:39–52)

[Further Woes]


(cf. 13:34–35)

[Lament for Jerusalem]



A Widow’s Pittance




Temple’s Destruction




On the End Time




Desolating Sacrilege



False Messiahs




Coming Son of Man



Exhortation to Watch


[Need for Watchfulness]


[Eschatological Parables]


[Teaching in the Temple]




Plot to Kill Jesus



(cf. 7:36–50)

Woman Anoints Jesus




Judas Plots Betrayal




Preparation for Passover




The Last Supper

(cf. 20:24–28)

(cf. 10:41–45)


Dispute Over Precedence




Peter’s Denial Predicted


[Two Swords]




Prayer in Gethsemane




Judas Betrays Jesus


[Young Man Flees]




Jesus is Arrested

(cf. 26:69–75)

(cf. 14:66–72)


Peter Denies Jesus (Lk.)

(cf. 26:67–68)

(cf. 14:65)


Jesus Mocked and Beaten




Jesus Before Sanhedrin



(cf. 22:63–65)

Jesus Mocked and Struck



(cf. 22:56–62)

Peter’s Denial (Mt./Mk)




Jesus Taken to Pilate


[Death of Judas]




Jesus Before Pilate


[Jesus Before Herod]




Jesus is Sentenced



Jesus is Mocked




Simon of Cyrene




Jesus is Crucified




The Death of Jesus




The Burial of Jesus


[Guard at the Tomb]




The Empty Tomb

Part II

The Value of Formal Arguments from Order

Chapter 3

Two Formal Arguments from Order

In his detailed analysis and appraisal of The Revival of the Griesbach Hypothesis, Christopher M. Tuckett drew attention to an important point:

there are at least two quite distinct arguments from order and these should not be confused. There is the argument which appeals to the lack of agreement between Matthew and Luke against Mark. But there is also an argument which appeals to the disagreements in order, and claims that good reasons can be found for the changing of Mark’s order by Matthew and Luke, but not vice versa.

Tuckett was not the first to differentiate between two arguments from order for Markan priority. In 1924, J.F. Springer distinguished between what he called the ‘argument from corroboration’ and the ‘explanatory argument’. As he put it, the argument from corroboration was based on the observation that Mark’s order is almost always supported or corroborated by either Matthew or Luke or both so that they do not agree together against it. On the other hand, the explanatory argument aimed to explain Matthew’s and Luke’s disagreements with Mark’s order as intentional alterations on their part, thereby allowing the critic to interpret their respective orders as based largely on that of Mark’s Gospel.

These two arguments for Markan priority are distinct types of arguments. The first relies entirely on the formal pattern of concordance and divergence in the respective orders of the synoptic gospels. In other words, one’s evidence is derived from a table of synoptic parallels. One does not refer to the content of specific texts, only to the pattern of similarity and dissimilarity between the relative orders of each gospel. In the revised version of his study, ‘The Priority of Mark’, G.M. Styler referred to such arguments as ‘formal arguments’, distinguishing them from arguments based on the detailed evaluation of parallel texts requiring ‘judgments about probabilities’. A formal argument from order, then, is an inference based on the pattern of agreement and disagreement in order of pericopes between the first three gospels without reference to, or comparative evaluation of, the material content of parallel texts. This chapter analyses two formal arguments from order that have been advanced, respectively, for the theory of Markan priority and for Griesbach’s theory of Markan posteriority.

Proponents of both the Markan hypothesis and the two-gospel hypothesis generally agree on the pattern of agreement and disagreement in order between the synoptic gospels, which may be summarized as follows. Where all three gospels share the same material (the triple tradition), they often follow the same relative sequence, relative because each sometimes interrupts the sequence of shared material by inserting unique materials or material shared between only two of the three gospels. Nevertheless, Matthew’s order of triple-tradition material sometimes conflicts with Mark’s and Luke’s orders, and Luke’s order of triple-tradition material sometimes disagrees with Matthew’s and Mark’s orders. Put differently, where Matthew’s order of triple-tradition material differs, Mark’s and Luke’s orders agree; and where Luke’s order of triple-tradition material differs, Matthew’s and Mark’s orders agree. Outside the triple tradition, Matthew and Mark agree in the relative sequence of material shared between them alone, and Mark and Luke agree in the relative sequence of material shared between them alone. But Matthew and Luke often do not agree in the relative sequence of material shared between them alone. Indeed, a significant proportion of the Matthew-Luke double tradition (Q) has no common order.

Johann Jakob Griesbach was among the first to interpret these observations as evidence for one particular theory of synoptic relations:

Mark compiled his whole work (apart from about twenty-four verses which he added from his own sources …) from the works of Matthew and Luke in such a manner that … he retained the order observed by Matthew in such a way, that wherever he forsakes it he sticks to the path of Luke and follows him and the order of his narrative step by step …

In ‘Modern Developments of Griesbach’s Hypothesis’, William R. Farmer noted that ‘the best known argument [from order] is associated with the name of Streeter and is based upon the observation that Matthew and Luke never agree against Mark in the order they give to pericopes’. He then continued:

A second argument from order … is associated with the name of Griesbach and is based on the observation that whenever the order of Mark is not the same as that of Matthew, it follows the order of Luke, i.e. that Mark has no independent chronology. Griesbach held that this is best explained if Mark is third and is dependent for his order on Matthew and Luke. I concur with this view and regard it as a weighty consideration in favour of Mark being third. Griesbach’s point may be simply put this way: all that is needed to understand the order of events in Mark is that given in Matthew and Luke. Mark’s order shows no independence of Matthew and Luke (excepting the single case of his ordering of the cleansing of the temple). This seems explicable only by a conscious effort of Mark to follow the order of Matthew and Luke. Neither Matthew nor Luke could have achieved this alone. They would have had to conspire with one another or find some other way to contrive this chronological neutering of Mark, i.e. robbing his chronological independence. Mark on this view can only be third and must have known Matthew and Luke.

Although Farmer and Tuckett each characterized the formal argument from order for Markan priority as based on the absence of agreement between Matthew and Luke against Mark, this is actually a secondary observation contingent upon the more basic datum that in the so-called Markan tradition, whenever Matthew’s and Mark’s orders disagree, Mark’s and Luke’s orders agree, and whenever Mark’s and Luke’s orders disagree, Matthew’s and Mark’s orders agree. In other words, the absence of agreement in order between Matthew and Luke against Mark in the so-called Markan tradition is a corollary of the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke where all three do not agree in relative order. Though biased in its formulation, B.H. Streeter’s influential statement of the formal argument from order for Markan priority reflects the secondary and contingent nature of the observation about the absence of agreement in order between Matthew and Luke against Mark by noting this last:

The order of incidents in Mark is clearly the more original; for wherever Matthew departs from Mark’s order Luke supports Mark, and whenever Luke departs from Mark, Matthew agrees with Mark. The section Mk. 3:31–35 alone occurs in a different context in each gospel; and there is no case where Matthew and Luke agree together against Mark in a point of arrangement.

K. and S. Lake employed a similar version of this particular inference in their Introduction to the New Testament. Defending Matthew’s and Luke’s dependence on Mark’s Gospel, they wrote:

The method of proving this is complicated in detail but simple in principle. The material common to Matthew, Mark and Luke is arranged in three parallel columns in such a way that when the same story comes in all three gospels the three forms of it are placed side by side. Attention is first paid only to the order in which the paragraphs are arranged and not to their wording. It is then clear that the order of passages found in all three gospels is sometimes Matthew and Mark against Luke and sometimes Luke and Mark against Matthew, but never Matthew and Luke against Mark. This is prima facie evidence that both Matthew and Luke copied Mark.

According to Humphrey Palmer, this type of argument, which relies on ‘the comparative absence of one type of variation among three manuscripts or documents’, has been used in the two fields of text and source criticism. However, such an argument or inference is only valid for determining ‘mediation’ between documents and cannot demonstrate the direction of dependence between them. With regard to the evidence that Streeter and the Lakes adduced, this simply means that in the triple tradition Mark’s Gospel is more closely related to Matthew and Luke than the latter two are to each other.10 Another way to express this is to say that within the triple tradition Matthew and Luke generally agree in order only insofar as they both agree with Mark’s order. However, Mark’s ‘mediating’ role in the triple tradition can be explained by Markan priority, Markan posteriority, or by Mark being secondary to Matthew or Luke and prior to the other. It was not until 1951 that critics generally began to appreciate this fundamental limitation to Streeter’s formal argument from order.

The Refutation of Streeter’ s Formal Argument from Order

The year 1951 is important in the history of synoptic source criticism because it marks the publication of B.C. Butler’s book, The Originality of St Matthew: A Critique of the Two-Document Hypothesis. Butler’s exposure of what he called the ‘Lachmann fallacy’ provoked a reappraisal of Streeter’s formal argument from order for Markan priority. Although Butler did not prove Markan priority to be untenable, he did demonstrate that Streeter’s inference from order was inconclusive. Streeter had argued that the only theory capable of explaining the phenomenon of order was Markan priority. Although Butler did not disprove Streeter’s theory, he was able to show that Streeter’s argument was fallacious.

Butler disclosed the inconclusiveness of Streeter’s argument from order by demonstrating that the particular argument he used was valid only on different presuppositions. The focus of Streeter’s argument was the absence of agreement in order between Matthew and Luke against Mark. Had Streeter assumed that all three synoptic gospels derived from a common source, this datum would imply that Mark’s order of arrangement is probably closest to that of the original source. The reason for this is that two of the synoptists are unlikely to have rearranged the order of their common source in precisely the same ways. If two of the synoptics agree against the order of the third, the two in agreement probably represent the original order of the shared source. Since Matthew and Luke hardly ever agree against the order of Mark, it would have been legitimate for Streeter to infer that Mark’s order most faithfully preserves the order of the (assumed) common source.

However, Streeter abandoned the idea of a common source for the synoptic gospels. Butler pointed out that having abandoned this assumption, Streeter could no longer appeal to the absence of agreement in order between Matthew and Luke against Mark as evidence for one particular source theory. His own conclusion was that ‘Mark is necessarily the connecting-link between Matthew and Luke … but not necessarily the source of more than one of them’. In short, he demonstrated that it is impossible to make definitive judgments about which of the first three gospels is earliest on the inconclusive evidence Streeter had adduced.

Although Butler’s refutation of Streeter’s inference from order is well known and often cited, little attention seems to have been given to H.G. Jameson’s work, The Origin of the Synoptic Gospels: A Revision of the Synoptic Problem. Jameson’s remarks about Streeter’s inference from order are interesting for a number of reasons, not least because, if accurate, they answer a more recent objection to Streeter’s formal argument from order.

Jameson focused on Streeter’s observation that whenever Matthew fails to agree with Mark’s order, Luke agrees with Mark and vice versa, together with the more general statement that Matthew and Luke do not agree together in order against Mark. He concluded that ‘these are really quite insignificant facts’. He perceived that Streeter’s observations could be explained simply by noting that Matthew and Luke disagree with Mark’s order in different contexts. In his view, ‘two familiar facts’ explain Streeter’s observations: (1) that Matthew and Mark after the dislocations of order in the early chapters, agree throughout the rest of their course, and (2) that Luke, when he is following Mark, scarcely ever deserts his order at all except towards the close’.16 Jameson found it odd that Streeter should attach such importance to his observations about pericope order. To his mind, Matthew’s and Luke’s failure to agree against Mark’s order was insignificant because Matthew-Mark disagreements in order and Mark-Luke disagreements in order occur in different contexts. Given this, it is unlikely that Matthew and Luke would differ from Mark’s order at the same point. As Jameson said, ‘It is evidently very unlikely, under these conditions, that variations in order in (1) and (2) should coincide’. In other words, since Matthew’s Gospel differs from Mark’s order only in its early chapters and since Luke’s Gospel differs from Mark’s order primarily in its later chapters, it is unlikely that Matthew and Luke would agree in order against Mark. Jameson’s point was that absence of agreement in order between Matthew and Luke against Mark can only be significant if Matthew and Luke diverge from Mark’s order in the same Markan contexts.

The Neo-Griesbachian Objection to Streeter’s Inference from Order

Recent advocates of the Griesbach or two-gospel hypothesis contend that the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke can only be explained on Markan priority terms by supposing that Matthew and Luke conspired to produce this pattern. A popular way of expressing this objection has been to question the plausibility of two writers working independently from the same source, hence without knowledge of each other’s editorial decisions, yet inadvertently producing the pattern of alternating agreement in order. Farmer originally raised this objection by claiming:

On the Markan hypothesis, within Streeter’s terms, it is possible to explain how Matthew and Luke would sometimes independently reproduce the same order and content for their material through their use of Mark. But this hypothesis would afford no explanation for Luke’s following the order and content of Mark whenever Matthew deviated from Mark, and Matthew’s following the order and content of Mark whenever Luke deviated from Mark. Since on this hypothesis Matthew has no knowledge of what Luke has done, he could not so consistently have supported Mark’s order if he had wanted to, and the same holds true for Luke.

To clarify his point, Farmer added:

The problem of Markan order can be posed this way: it is as if Matthew and Luke each knew what the other was doing, and that each had agreed to support Mark whenever the other departed from Mark. Such concerted action is excluded by the adherents of Markan priority in their insistence that Matthew and Luke were completely independent of one another. Streeter’s statement, ‘The relative order of incidents and sections in Mark is in general supported by both Matthew and Luke; where either of them deserts Mark, the other is usually found supporting him’, was a tour de force’, by which a serious problem for the Markan hypothesis was converted into an argument in behalf of the priority of Mark.

This objection to Streeter’s inference from order, namely, that Matthew and Luke must have conspired to produce the pattern of alternating agreement in order, has been repeated time and again by critics who have abandoned the theory of Markan priority in favour of Markan posteriority. For example, in 1970, at the Pittsburgh Festival on the Gospels, David L. Dungan asked:

How did it happen that the whole of Mark, almost, is covered precisely by one or the other or both Matthew and Luke, acting as if in divinely foreordained harmony and invisible guidance (for, as we all know, neither was aware of what the other was doing)? Just when Luke goes off into a special passage, there appears at Mark’s side faithful Matthew, as if by magic, and just when Matthew suddenly departs on an errand of his own, in the nick of time back comes Luke, as if in response to a providential bath qol. How is it possible?

In his History and Criticism of the Marcan Hypothesis, Hans-Herbert Stoldt asked how Matthew’s and Luke’s alternating ‘accompaniment’ to Mark’s order could be explained on Markan priority terms, then retorted:

It is completely inexplicable—unless through a transcendental contact. What a mysterious understanding would have had to exist between the two for Luke to have known exactly when Matthew stopped accompanying Mark; that he then should have jumped in, at that same moment and without being told, in order to assume the accompaniment in place of Matthew, until after some time he stopped again, to be replaced by Matthew. This would have had to go on, back and forth, throughout the entire Gospel of Mark, from the beginning to the end. What magical events could have caused this repeated exchange of roles, and in uneven sequence and length at that? What utterly enigmatic understanding would have prompted the first and third Evangelists to sense, without knowledge of each other, when the other departed from the narrative sequence of Mark, and what uncanny parapsychological contact could, from time to time, have sent out the magical impetus for them once more to take their turns accompanying Mark? This whole conception is—well, let us just say: scarcely believable.

Harold Riley reiterated this same objection, albeit with less hyperbole:

at every point where Matthew ceases to follow Mark’s order, whether for a short or longer period, Luke continues in it; and wherever Luke ceases to follow Mark’s order, Matthew in his turn continues in it. There is surely an inescapable conclusion to be drawn from this. If Matthew and Luke were dependent on Mark for the order of events, they must have agreed together that they would do this. Without constant collaboration, the result would be quite impossible. That they followed such a course is incredible, and therefore the conclusion cannot be avoided that the hypothesis that they were dependent on Mark cannot be sustained.

According to Frans Neirynck, ‘The “conscious intention” behind the alternation and the impossibility of a “concerted action” of Matthew and Luke is one of the traditional motifs in Griesbachian literature’. Stated in the manner that Farmer, Dungan, Stoldt, Riley and others have objected to the theory of Markan priority based on the pattern of alternating agreement in order,25 this datum seems to cast doubt on the tenability of the Markan hypothesis. One has to ask, however, whether it is really so difficult to imagine Matthew’s and Luke’s independent use of Mark’s Gospel resulting in one of the two generally agreeing with Mark’s order whenever the other diverged from it. In short, is it really so difficult to imagine Matthew and Luke using Mark’s Gospel in such a way that whenever one of them transposed a Markan pericope, the other remained in agreement with Mark’s order at that point? As Streeter observed, ‘Matthew reproduces 90% of the subject matter of Mark in language very largely identical with that of Mark; Luke does the same for rather more than half of Mark’. In other words, Matthew has parallels to nearly the whole of Mark’s Gospel and Luke has parallels to over half of Mark. Given the high percentage of Markan material paralleled in both Matthew and Luke coupled with the high degree of agreement in order in the triple tradition, one is less easily persuaded by recent argumentation from within the two-gospel camp.

Assuming Markan priority, the high level of triple agreement in order can be interpreted to indicate a general tendency on the part of Matthew and Luke to respect Mark’s sequence of incidents. Coupled with the observation that a large proportion of Mark’s content is paralleled in both Matthew and Luke, Matthew’s and Luke’s conservatism vis-à-vis Mark’s order could explain the pattern of alternating agreement in order that advocates of the two-gospel hypothesis find so difficult to explain on Markan priority terms. Stoldt’s remark about the ‘mysterious’ and ‘enigmatic’ understanding demanded by Luke’s ‘jumping in to accompany’ Mark’s order precisely when Matthew abandoned it and vice versa completely ignores the considerable amount of triple-tradition agreement in order. Such a remark could only have weight if threefold agreement in order were the exception rather than the rule. Since on the Markan hypothesis Matthew and Luke generally ‘accompany’ Mark when they utilize his materials, to contend, as Stoldt does, that ‘this [alternating accompaniment] would have had to go on, back and forth, throughout the entire Gospel of Mark, from the beginning to the end’ is to exaggerate. There is certainly a pattern of alternating agreement in order between Matthew-Mark and Mark-Luke, but it is not nearly as extensive as the pattern of threefold agreement in order. Stoldt allowed his rhetoric to hyperinflate without appropriate qualifications. Assuming Markan priority, Matthew and Luke did not need ‘magical impetus’; they simply maintained an editorial policy of generally agreeing with Mark’s order when they used Mark’s materials. When one transposed a Markan pericope, it happened infrequently enough that the other almost always continued to support Mark’s order. This more balanced analysis of the pattern of alternating agreement in order contradicts Riley’s contention that ‘without constant collaboration, the result would be quite impossible’.

Responding to Farmer’s view that the pattern of alternating agreement in order is readily explicable on the theory of Mark’s dependence on Matthew and Luke, but inexplicable on the Markan hypothesis unless Matthew and Luke conspired to produce it, Neirynck replied:

The significance of the phenomenon … may become questionable with a more concrete approach. The basic statement remains the common order Mark-Matthew and Mark-Luke … Emphasis on the alternating support seems to imply that agreements and disagreements with the relative order of Mark are treated as comparable quantities. In fact, the disagreement against Mark is the exception and the absence of concurrence between Matthew and Luke is less surprising than the somewhat misleading formulation ‘whenever the other departs’ may suggest.

In light of Neirynck’s reply to Farmer, Dungan responded to Neirynck, particularly to his remark: ‘the basic statement remains the common order Mark-Matthew and Mark-Luke’. Much of Dungan’s critique is valid, particularly his criticism of Neirynck’s methodological starting point, which Dungan designated the ‘Lachmann Gambit’.30 However, on one point his criticism is actually more applicable to advocates of Mark’s posteriority:

from a Griesbachian point of view, Neirynck still misstates the Neo-Griesbachian approach with regard to the phenomena of order: we look at the combination of triple agreement in order and alternating support (especially the sudden Markan leaps from one side to the other) which together comprise the whole set of data to be accounted for.

Two points may be made about this statement. First, this seems to be the very point Neirynck tried to make when he remarked, ‘the disagreement against Mark is the exception and the absence of concurrence between Matthew and Luke is less surprising than the somewhat misleading formulation “whenever the other departs” may suggest’. What he seemed to recommend is that one should view the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke in conjunction with the large amount of triple-tradition agreement in order. If so, Dungan’s statement reiterated Neirynck’s advice, albeit from a different viewpoint.

A second point about Dungan’s statement is that while he averred that the ‘Neo-Griesbachian approach’ is to ‘look at the combination of triple agreement in order and alternating support … which together comprise the whole set of data to be accounted for’, his own and others’ formal argument from order against Markan priority and in favour of Mark’s posteriority completely ignores the importance of the large amount of triple-tradition agreement in order. One looks in vain among the neo-Griesbachian formal arguments reproduced above for any acknowledgment that the disproportionately large amount of triple-tradition agreement in order is relevant to their concern about how, independently of each other, Matthew and Luke could in turn provide almost constant ‘support’ for Mark’s order of incidents, even when one chose to follow a different order. As another example, while reviewing The Order of the Synoptics by Bernard Orchard and Harold Riley, Dungan commented forcefully on Riley’s formal argument from order, quoted above:

Riley bases his entire discussion on an assertion that has never yet been refuted to my knowledge: ‘at every point where Matthew ceases to follow Mark’s order, whether for a short or longer period, Luke continues in it; and wherever Luke ceases to follow Mark’s order, Matthew in his turn continues in it’ (p. 7; cp. p. 10f.). It is fundamental to observe the only two conclusions that can be drawn from this observation of the literary data: (a) either Mark was written first and then Matthew and Luke agreed to do this, which is absurd (and ruled out by the Two Document Hypothesis) or (b) Mark was written last and used one and then the other successively, thus producing this alternating support between Matthew and Luke respectively. All who think that this statement also supports the bland conclusion that ‘Mark is merely the middle term’ between Matthew and Luke (e.g., Neirynck and Tuckett) are totally mistaken in a fundamental point of logic.

Neither Riley’s observation nor Dungan’s inference based upon it indicates that either was aware of highlighting the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke without reference to the more prominent triple agreement in order. Using a graphic analogy, one can perceive this interrelated evidence in proper perspective only by superimposing the pattern of alternating agreement in order against the backdrop of the extensive agreement in order between all three synoptic gospels.

The objection that the pattern of alternating agreement in order is inexplicable on the Markan hypothesis is essentially an expression of incredulity that Matthew and Luke should have failed to transpose the same Markan pericope. Farmer put it neatly when he wrote, ‘The fact that both Luke and Matthew frequently deviate from Mark, either in order or by omission of Markan material, raises the question of their failure to deviate from Mark’s order or to omit his material more often at the same place than they do’. Two related factors undergird this claim: (1) Farmer contends that Matthew and Luke disagree with Mark’s order at points; and (2) he regards omission of material as similar to, if not an actual instance of, disagreement in order. Taken together, these constitute frequent deviations from Mark by Matthew and Luke. Indeed, his contention that Matthew and Luke frequently deviate from Mark seems to rely on the view that omission is as much a case of order divergence as transposition. In any case, for Farmer, omissions and transpositions of Mark’s material are twin phenomena requiring explanation on Markan priority terms.34

While at this point in The Synoptic Problem Farmer treated the phenomena of order and omission—that is, absence of shared content—together, they are nevertheless distinct phenomena and can be considered independently. Indeed, elsewhere Farmer has reiterated this same basic objection to the Markan hypothesis without reference to the issue of omission.

In his critique of Streeter’s inference from order, Farmer revived Jameson’s criticisms of Streeter’s arguments by reproducing them in an appendix to The Synoptic Problem. His appeal to Jameson’s refutation of Streeter’s inference from order is ironic because Jameson’s observations not only reveal the inconclusiveness of Streeter’s argument, but also neutralize Farmer’s observations about the pattern of alternating agreement in order. The reason for this is that if Jameson’s observations are correct, they nullify any formal argument based solely on the pattern of alternating agreement in order, including the claim by advocates of the two-gospel hypothesis that on the assumption of Markan priority the pattern of alternation requires what Jameson regarded as an ‘absurd supposition’, namely, that Matthew and Luke conspired to produce this pattern.

Although proponents of the two-gospel hypothesis have understood the pattern of alternating agreement to imply that Matthew and Luke must have collaborated to produce this pattern if they both used Mark’s Gospel, Jameson’s observations apply equally to their insistence upon the significance of alternating agreement in order. While he did say that ‘the only thing which could give [Streeter’s observations] any significance would be the absurd supposition that Matthew and Luke had agreed together that one or other of them would always “support” Mark, even if the other deserted him’, this qualification to his overriding judgment that the pattern of alternation is really insignificant is inconsistent with his fundamental observation, namely, that Matthew-Mark disagreements in order occur in the first half of Mark’s Gospel and Mark-Luke disagreements in order occur in Mark’s second half. The only ‘agreement’ that Matthew and Luke would have had to make to account for this phenomenon is that one would disagree with Mark’s order only in his early chapters and the other would disagree with Mark’s order only in his later chapters. One can think of reasons why two writers using the same source might choose, independently of each other, to rearrange the sequence of incidents in that source in different contexts. In any case, it seems unnecessary to claim that Matthew and Luke must have ‘agreed to support Mark whenever the other departed from Mark’ (Farmer) because this objection to the Markan hypothesis suggests that Matthew and Luke had constantly to check with each other to ensure that one didn’t diverge from Mark’s order at the same point as the other.

Assuming for the moment the accuracy of Jameson’s observation, in response to Farmer’s contention that ‘since on [the Markan] hypothesis Matthew has no knowledge of what Luke has done, he could not so consistently have supported Mark’s order if he had wanted to, and the same holds true for Luke’, a Markan priorist has only to point out that the consistent support for Mark’s order by Matthew or Luke or both is the natural result of the high degree of Markan pericopes paralleled in Matthew’s and Luke’s Gospels coupled with the unembarrassing coincidence that Matthew’s disagreements with Mark’s order happen to occur in the first half of Mark’s Gospel, whereas Luke’s disagreements with Mark’s order occur in the second half of Mark. In other words, the pattern of alternating agreement in order is the result of a combination of two facts: first, that so much of Mark’s material is paralleled in either Matthew or Luke or both; and second, that Matthew and Luke did not disagree with Mark’s order in the same parts of Mark. There is no need to appeal to ‘magic’ and ‘divine guidance’ (Dungan), ‘transcendental’ or ‘parapsychological contact’ (Stoldt), or ‘constant collaboration’ (Riley) between Matthew and Luke to account for the alternating agreement in order between Matthew-Mark and Mark-Luke. While neo-Griesbachians have applauded Jameson’s demonstration that Streeter’s formal argument from order was inconclusive, they have failed to see that, if correct, his observations also refute their argument against Markan priority based on alternating agreement in order.39

Re-evaluating the Neo-Griesbachian Formal Argument from Order

Of Mark’s 100 pericopes (excluding Mk 1:1 and Mk 16:9–20), 90 have parallels in Matthew’s Gospel,41 and 75 have parallels in the Gospel of Luke. Given the high percentage of Markan pericopes paralleled in Matthew’s Gospel (90%) coupled with the relative infrequency of Matthew’s disagreements with Mark’s order of shared materials, and given the reasonably high percentage of Markan pericopes that are paralleled in Luke’s Gospel (75%) coupled with the relative infrequency of Luke’s disagreements with Mark’s order of shared materials, it is hardly too coincidental on the Markan hypothesis that when Matthew diverged from Mark’s order, Luke supported Mark’s order, and vice versa. There is no need on the Markan hypothesis to resort to scenarios involving collusion or some kind of mystical contact to explain the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke. This is especially the case if one accepts Jameson’s observation that order differences between Matthew’s and Mark’s Gospels and between Mark’s and Luke’s Gospels occur in different parts of the Gospel of Mark.

It has been noted that although Jameson revealed Streeter’s argument from order to be inconclusive, his observations, if accurate, also rescue Streeter’s inference from Farmer’s objection to it. Jameson’s observations do not confirm Mark’s priority, but they do allow that Markan priority provides one possible explanation of the pattern of similarity and dissimilarity in order between the first three gospels. This is because most of Mark’s materials have parallels in either the Gospel of Matthew or Luke or both, and because disagreements in order between Matthew’s and Mark’s Gospels and those between Mark’s and Luke’s Gospels occur in different contexts. If Jameson was right, it stands to reason that where Matthew disagrees with Mark’s order, that is, in his first 13 chapters, Luke will probably agree with Mark’s order because Luke’s disagreements with Mark’s order occur largely in his later chapters. On the other hand, where Luke disagrees with Mark’s order, that is, in his later chapters, Matthew will probably agree with Mark’s order because he disagrees with Mark’s order only once after Mk 6:13. In response to Farmer’s question—’Since both [Matthew and Luke] frequently desert Mark … and since neither knows what the other is doing, why do not their desertions of Mark coincide more frequently?’—one might appeal to Jameson’s observation to defend the view that it is not difficult to suppose that independently of each other Matthew and Luke happened not to diverge from Mark’s order at the same point.

The obvious question is, if Jameson was correct, why has no Markan priorist used his observations against advocates of the two-gospel hypothesis? Tuckett did appeal to Jameson’s observations to defend his own and Neirynck’s judgment that Farmer’s argument from alternating agreement overestimates Matthew’s and Luke’s respective disagreements with Mark’s order. One would have thought this an easier way for Tuckett to defend the Markan hypothesis than by devising a mathematical model and submitting to statistical analysis the neo-Griesbachian objection ‘that the failure of Matthew and Luke to agree in changing the order of any Markan pericope is too much of a coincidence to be credible’.45 While there are difficulties with his mathematical model, Tuckett attempted to demonstrate that it would have been relatively predictable for Matthew and Luke to transpose a number of Markan pericopes, yet fail to rearrange the same ones. Indeed, according to Tuckett, there is a greater probability that Matthew and Luke would not have transposed any of the same Markan pericopes, which contradicts the expectation by advocates of the two-gospel hypothesis that Matthew and Luke would probably have rearranged at least some of the same Markan pericopes given their respective tendencies to transpose materials from their Markan sequence. Tuckett’s rejoinder to Farmer and others dismisses the suggestion of collusion between Matthew and Luke to account for the pattern of alternating agreement in order. But his complicated rebuttal is superfluous if Jameson’s observations can be corroborated.

If Tuckett could find no use for Jameson’s observations in defence of Mark’s priority, perhaps he disputes their accuracy. This seems to be the case because each pericope that Tuckett regards as a Lukan transposition is from Mark’s first six chapters, the section from which Matthew allegedly made most of his transpositions. Werner Georg Kümmel agrees with Tuckett that after Mk 6:7 ‘Mt. and Lk. practically never deviate from Mk’s sequence’. On the other hand, Neirynck considers five of Luke’s nine alleged transpositions as occurring after Mk 6:13.

Despite disagreements about Luke’s transposition of Markan pericopes after Mk 6:13, Neirynck, Tuckett and Kümmel nevertheless concur that Luke transposed four Markan pericopes before Mk 6:13. Assuming Markan priority, then, there are at least four Lukan transpositions within the same section of Mark’s Gospel from which most of Matthew’s rearrangements were also made. Although it is understandable why Jameson was reluctant to treat these as Lukan transpositions of Markan materials, his basic observation is not accepted by recent advocates of Markan priority. This may explain why defenders of the Markan hypothesis have not appealed to it to rebut Farmer and others.

Where does this leave us? If Jameson was incorrect that order differences between Matthew-Mark and Mark-Luke occur in different parts of Mark’s Gospel, does this mean that Farmer’s question is still a thorn in the side of Markan priorists? Tuckett clearly does not think it problematic that Matthew and Luke failed to transpose any of the same Markan pericopes. But his calculations were based on a questionable mathematical model. Apart from the assumption of random selection of materials, which cannot be assumed in the case of the synoptists, he then only calculated the probability of Matthew’s and Luke’s failure to coincide in transposing 6 and 4 pericopes respectively from the 80 Markan pericopes prior to Mk 14:1, thereby excluding the differences in order between Mark’s and Luke’s later chapters. This reduction of the total ‘sample space’ is suspect because it fails to take the whole of Mark’s narrative into account, from which Matthew and Luke could have transposed any particular pericope.

While I am not confident about the utility of statistical analysis for resolving the synoptic problem, it seems worthwhile to repeat Tuckett’s test, once keeping in view the entire Gospel of Mark and once limiting the parameters to the material before Mk 6:14 because this is the point at which Matthew is often alleged to have ceased transposing materials from their Markan contexts.

According to my subdivision of the synoptic materials, there are 100 pericopes in Mark’s Gospel. On my arrangement of parallel pericopes,53 and assuming Markan priority, Matthew transposed 7 pericopes or sections. (1) Mt. 4:23–5:2; (2) Mt. 8:2–4 (cf. Mk 1:40–45); (3) Mt. 8:18, 23–34 (cf. Mk 4:35–5:20); (4) Mt. 9:18–26 (cf. Mk 5:21–43); (5) Mt. 10:2–4 (cf. Mk 3:16–19); (6) Mt. 10:5–15 (cf. Mk 6:6–13); and (7) Mt. 10:17–22 (cf. Mk 13:9–12). Luke, on the other hand, transposed 13 pericopes: (1) Lk. 3:19–20 (cf. Mk 6:17–18); (2) Lk. 4:16–30 (cf. Mk 6:1–6); (3) Lk. 5:1–11 (cf. Mk 1:16–20); (4) Lk. 6:12–16, 17–19 (cf. Mk 3:7–12, 13–19); (5) Lk. 7:36–50 (cf. Mk 14:3–9); (6) Lk. 8:19–21 (cf. Mk 3:31–35); (7) Lk. 10:25–28 (cf. Mk 12:28–34); (8) Lk. 11:14–23 (cf. Mk 3:22–27); (9) Lk. 13:18–19 (cf. Mk 4:30–32); (10) Lk. 22:21–23 (cf. Mk 14:18–21); (11) Lk. 22:24–30 (cf. Mk 10:41–45); (12) Lk. 22:56–62 (cf. Mk 14:66–72); (13) Lk. 22:63–65 (cf. Mk 14:65). The probability that Matthew and Luke transposed this number of pericopes respectively from Mark’s 100 pericopes but happened not to rearrange the same one is approximately 0.37.

There are on my subdivision of the material 33 pericopes before Mk 6:14. On the Markan hypothesis, Matthew transposed 6 pericopes or sections: (1) Mt. 4:23–5:2 (cf. Mk 1:39; 3:7–8); (2) Mt. 8:2–4 (cf. Mk 1:40–45); (3) Mt. 8:18, 23–34 (cf. Mk 4:35–5:20); (4) Mt. 9:18–26 (cf. Mk 5:21–43); (5) Mt. 10:2–4 (cf. Mk 3:16–19); and (6) Mt. 10:5–15 (cf. Mk 6:6–13). Luke also transposed 6 pericopes: (1) Lk. 4:16–30 (cf. Mk 6:1–6); (2) Lk. 5:1–11 (cf. Mk 1:16–20); (3) Lk. 6:12–16, 17–19 (cf. Mk 3:7–12, 13–19); (4) Lk. 8:19–21 (cf. Mk 3:31–35); (5) Lk. 11:14–23 (cf. Mk 3:22–27); and (6) Lk. 13:18–19 (cf. Mk 4:30–32). The probability of 2 redactors independently transposing 6 pericopes each from a total of 33 but never transposing the same one is approximately 0.27.

One could not be as confident about these percentages as Tuckett was about his original calculations. Even so, Markan priorists would probably experience little discomfort from these results. In each instance the probability of Matthew and Luke making their respective transpositions and failing to relocate the same Markan pericope is relatively high, nearly 40 per cent (0.37) when the whole of Mark is considered and over 25 per cent (0.27) when only Mk 1:1–6:13 is considered. However one manipulates the data, it is difficult to show that Matthew’s and Luke’s failure to transpose the same Markan pericope is especially problematic on the Markan hypothesis. As a further example, Styler also resorted to a statistical argument in response to the neo-Griesbachian contention that, as he put it,

if both Matthew and Luke were drawing on Mark, usually following his sequence, but sometimes changing the order and sometimes making omissions, then, since ex hypothesi they are acting independently, they ought, according to statistical probability, to coincide, at least occasionally, in making an alteration at the same time. But this happens so rarely as to cast serious doubt on this hypothesis. When one ‘alters’ or ‘omits’, the other so regularly ‘supports’ Mark as to defy any reasonable random expectation.

He then wrote, in a footnote:

Any reduction to statistics will be rough, since the identification of separate paragraphs may be arbitrary. On a count based on the table of contents in Huck-Lietzmann’s synopsis from Mk 1:16 to 6:44, the figures are as follows: out of 32 paragraphs in Mark, Matthew has 21 in the Markan order, omits 4, and has 7 in a different order; Luke omits 1, moves 1, and has remote parallels to 3 or 4—and all of these belong to Matthew’s 21 paragraphs which are in Markan order, and none to his 7 (or 11). If their procedure is random, I reckon that the chances against this result are about 10:1. But there is one paragraph in this section (the parable of the seed growing secretly) which both Matthew and Luke omit, as well as many individual sentences. Further, the assumption that Matthew and Luke are making alterations at random is unwarranted. If it is conceded that Matthew has altered the order of seven miracle stories in order to form a compact group—so that this is seen as one deliberate act, and not seven random ones—the statistics cease to be alarming to supporters of the two-document hypothesis.

Despite repeated efforts by advocates of the two-gospel hypothesis to refute Mark’s priority on the basis of the pattern of alternating agreement in order, Markan priorists have not been convinced. This is for good reason, although no defender of Markan priority has stated why in clear and unambiguous terms, namely, that the phenomenon of alternating agreement in order has not been analysed alongside and in relation to the general agreement in order within the triple tradition. In my view, formal arguments based on the pattern of alternating agreement in order or the absence of agreement in order between Matthew and Luke against Mark should be consigned to the scrapheap of ideas. At the formal level, things are as Butler left them.64 One can neither argue for Markan priority nor dismiss the Markan hypothesis solely on the basis of the formal pattern of agreement and disagreement in order between the synoptic gospels. Advocates of Markan priority have learned not to appeal to the absence of agreement in order between Matthew and Luke against Mark as evidence for their hypothesis. Similarly, neo-Griesbachians should discontinue raising the question whether Matthew and Luke could transpose as many Markan pericopes as they allegedly did on the Markan hypothesis without coinciding in their redactional decisions of this specific type.

Chapter 4

The Phenomenon of Correlation

Although an inference based on the relative absence of agreement in order between Matthew and Luke against Mark played a prominent role in the history of the Markan hypothesis, it was shown to be inconclusive by B.C. Butler. In the previous chapter I argued that the argument from alternating agreement in order between Matthew-Mark and Mark-Luke, which features in recent argumentation by advocates of the two-gospel hypothesis, is also inconclusive. However, a third formal argument has not been adequately assessed in the scholarly literature, which is surprising because this particular argument appears capable of tipping the scales one way or the other in favour of Mark’s priority or posteriority.

The phrase ‘phenomenon of correlation’ is used here as shorthand for what some perceive to be a positive correlation between agreement in order and agreement in wording in two pairs of the synoptic gospels. In his ‘New Introduction to the [Synoptic] Problem’, William R. Farmer drew attention to three literary phenomena that in his view were more readily explained by Mark’s dependence on Matthew and Luke than by Markan priority. The first was the phenomenon of order and the second was the collective weight of the so-called ‘minor agreements’ of Matthew and Luke against Mark. The third was the phenomenon of correlation:

There is a third literary phenomenon which has seldom been noted, but which is also more readily explicable when Mark is third than in any other position. This is the strange positive correlation of order and degree of similarity between Matthew and Mark on the one hand and Luke and Mark on the other. That is, Mark tends to agree more closely with Matthew when they follow an order different from Luke, but more closely with Luke when they follow an order different from Matthew.

Farmer noted that this phenomenon was particularly noticeable in the first half of Mark’s Gospel. He then pointed out that this correlation between order and wording was difficult to explain on any theory of synoptic relations other than Griesbach’s. It is especially difficult to explain on the Markan hypothesis, he claimed, because of Matthew’s and Luke’s alleged independent use of Mark’s Gospel (or Ur-Markus). According to Farmer,

For since Matthew had no knowledge of Luke’s redactional use of Mark, there is no way he could have known to begin copying the text of Mark more closely where Luke’s order was different from that of Mark. Conversely, there is no way in which Luke could have known to begin copying the text of Mark more closely at the point where Mark’s order and that of Matthew departed from one another.

In short, if Matthew and Luke borrowed from Mark independently of each other, neither would know to copy Mark’s text more closely at the point where the other had decided to transpose or omit a Markan pericope or section.

On the other hand, there is no difficulty explaining this phenomenon if Mark was dependent on the Gospels of Matthew and Luke. As Farmer observed, it would have been natural for Mark to adhere more closely to the text of the gospel whose order he was following at the time. He also noted that Mark need not have adopted this procedure without ever deviating from it. ‘Indeed’, as Farmer conceded, ‘the phenomenon is ambiguous enough to indicate that if in fact Mark was third, he did not follow this pattern with absolute consistency’. Despite this ambiguity, however, he considered the positive correlation between agreement in order and the level of shared vocabulary to be significant enough to undermine the Markan hypothesis and support Mark’s posteriority.

Assuming that a positive correlation does exist between agreement in order and verbal correspondence, one can certainly explain this phenomenon more easily by Mark’s dependence on the Gospels of Matthew and Luke than on the Markan hypothesis. The question is whether there is such a correlation. Although Farmer referred his reader to the final chapter of his book for ‘the specific confirmation of this thesis’, his discussion is largely impressionistic. He provided no statistical confirmation that a significant correlation exists between agreement in order and agreement in wording. While Farmer did not think it an easy task to verify the phenomenon of correlation, it is surely worth the effort to try. My aim in this chapter is to verify or falsify Farmer’s phenomenon of correlation.7

Although the phenomenon of correlation has not received the attention it properly deserves, G.M. Styler did refer to it briefly in the revised and expanded version of his study, ‘The Priority of Mark’. After concluding his summary of Mark’s alleged compositional procedure on Griesbach’s hypothesis by noting that ‘his language tends to be closer to that of the source whose order he is following when there has been a divergence’, Styler commented in a footnote:

Farmer … claims that this is a recognizable trend, but no more. If it is established that there is such a trend it must be noticed that it is no embarrassment to Marcan priorists: it could well be due to a tendency on the part of (e.g.) Matthew to depart further than usual from Mark’s text when he abandons Mark’s order. Marcan priorists would indeed be embarrassed if it were shown that (e.g.) the language of Mark and Luke were regularly closer in those passages where Matthew has a deviant order than in those passages where all three have the same sequence. I have examined [Bruno] de Solages’ figures for passages in both classes, and cannot detect any trend in this ’embarrassing’ direction.

One wonders why, on the Markan hypothesis, Matthew and Luke would be inclined to depart further from Mark’s wording when they chose to transpose a Markan pericope, but the crucial point is Styler’s concession that it would be difficult, assuming Markan priority, to explain a consistently closer correspondence in wording between Matthew-Mark or Mark-Luke when the third gospel in either case disagrees in narrative sequence.

There are two ways of examining the data to see whether a positive correlation exists between agreement in order and agreement in wording when either Matthew or Luke diverges from the common sequence. One is to examine the material in the form of a running commentary on the pattern of sequential parallelism and verbal correspondence. This is the kind of verification Farmer attempted in his chapter on Mark’s redaction of his materials, and it has been supplemented by Harold Riley in two publications.11 Although this method of corroborating the phenomenon of correlation is imprecise, circular and subjective, it is a necessary corrective to statistical analysis of the data. Before attempting a statistical analysis of the data, then, Farmer’s and Riley’s comments need to be evaluated.

Farmer’s and Riley’s Attempts to Verify the Phenomenon of Correlation

According to Farmer, Mk 1:1–20 is primarily dependent on Mt. 3:1–4:22, even though Mark conflated Luke’s text with Matthew’s in Mk 1:2–4 and 7–13. Mark 1:4 and 7 actually have more in common with Luke’s parallels than with Matthew’s, but at this point Farmer’s observations about a closer correspondence between Mark’s and Matthew’s texts are accurate for the most part, particularly in Mk 1:5–6, 9–11 and 13c–20. In a later context, however, Farmer noted:

Whenever Mark comes to a series of passages in Matthew and Luke where they both have the same literary units in the same order, his text does not tend to be uniformly closer to that of one of his predecessors than it is to that of the other. In other words, Mark’s text tends to be closer to that of the Gospel whose order he is following, only when the other Gospel has the same material in quite another order.

Given this proviso, it is not within Mk 1:1–20 that one has the opportunity to examine the phenomenon of correlation, but in the next section, Mk 1:21–3:19, where Mark allegedly followed Luke’s order of events. Although Farmer did not contend that Mark’s text is always closer to Luke’s in this section, he did state that ‘Mark’s text is significantly closer to that of Luke in Mark 1:29–31; 1:40–45; 2:1–12; 2:23–28’. He said the same for Mk 3:1–6. Do the data support these claims?

Using the N-A26 text, of the 44 words in Mk 1:29–31, 17 are unique to Mark’s text and 7 are common to all three gospels in identical form. A further six words occur also in Mt. 8:14–15 in identical form, but two of these are the word καὶ and a further two, ὁπυρετός, are also shared with Lk. 4:39, except that Luke uses the dative rather than nominative case. In short, the only significant detail shared only between Mark’s and Matthew’s accounts of the healing of Simon Peter’s mother-in-law is that Jesus is reported to have taken hold of her hands. In addition to the words shared between all three in identical form, Mark also shares 11 words in identical form with Luke’s text and a further one word in a different case. Of these 12 words, one also appears in Matthew’s text in a different case. In sum, Matthew and Mark share a total of 13 words in identical form and a further 5 in a different form, while Mark and Luke share 18 words in identical form and a further 6 in a different form. Mark’s and Luke’s texts are closer, but one cannot say they are significantly closer than Mark’s account is to Matthew’s.

Of the 99 words in Mk 1:40–45, 41 are unique and 32 are shared between all three gospels in identical form. Of the remaining 26 words, 5 are common to all three, although 2 have the same form in Matthew and Mark and one has the same form in Mark and Luke. Of the remaining 21 words, Mark shares 4 words in identical form with Matthew and a further 3 in dissimilar form, whereas he shares 10 words in identical form with Luke and a further 4 in dissimilar form. In total, Mark shares 38 words in identical form and 6 words in dissimilar form with Matthew, while he shares 43 words in identical form and 8 words in dissimilar form with Luke. Again, Mark is closer to Luke than to Matthew, but not significantly closer.

Although the slightly larger number of verbal correspondences between Mark and Luke in this passage is not significant, one could argue that the nature of their correspondences is significant. Only Mark and Luke speak of the leprosy departing from the leper, although Mark also adds that the leper was cleansed, which is how Matthew described the cure.16 In addition, only Mark and Luke record the words περὶ τοῦ καθαρισμοῦ σου in Jesus’ instructions to the healed leper. However, Mark and Luke used 99 and 98 words respectively, in contrast to Matthew’s 52 words. Furthermore, it is in Mk 1:45 and Lk. 5:15–16, which have no parallel in Matthew’s account, that a number of verbal and thematic parallels occur, particularly the message about Jesus, his presence in the wilderness and his popularity with the people.

Of the 196 words in Mk 2:1–12, 81 are unique and 52 are shared between all three gospels in identical form. A further 17 are also common to all three, 10 in the same form in Matthew and Mark and 3 in the same form in Mark and Luke. Of the remaining 46 words, Mark shares 13 in identical form with Matthew and a further 4 in a dissimilar form, while he shares 20 words in identical form with Luke and a further 9 in dissimilar form. Mark therefore shares with Matthew a total of 75 words in identical form and a further 11 in a different form, and he shares with Luke a total of 75 words in identical form and a further 23 in dissimilar form. Mark is certainly closer to Luke than to Matthew, but is he significantly closer to Luke? One notes that only Mark and Luke record in almost identical wording the query, ‘Who is able to forgive sins except God alone?’ Mark’s and Luke’s accounts also conclude similarly in contrast to Matthew’s account. On balance, it is difficult to attach source-critical significance to Mark’s and Luke’s unique similarities.

Of the 108 words in Mk 2:23–28, 29 are unique and 42 are shared between all three gospels in identical form. A further 17 are also common to all three, although 6 of these have the same form in Matthew and Mark and 6 have the same form in Mark and Luke. Of the remaining 20 words, Mark shares 5 with Matthew in identical form and one in dissimilar form, whereas he shares 12 words with Luke in identical form. In sum, Mark shares with Matthew 53 words in identical form and 12 in dissimilar form, whereas he shares with Luke 62 words in identical form and 11 in dissimilar form.

Of the 95 words in Mk 3:1–6, 26 are unique and 15 are shared between all three gospels in identical form. Twelve more also occur in all three gospels, with eight of these having the same form in Matthew and Mark. Of the 42 remaining words, Mark shares 16 in identical form with Matthew and one more in a different form, whereas he shares with Luke 22 words in identical form and a further 4 in dissimilar form. Altogether, then, Mark shares 39 words with Matthew in identical form and 5 in a different form, while with Luke he shares a total of 37 words in identical form and 15 in dissimilar form. In terms of identical wording, Mark’s text is fairly evenly matched by both Matthew’s and Luke’s parallels, but Mark is somewhat closer to Luke if the total number of common words is emphasized, irrespective of word form. It is noteworthy that Mark’s and Luke’s texts are significantly close in wording at two particular points, neither of which has a Matthaean parallel: Mk 3:3//Lk. 6:8 and Mk 3:4b–5a//Lk. 6:9b–10a. On the other hand, Mk 3:6 has no word in common with Lk. 6:11, but is remarkably close to the wording of Mt. 12:14.

The table below summarizes verbal correspondence in these five pericopes:


Total Words in Common:




Mk 1:29–31 (44 words)



Mk 1:40–45 (99 words)



Mk 2:1–12 (196 words)



Mk 2:23–28 (108 words)



Mk 3:1–6 (95 words)



Mark is clearly closer to Luke’s text than to Matthew’s in each of these pericopes.

Moving on to Mk 4:35–5:43, Farmer first drew attention to Matthew’s parallels to the three pericopes in this section of Mark’s Gospel. He then asserted:

It is clear that these Matthean parallels have influenced the wording of Mark’s text. But it is also unmistakably clear that the text of Mark in this section is significantly closer to the text of Luke than to the text of Matthew. This is particularly so in Mk 5:1–20, the story of the Gerasene Demoniac, and 5:21–45, the story of Jairus’ daughter and the woman with the hemorrhage. This is a very clear example of the phenomenon of a positive correlation between agreement in order and agreement in wording …

More recently, C.S. Mann made a similar observation:

In the ensuing selection of miracle stories [i.e. Mk 4:35–5:43], it is clear that Mark took three units in succession from Luke 8:22–56. While there are Matthean parallels which have had some influence on Mark’s text, it is far closer to the Lukan examples—especially in the accounts of the demonpossessed, Jairus’ daughter, and the woman with a hemorrhage. In other words two features of conflation between two texts were at work: the use of order in one source, coupled with textual use of another.

Using Mk 5:1–20 as an example, of the 325 words in this pericope, 146 are unique and 47 are shared with both Matthew and Luke in identical form. A further 9 words appear in all three gospels, with 6 of these sharing exactly the same form in Mark’s and Luke’s texts. Of the remaining 123 words, Mark shares with Matthew 14 in identical form and 5 in a different form, whereas he shares with Luke 77 words in identical form and 26 in dissimilar form. There is clearly a significantly closer correspondence between Mark and Luke than between Matthew and Mark. However, the Markan priorist can retort that Matthew’s account is severely abbreviated, comprising only 135 words, and that Luke simply narrated this incident using nearly as many words (293) as Mark. The same observation applies also to the following pericope, Mk 5:21–45.

Mark 6:1–13 and parallels is a fascinating section because the phenomenon of alternating agreement in order between Matthew-Mark and Mark-Luke occurs twice in quick succession. At Mk 6:1, the agreement in order between Mk 4:35–5:43 and Lk. 8:22–56 ceases and agreement in order between Mt. 13:53–58 and Mk 6:1–6 begins, but this section of Matthew-Mark agreement in order ends at Mk 6:6 and a brief section of Mark-Luke agreement in order begins (Mk 6:7–13; Lk. 9:1–6). On Griesbach’s hypothesis of Markan posteriority, Mark makes two abrupt transitions between his two principal sources in this brief section—from Luke to Matthew at Mk 6:1, and from Matthew back to Luke at Mk 6:6. According to Farmer, ‘The close agreement between Mark and Matthew at this point [Mk 6:1–6; Mt. 13:53–58] is to be compared with Mark’s close agreement with the text of Luke in the preceding passages, where in following Luke’ s order he had to deal with Matthean parallels which were in quite a different order’.

An important question is whether Matthew’s verbal parallels to Mk 6:1–6 are more numerous than in his parallels to the three pericopes in Mk 4:35–5:43. As the table below reveals, the proportion of words in Mt. 13:53–58 with Markan parallels is higher than in any of his parallels to Mk 4:35–5:43, albeit only marginally so in Mt. 8:18, 23–27, Matthew’s parallel to Mk 4:35–41.


Total in Matthew

Common Words


Mk 4:35–41 (118 words)




Mk 5:1–20 (325 words)




Mk 5:21–43 (373 words)




Mk 6:1–6 (125 words)




One should note that Mt. 13:53–58 uses a significantly higher proportion of the number of words used in Mk 6:1–6 than in any of the pericopes in Mk 4:35–5:43. Mark’s version of Jesus’ rejection in Nazareth is relatively brief, however, so it would have been difficult to abbreviate it much further. Once again, the length of Matthew’s version of this incident has a bearing on how one interprets the evidence.

Perhaps the decisive factor for Markan priorists is that in relation to this instance of alternating agreement in order, they have a ready explanation for Luke’s rearrangement of this episode, irrespective of whether he expanded Mark’s text or used another version of the story. Nevertheless, the correspondence in wording between Matthew and Mark in this instance is significantly higher than in places where Matthew and Mark do not share the same relative sequence.

In Riley’s overview of Mark’s alleged dependence on the Gospels of Matthew and Luke, he attached particular significance to the sequential parallelism between Mk 1:21–3:19 and Lk. 4:31–6:19. After discussing the higher level of verbal correspondence between Mark’s and Luke’s texts in a number of parallel pericopes in this section (for example, Mk 2:1–12, 5:1–20 and 5:21–45), he remarked:

If any one of the narratives of Mk 1:21–3:19 is considered in isolation, it can be argued that Luke is dependent on Mark, not Mark on Luke; but what calls for explanation is the fact that the closeness of their narratives is associated with their being in the same sequence. Markan priority gives no answer to this; it is Mark’s dependence on Luke and Matthew that is confirmed.

It is true that in all but one of the pericopes in this section of Mark’s Gospel, namely, Mk 1:32–34, there is greater verbal correspondence between Mark and Luke than between Matthew and Mark. The Markan priorist may respond, however, in at least two ways. First, where there is a Matthaean parallel to Mark’s and Luke’s texts in this section, it is usually considerably briefer. The exceptions are Mt. 9:9–13, which has only one word less than Lk. 5:27–32, and Mt. 12:1–8, which is longer than either Mk 2:23–28 or Lk. 6:1–5 because of a unique Matthaean inclusion, Mt. 12:5–7. Where Matthew abbreviated Mark’s account, assuming Markan priority, and Luke was nearly as expansive as Mark (and sometimes more so), one would expect a higher level of verbal correspondence between Mark and Luke than between Matthew and Mark. In contrast to Riley, Markan priorists will stress Matthew’s brevity rather than sequential parallelism to explain the higher correlation between Mark’s and Luke’s wording in this section.

A second response to the alleged significance of the phenomenon of correlation is Styler’s contention that even if there is a positive correlation between agreement in order and agreement in wording, it could be the result of a tendency on Matthew’s or Luke’s part to deviate more than usual from Mark’s wording when rearranging the relative sequence of Mark’s incidents. Why this might be so was not explained, even though there is no obvious reason why this should be so. Taken together, however, these two responses reduce the probative force of Farmer’s and Riley’s observations.

A Statistical Test of the Phenomenon of Correlation

A more objective test of the phenomenon of correlation must be carried out, and Styler indicated how this might be done. In ‘The Priority of Mark’ he conceded that ‘Markan priorists would indeed be embarrassed if it were shown that (e.g.) the language of Mark and Luke were regularly closer in those passages where Matthew has a deviant order than in those passages where all three have the same sequence’. Following Styler’s lead, I have devised the following statistical test of the phenomenon of correlation:

Step 1    is to establish a statistical profile of the correspondence in wording between Matthew-Mark and Luke-Mark, initially restricting attention to those parts of the synoptic gospels in which all three agree in order.

Step 2    is to establish a statistical profile of the correlation in wording between Matthew and Mark when Luke disagrees in order, and between Luke and Mark when Matthew disagrees in order.

Step 3    is to establish a statistical profile of the correlation in wording in both the Matthew-Mark twofold tradition and the Mark-Luke twofold tradition.

If the Matthew-Mark correspondence in wording or the Luke-Mark correspondence in wording in either Step 2 or Step 3 is significantly closer than in the triple tradition where all three gospels share the same narrative sequence (Step 1), this would be inexplicable on the Markan hypothesis. The reason why a significantly higher verbal correspondence between Matthew-Mark or Mark-Luke in Steps 2 and 3 than in Step 1 would be damaging to the Markan hypothesis is that Luke’s divergence from Mark’s order, either by transposition or omission, should not affect Matthew’s use of Mark’s wording. Likewise, Matthew’s transposition or omission of a Markan pericope should have no observable impact on Luke’s reliance on Mark’s vocabulary. Of course, one must allow for coincidence, but coincidence could not explain a consistently higher verbal correlation between Matthew-Mark and Mark-Luke in Steps 2 and 3.

Farmer was right to warn about the difficulty of testing the phenomenon of correlation. A preliminary difficulty is that this phenomenon can only be tested on the basis of agreement about what pericopes are ‘out of order’, and as C.M. Tuckett has pointed out, ‘the precise nature of the disagreements in order … cannot be uniquely specified’. This statistical test is based on my own table of synoptic parallels,27 which was constructed with a clear understanding of the many difficulties involved.

Another difficulty emerges as soon as one begins to quantify the data. For example, should one count all the words in Matthew’s and Luke’s parallels to Mk 1:12–13? If so, this and other similar decisions will have a significant impact on the total profile of the proportion of Matthew’s and Luke’s words that correspond to Mark’s vocabulary in the triple tradition where all three gospels agree in order. Likewise, should one include all of Matthew’s fulfilment quotations from the Hebrew Bible? Assuming Mark’s priority, Matthew obviously added these to his Markan source rather than deriving them from it.

Because of this difficulty, I have chosen to conduct the statistical test twice. In the first test, I include all words in Matthew’s and Luke’s parallels to Mark’s pericopes, except for a small number of parallels that seem clearly atypical. In the second test, I restrict Matthew’s and Luke’s materials to verses or, in some cases, part-verses that correspond to Markan material.29

Judgments about whether triple-tradition materials are in order or otherwise are based on my arrangement of parallel pericopes. The statistics are based on the N-A26 text as found in the thirteenth edition of Synopsis Quattuor Evangeliorum, edited by Kurt Aland. In the tables that display the statistical data,31 the first column denotes the pericope number in Synopsis Quattuor Evangeliorum (SQE). The second and fourth columns display the total number of words in Matthew’s and Luke’s pericopes respectively. The third and fifth columns display, respectively, the number of words in Matthew’s and Luke’s pericopes that correspond to Mark’s wording. In each instance, the first figure denotes the number of words that are identical in form to Mark’s wording, while the second figure is the number of words shared with Mark but not in exactly the same form. The sum of the two figures gives the total number of words that Matthew or Luke shares with Mark in a particular pericope, hence the plus sign (+) between them.

In Test 1, Step 1, the correspondence in wording between Matthew and Mark ranges between 0.48 and 0.57, with the lower end of the range representing exact verbal agreement. The correspondence in wording between Luke and Mark is 0.40–0.48.33 In Step 2, the verbal correspondence between Matthew and Mark is 0.41–0.51, somewhat less than in Step 1, but the verbal correspondence between Luke and Mark is 0.40–0.52, which is identical at the lower end of the range and slightly higher at the upper end. In Step 3, the correspondence in wording between Matthew and Mark is 0.40–0.48, which means that the upper end of the range in Step 3 is the same as the lower end of the range in Step 1. The correspondence in wording between Luke and Mark in Step 3 is 0.45–0.55, which is a full five per cent higher at each end of the range than the verbal correspondence between these two gospels in Step 1. Set out in tabular form, the results of Test 1 are as follows:




Step 1



Step 2



Step 3



In Steps 2 and 3 of Test 1, the verbal correspondence between Matthew and Mark is lower than in Step 1, but the reverse is true between Luke and Mark. Advocates of the two-gospel hypothesis will find little support in the statistics for Matthew-Mark, but they might view the statistics for Luke-Mark as corroborating their theory. On the other hand, proponents of the Markan hypothesis might argue that the statistical results for the two pairs of gospels cancel each other out because they reveal a certain capriciousness on Mark’s part if he borrowed from the Gospels of Matthew and Luke. In short, why would Mark adhere more closely than usual to Luke’s text when he was following Luke’s order alone, but less closely than usual to Matthew’s text when he was following Matthew’s order alone, especially since he generally adhered more closely to Matthew’s text than to Luke’s text (cf. Step 1)?

In Test 2, Step 1, the verbal correspondence is 0.54–0.63 for Matthew and Mark and 0.44–0.54 for Luke and Mark. In Step 2, the correspondence between Matthew and Mark is 0.51–0.63, which is hardly different from the range in Step 1, whereas the correspondence between Luke and Mark is nearly five per cent lower than for Step 1, 0.39–0.50. In Step 3, the correspondence between Matthew and Mark is 0.49–0.60, which is lower than for Step 1, and the correspondence between Luke and Mark is 0.45–0.55, which is only one per cent higher at each end of the range than for Step 1. In tabular form, the results of Test 2 are as follows:




Step 1



Step 2



Step 3



Styler noted that ‘Markan priorists would indeed be embarrassed if it were shown that (e.g.) the language of Mark and Luke were regularly closer in those passages where Matthew has a deviant order than in those passages where all three have the same sequence’. Neither Step 2 of Test 2 nor Step 3 causes embarrassment to Markan priorists because neither shows that Matthew’s and Mark’s vocabulary is significantly closer when Luke either disagrees in order or does not share the material; nor does either step show that Luke’s and Mark’s vocabulary is significantly closer when Matthew either disagrees in order or does not share the material. Although Test 2 is more precise because it excludes unique material in Matthew’s and Luke’s parallels to Markan pericopes, the results are less conclusive than those for Test 1.

In view of the difficulties associated with devising an objective test of the phenomenon of correlation coupled with the inconclusiveness of the results, it seems prudent not to attach source-critical significance to this phenomenon.

Conclusion and Preview

Thus far, I have reviewed and evaluated three formal arguments: (1) the argument for Markan priority based on the absence of agreement in order between Matthew and Luke against Mark; (2) the argument for Markan posteriority based on the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke; and (3) the argument for Markan posteriority based on the phenomenon of correlation between agreement in order and agreement in wording. Each is inconclusive and therefore cannot discriminate between the relative merits of the Markan hypothesis and the two-gospel hypothesis. Butler was right that Streeter’s inference from order was fallacious, but Tuckett is also right that ‘if the three gospels are directly related to each other, then any argument from order … is in itself, at the purely formal level, logically inconclusive’.

If formal arguments from order are inconclusive, this does not imply that the phenomenon of order is insignificant. Indeed, the pattern of agreement and disagreement in order between the first three gospels may well be important if one employs a different type of argument. In the introduction to the commentary on the Gospel of Matthew in the ICC series, W.D. Davies and Dale C. Allison write:

The problem with many versions of the argument from order is that they remain abstract. In order to break through the impasse in the discussion, one must inquire which is more credible, the reworking of Mark’s order by Matthew and Luke or the reworking of Matthew and Luke by Mark.

In the remaining section of this study, I inquire which of these two theories, Markan priority or Markan posteriority, provides the more credible explanation for the phenomenon of order by using what I call a compositional argument from order. In contrast to formal arguments from order, a compositional argument from order aims to provide plausible reasons for disagreements in order by comparing the relevant pericopes and their parallels in their respective contexts and judging, on redaction-critical grounds, which of the synoptists is most likely to be responsible for such disagreements—whether on the Markan hypothesis by Matthew and Luke rearranging or transposing Mark’s pericopes, or on the two-gospel hypothesis by Mark alternating between the two partially different sequences of Matthew and Luke.42 While extensive agreement in the order of shared materials may indicate a literary relation between documents, it cannot indicate the direction of dependence between them. The specific nature of the relation between two or more interrelated documents is more likely to be ascertained by attending to disagreements in order within the broader pattern of agreement in narrative sequence. Consequently, in relation to the synoptic gospels, compositional arguments from order naturally focus on disagreements rather than the more extensive agreements in order. As E.W. Lummis pointed out, ‘No dependence-theory … can be considered satisfactory unless it either accounts for the order in the dependent gospel or at least discloses some reason for the disturbance of order, and indicates some principle of arrangement in the secondary document’.

In this connection, the remainder of this book may be regarded as an attempt to evaluate the ongoing discussion initiated by Frans Neirynck and David L. Dungan at the Jerusalem Symposium on the Interrelations of the Gospels in 1984. In his overview of ‘The Two-Source Hypothesis’, Neirynck remarked:

The argument from order for Markan priority is nothing more, and nothing less, than the demonstration that the differences of the order in Matthew and Luke receive a plausible explanation as changes of Mark which are consistent with the general redactional tendencies and the compositional purposes of each gospel. It is clear that in this area, as in any other area of the synoptic problem, our method should be a joint effort of source criticism and redaction criticism (or composition criticism).

Subsequently, in discussing ‘Matthew 4:23–5:2 and the Matthean Composition of 4:23–11:1’, he attempted to demonstrate that Matthew’s rearrangements of Mark’s order are limited to the section Mt. 4:23–11:1 and are explicable on redactional and compositional grounds. In response to Neirynck’s source-critical discussion of Mt. 4:23–5:2, Dungan and the two-gospel team presented an alternative analysis of the same synoptic texts discussed by Neirynck in which they provided ‘a very brief redactional/theological analysis to indicate Mark’s motivation for including or excluding material from his sources at each step along the way’.

A difficulty one encounters when using a compositional argument from order to determine whether Mark’s Gospel is prior or posterior to the Gospels of Matthew and Luke is that these two theories differ in their assumptions about the compositional procedures of the secondary evangelists. This means that the reasons given for Matthew’s and Luke’s transpositions on the Markan hypothesis may be of a different kind from those given for Mark’s transitions between the Gospels of Matthew and Luke on the two-gospel hypothesis. This makes it difficult to compare and to evaluate the relative strengths and weaknesses of the respective arguments for each theory. As a result, the chapters in Part III occur deliberately in four successive steps. First, to help to avoid anachronistic judgments about how the synoptists went about the task of writing their gospels, Chapter 5 examines compositional conventions at the time when the gospels were written. Chapter 6 examines three Lukan pericopes that occur in different relative contexts from their parallels in the Gospels of Matthew and Mark to ascertain whether they are best understood as Luke’ s modifications to the relative order followed by Matthew and Mark. Evidence suggestive of Lukan dependence on the Gospel of Matthew is also discussed. Chapter 7 examines disagreements in order between Matthew and Mark and evaluates source- and redaction-critical arguments that have been offered to explain them. Chapter 8 reviews the recent history of the two-gospel hypothesis and evaluates reasons for Mark’s alleged compositional procedure in various contexts where he is alleged to have ceased following the order of one gospel for that of the other. In this way, the following hypotheses are tested: Luke’s dependence on Mark; Luke’s dependence on Matthew; Matthew’s dependence on Mark; and Mark’s dependence on Matthew and Luke.

Part III

Compositional Arguments from Order: A Reappraisal

Chapter 5

Compositional Conventions in the First Century ce

In this chapter my purpose is to examine various factors relating to the use of source material at the time when the Gospels of Matthew, Mark and Luke were written, which is generally conceded to have been during the second half of the first century of the Common Era. These factors include ancient conventions of reading and writing, the writing equipment used and the compositional habits that a writer probably learned, particularly with regard to ways in which existing materials were used in producing a new work. To begin with, however, one needs to be aware of what one critic has described as the ‘oral environment’ of the first century ce.

The ‘Oral Environment’ of the First Century ce

In an insightful study, ‘Omne Verbum Sonat: The New Testament and the Oral Environment of Late Western Antiquity’, Paul J. Achtemeier stressed the significance of what he considered to be a neglected feature in much New Testament research, namely, that the culture within which most, if not all, of the New Testament writings were produced and disseminated was one of ‘high residual orality which nevertheless communicated significantly by means of literary creations’. In his discussion, he borrowed from Walter Ong the definition of ‘residual orality’ as ‘habits of thought and expression … deriving from the dominance of the oral as a medium in a given culture’.

Achtemeier focused both on the impact of a predominantly oral environment on the writing and reading of first-century documents in general and on verbal cues in New Testament texts that enabled readers and hearers to understand their structure and organization. Although he referred to other related issues, he was primarily concerned to understand the environment in which the New Testament was written and to illuminate, by reference to specific New Testament texts, how documents assisted readers and hearers to understand their structure and organization by the presence of oral/aural ‘signposts’.

Achtemeier first reviewed how documents were produced and read in the Hellenistic era. Citing Papias’s and Seneca’s preference for oral over written records, he stressed that ‘ancient culture remained committed to the spoken word’. He also commented on conventions of writing that contributed to this preference:

In addition to a cultural bias in favor of the oral over the written, the sheer physical nature of the written page in classical antiquity militated against its ease of reading and in that way also contributed to the culture’s reliance on the oral mode in communication. The written page consisted entirely of lines each containing a similar number of letters, lines that ended and began irrespective of the words themselves. Documents were written without systematic punctuation, without indications of sentence or paragraph structure, indeed without separation of the letters into individual words. As a result, no visible indications presented themselves to the ancient readers that would have rendered them aid in their attempt to discern the structure, and hence the meaning, of the piece of literature they confronted.

Although well aware of the variety and wide distribution of written materials during the period when the New Testament was written, Achtemeier remarked, ‘The existence of such wide varieties of written material … should not mislead us with respect to the essential orality of that culture, an orality demonstrated both in the manner by which literature was produced and in the manner in which it was read. Both were predominantly, indeed exclusively, oral’. By this he meant that written composition involved dictation, even when writing one’s own material. We do not normally describe the act of writing in one’s own hand as dictation, but according to Achtemeier,

the oral environment was so pervasive that no writing occurred that was not vocalized. That is obvious in the case of dictation, but it was also true in the case of writing in one’s own hand. Even in that endeavor, the words were simultaneously spoken as they were committed to writing, whether one wrote one’s own words or copied those of another … In the last analysis, dictation was the only means of writing; it was only a question of whether one dictated to another or to oneself.

Similarly, reading was also generally done aloud. The wealthy often had a slave to read to them and writers often had their works ‘read’ by reading them aloud at a public gathering or by having someone else read them aloud to an audience. Referring to the episode in Acts 8:26–39, in which Philip heard the Ethiopian reading from the book of Isaiah, and to Augustine’s speculations about the reasons for Ambrose’s seemingly unique habit of reading silently, Achtemeier ventured the view that even when the ancient reader read alone, he or she invariably read aloud: ‘Reading was therefore oral performance whenever it occurred and in whatever circumstances. Late antiquity knew nothing of the “silent, solitary reader”.’

Achtemeier’s contention that ‘late antiquity knew nothing of the “silent, solitary reader” ‘ has not gone unchallenged. Frank D. Gilliard has demonstrated that the practice of silent reading in antiquity was not as unusual as Achtemeier claimed. He also contested Achtemeier’s claim that ‘no writing occurred that was not vocalized’, but this did not lead him to repudiate Achtemeier’s principal contention about the predominantly ‘oral environment’ in which the New Testament documents were written and originally read.

Apart from the question of whether reading was generally done aloud or in silence, Achtemeier noted that whether one read alone or to a group, the task of reading was difficult. A major reason for the difficulty was that the format of written texts ‘conveyed virtually no information about the organization and development of the content it intended to convey’. Of course, as he noted, the purpose of a document is to convey information. Hence, first-century writers had to find alternatives to visual markers within their manuscripts both to organize the material and to assist in conveying their intended meaning. ‘The alternative to visual structuring of a manuscript to indicate organization of meaning is to include oral indications of structure within the material’.13 Examples of oral/aural techniques include repetition, introductory and concluding formulae, and inclusio, in which a literary segment begins and ends with a similar formula. Because such techniques were meant to be heard, methods of organizing written materials were based primarily on sound rather than sight.

Based on his analysis of the predominantly oral environment in which the New Testament documents were written, Achtemeier concluded that these documents are ‘oral to the core’ and that ‘to be understood, the New Testament must be understood as speech’. After providing some examples of oral structural indicators in both the gospels and New Testament epistles, he noted some implications for the way New Testament critics go about their work, each of which impinges on the practice of source criticism. His observations have potentially far-reaching impact:

One wonders if it can so quickly be assumed that where there are discrepancies or inconsistencies in a Gospel or a letter, it is the result of the combination of divergent written sources. It may well be the case that such inconsistencies are the result of the need to provide oral/aural clues to the one who listens to the document. Of course the New Testament documents were written down, but they were written, and would be read … in a way far different from that to which we are accustomed, and much closer to an oral than to a print environment. It may well be the case that the inconsistencies one can find, say, in the Gospel of Mark are more likely to be due to the orality of that document, and hence the need to provide oral clues for its understanding, than to its author’s combination of various written sources.

Achtemeier illustrated his point by noting that in Mk 4:1–2 Jesus gets into a boat to avoid being crushed while teaching the crowds; the boat is ignored at the end of the parable of the sower in Mk 4:9 and the switch of scene in Mk 4:10, where Jesus is alone with his disciples; however, Jesus is assumed to be in the boat ‘on the evening of that same day’ in Mk 4:35. According to Achtemeier, the reference to the boat in Mk 4:35–36 may well be an oral inclusio that indicates the conclusion of the parabolic section begun in Mk 4:1–2. If he is right, Mark sacrificed narrative consistency to help his readers and hearers understand the thematic organization of his Gospel.18

Another implication of the essentially oral environment in which the gospels and other New Testament writings were written is that writers depended much more on memory than we are wont to do when quoting from a written source. This may have been partly because of the awkwardness of locating a particular passage within a roll. Passages could be marked, but, as Achtemeier noted, without markings ‘one would have to recognize at each point in the search where one was in relation to the desired passage in order to find it. The intimate knowledge of a writing that was required to make location of a specific passage possible would thus virtually obviate the need to do so’.20 The practice of direct copying was widespread; this was how most books were published in multiple copies. Obviously, a gospel writer, for example, could copy directly from available sources, but this seems not to have been standard practice for authors of new works.22 On the other hand, an author who dictated to a scribe was free to quote directly as he or she read. However, according to Achtemeier, ‘In light of the pervasive orality of the environment, and the physical nature of written documents, references were therefore much more likely to be quoted from memory than to be copied from a source’.

Achtemeier’s views are consistent with those of others who have explored how ancient books were written and read. In particular, the features he identified as being part and parcel of a predominantly oral environment—including audible dictation in the process of writing, even when transcribing oneself; the standard practice of reading aloud even when alone; the importance of memory, which was cultivated as part of one’s education; the absence of slavish dependence on written sources; and the presence of oral/aural markers in written documents—help to guard against anachronistic views about how the gospels were written.

Writing Equipment and Compositional Habits in Antiquity

Although the first century was a predominantly ‘oral environment’, reading and writing were widespread. According to Colin H. Roberts, ‘The world into which Christianity was born was, if not literary, literate to a remarkable degree; in the Near East in the first century of our era writing was an essential accompaniment of life at almost all levels …’ George Kennedy concurs: ‘The New Testament could not have been written at a time of greater literacy, education, or understanding’,26 But what were the social, cultural and physical constraints within which a first-century writer wrote? And what were the writer’s tools and materials?

In The Griesbach Hypothesis and Redaction Criticism, Sherman E. Johnson stressed the paucity of our information about such matters and implicitly warned against over-confident conclusions:

Since we work with written documents, we have to use literary methods as far as they will take us. Source criticism, form criticism, and redaction criticism have proved to be essential tools for reconstructing the story of Jesus and of primitive Christianity. But there are variables that make many conclusions tentative. We do not know just how a later evangelist would rewrite an earlier gospel. Did he have the convenience of a codex or did he have to unroll a scroll from time to time? And did he always look at a written text or might he sometimes have depended on his memory of it? There is the possibility, too, that at some points he depended on an oral tradition known to him.

These observations are sagacious; there is much about the ancient world we do not know, including certain details about how authors went about their work. T.C. Skeat, an authority on ancient manuscripts, warned that ‘few subjects are more obscure than the methods of ancient book-production’. Johnson’s and Skeat’s observations cannot be gainsaid. However, one suspects that in New Testament research in general and in gospel criticism in particular, scholars have not paid sufficient attention to the information we do have about ancient writers and their habits, which is not negligible. Indeed, based on available evidence, it is possible to answer at least one of Johnson’s questions with reasonable confidence.

In 1911, William Sanday published a study on ‘The Conditions under which the Gospels Were Written, in their Bearing upon Some Difficulties of the Synoptic Problem’. His concern was to strengthen the case for the two-document hypothesis as much as to explore how the synoptists went about the task of writing their gospels. Indeed, he opened his essay by admitting, ‘We assume what is commonly known as the “Two-Document Hypothesis” ‘.30 A little later he advised:

the main purpose of the present essay is to suggest that in the particular direction which I am going to follow is to be found the simplest and most satisfactory solution of a group of difficulties which on a comparison of the Three [Synoptic] Gospels are raised by the points in which they differ.

Despite Sanday’s bias, his essay contains valuable information on some of the variables and physical constraints associated with the writing of the gospels.

In the first part of his essay, Sanday examined the various differences between the first three gospels. He provided examples of the same or similar words used in different senses, having a different reference or being spoken by different people. He also noted instances of material appearing in different forms, for example, in direct speech in one gospel and as part of the narrative in another, or as a question in one gospel and as a statement in another. He emphasized instances in which, for example, Mk 10:18 reads, ‘Why do you call me good?’ while the parallel in Mt. 19:17 reads, ‘Why do you ask me about the good?’ Finally, he referred to transpositions of phrases within sentences, transpositions within sections and transpositions of whole sections. His general remark was that such illustrations indicate the freedom with which the synoptists reworked their material as they incorporated it into their own narratives. He then reiterated his working assumption that ‘in all the cases in which Mk is involved we believe his version to be the original, and the variants in the other Gospels are deviations from the original’.

Having said this, however, Sanday alerted his reader to two phenomena that are difficult to reconcile with the theory of Markan priority: secondary features in Mark’s Gospel, including Matthew’s and Luke’s agreements against Mark in triple-tradition material, and Matthew’s and Luke’s omission of Markan materials, particularly Luke’s omission of Mk 6:45–8:26. Sanday’s discussion of ‘the conditions under which the Gospels were written’ was clearly no disinterested investigation; rather, it was aimed at reconciling these anomalous data with the two-document hypothesis. Indeed, ‘Towards a Resolution of Difficulties for the Two-Document Hypothesis’ might have been a more accurate subtitle for Sanday’s essay than the subtitle he actually used.

Sanday’s first point, under the heading ‘Psychological Conditions’, was that the synoptists were not copyists but historians. He did not mean to categorize the gospel writers as historians of the same calibre as other historians of the period. Rather, his point was to underscore the relative freedom of the gospel writers vis-à-vis their sources. According to Sanday,

[This freedom] shows us the Evangelists, not as painfully transcribing the older texts on which they relied (such as Mk and Q), or feeling themselves in any way called upon to reproduce them verbally, but as setting to work in a spirit independent and yet on the whole faithful, not punctilious and yet not wilfully capricious and erratic, content to tell their story very much as it came, sometimes in the words of their predecessors and sometimes in their own.

Sanday then proceeded to discuss the ‘external conditions’, namely, physical aspects of first-century life, that impinged on the task of writing. He rightly warned against anachronistic ideas about how ancient authors wrote their books. We are unlikely to imagine first-century writers sitting down to computers or word-processors, but it is difficult not to envisage them setting their source materials on a desk or table so that they were within arm’s reach for consultation. Tables were available in the first century, but according to Sanday they were used for eating and paying out money, or as stands for vases and statues, not for writing. Desks were also available, but they had a small surface area and were nothing like our school or office desks. In Sanday’s words, ‘Sometimes the writer sits at such a desk, more especially in the later examples from the fourth century onwards, when the codex, or book proper, had superseded the roll. But in the earlier examples the writer is usually represented with the roll open simply upon his knees’.

Sanday’s judgment that tables and desks were not normally used for writing has not been refuted. The papyrologist Eric G. Turner opined that ‘the Greek scribes no doubt at first sat cross-legged like the Egyptian, supporting the writing surface on the garment tightly stretched across their knees’. Skeat advised against building too much on scanty data from antiquity, but nevertheless remarked that ‘there is virtually no evidence for the use of chairs, tables or desks, and it would appear that the scribe sat on a stool or even on the ground and rested the section of the roll on which he was writing on his knee, holding the remainder of the roll with his free hand’. Bruce Metzger has confirmed this judgment, noting also that for less extensive tasks, scribes wrote while standing:

It is a fact … that a variety of evidence supports the conclusion that in antiquity scribes were not accustomed to write on tables or desks. On the contrary, an accumulation of artistic, archaeological, and literary evidence indicates that when a scribe was making relatively brief notes on a wax tablet or on a sheet of papyrus or parchment, he would usually stand and write while holding the writing material in his left hand. When a scribe had a more extensive task, such as the copying of a rather lengthy manuscript, he would sit, occasionally on the ground but more often on a stool or bench, supporting the scroll or codex on his knees, which were sometimes raised the higher by the use of a footstool or dais under the scribe’s feet.

The physical nature of books during the time when the gospels were produced also suggests that writing habits differed markedly from our own. It is sometimes suggested that Christians may have used the codex form from the outset. For example, Kurt and Barbara Aland asserted, ‘All the literature of the period was written on scrolls (including Jewish literature, with the sole qualification that leather was used for the Holy Scriptures); yet apparently from the very beginning Christians did not use the scroll format for their writings, but rather the codex’. However, most scholars think that the gospels were first written on rolls. This was certainly the view of Sanday, who wrote of the gospels:

They were rolls, and rather lengthy rolls, with the writing in short vertical columns across them, as a rule less than a foot high. They were therefore rather cumbrous, and not quite easy to keep open at a particular place … There are many representations of a writer or student making use of books (i.e. of rolls); but to the best of my belief these are always, or almost always, contained in a sort of round canister (capsa) or square box (scrinium) which stands upon the ground … Under such conditions it is not at all likely that the roll would be taken out and referred to more often than could be helped.

Cognizance of the conditions under which the synoptists probably wrote undoubtedly takes us some way towards appreciating how they created their own narratives out of existing written materials. Sanday’s remarks are apposite:

A modern … would have the document he was using constantly under his eye. There would be hardly any interval of time between the perusal of its text and the reproduction of it in writing … With the ancient writer it would be otherwise. He would not have his copy before him, but would consult it from time to time. He would not follow it clause by clause and phrase by phrase, but would probably read through a whole paragraph at once, and trust to his memory to convey the substance of it safely from the one book to the other.

Sanday rightly warned against applying rigid rules to the way that books were used or produced, and he urged that allowances be made for first-century writers’ idiosyncracies. Much of what is possible for an author now was possible for an author then, even if more difficult in some respects. For example, it may not have been customary to use weights to keep a scroll open to a particular section, but it was possible. Ancient copyists had to keep scrolls open at a particular place so that they could copy what was in the book they were copying. The synoptists were not copyists, as Sanday observed, but they were capable of using techniques employed by copyists, even if they adapted those techniques to their own needs.

After considering the physical constraints under which first-century writers worked, Sanday sought to explain two of the difficulties for the Markan hypothesis mentioned earlier by recourse to some of these ‘conditions’. The first of these was the number of Matthew-Luke agreements against Mark in triple-tradition material. Sanday accepted C.H. Turner’s suggestion that some of these agreements may be the result of independent yet obvious corrections of Mark’s style and phraseology, while others may simply be present because we have not yet recovered the ‘true text’ of either Matthew’s or Luke’s Gospel. He also recommended that some of these agreements against Mark’s text are the result of overlaps between Mark and Q. In his view, however, the majority of Matthew-Luke agreements against Mark are the result of Matthew’s and Luke’s use of a recension of the text of Mark’s Gospel that is different from the text from which the extant manuscripts of the Gospel of Mark derive. In other words, the copies of Mark’s Gospel used by Matthew and Luke were probably from a different ‘line of descent’ than those copies that are extant and form the basis of the canonical text of Mark’s Gospel. Neither line of descent was free from errors and alterations, of course, but the ‘family’ of copies of Mark’s Gospel no longer represented in the textual tradition was influenced by a copyist who ‘corrected’ and ‘improved’ his or her copy of Mark’s Gospel before it was (or its descendants were) used by Matthew and Luke. In Sanday’s words,

We can form two interesting inferences as to the divergent families or lines of descent derived from St. Mark’s autograph. One is … that the parent of our extant authorities was very near to the autograph, and represents it closely. The other is, that on the line of perhaps four or five copies intervening between St. Mark’s autograph and the copies used by Mt. Lk. one at least must have been the work of a person with literary tastes and habits, who did not hesitate to improve the text before him and make it more correct and classical. This process of improvement went so far that I have ventured to call it a ‘recension’.

Sanday seems to have thought that by demonstrating that the synoptists were not slavish copyists, but authors who exhibited freedom in relation to their sources, he could legitimately explain how errors, alterations, ‘corrections’ and ‘improvements’ found their way into one line of copies of Mark’s Gospel. He seems also to have blurred the distinction between copyists and authors. Authors may well have relied upon memory as they recorded the substance of a passage they had recently read, but that is not how a copyist worked. Something like what he suggested may have occurred, but it does not necessarily follow from his discussion of ‘the conditions (whether psychological or external) under which the Gospels were written’.

More plausible is his discussion of Luke’s reason for allegedly omitting Mk 6:45–8:26. In the main, Sanday concurred with the reasons put forward by John Caesar Hawkins later in the same volume, provided Luke found himself in a position where he was forced to omit something. Sanday considered that one of the physical constraints within which first-century authors laboured was the length of papyrus rolls, which came in more or less standard lengths. Since Luke’s writings are among the longest in the New Testament, he reasoned that Luke was forced to omit some of his materials to fit his Gospel on a standard size roll. ‘I have little doubt’, he wrote, ‘that St. Luke was conscious of being pressed for space, and that he felt obliged to economize his materials’.

Sanday concluded by writing, ‘The essay will have served its purpose if it enables any of its readers to form for themselves a more exact conception of the processes which gave shape to the Gospels as we have them, and of the influence of various kinds to which they were due’. Although his purpose was explicitly more biased than this final statement indicates, Sanday’s crucial insight was that before making judgments about how the synoptists made use of earlier materials in the writing of their gospels, an indispensable first step is to examine the physical constraints within which first-century writers worked, what habits and customs they were likely to follow and what techniques they are likely to have learned.

Compositional Conventions

In The Making of Luke—Acts, Henry Joel Cadbury remarked, ‘Few writers and fewer of their readers realize how much the composition of books is determined by group habits. No writing is the result of free and untrammeled choice. It is a process hemmed in with the compulsions of convention’. While discussing how classical authors made use of materials at their disposal, Cadbury emphasized their unwillingness to name sources, their tendency to paraphrase rather than to quote directly and their inclination to borrow from one source at a time.49 However, he was reluctant to conclude that all authors used the same methods:

The methods of several authors differ, and even one author, as Josephus shows us, handles some parts of his material much more freely than other parts. Sometimes a sentence or section of the source was read and recast; sometimes (and this on the whole is the method of Luke) the sentence structure and even most of the wording were retained. In the former case exchange of synonyms is less striking than in the latter, but in neither case must the change be considered to have been always deliberate, or the substitution always significant. Even changes which look like conscious improvement in style or diction are not like the proof-reader’s blue pencilings, but come naturally in the process of paraphrase.

At the 1977 Colloquy on the Relationships among the Gospels, George Kennedy further advanced our understanding of compositional conventions in antiquity by commenting on the writing habits of Plutarch and Suetonius, two biographers roughly contemporaneous with the gospel writers: ‘Plutarch and Suetonius are important for the criticism of the gospels primarily as a standard of comparison. They illustrate what constituted learned and literary biography of the times and how the materials were assembled’. His observations on Plutarch are of particular interest for the light they shed on an ancient author’s methods of research and composition:

Recent studies have made clear that, even though he read widely and deeply in both Greek and Latin, he should not be regarded as a systematic researcher. He read works through and took notes on whatever interested him; he did not ordinarily take a topic and pursue it back through references in a variety of sources. The elder Pliny and other ancient scholars probably worked in a similar way. When they were ready to write systematically, they used their memories and their notes, only occasionally going back to the original. One reason for this was that published works were usually cumbersome scrolls, while notes were often in the more convenient form of a codex.

Kennedy noted that Suetonius worked in a similar fashion, with the important distinction that he occasionally included verbatim quotations from Julius Caesar and Augustus taken from notes—’which, unlike Plutarch’s, were arranged topically rather than chronologically’—based on official documents that were available to him as a staff-member of the emperor Hadrian.

Cadbury’s and Kennedy’s observations have largely been confirmed in a series of studies by F. Gerald Downing, who has attempted to present as full a picture as possible of how first-century writers worked, particularly when incorporating material from pre-existing sources. His purpose has been to refute the view that an apparently complex procedure may yet account for the composition of the gospels and thus cannot be discounted simply because we are unable to make sense of it. More specifically, in a manner reminiscent of Sanday, his intention has been to demonstrate that what he regards as the relatively simple compositional procedure envisaged by the two-source theory is most consistent with what can be determined about compositional conventions in the first century. According to Downing,

We are in a position to tell with a considerable degree of certainty what compositional procedures for making use of existing writings would have been readily available in the first century. We can tell on the basis of many examples of practice and some indications of theory: even the most highly literate and sophisticated writers employ relatively simple approaches to their ‘sources’. We are also able to take into account the sort of compositional exercises people are likely to have been taken through in their elementary education in Greek.

For Downing, such evidence favours the two-source hypothesis over competing theories, which in his view envisage more complex compositional procedures.

Downing has not suppressed his partiality for the two-source theory. As a result, one cannot help but wonder whether data are ignored because they do not support his preferred source theory. Nevertheless, his willingness to pursue this line of research is commendable. As he noted in ‘Compositional Conventions and the Synoptic Problem’, ‘the long debate on the sources of the Synoptic Gospels seems to have been conducted without paying much or any attention to this issue of whether any indications of “sensible” compositional procedures in the first century ce are available’. Also significant in his view is what is known about elementary education during the Hellenistic period, particularly from examples of written exercises: ‘The procedures are always so similar that it would be absurd to suppose without massive supporting evidence that the New Testament evangelists could have learned to write Greek and cope with written source material at all while remaining outside the pervasive influence of these common steps toward literacy’.58

Despite difficulties associated with comparing the literary procedures of the synoptists with those of classical authors, it is a worthwhile exercise even if such comparisons provide only partial guidance to the synoptists’ use of sources and method(s) of composition. As Sharon Lea Mattila has observed, ‘Taken in the abstract, any type of compositional method could be proposed—and this is precisely where the major problem lies. Clearly, it is imperative to introduce parameters into source-critical analysis, to establish “the boundaries between reasonable certainty and ingenious speculation”.’60

Referring to some of the more sophisticated authors of ancient times, including Livy, Plutarch, Hieronymus of Cardia, and Lucian, Downing maintained that literary composition in antiquity was a relatively uncomplicated procedure. For example, when using two sources, Livy generally alternated between them:

Mostly Livy will alternate intact but paraphrased blocks of Polybius with blocks of his (now lost!) Roman source, ‘now composing with great care and concentration, now adapting the source in front of him rapidly and mechanically’. Such conflation of two parallel accounts as does appear is occasional and ad hoc, lacking … any sign of any prior analytical ‘unpicking’.

Regarding Plutarch, Downing quoted D.S. Russell as saying, ‘The process of composition is likely to have involved much less “paper-work” than a modern scholar likes to think’. He also relied on studies of Plutarch’s writing habits by C.B.R. Pelling.64 According to Pelling, Plutarch had access to a variety of sources but borrowed largely from a single source, which was typical: ‘Time and again, we find Greek and Roman historians claiming a wide range of reading, and deserving to be believed; yet, time and again, we find them demonstrably basing their narrative of individual episodes on a single source’.

Downing also quoted Pelling’s sketch of an ancient writer at work:

A writer would not normally refer back to [earlier] reading to verify individual references, and would instead rely on his memory, or on the briefest of notes … Stray facts and additions would be recalled from the preliminary available reading, but it would be a very different matter to recall the detail of an episode’s presentation … Such a procedure seems less perverse in view of the physical difficulties of working with papyrus rolls … [with] non-existent or rudimentary … indexing, chapter-headings, line- and column-numbering … Even if, for example, a slave held a second roll for an author to compare accounts, or the author himself used a book-rest, combining versions would still be awkward.

He then stated his view of the implication of this evidence:

Even had one of our evangelists wanted to emulate the well-staffed and well-equipped compositional procedures of a sophisticated literary figure, nothing would have suggested that he should begin by analysing his source material, nor on that or any other basis that he should plan some complex conflation of his sources. Only the ‘Streeterian’ model fits at all in the first century.

Downing then turned his attention to the Jewish historian Josephus, whose use of Jewish Scripture ‘looks very like the procedures suggested by the classical scholars cited …’ In an earlier study of parallels between Josephus’s use of Scripture in his Antiquities and the redactional tendencies of the synoptists, Downing sought to show that Josephus’s ‘stated aims and many of his more obvious tendencies are closely paralleled in Luke among the synoptists’. Among Josephus’s redactional tendencies, Downing noted that he felt free: (1) to omit (for various reasons, including discrepancies, repetition, narrative continuity, the desire to avoid, where possible, accounts of miraculous or magical events and what Downing called ‘inappropriate theology’ and ‘the apologetically awkward’); (2) to add details and especially speeches; (3) to rearrange his materials ‘to create a fresh order of events, sometimes for the sake of coherence, sometimes simply to allow the narrative to flow’;70 (4) to assemble together similar types of material or material relating to a single theme; and finally, (5) to conflate. Concerning conflation, Downing claimed that where Josephus’s sources largely agreed, he ‘happily follows’, although he might combine ‘minor divergent items’ or choose between them on the criteria of ‘overall harmony, piety, moral uplift, apologetic impact, and so on’. Where his sources conflicted more seriously, Josephus followed what he took to be the older and fuller source. When accounts of the same series of events in his sources conflicted in detail, Josephus wrote his own version of the events in question, only borrowing some items from each source.72 In short,

The keynote of Josephus’ method is … ‘simplicity’, and simplicity seems to be a major part of his aim. Where his sources are straightforward he is happy just to paraphrase; where a single source seems illogical, he tidies it up; and if he has two sources that will not readily combine, he makes up a third account of his own, blithely ignoring large parts of both. But it remains a ‘version’, quite clearly. There is no major invention, no major allusiveness.

Downing has not questioned the general consensus that there is a literary relation between the synoptic gospels. In his view,

All the studies so far cited would still warrant the judgment that the Synoptic Gospels display a ‘literary’ relationship to one another. Matthew and Luke might just possibly have been relying on memory (as Russell and Pelling both suggested Plutarch might). But that would have to have amounted to a recalling of Mark as a memory available only in one form, a finely memorised document, to allow for the often very close reproduction of one synoptic writer by another, both in wording in distinguishable passages and in the order of such passages. Memory would not seem impossible; but, for the frequent verbatim and near-verbatim identity that obtains, the presence of the actual document before the dependent writer is still more likely.

He made a complementary point in his earlier study on Josephus. After noting that Josephus rarely borrowed word for word, he asserted:

It is not the divergencies among the synoptics (or even between them and John) in parallel contexts that are remarkable; it is the extraordinary extent of verbal similarities. The question is, Why were they content to copy so much? rather than, Why did they bother to change this or that? The procedure is not however, mechanical, and there are considerable divergencies.

Convinced that the relation between the synoptic gospels is literary, Downing focused upon conflation as decisive for determining the relative probability of a source theory. In his view, every literary source theory implies some conflation on the part of the secondary writers:

every attempt at a ‘literary’ explanation of the relationship between the first three Gospels does … imply the somewhat unusual and not a little adventuresome process of conflation of some kind. Either Mark has complicatedly part unpicked and then (re-)conflated Matthew and Luke (Griesbach), or Luke has complicatedly unpicked and (re-)conflated Mark and Matthew (Farrer); or Matthew and Luke in a much more simple fashion, without any ‘unpicking’, have independently conflated Mark and ‘Q’.

Downing’s point is that the more complex the conflation, the more unlikely the procedure. He hinted at which of the competing hypotheses listed are improbable on this criterion both by using the awkward term ‘complicatedly’ and by describing the respective dependent authors’ literary procedure as ‘unpicking’ and ‘(re-)conflating’. He did not clarify what he meant by these terms, but what he seems to have meant by the term ‘unpick’ is to analyse one’s source(s) to determine what comes from one source and what comes from another. On both Griesbach’s and Farrer’s hypotheses, the synoptist who wrote third would recognize the probable use of one of the earlier gospels by the author of the other. According to Downing, on Griesbach’s hypothesis, Mark first established what Luke borrowed from Matthew and then conflated material from both Matthew and Luke, but in doing so omitted much of what they have in common. On Farrer’s theory, Luke first determined what Matthew borrowed from Mark and then conflated both Matthew and Mark, omitting in the process some of the material he found in both Matthew and Mark.

Downing provided an example of what he meant by ‘unpicking’ in his two-part study on Josephus and the synoptists. After reviewing Luke’s Gospel on the assumption that Luke conflated Mark and Q, and urging the plausibility of Luke’s reliance on these sources using compositional procedures similar to those used by Josephus, Downing turned his attention to the relative plausibility of Farrer’s and Griesbach’s hypotheses. On Farrer’s theory, Luke occasionally came upon passages where Matthew copied Mark almost verbatim. Although Luke sometimes copied Mark closely and more frequently copied Matthew closely, in those places where Matthew provided what Downing called a ‘ready-made conflation’ by copying Mark closely, Luke often failed to use that material. On Griesbach’s hypothesis, Mark clearly omitted numerous passages in the Gospels of Matthew and Luke that are closely similar in wording. In Downing’s view,

The near identity of these passages with each other would make it easy for Mark to excise them; but it is hard to imagine why he should; and in sequences such as the Baptism and Temptation, the Beelzebul Controversy, the Mission Addresses, and the Apocalypses, it is as difficult to imagine how he could, as it was for Luke [on Farrer’s theory]. This imagined Mark, intent on conflation, rejects (as too easy?) every pericope where an all-but total conflation is there ready for him; and on finding sequences where close parallels are mixed with divergent and with quite distinct matter goes to great pains to pick out the latter, sometimes to copy it exactly.

For Downing, therefore, to ‘unpick’ is to ‘disentangle’ or ‘unravel’80 sources within a source, that is, to engage in source criticism. His contention that Luke (on Farrer’s theory) and Mark (on Griesbach’s theory) must first have ‘unpicked’ the sources within their sources was based on what he regarded as an unlikely omission of materials, which on either theory was common to both source gospels.

Whereas Downing’s earlier study concentrated on Josephus’s use of sources, in ‘Compositional Conventions and the Synoptic Problem’ he devoted attention to Plutarch’s account of Camillus’s siege of the Etruscan town Veii, which according to Downing is the only other instance of extensive conflation in classical literature from around the time of the first century ce. In his view, Plutarch’s account is a conflation of two earlier accounts of the siege of Veii by Livy and Dionysius of Halicarnassus. In suggesting this, however, he disagreed with the general consensus among classicists:

there seems to be considerable resistance among classical scholars to any admission that anything even as complicated as this [conflation of two sources] would have been attempted and a strong preference for ascribing the parallels to the influence of the ‘lost’ common source of Dionysius and Livy, rather than supposing that a first century author might have written with his eye on two texts.

Although Downing considered the evidence to indicate that Plutarch conflated two sources, he contended that even if he were wrong, ‘then the argument for supposing that a first-century use of sources was very simple and quite rough and ready is … even stronger than my own study of Josephus would suggest’.

Downing’s analysis of Plutarch’s account of the siege of Veii led him to defend the conclusion that it is the product of conflation:

It would seem to me quite clear that Plutarch is conflating, but very simply and somewhat arbitrarily, without much very close attention to the texts. He certainly shows no sign of any interest in ‘unpicking’ the changes Livy and Dionysius may have made in their common (lost) source, and in fact no sign of having laid them at all closely side by side. Where they agree, he follows (unless the story line is particularly weak); where they can be taken as supplementing each other, he allows them to; where they entirely disagree, he simply follows one; where they contradict in detail in an otherwise similar episode, he makes up his own version. All this matches precisely … what we are told about the exercises in writing that Plutarch is likely to have done as a lad.

He also presented evidence against the view that Plutarch relied solely on the non-extant source used by both Livy and Dionysius, but it is unnecessary to assess his argument in support of the view that Plutarch did sometimes resort to conflation. Though Downing was convinced that Plutarch did conflate in this instance, in his view ‘the overriding point is that even if the somewhat difficult procedure of conflation seems to afford the more plausible hypothesis, it is still a very simple process of conflation that is at issue, and there is no contemporary analogy at all for anything more complicated still’.

The remainder of Downing’s article on ‘Compositional Conventions and the Synoptic Problem’ attempted to demonstrate, as do his other articles mentioned above, how closely the two-source theory matches literary conventions of the time when the gospels were written, yet how novel and unconventional the literary procedures assumed on alternate source theories like Griesbach’s and Farrer’s would have been in the first century ce. In Downing’s view,

The simplest alternatives to the Two-Source hypothesis have whoever is third (Mark for Griesbach, Luke for Farrer) initially unpicking and then reassembling material in the other two … Even where such unpicking and reassembly might seem relatively easy (say, in whole narrative units), there is no contemporary analogy. No other writer of the time runs backward and forward in two sets of materials in the way the hypothetical third of the Gospel writers must be supposed to have done (on these two competing hypotheses).

Attending first to Downing’s criticism of Griesbach’s hypothesis, one may ask whether Mark must have ‘unpicked’ and then ‘reassembled’ material from the Gospels of Matthew and Luke to create his own narrative. Furthermore, must Mark have run ‘backward and forward in two sets of materials’? On Griesbach’s theory, Mark alternated between two principal sources, but he almost never backtracked in either one; he moved steadily forward in both sources. It is incorrect, therefore, to say that Mark ‘runs backward and forward in two sets of materials’. He admittedly omitted lengthy sections from both sources, sometimes because he continued to progress forward in his sources rather than backtrack to recover omitted sections. His omission of ‘ready-made’ conflations such as the temptation accounts may seem puzzling, but would not have been difficult. At issue for Downing is the simplicity of the conflation involved; on this point Griesbach’s Mark follows a relatively simple and uncomplicated procedure, something Downing conceded.

In this connection, it is worth noting David L. Dungan’s historical analogy for Mark’s compositional procedure on Griesbach’s terms. In his view, Mark’s procedure ‘seems to resemble very much Arrian’s description of his procedure when he sat down around the middle of the second century to compose an account of Alexander’s exploits’. In none of the articles under review did Downing refer to Arrian. Like Charles H. Talbert, he may have thought that Arrian’s description of his literary procedure does not support Griesbach’s hypothesis. For Dungan and other neo-Griesbachians, however, Arrian’s description of his use of two sources is similar enough to Mark’s compositional method on Griesbach’s theory to warrant the comparison.91

As for Mark’s alleged procedure of ‘initially unpicking and then reassembling material’ in Matthew’s and Luke’s Gospels, Downing’s objection is reminiscent of E.A. Abbott’s reductio ad absurdum argument against Griesbach’s hypothesis:

To embody the whole of even one document in a narrative of one’s own, without copying it verbatim, and to do this in a free and natural manner, requires no little care. But to take two documents, to put them side by side and analyse their common matter, and then to write a narrative, graphic, abrupt, and in all respects the opposite of artificial, which shall contain every phrase and word that is common to both—this would be a tour de force even for a skilful literary forger of these days, and may be dismissed as an impossibility for the writer of the Second Gospel.

Downing did not claim that Griesbach’s theory envisages Mark attempting to incorporate every word and phrase common to the Gospels of Matthew and Luke; indeed, his objection centred on what Mark must have omitted from the Gospels of Matthew and Luke. Nevertheless, Abbott’s description of Griesbach’s Mark setting two sources side by side and analysing their common matter sounds like Downing.

In response to Abbott, William R. Farmer and E.P. Sanders pointed out that apart from being inaccurate, Abbott’s argument also misconstrued what Mark must have done to achieve his conflation of the Gospels of Matthew and Luke.94 As Sanders put it,

If Mark had conflated Matthew and Luke, he would not have had to analyze their common matter and labor to include it. He could simply have copied first one then the other, thereby automatically including what was common to them, excluding any chance that they would agree together against him, and also creating agreements with each of them against the other.

Of course, how closely Mark ‘copied’ either of his principal sources would affect the extent of agreements between Matthew and Luke against his own account.

Farmer’s and Sanders’s responses to Abbott’s argument against the Griesbach hypothesis also serve to blunt Downing’s point about Mark’s alleged ‘unpicking’ of his two main sources on Griesbach’s hypothesis. Given the existing similarities between the Gospels of Matthew and Luke, Mark did not need to ‘unpick’ passages in which Luke might be thought to have borrowed from Matthew’s Gospel and then ‘reassemble’ them. Where Matthew and Luke agree to a significant extent in parallel passages, Mark’s use of either one would ensure general agreement with both. The degree to which he borrowed closely or freely would determine the degree to which he agreed or disagreed with either of his sources.

In any case, Downing’s primary concern was why, on the Griesbach or two-gospel hypothesis, Mark omitted common material that occurred in the same context and in the same relative sequence in Matthew and Luke. Not only did Farmer suggest that Mark followed the common order of his two principal sources whenever possible, but Downing also maintained that this was conventional for writers in antiquity: ‘We would expect a contemporary writer to choose the “common witness” both for ease and for security. It is what Josephus and (by my account) Plutarch do for ease at least, and what Tacitus, for instance, says he does for safety …’97

The only example Downing offered to illustrate his objection to Mark’s compositional procedure on Griesbach’s hypothesis comes from the beginning of the Gospel of Mark:

in the baptism and temptation narrative, for instance, [Mark] would reject John’s preaching and the encounters with the adversary, where an extensive conflation is ready-made; he would reproduce most closely the account of John’s dress and his baptism of those who confessed, from Matthew only (while omitting the one phrase there that Matthew and Luke share), but then make his own version of much of the rest.

On Griesbach’s theory, Mark clearly omitted Matthew’s and Luke’s almost identical account of John’s preaching of repentance (Mt. 3:7–10; Lk. 3:7–9; perhaps also Mt. 3:12; Lk. 3:17) and most of their accounts of Jesus’ encounter with the devil in the desert (Mt. 4:1–11; Lk. 4:1–13). If Mark adhered strictly to an editorial policy of incorporating everything common to Matthew and Luke, these omissions are anomalous, but no Griesbachian has thought that Mark adhered to this policy inflexibly. Furthermore, in each case one can offer plausible reasons why Mark chose to omit. For Mark, John’s preaching of repentance may have focused too much attention on John’s message, whereas in this opening section he was leading up to Jesus’ call for repentance in Mk 1:14–15. In addition, the conflicting sequences in the two temptation accounts may have been reason enough not to try to conflate them.

Downing also mentioned Mark’s failure to include the one phrase that Matthew and Luke share in their respective baptism accounts. Presumably the phrase is καὶ πυρί in Mt. 3:11 and Lk. 3:16. However, apart from uncertainty about what John might have meant by saying that the one who was coming after him would baptize with fire, the omission of this phrase along with John’s apocalyptic message in Mt. 3:12 and Lk. 3:17 is of a piece. On the Griesbach or two-gospel hypothesis, what Mark incorporated and what he omitted may sometimes seem strange, but to achieve what he achieved Mark did not have to ‘unpick’ and ‘reassemble’. Nor in omitting ‘ready-made conflations’ in Matthew and Luke did he have to overcome ‘considerable difficulties in doing so’, at least in Mk 1:1–13. Indeed, allegedly anomalous results such as ‘rejecting close parallels’ and ‘preferring the unique’ might well indicate Mark’s use of a simple method of conflation whereby he borrowed from one source at a time, which is what Downing contends was the standard compositional procedure of the time. His criticism of the Griesbach hypothesis is therefore less convincing than it at first appears.

Like an earlier study from the mid-1960s, Downing’s article, ‘A Paradigm Perplex: Luke, Matthew and Mark’, attempted to rebut Farrer’s theory. Whereas his earlier study was written in response to Austin Farrer’s ‘On Dispensing with Q’,102 in ‘A Paradigm Perplex’ he responded to the two-volume work by Michael D. Goulder, Luke—A New Paradigm. Downing’s appraisal of Goulder’s work begins by reiterating his view that all recent efforts to instate alternative hypotheses to the two-source theory, including Boismard’s multiple-stage theory, the Griesbach hypothesis and Farrer’s theory, ‘are vitiated by anachronistic assumptions concerning available compositional techniques in the first century’.104 To evaluate Goulder’s ‘portrait’ of Luke as an author, he compared what Goulder wrote about Luke’s compositional techniques with ‘what our surviving ancient sources suggest that a writer at the time attempting to “reconcile” such works as Mark and Matthew (or Luke and Matthew, or suggested earlier editions of these) might have been expected to do on the basis of common education and common literary practice’.

Downing’s criticisms of Goulder are pertinent only insofar as they relate to compositional conventions of the first century ce. In large part, he repeated points made in earlier studies, especially ‘Compositional Conventions and the Synoptic Problem’. On the practice of conflation, he remarked that, when attempted, it was carried out as simply as possible: ‘any close conflation of, say, two parallel accounts of the same event would be very uncommon: there would be little if any precedent to suggest it to Goulder’s Luke (or to Boismard’s Luke or Farmer’s Mark)’. In addition, exact copying of a source, though obviously possible, was not standard in the composition of a new work: ‘In those places where close visual attention to a single source seems discernible the result will often be a paraphrase rather than word-for-word quotation’.108

Downing allowed that Luke may have manipulated two or more scrolls (not codices) at a time, either writing himself or dictating to a scribe. In his view, however, while there are analogies for the concurrent use of two sources in antiquity, the result in Luke’s case, assuming Farrer’s theory, is without precedent. In his view, what Luke accomplished is ‘almost the exact opposite of what was done by the only scholars who could have afforded him a model. As did Tatian later, they collated scrolls to minimise divergencies. Goulder’s Luke is actually vastly increasing them’. Accepting that Matthew also reordered his sources on the two-source theory, Downing maintained that it was a much more simple procedure than Luke’s conflation of his two main sources on Goulder’s hypothesis. However, as Goulder pointed out in response to Downing, the question of which source theory is simplest is fraught with ambiguity:

The play Downing makes with ‘simple’ and ‘complex’ is artificial. We have both to make numerous explanations to account for considerable texts, and each time we are both involved in hypotheses. If these are consistent, that does not matter. My claim to be ‘simple’ is solely that I do not hypothesize any lost documents (Q, M, L). Nor do I suggest that my theory is more probable than Downing’s because of this. It is just much easier for him to avoid any problem by supposing such a text in the lost document, while I have to give an account on the basis of a text in front of me. If I can do this plausibly, I have the right of first consideration.

As for Downing’s point that Goulder’s Luke flies in the face of ancient literary convention because his narrative exacerbates rather than reduces disagreement between his sources, all source theories involving direct dependence between the gospels face the same criticism. On the two-source theory, for example, Matthew’s rearrangement of Mark’s order is not the result of his adherence to the order of Q; furthermore, on the commonly accepted assumption that Luke retained the original order of Q better than Matthew, Matthew’s rearrangement of Q-material is almost as radical as Luke’s rearrangement of the same material on Goulder’s hypothesis. Even if it could be shown that one of the synoptists introduced more discrepancies in the wording or arrangement of his source(s), the result would be unlikely to assist in determining which source theory is most plausible.

The primary value of Downing’s series of studies discussed above is not his attempt to show that the two-source theory conforms better to conventional literary practice in antiquity than do competing source theories, but his more basic reminder that source criticism of the gospels should not be conducted without an appreciation of classical conventions of writing and use of sources. For example, in ‘A Paradigm Perplex’ he drew attention to the physical constraints under which Luke and the other synoptists are likely to have produced their gospels. He also appealed to what an ancient learned about writing techniques from educational exercises, namely, ‘an ability to paraphrase, précis, expand and simply omit’.113 He was also able to show by reference to studies of other ancient writers’ redactional tendencies that when an author decided to alter what he or she found in available source material, ‘the criteria for change, for what counts as improvement, are narrative coherence, plausibility, interest, clarity, religious piety and propriety, and informal “political” apologetic’. While Downing’s attention to literary conventions in antiquity may not ultimately prove or disprove the viability of any source theory, he has helped to provide a better understanding of how the gospel writers probably went about their work, which in turn may help to preclude anachronistic ideas from future theories of the mechanics of producing a gospel in the first century ce.

What, then, makes up our mental picture of a gospel writer at work? First, he or she did not rely slavishly on written source material. While written sources might be consulted, direct copying was the exception rather than the rule. After consulting one or more sources, the author usually recounted the substance of what he or she had read in his or her own words, sometimes paraphrasing, summarizing, expanding or omitting. Occasionally the author related part of the narrative from memory. In short, much of the time the author relied heavily upon his or her short-term or long-term memory. Alongside written sources, we should also accept the influence of oral tradition. We should also allow for the use of a wooden tablet for note-taking, or even for drawing up an outline of narrative and didactic sequences. The writer may have a bookstand or two, but may not have had even one. He or she may have done the writing, but may also have dictated. In any case, whoever did the writing was probably seated, either on the ground or on a stool, and probably wrote on a roll of papyrus resting on his or her knee or on something supported on both knees. Even when written by the author’s own hand, the words were probably mouthed audibly as they were transcribed. The writing process is likely to have been an oral ‘performance’ of the section of gospel narrative transcribed at a given time. The author wrote without dividing the words and with little or no use of punctuation; nor did he or she use headings or breaks between sections. This is the picture of an ancient writer’s modus operandi that should be kept in mind while reading Chapters 6, 7 and 8.

Chapter 6

The Disagreements in Order in Luke 3:1–5:11

In this section of his Gospel, Luke disagrees with the relative order of both Matthew and Mark on three occasions: Lk. 3:19–20, Lk. 4:16–30 and Lk. 5:1–11. Apart from the addition or omission of unique materials, Lk. 3:19–20 and 4:16–30 are the first disruptions in the relative order of shared (or functionally parallel) materials in the so-called Markan tradition, as the table below shows.




Pericope Heading




John the Baptist



John’s Preaching


John’s Ethical Exhortations




The Coming One

(cf. 14:3–4)

(cf. 6:17–18)


John is Imprisoned




Jesus’ Baptism

(cf. 1:1–17)


Luke’s Genealogy




Jesus is Tested




Jesus in Galilee

(cf. 13:53–58)

(cf. 6:1–6a)


Rejection in Nazareth



(cf. 5:1–11)

First Disciples Called



Jesus in Capernaum


(cf. 1:39)

(cf. 4:44)

Matthaean Summary


(cf. 6:20–49)

Sermon on the Mount


(cf. 1:22)


Transitional Summary


(cf. 1:40–45)

(cf. 5:12–16)

Jesus Heals a Leper


(cf. 7:2–10)

Centurion in Capernaum




Simon’s Mother-in-law




Jesus Heals Many

(cf. 4:23)



Preaching in Galilee

(cf. 4:18–22)

(cf. 1:16–20)


First Disciples Called

(cf. 8:2–4)



Jesus Heals a Leper

Luke 3:19–20, Herod’s Imprisonment of John the Baptist

Luke 3:19–20 occurs within a section of synoptic parallels corresponding to the second of Bo Reicke’s 12 blocks of synoptic material subdivided on thematic and topographic grounds. Reicke’s block 2, concerned with ‘Christ’s Baptism and Related Events’, comprises everything in the table above before Lk. 4:16–30, which he placed at the beginning of block 3, ‘Activity in and near Capernaum’. According to Reicke,

Within this complex [i.e. block 2] there is merely a quantitative difference between the Gospels, either implying that Mark has sometimes presented a shorter version or that Matthew and Luke have included more of the relevant material. On practical accounts the extant traditions were occasionally abbreviated by Mark or supplemented by Matthew and Luke. Yet the common structure is obviously essential, seeing that context-parallel triple traditions serve as the indispensable fundament of the narrative complex and receive support from double and single traditions which are consistently related to the matters treated in the common traditions. Everything forms a triple accord, in which given themes have been followed through in a harmonious way. Owing to this coherence, block 2 is unique within the synoptic material.

Reicke displayed Lk. 3:19–20 as Sondergut supplementing Luke’s account of John the Baptist’s preaching, but he did not indicate that this brief passage has what he called ‘alibi analogies’ in both Matthew and Mark in later contexts, namely, Mt. 14:3–4 and Mk 6:17–18. Luke 3:19–20 is clearly ‘related’ to what Luke has said about John the Baptist. However, it contradicts Reicke’s assertion that ‘everything forms a triple accord, in which given themes have been followed through in a harmonious way’. Luke’s decision to supplement his section on John the Baptist by narrating Herod’s imprisonment of John at this point creates disharmony within what Reicke called ‘the indispensable fundament of the narrative complex’ by removing John from the scene before his role within the narrative is properly concluded, that is, before Jesus is baptized. As a result, in Luke’s Gospel Jesus is baptized anonymously. One assumes that Luke intended his readers or hearers to understand that Jesus was baptized by John, but this assumption stems from a knowledge of the other gospels, not from Luke’s account at this point. Indeed, nowhere in his Gospel does Luke state explicitly that Jesus was baptized by John.

Luke tells us more about John the Baptist than the other evangelists, partly because he relates events surrounding his birth and partly because he refers to the influence of John’s reputation in Acts. Only Luke provides information about John’s parents, including their names, tribal affiliation and the details that his father was a priest and that his mother was infertile. Only Luke has John’s birth announced by a heavenly messenger, who says to Zechariah:

Your wife Elizabeth will bear you a son, and you shall name him John; you will have joy and delight, and many will rejoice at his birth. For he will be great in the Lord’s sight, he will drink no wine or strong drink, he will be filled with the Holy Spirit from the moment of his birth, and he will turn many of the Israelites to the Lord their God. He will go forward in God’s sight in the spirit and power of Elijah, to turn the minds of fathers to their children, to turn the disobedient to the wisdom of the righteous, to make ready for the Lord a people thoroughly prepared (Lk. 1:13–17).

Only Luke says that John and Jesus were related (Lk. 1:36); only Luke relates John’s birth, circumcision and naming (Lk. 1:57–66); and only Luke provides some of John’s explicit ethical exhortations (Lk. 3:10–14).

As already indicated, only Luke relates the imprisonment of John before the baptism of Jesus, and he alone recounts the baptism of Jesus without mentioning John. Mark says explicitly that Jesus was baptized in the Jordan by John (ἐβαπτίσθη εἰς τὸν Ἰορδάνην ὑπὸ Ἰωάννου). Matthew goes so far as to record a conversation between John and Jesus at the time of his baptism (Mt. 3:13–15). By contrast, Luke records no direct encounter between John and Jesus; the closest the two come to meeting is when the prenatal John leaps in his mother’s womb at the sound of the pregnant Mary’s greeting in Lk. 1:41 (cf. Lk. 1:44)!

After mentioning that John was imprisoned by Herod at Lk. 3:20, Luke refers directly to John only once in his Gospel. In Lk. 7:18, John sends two disciples to ask Jesus whether he really is ‘he who is coming’. Luke does not record that John was in prison at the time, in contrast to Matthew’s explicit remark: ‘John heard in prison about the deeds of the Messiah’ (Mt. 11:2). Elsewhere Luke refers to John indirectly or in passing, for example, when he notes the custom of John’s disciples in Lk. 5:33 and 11:1, or when he uses John’s teaching as a foil for Jesus’ teaching in Lk. 7:24–35, or when he relates that people considered Jesus to be John redivivus (Lk. 9:7, 19), or when he records that Herod had beheaded John (Lk. 9:9). He also uses John as the dividing line between the time of the Law and the Prophets and the time when the Good News of the Kingdom of God is proclaimed (Lk. 16:16). The nature of John’s baptism crops up from time to time in Luke’s Gospel such as at Lk. 20:1–8, but is brought up frequently in Acts (1:5; 10:37; 11:16; 13:24; 18:24; 19:7). It is only at the beginning of Acts that we discover that Luke also shares the tradition that John baptized Jesus (see Acts 1:22). While Luke omitted John from his account of Jesus’ baptism, he clearly knew that John baptized Jesus and intended his reader or hearer to assume that Jesus was baptized by John. His decision to recount John’s imprisonment before his account of Jesus’ baptism obviously created a difficulty for Luke, which he resolved by narrating Jesus’ baptism without reference to John. Why did Luke do this?

In The Style and Literary Method of Luke, Henry Joel Cadbury contended that Lk. 3:19–20 is one of two instances in which Luke’s motive was ‘clearly the desire to conclude at once a subject when it has been introduced’. Joseph A. Fitzmyer, who asserted not only that Luke’s transpositions of Markan episodes are ‘readily explicable in terms of Lukan composition’ but also that ‘in each instance one can detect a clear reason why Luke has made the transposition’, considered that ‘the imprisonment of John the Baptist (Mark 6:17–18) is moved up by Luke to 3:19–20 in an effort to finish off the story of John before the ministry—and even the baptism!—of Jesus’. Michael D. Goulder commented more fully:

Luke sealed off each of his previous stories, sending the participants back home, or keeping John in the deserts [sic] until the day of his showing forth to Israel. Now he seals off the Baptist’s preaching; indeed he seals off the Baptist, by bringing forward his imprisonment from Mt. 4:12; 14:1–4/Mk 1:14; 6:17. This is in one way an illogical thing to do, for he still has to have Jesus baptized, which requires the Baptist’s presence; but Luke is not deterred by that, and records the baptism in a genitive absolute without mentioning the baptizer (v. 21). He has kept the stories in compartments hitherto, John-Jesus-John-Jesus-John, and he wishes to draw the John sequence to a firm close. The motive is not just neatness, though. He wishes to keep John in a clear secondary place, and this can be done effectively by putting him in prison as early as possible; and it may be that he has also in mind that John’s Question from Prison has to come earlier in his narrative than in Matthew’s at 7:18f.

Hans Conzelmann argued that Luke related the imprisonment of John the Baptist—and hence the end of his public ministry—as early as he did to differentiate between the old era culminating in John’s mission and the new era inaugurated by Jesus’ mission. In his view,

The reference to the imprisonment in [Lk.] 3:19f. divides the section concerning John from the section concerning Jesus in the sense of drawing a distinction between the epochs of salvation, for which [Lk.] 16:16 provides the clue. Now the way is open for the story of Jesus. The fact that the activity of the two still overlaps cannot be entirely eliminated, but Luke deprives it of any real significance. According to [Lk.] 3:21f. Jesus is baptized as one of the people, like everyone else. Luke excludes any suggestion that John plays an important part in the incident. This is in keeping with his whole conception of the significance of John.

In his commentary on Luke’s Gospel, Eduard Schweizer was more tentative: ‘Luke anticipates John’s arrest … perhaps in order to distinguish his role from that of Jesus’. Walter Wink’s readjustment to Conzelmann’s schema of Heilsgeschichte to include John the Baptist’s public ministry as preparation within the period of fulfilment rather than as the culmination of the period of promise (the Law and the Prophets) did not lead him to suggest another reason for Luke’s decision to recount John’s imprisonment prior to the baptism of Jesus. However, if Luke’s primary reason for recounting John’s imprisonment as early as he did was to differentiate between John’s role and that of Jesus, or between two eras of redemptive history, surely John’s words in Lk. 3:15–16 would have sufficed. Here Luke addresses the question of whether John was the messiah and has John point forward unambiguously to ‘one stronger than I, whose sandal strap I am not worthy to untie’. One might argue that the fact that Luke, like John the evangelist, placed John the Baptist’s testimony to Jesus within the context of public speculation about whether or not he was the expected messiah is suggestive of Luke’s motive for relating John’s imprisonment before Jesus is baptized. It is conceivable that Luke recounted John’s imprisonment before Jesus’ baptism, which thereby precluded any explicit statement that John baptized Jesus, so as to deny his readers and hearers any opportunity of ascribing greater importance to the baptizer (John) than to the one who was baptized (Jesus).

In Profiles of a Rabbi, Bruce Chilton maintained that Luke’s reason for placing Herod’s imprisonment of John before his account of the baptism of Jesus is illuminated by the story of Apollos in Acts 18:24–19:7. According to Chilton, the account of Apollos and others who ‘knew only the baptism of John’ indicates that ‘Acts takes some pains to stress that John was a purely preparatory figure in relation to Jesus. On this basis, the placement of the notice of John’s imprisonment in the Gospel is explicable, although the presentation—explicable or not—appears clumsy in comparison with Matthew and Luke [sic]’. In short, Luke’ s unique placement of John’s imprisonment before the baptism of Jesus serves to guard against possible misunderstanding of the relation between John and Jesus.

Advocates of the two-gospel hypothesis offer a similar reason for Luke’s redaction at this point, despite their contention that Luke’s primary source is the Gospel of Matthew. In their view, Luke conflated elements from Mt. 4:12 and Mt. 14:3–12 in his brief account of John’s imprisonment:

The point of this conflation is to complete the process of separation between Jesus and John for Luke’s readers/hearers, so that the ministries of John and Jesus do not overlap. Luke achieves this result by referring here to John’s imprisonment (cf. Mt. 14:3–12) as if John’s public activities were curtailed prior to Jesus’ baptism. Luke pointedly omits any mention that it was John who baptized Jesus. Thus, Luke concludes ‘the period of the Law and the Prophets [which was until John]’ (Lk. 16:16).

Commenting on Lk. 3:19–20, Darrell L. Bock noted, ‘Much speculation exists about Luke’s motive for this chronological transposition, including the argument that he wished to separate totally the period of John from the period of Jesus’ ministry’. Heeding Fitzmyer’s warning not to read too much theological significance into Luke’s placement of this event, however, he opined, ‘The order probably represents a literary preference to present John and then focus on Jesus’.16

Whatever motivated Luke to recount Herod’s imprisonment of John before the baptism of Jesus, his arrangement of material is secondary at this point. Since in Acts 1:21–22 he records that Jesus was baptized by John, his account of John’s imprisonment in Lk. 3:19–20 is a chronological anticipation, though not narrated as such, just as the account of John’s death in Mk 6:17–29 is recounted as a ‘flashback’. The relative sequence of Lk. 3:19–20 is secondary because it presupposes the more chronological sequence of events reflected in Matthew’s and Mark’s Gospels.

An obvious rearrangement of traditional or historical chronology, as in Lk. 3:19–20, does not necessarily imply posteriority to a gospel that retains the traditional or historical chronology. It is possible that Luke’s account is earliest, and that Matthew’s and/or Mark’s versions are later attempts to ‘correct’ a chronological discrepancy. However, given the broad consensus that the first three gospels are related to some extent in a literary way, the presumption must be that Luke’s sequence of incidents at this point is secondary to that of Matthew and Mark.

Beyond this, does Lk. 3:19–20 provide any clues about which of the other two gospels Luke may have used? To determine this, one must examine Lk. 3:19–20 in its wider context, namely, Lk. 3:1–4:15 and parallels. While Mt. 14:3–4 and Mk 6:17–18 are often displayed as the parallels to Lk. 3:19–20, they are not the only relevant parallels. Like Luke, both Matthew and Mark refer to the imprisonment of John before the beginning of Jesus’ public work. Immediately following their respective accounts of the testing of Jesus in the wilderness, they each record the arrest of John in connection with Jesus’ return to Galilee (Mt. 4:12; Mk 1:14). Matthew’s use of the verb ἀναχωρέω in Mt. 4:12 even suggests that Jesus withdrew or retreated into Galilee because he feared a fate similar to John’s.

Luke seems to make no connection between the arrest of John the Baptist and Jesus’ return to Galilee. However, his use of the verb ὑποστρέφω at Lk. 4:14 raises interesting questions. Among the gospel writers only Luke uses this verb. It occurs 21 times in his Gospel and 11 times in Acts. Elsewhere in the New Testament, this verb occurs three times (Gal. 1:17, Heb. 7:1, 2 Pet. 2:21). Luke uses ὑποστρέφω five times before using it at Lk. 4:14, once at Lk. 4:1. It means to ‘return’, ‘turn back’ or ‘turn away’. Commenting on Luke’s use of ὑποστρέφω at Lk. 4:1, Fitzmyer noted,

The verb hypestrepsen can mean either ‘returned’ (i.e. to Nazareth or Galilee) or ‘withdrew, turned aside’ … But since Luke has not mentioned earlier that Jesus came from Nazareth (contrast Mk 1:9) or from Galilee (contrast Matt 3:13), there is little reason to read the first sense into Luke’s use of the verb here. See further [Lk.] 4:14, where the term of the withdrawal will be specified.

At Lk. 4:14, Fitzmyer translated ὑποστρέφω to mean ‘withdraw’. Frans Neirynck and Frans Van Segbroeck regard Luke’s use of ὑποστρέφω at Lk. 4:14 as a synonym or substitute for Mark’s use of the verb ἔρχομαι at Mk 1:14. If Fitzmyer is right, however, Luke’s use of ὑποστρέφω at Lk. 4:14 could be regarded as a synonym of ἀναχωρέω at Mt. 4:12.

The difficulty Fitzmyer pointed out for interpreting ὑποστρέφω to mean ‘return’ at Lk. 4:1 warrants further consideration. If understood in this sense at Lk. 4:1, one naturally asks to what place Luke understood Jesus to be returning. As Fitzmyer noted, Luke provides no answer to this question because, unlike the other synoptists, he did not previously advise from where Jesus came to be baptized. The same difficulty applies to Lk. 4:14. One could argue along the following lines that Luke’s use of ὑποστρέφω at Lk. 4:1 and 4:14 reveals his dependence on either Matthew or Mark or both. Influenced by Mt. 3:13 or Mk 1:9 (or both) to think of Jesus leaving Galilee to be baptized by John in the Jordan, but forced to modify his account of Jesus’ baptism because he had already narrated John’s imprisonment, Luke omitted the reference to Galilee and to the Jordan in Lk. 3:21 but presupposed these references in his source material at Lk. 4:1 and 4:14. According to advocates of the two-gospel hypothesis, ‘Luke’s reference to Jesus’ “return from the Jordan” (Lk. 4:1) shows Lukan dependence on Mt; Luke does not mention earlier in his gospel that Jesus was baptized “in the Jordan” but Mt. does (cf. Mt. 3:13)’. Neo-Griesbachians presuppose Luke’s use of Matthew rather than Mark, but if ὑποστρέφω means ‘return’, Luke’s dependence on Mark could also explain his choice of terminology at this point.

On the other hand, if (with Fitzmyer) one accepts that by using ὑποστρέφω at Lk. 4:1 and 4:14, Luke did not intend to convey the sense of ‘return’, one could argue that his terminology at Lk. 4:14 reflects the influence of ἀναχωρέω at Mt. 4:12, despite the fact that Luke has already mentioned John’s arrest and therefore cannot explicitly link it to Jesus’ ‘withdrawal’ to Galilee.

Neither conjecture is particularly strong. However, both are consistent with, and therefore supportive of, Luke’s posteriority to Matthew or Mark or both, which is the inescapable conclusion based on the chronology of John’s imprisonment in Luke’s Gospel relative to the placement of this incident in Matthew’s and Mark’s Gospels. Luke 3:19–20 provides evidence of Luke’s posteriority, even if one cannot say categorically whether Luke is posterior to Matthew or Mark or, indeed, both.

Luke 4:16–30, Jesus at Nazareth

Luke 4:16–30 immediately follows the structural division at Lk. 4:14–15, which is a transitional or ‘bridging’ summary that ‘gives an overview of the Galilean ministry of Jesus’.28 Fitzmyer noted that Lk. 4:14–15 contains three Lukan motifs: (1) Jesus begins his ministry ‘armed with the Spirit; (2) Jesus’ public activity is described in terms of teaching; and (3) Luke’s theme of ‘universality’ is sounded in the statement that Jesus was ‘praised by all the people’. That Lk. 4:16–30 follows this important transitional summary suggests its significance.

Most scholars view Lk. 4:16–30 as ‘programmatic’, by which they mean that Luke placed this episode at the beginning of Jesus’ ministry for thematic rather than chronological reasons. ‘It has become axiomatic in studies of Luke’, according to Joel B. Green, ‘that Jesus’ sermon at Nazareth (4:16–30) is programmatic for our understanding of the mission of Jesus in the Gospel’.

B.H. Streeter made this remark about Lk. 4:16–30:

St. Luke knew that this was not the first public act of Christ, for he makes Him allude to previous work at Capernaum (4:23), but he puts it before any other act of His public ministrations because he sees in it an epitome of the whole Gospel and the key to its destined reception by Jew and Gentile respectively.

J.M. Creed made a similar point when commenting on this passage:

Its real function is to introduce the main motifs which are to recur throughout the Gospel and Acts, and this it does with great effect: the Gospel to the poor is preached by Jesus in his own home and rejected. The rejection by Nazareth foreshadows the rejection by the Jewish people and the subsequent universal mission of the Church.

In The Making of Luke—Acts, Cadbury noted, ‘A similar story was indeed to be found in Mark’s account of Jesus’ rejection in his home town, but in Luke it is placed as a program at the outset of Jesus’ ministry. In modern terms it is a “keynote speech” ‘. While discussing Luke’s alleged appropriation of Mark’s arrangement of material, Rudolf Bultmann referred to Lk. 4:16–30 as the most important of Luke’s ‘occasional small transpositions of the text of Mark’ and spoke of Luke displacing the story of Jesus’ visit to Nazareth and reshaping it to form a ‘programmatic entrance’.

In defence of the unity of Luke—Acts, Robert Maddox observed:

the mission of Jesus begins with a scene in which the rejection of the message of salvation by the Jews and its acceptance by the Gentiles is anticipated, and the mission of Paul ends with a scene in which this is declared to be an established fact (Luke 4:16–30; Acts 28:17–28). This looks like a deliberate, structural element.

Maddox also offered other reasons why Luke placed his account of Jesus’ rejection in Nazareth to open Jesus’ ministry: ‘From the purely narrative point of view, it gave a specific and graphic setting for the beginning of Jesus’ work’. He also noted how the many references to the Holy Spirit in the material prior to Lk. 4:16–30 lead up to Jesus’ words in Lk. 4:18, ‘The Spirit of the Lord is upon me’. For Maddox, Luke’s stress on the presence and activity of the Holy Spirit is a sign of the eschatological fulfilment of God’s promises.

According to E. Earle Ellis, Lk. 4:16–30 ‘is placed at this juncture in the Gospel for a thematic purpose. Chronologically, it can be located only very generally in the Galilean mission’. In his view,

The rejection of Messiah, alluded to earlier (2:34f.), is the keynote of this ‘inaugural’ address of the Galilean mission … It is a tragic story, and Nazareth is a preview of what is to happen later in all areas and among all classes of Jews (cf. 9:52f.). It is here also that Jesus first hints that the benefits of the gospel will be given to the Gentiles (26f.).

Marshall regards Lk. 4:16–30 as programmatic because ‘it contains many of the main themes of Lk.—Acts in nuce, Fitzmyer concurs: ‘It is an important episode in the Lukan Gospel, foreshadowed in a sense in Simeon’s oracle in the infancy narrative (2:34) and foreshadowing in a way the account of the entire ministry that is to follow’. A fuller statement by Fitzmyer sums up why many regard Lk. 4:16–30 as programmatic:

The Lukan story, transposed to this point in the Gospel, has a definite programmatic character. Jesus’ teaching is a fulfillment of OT Scripture—this is his kerygmatic announcement (the Lucan substitute for the omitted proclamation of Mark 1:14b–15). But that same teaching will meet with success and—even more so—with rejection. Luke has deliberately put this story at the beginning of the public ministry to encapsulate the entire ministry of Jesus and the reaction to it. The fulfillment-story stresses the success of his teaching under the guidance of the Spirit, but the rejection story symbolizes the opposition that his ministry will evoke among his own. The rejection of him by the people of his hometown is a miniature of the rejection of him by the people of his own patris in the larger sense.

Similarly, Green has identified four factors that underscore the significance of Lk. 4:16–30 within Luke’s Gospel. First, by contrast with the Gospels of Matthew and Mark, Jesus’ ministry in Nazareth occurs at the very beginning of Jesus’ public work. Second, the content of Jesus’ ‘sermon’ at Nazareth illustrates his synagogue teaching generally, which, although mentioned fairly often (cf. Lk. 4:15, 31–37, 44; 6:6; 13:10–17), is not recorded elsewhere. Third, summaries of Jesus’ ministry in Luke—Acts (e.g. Lk. 7:18–23; Acts 10:34–43) refer back to the incident in Nazareth as being ‘paradigmatic’ for Luke’s presentation of Jesus’ mission. Finally, the content of Jesus’ ‘sermon’ is decisive for understanding Luke’s interpretation of Jesus’ public ministry. As Green put it, ‘In these words the shape of Jesus’ ministry in the Third Gospel is given form’.

The programmatic nature of Lk. 4:16–30 stands behind most explanations for Luke’s divergence from the shared order of Matthew and Mark at this point in their narratives. For example, according to Fitzmyer, ‘Jesus’ visit to Nazareth (Mark 6:1–6) is transferred by Luke to the beginning of Jesus’ Galilean ministry (4:16–30) to serve a programmatic purpose: it presents in capsule form the theme of fulfillment and symbolizes the rejection that will mark the ministry as a whole’. Likewise, for Christopher M. Tuckett:

The transfer of the story of the rejection at Nazareth from Mark 6:1–6 to Luke 4:16–30 needs little explanation, given the very important programmatic significance which the story clearly has for Luke’s two-volume work. The change in order is thus part of the (redactional) stress which is clearly laid on it (though perhaps the altered position may also be due in part to the presence of a rejection story in Luke’s source at this point).

However, as Christopher J. Schreck noted in his Forschungsbericht on this passage, ‘Scholars can agree wholeheartedly on the programmatic nature of the passage and still fiercely disagree on the nature of the program’. He also underscored the continuing significance of this pericope for understanding the nature of Luke’s source material and compositional strategy:

the pericope continues to demand the attention of contemporary researchers because Luke 4:16–30, with its added special material and distinctively prominent location vis-à-vis the other Synoptics, is an unavoidable test case for hypotheses on the larger question of Lukan tradition and redaction. Indeed, the larger section, Luke 3:1–4:30, of which the Nazareth pericope is the culminating scene, can be considered a proving ground for theories of Lukan composition, since it compounds a number of the most vexing issues in Lukan source and redaction criticism.

On the source question, Schreck advised: ‘The fundamental question which constitutes the great divide in criticism of the passage seems to be whether or not Luke 4:16–30 is to be considered basically dependent upon Mark 6:1–6a’. At the extremes of the spectrum are Heinz Schurmann, who maintains that Lk. 4:16–30 is independent of Mk 6:1–6, and Conzelmann, who regarded Lk. 4:16–30 as the result of Luke’s redaction of Mk 6:1–6. Scholars apart from Schürmann who have advocated Luke’s independence of Mark in this pericope include Alfred Plummer, Cadbury,52 Marshall, Schweizer54 and Francois Bovon. After reviewing ‘the prevailing view’ (Lukan dependence on Mk 6:1–6) and ‘a minority opinion’ (Lukan independence), Schreck listed the ‘Markan elements’ in Lk. 4:16–30:

Among the motifs which Luke 4:16–30 has drawn from Mark’s account of the Nazareth visit can be listed: 1. Jesus’ visit to his own hometown (Mark 6:1/Luke 4:16; 4:23); 2. teaching/speaking in the synagogue on the sabbath (Mark 6:2/Luke 4:16); 3. the synagogue audience’s reaction of astonishment/wonderment (Mark 6:2/Luke 4:22); 4. the question about Jesus’ origins (Mark 6:3/Luke 4:22); 5. Jesus’ response: the saying about the prophet (Mark 6:4/Luke 4:24); 6. the impossibility/refusal to work (many) miracles in Nazareth (Mark 6:5/Luke 4:23ff.); 7. rejection (Mark 6:3/Luke 4:28–29). Moreover, by his transposition of the scene forward from the Markan order, Luke has made the Nazareth visit a programmatic proclamation of Jesus’ message and mission, thus taking over, by position and function as well as by content, motifs from Mark 1:14–15. Among these can be numbered: 1. Jesus’ return to Galilee (Mark 1:14/Luke 4:14); 2. Jesus’ proclamation of the gospel (Mark 1:14–15/Luke 4:18); 3. Jesus’ announcement of fulfillment (Mark 1:15/Luke 4:21).

Schreck’s near-exhaustive review of recent literature on this pericope led him to conclude that the prevailing view is still that Lk. 4:16–30 is primarily, if not solely, the result of Luke’s redaction of Mk 6:1–6a.

As Schreck noted, critics who presuppose Markan priority, including those who advocate Farrer’s theory, usually regard Lk. 4:16–30 as a freely redacted version of Mk 6:1–6a. For example, John Drury, who accepts Farrer’s view that Matthew was one of Luke’s principal sources, does not envisage Luke drawing from Matthew at this point. On the contrary, in his view, Luke developed the note of fulfilment in Mk 1:15 into ‘the Nazareth manifesto’ by recalling elements of the stories of Elijah and Elisha in 1 Kings 17 and 2 Kings 5, but borrowed the theme of opposition from Mk 6:1–6 and ‘dramatised’ the proverbial saying in Mk 6:4 about the prophet in his own country by recording an attempt to kill Jesus.

The same cannot be said for Goulder, who also advocates Farrer’s theory. In Goulder’s view, various data suggest Luke’s reliance on both Matthew and Mark at this juncture. Since the majority of critics agrees that Luke borrowed from Mark, only Goulder’s reasons for Luke’s dependence on Matthew require attention. First, he suggests that φήμη in Lk. 4:14, which occurs nowhere else in Luke—Acts, is a ‘reminiscence’ of Mt. 9:26, which records a statement similar to that in Lk. 4:14. Second, he avers that in Lk. 4:15 ‘the presence of the referentless αὑτῶν, meaning “the Jews’ synagogue”, betrays Luke’s dependence on Matthew. Luke never has this expression elsewhere, but it is a favourite of Matthew’s …’ Third, Luke’s dependence on Matthew is revealed by the place name Nazareth in Mt. 4:13 and Lk. 4:16. As Goulder points out, ‘Not only is Nazareth common to the two contexts, against Mark, but the spelling Ναζαρα [sic] is unique to these two passages; Matthew has the name in a redactional expansion of Mk 1:14, and the coincidence of placing with Luke cannot be accidental’.

The form of Ναζαρά at Lk. 4:16, which coincides with no other spelling of Nazareth in the New Testament apart from Mt. 4:13, has provoked much speculation. For example, Schweizer conjectured that ‘Matthew 4:13 suggests a pre-Lukan narrative in which Jesus turned from “Nazareth” to the “Gentiles” at Capernaum … because both use the unique form “Nazara” for Jesus’ home town’. Cadbury’s point about Luke’s penchant for variation, especially for variant spellings of place names like Jerusalem,63 should caution critics not to place more weight on the variant spelling of Nazareth at Lk. 4:16 than it can properly bear. Nevertheless, Luke’s six references to Nazareth constitute half the total number of references to this Galilaean town in the entire New Testament and his five other references in Lk. 1:26, 2:4, 2:39, 2:51, and Acts 10:38 all use the uniform spelling Ναζαρέθ.

In a study entitled ‘Nazareth: A Clue to Synoptic Relationships’, William O. Walker, Jr, argued not only that Matthew’s Gospel reflects a tradition that placed Jesus’ rejection at Nazareth immediately after his temptations in the wilderness, but also that Mt. 4:12–17 and Lk. 4:14–32 are both dependent on common (or similar) source material. In support of this position, he identified seven ‘clues’. The first is Matthew’s reference to Jesus leaving Nazareth in Mt. 4:13. In Walker’s view,

It would appear, therefore, according to Matthew, that Jesus did, in fact, go to Nazareth after his temptations (as in Luke), but nothing is said in Matthew about anything that happened while he was there—there is only the bare mention of his leaving. The Greek root of the participle referring to his leaving, however, is a compound root, carrying the idea of ‘leaving behind’, ‘abandoning’, ‘forsaking’, or ‘deserting’. Thus, there may well be more than meets the eye in Matthew’s apparently passing reference to Jesus leaving Nazareth. The root suggests some strong reason for leaving—perhaps a reason such as rejection by the people of Nazareth (as is spelled out at this point in the Lukan narrative).

Walker thought it more likely that Matthew was familiar with a source that recounted Jesus’ rejection in Nazareth at this point, but chose to ignore it in favour of another rejection account elsewhere, than that Luke used Matthew and seized upon the reference to Jesus ‘abandoning’ Nazareth for Capernaum in Mt. 4:13 as an opportunity to include his rejection story at this point in his Gospel. He reasoned that Luke’s dependence on Matthew fails to explain Matthew’s obscure reference to Jesus’ departure from Nazareth. One might reply that Luke’s dependence on Matthew need only explain satisfactorily Luke’s decision to recount his rejection story at this point, which it does; Lukan dependence on Matthew need not explain Matthew’s reference to Jesus’ departure from Nazareth. The two hypotheses are not mutually exclusive; Matthew’s familiarity with a hypothetical source containing an account of Jesus’ rejection in Nazareth at the beginning of Jesus’ Galilaean ministry does not necessarily imply Luke’s familiarity with the same source, nor does it preclude Luke’s dependence on Matthew. In any case, for Matthew, Jesus’ ‘retreat’ (ἀναχωρέω) to Galilee in Mt. 4:12 may naturally have implied a sojourn in his home town (cf. Mt. 2:23) before choosing to settle in Capernaum.

Furthermore, Walker may have relied too much on the suggestive power of the participle καταλιπὼν in Mt. 4:13. The compound verb καταλείπω is used 24 times in the New Testament, 18 times in the four gospels and Acts. Matthew uses καταλείπω on four occasions: 4:13; 16:4; 19:5; 21:17. In Mt. 16:4 and 21:17 it is used to describe Jesus leaving a scene of controversy. But in Mt. 19:5 καταλείπω is used of a man leaving his parents to be joined to his wife. Mark uses the same verb in his parallel pericope, Mk 10:2–12. There is a sense of finality about leaving one’s parents to marry. The same sense is conveyed by Luke’s choice of καταλείπω to describe Levi leaving everything to follow Jesus in Lk. 5:28. The same may be true of Mt. 4:13. In short, Matthew may simply have intended to convey a sense of finality about Jesus’ departure from Nazareth to take up residence in Capernaum.

On the other hand, of the 18 occurrences of καταλείπω in the gospels and Acts, only Matthew uses this term to describe Jesus departing after engaging in controversy with members of the Jewish religious establishment. On both occasions almost identical phraseology is used: καὶ καταλιπὼν αὑτοὺς ἀπῆλθεν (Mt. 16:4); καὶ καταλιπὼν αὑτοὺς ἐξῆλθεν … (Mt. 21:17). This might support Walker’s argument, although Matthew was clearly capable of using the term in a neutral sense. For Walker, ‘some type of relation between Matthew and Luke appears clear at this point—a relation in which Luke reflects a primary and Matthew a secondary version of the tradition’, but it would seem judicious, on the basis of the evidence adduced, to settle for his preliminary observation without claiming that either Matthew or Luke reflects a primary or secondary version of the tradition.

Walker’s second clue is the common form of the place name Nazareth. As he points out, ‘The name occurs twelve times in the Gospels and Acts and nowhere in the remainder of the New Testament. The references are: Ναζαρέτ (Matt 2:23; Mark 1:9; John 1:45, 46); Ναζαρέθ (Matt 21:11; Luke 1:26; 2:4, 39, 51; Acts 10:38); and Ναζαρά (Matt 4:13; Luke 4:16)’. Matthew shows no preference for any one spelling; on each occasion he spells Nazareth differently. Luke, on the other hand, clearly prefers the spelling Ναζαρέθ (five times); only at Lk. 4:16 does he use a different spelling. In contrast to Goulder, Walker considered this to be best explained by mutual dependence on a common source other than Mark.

Third, there is Matthew’s passing reference to the arrest of John the Baptist in Mt. 4:12. According to Walker,

It is barely possible that Luke, seeing Matthew’s obscure allusion at this point, would introduce an account of John’s arrest earlier in his own narrative, thus making Matthew’s reference intelligible; this would imply, however, that Luke then eliminated the reference to John’s arrest at the point where Matthew has it, thus, in fact, removing the need for introducing his own account of the arrest earlier in the narrative. More likely, Matthew is here using a source that (like Luke) reports the arrest of John before it reports the beginning of Jesus’s public ministry.

Advocates of Markan priority would no doubt concur. They contend, however, that the reference to John’s arrest in Mt. 4:12 is taken from Mk 1:14, not some other source. Moreover, on this theory Luke’s account of John’s imprisonment in Lk. 3:19–20 is not intended to make Mark’s reference to John’s arrest intelligible, but rather to make a clear demarcation between the mission of John and the mission of Jesus. The same applies even if one considers that Luke was familiar with Matthew’s Gospel. An additional hypothetical non-Markan source is not required to explain Luke’s report of John’s arrest in Lk. 3:19–20.

For Walker, a fourth clue to Matthew’s and Luke’s joint dependence on a non-Markan source at this point in their respective accounts is that both refer to Jesus going to Capernaum upon leaving Nazareth (Mt. 4:13; Lk. 4:31). In addition, both characterize Capernaum in some way, either as located by the sea in the region of Zebulun and Naphtali (Matthew) or as a city of Galilee (Luke). Furthermore, in contrast to Mark, both Matthew and Luke record Jesus going to Capernaum before the call of the first disciples; Mark delays his first mention of Jesus’ entrance into Capernaum until after the call of the first disciples in Mk 1:16–20.

A fifth clue is that both Matthew and Luke refer to Galilee twice, first immediately following their respective accounts of Jesus being tempted in the wilderness (Mt. 4:12; Lk. 4:14) and later in connection with Capernaum (Mt. 4:15; Lk. 4:31). The earlier references are clear contextual parallels and are paralleled also by Mk 1:14. Walker considered the two references in each to be ‘parallel as regards content’, but the importance of this is unclear.

Walker’s sixth clue is that both Matthew and Luke make use of a fulfilment-of-prophecy formula and both appeal to a passage from Isaiah (although not the same one). He accepted that Luke’s appeal to prophecy occurs as part of his rejection account while Matthew’s is connected with Jesus’ move to Capernaum. However, according to Walker,

If Matthew is familiar with an account similar to Luke’s … but has chosen to eliminate the actual rejection story at this point, he may well have decided … to preserve an appeal to prophecy (indeed, to the prophet Isaiah), changing the specific passage in light of his omission of the rejection story.

This is a tenuous conjecture. Matthew’s fulfilment-of-prophecy motif is a well-established characteristic even by this early point in his Gospel. Indeed, Mt. 4:14–16 is the seventh occasion in which Matthew describes an event in the life of Jesus as fulfilling a prophecy: 1:22–23; 2:5–6; 2:14–15; 2:17–18; 2:23; 3:3; and 4:14–16.

Finally, Walker noted that both Matthew and Luke conclude this segment of their Gospels by referring to Jesus preaching (Mt. 4:17) or teaching (Lk. 4:31b–32). He then asserted,

On the basis of these seven clues, it is possible to build a very strong case that Matthew and Luke are using common (or similar) source material for the parallel passages under consideration and, indeed, that Luke reflects the source material more closely than does Matthew. Specifically, it appears that Matthew was familiar with a rejection story similar or identical to that included by Luke and that, in Matthew’s source material, this rejection story appeared at the same point in the narrative as it now appears in Luke.

Walker’s seven clues suggest some relation between Matthew and Luke at this point, but they hardly constitute a ‘very strong case that Matthew and Luke are using common (or similar) source material for the parallel passages under consideration’. While similar in some respects to Schürmann’s view that an alternate source stands behind Matthew and Luke at this point, Walker’s proposal contains too many inconclusive conjectures. His third, fifth, sixth and seventh clues are particularly weak, and his first is only suggestive. Only his second and fourth clues—one concerned with Jesus’ departure from Ναζαρά, the other with Jesus’ entry into Capernaum—are compelling. Together they constitute a significant agreement in content between Matthew and Luke against Mark.

More recently, David B. Peabody has argued that Walker’s seven ‘clues’ can be explained equally well, if not better, by assuming that Luke used Matthew. In a study entitled ‘Repeated Language in Matthew: Clues to the Order and Composition of Luke and Mark’, Peabody first identified two presuppositions: that the gospel writers used scrolls rather than codices and that the gospel autographs had no chapter, verse or other formal indicators of structure. He then asked, ‘In the absence of such conveniences, in the earliest Christian communities, how might cross references within and among the gospels have been made?’ While allowing that the second and third synoptists may have marked the copy or copies of the gospel(s) in their possession, Peabody identified a Matthaean literary feature that may have served as an ‘internal cross referencing system’ for a later writer:

In this paper I intend to use certain linguistic clues within the text of Matthew to hypothecate about some of the literary history (Redaktions-geschichte) of the Synoptics. Specifically, assuming the Two Gospel Hypothesis, I will attempt to explain some of Luke’s and Mark’s compositional procedures assuming they have joined together two or more separated literary contexts from Matthew. All of the separated literary contexts in Matthew that I will discuss contain not only similarities in content, but also similarities or identities in language. By contrast, neither the literary contexts in Mark nor those in Luke, parallel to these units in Matthew, contain much, if any, of this repeated language of Matthew. I will argue that Luke and/or Mark may have utilized some of these similarities and/or identities in language within the text of Matthew to point them to separated literary contexts in Matthew. I will then go on to explain how Luke and/or Mark would have united these separated literary contexts in Matthew into new literary units in their texts parallel to Matthew.

For Peabody, Luke’s and/or Mark’s dependence on Matthew better explains Matthew’s repeated language than does Matthew’s dependence on Mark (or Luke) because he considered it unlikely that after dividing a literary unit from his source and placing part of that unit in a different context, Matthew would then introduce ‘similar and/or identical language into each of his, newly separated, literary contexts’. More specifically, he found it difficult to understand why Matthew would divide literary units from his source and then re-introduce ‘literary clues’ to their original unity in each instance. ‘I find this hypothetical literary procedure less likely’, he wrote, ‘than a literary procedure whereby a later author would choose to unify separated literary contexts on the basis of a common theme and shared language within those contexts’.78

In short, Peabody argued that Matthew’s repeated phraseology is original to Matthew—perhaps serving as an aural indicator of structure and narrative development—and that Luke or Mark (or both) occasionally united or transposed Matthaean materials on the basis of these repeated phrases. He considered this a more feasible explanation than that Matthew divided literary units in Mark or Luke only to reveal their original unity by employing similar or identical phrases in both instances.

The first cluster of pericopes that Peabody analysed was Lk. 4:16–30, Mt. 4:23–7:29 and Mk 1:21–22. He sought to show that similarities between Lk. 4:16–30 and both Mt. 4:23–7:29 and 13:53–58, but not Mk 1:21–22 and 6:1–6, are suggestive of Luke’s dependence on Matthew rather than Mark at this point.

Peabody noted that Matthew introduced his Sermon on the Mount after only four of the disciples had been called. This suggested to him that when Luke arrived at Matthew’s introduction to the Sermon on the Mount (Mt. 4:23–5:2), he decided not to include his own shorter version of the same sermon (Lk. 6:20–49) until after he had recounted the call of all 12 disciples, which he does in Lk. 6:12–16. Peabody contended that Luke was aware of the context of Matthew’s Sermon because he chose to include an alternative sermon, the programmatic sermon in Nazareth, in approximately the same context as Matthew’s Sermon on the Mount. However, while the context for Luke’s Nazareth sermon was suggested by the context of Matthew’s Sermon on the Mount, its content was based on Matthew’s account of the rejection of Jesus in his own territory in Mt. 13:53–58. According to Peabody,

The separate literary contexts that contain Matthew’s Sermon on the Mount (Matt 4:23–7:29) and Matthew’s story of Jesus’ rejection in his own country (Matt 13:53–58) could be linked in the reader’s mind by the similar language that concludes Matthew’s Sermon on the Mount (Matt 7:28) and immediately precedes Matthew’s story of Jesus’ rejection in his own country (Matt 13:53).

The similar language Peabody referred to is the slightly varied formula that appears at five junctures in Matthew’s Gospel to indicate the end of a major discourse: Καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησονς … (Mt. 7:28; 11:1; 13:53; 19:1; 26:1). In Peabody’s view, this formulaic expression at the end of both Matthew’s Sermon on the Mount and his Parables Discourse, which immediately precedes his account of the rejection of Jesus in his own territory, might explain Lukan parallels to both Mt. 7:28–29 and Mt. 13:53–58 in Lk. 4:16–30.

Peabody continued by noting that the formulaic expressions that ‘potentially link’ Matthew’s two passages are absent from Mk 1:21 and 6:1. In his view,

It is, therefore, less obvious why Luke would have seen the two separated contexts in Mark as related and brought them together (Mark 1:21–22 and 6:1–6) than it is to see how Luke would have seen the parallel contexts in Matthew as related and brought them together (Matt 4:23–7:29 and Matt 13:53–58).

Given the relatively frequent repetition of the formula at Mt. 7:28–29 and 13:53, however, one is forced to ask why Luke would think the passages preceding one of them and following another were ‘related’ in some way. Furthermore, given the obvious function of the formula as a transition between sections of discourse and narrative, why would Luke consider it reasonable to conflate two passages simply because they occurred in close proximity to two instances of a transitional formula?

After displaying the relevant texts side by side to indicate possible literary dependence, Peabody argued that while Luke changed the setting of his Nazareth sermon from ‘Jesus’ own country’ to ‘Nazareth’, he betrays dependence on Matthew by using the same word as Matthew used, πατρίς, in Lk. 4:24. He also suggested that the reference to Ναζαρά in Mt. 4:13 influenced Luke’s decision to locate his first sermon at Nazareth. Peabody then turned his attention to Mark:

If Mark had access to both Matthew and Luke, when he came to Matthew 4:23–7:27/Luke 4:16–30, he was faced with two different sermons, Matthew’s Sermon on the Mount and Luke’s Sermon in Nazareth. Rather than try to conflate these two sermons, Mark chose to omit both and provide a substitute summary. Mark does, however, give evidence of having known a sermon at this point in his gospel by providing a Markan summary passage about Jesus entering ‘Capernaum’ and ‘teaching in the synagogue’ (Mk 1:21).

When Johann Jakob Griesbach first proposed that Mark used Luke’s Gospel as well as Matthew’s, he noted that Mark omitted Matthew’s Sermon on the Mount because it was ‘too verbose’ and because the Jewish character of much of the material in it made it unsuitable for his audience. He said much the same thing about Luke’s sermon in Nazareth, although he also noted that Mark chose to recount Matthew’s version of Jesus’ rejection at Nazareth in a more similar context.84 Peabody argued that the presence of two different sermons at the same relative point in his sources suggested that Mark’s best editorial option was to omit both rather than to conflate them and to substitute a summary as he had done at Mk 1:12–13 when faced with the conflicting sequence of the temptations in Mt. 4:1–11 and Lk. 4:1–13.

Peabody next turned his attention to a formulaic expression that appears in near identical form in Mt. 7:28 and 22:33, but in more varied form in Mk 1:22 and 11:18. A similar expression occurs once in Luke’s Gospel at Lk. 4:32. According to Peabody, given that this ‘formula’ is repeated in the Gospels of Matthew and Mark, but occurs only once in Luke’s Gospel where Luke is parallel to both Matthew and Mark, its presence in Luke’s Gospel is best explained by copying. In short, ‘Luke is secondary to Matthew or Mark on the basis of this evidence alone’. Here Peabody’s argument relied on Farmer’s redactional criterion, according to which a pericope that contains literary or linguistic features clearly attributable to another synoptist should be judged as secondary to its parallels provided the relevant literary or linguistic feature occurs only in pericopes that are close literary parallels.87

Peabody also argued that while nothing in Mk 1:22 or 11:18 could not be explained by Markan dependence on Matthew’s parallel texts, the plural subject ὄχλοι and the plural verb in both Matthaean texts cannot be explained by Matthew’s dependence on Mark’s texts, provided Mk 11:18 is parallel to Mt. 22:33. ‘It would appear’, according to Peabody, ‘that the text of Mark represents a fragmentary preservation of a formula traceable to the author of Matthew’. Alternatively, if Mt. 22:33 and Mk 11:18 are not parallels, ‘then the direction of literary dependence is inconclusive from these data alone inasmuch as both Matthew (22:33) and Mark (11:18) can use the formula in contexts independent of one another as well as in parallel with one another (Mark 1:22/Matt 7:28)’.

To conclude his discussion of Lk. 4:16–30 and potentially related pericopes, Peabody first asked how Matthew’s and Luke’s independent decisions to insert ‘extended, though different, sermons’ into the Markan framework at the same point (between Mk 1:21 and 22) is to be explained on the Markan hypothesis. In his view, this ‘represents a significant compositional agreement of Luke and Matthew against Mark’. However, Markan priorists would probably experience little discomfort as a result of this observation. Perhaps the first thing that could be said in defence of Matthew’s and Luke’s independent use of Mark’s Gospel at this point is that Mk 1:21–22 is the first point in Mark’s narrative that seemed suitable for expanding into a sermon. Given Matthew’s concern to record Jesus’ Sermon on the Mount as early as possible in Jesus’ ministry and Luke’s desire, for programmatic reasons, to relate the rejection of Jesus at the beginning of Jesus’ ministry, what seemed redactionally feasible to one also seemed feasible to the other. In short, their agreement against Mark at this point is explicable. However, when coupled with Walker’s observation that Matthew and Luke also agree together against Mark by referring to Jesus’ departure from Ναζαρά and his arrival in Capernaum in much the same contexts, Peabody’s point gains in strength.

After asking how Matthew’s and Luke’s ‘agreement’ to include a sermon at the same relative point in Mark’s outline can be explained on the Markan hypothesis, Peabody then reiterated his point that Luke’s text can be adequately explained on the grounds that Luke perceived repeated language in different contexts of Matthew’s Gospel as sufficient reason to unite different incidents in Matthew into a single literary unit in his own Gospel:

the text of Luke in these contexts can be explained on the basis of Luke’s use of repeated language within the text of Matthew to link together separated literary contexts in Matthew into a single literary context in Luke. Since the language that might be used to link separated literary contexts in Matthew does not appear in parallel texts of Mark, it cannot be used to help explain how Luke might have used Mark in these contexts.

A question concerning Peabody’s discussion of Lk. 4:16–30 and potentially related passages is: Why would Luke consider repeated language in Matthew’s Gospel to be a suitable reason for transposing material or amalgamating features of two separate incidents into one literary unit? This seems too mechanical a procedure for someone of Luke’s literary calibre. Furthermore, Peabody devoted a lengthy study to how cross-references within and between the gospels may have been made, but nowhere did he ask these fundamental questions: Why was a system of cross-referencing needed? Did the earliest readers feel the need for cross-referencing between the gospels?

While neither Walker’s nor Peabody’s argument is particularly compelling, an overriding impression from one’s reading of their studies is that before his Nazareth pericope Luke is much more similar to Matthew than to Mark. Despite their widely divergent accounts leading up to Jesus’ public ministry, Matthew and Luke each contain birth narratives while Mark does not. In addition, though quite different and placed in different contexts, Matthew and Luke each provide a genealogy prior to this point in their respective Gospels, whereas Mark provides none. Matthew’s and Luke’s accounts of John the Baptist’s preaching are more similar to each other than either is to Mark, primarily because each relates more of John’s preaching than does Mark (Mt. 3:7–10, 12; Lk. 3:7–9, 17). Matthew and Luke each record in strikingly similar detail three specific temptations of Jesus in the wilderness, albeit in a different order, while Mark prefers to give a summary statement that simply mentions Jesus being tested in the wilderness. Only Matthew and Luke record that at the outset of his ministry Jesus left Ναζαρά before going to Capernaum (Mt. 4:13; Lk. 4:16–30). Against Matthew and Luke, Mark records Jesus calling his first four disciples before entering Capernaum. Finally, only Matthew and Luke provide any specific teaching of Jesus at this point in their Gospels. Restricting attention to these early sections of the synoptics, one must admit that Luke’s dependence on Mark is capable of explaining the phenomena, but Luke’s dependence on Matthew is more compelling. Certainly one need not appeal to Q or other hypothetical sources common to Matthew and Luke. Drury highlighted the non-Markan material common to Matthew and Luke between the preaching of John the Baptist and the temptations, which occur in the same order. As a result, he rightly pointed out that ‘there is no need of the Q theory up to the end of the temptation story’. On his own theory, however, there is also no need to appeal to Luke’s use of Mark. Of course, after the threefold agreement in relative order at Mt. 4:12–17, Mk 1:14–15 and Lk. 4:14–15, the respective relations between the synoptics become more complex. But the remarkable similarities between Matthew and Luke in their respective accounts of the beginning of Jesus’ public ministry should not be overlooked.

It is now time to examine afresh certain features of Lk. 4:16–30, with particular attention to what it reveals about Luke’s potential relation to Matthew or Mark or both. As Schreck’s discussion of this pericope reveals, Luke’s idiosyncratic form and placement of this story have stimulated much discussion and disagreement. Some of the major points of dispute are: (1) whether Luke had access to a non-extant source, in which a rejection story occurred at the beginning of Jesus’ ministry;94 (2) the significance of Ναζαρά at Lk. 4:16; and (3) various difficulties associated with Lk. 4:23, including its apparently incongruous relation to the material immediately preceding it, the meaning and suitability of the proverb Jesus quotes, the meaning of the verb ἐρεῖτέ and the source-critical significance of the reference to Capernaum. What follows focuses on the final two points associated with Lk. 4:23.

A preliminary point is that even if Lk. 4:16–30 is not dependent on Matthew’s or Mark’s version of Jesus’ rejection in Nazareth, it is nevertheless Luke’s substitute for at least one of their accounts. Luke does not record a similar rejection story in a later context, even though he parallels Mark immediately before and immediately after Mark’s account of Jesus’ rejection in his ‘home town’ and shortly thereafter parallels both Matthew and Mark briefly (Lk. 9:7–17) before a block in which only Matthew and Mark share material (Mt. 14:22–16:12; Mk 6:45–8:21). Provided Luke was familiar with an account that placed Jesus’ rejection in Nazareth nearer to the end than to the beginning of his Galilaean ministry, Lk. 4:16–30 can be regarded as a Lukan transposition, even if derived from another source.

The crucial question is: Was Luke familiar with a narrative sequence in which Jesus’ rejection in Nazareth or his ‘home town’ occurred later in the Galilaean ministry rather than at the beginning? Luke 4:23 suggests he was. Although Luke records the incident in the Nazareth synagogue as the first specific event in Jesus’ public mission, he records Jesus as saying: Πάντως ἐρεῖτέ μοι τὴν παραβολὴν ταύτην· ἰατρέ, θεράπευσον σεαυτόν ὅσα ἠκούσαμεν γενόμενα εἰς τὴν Καφαρναοὺμ ποίησον καὶ ὡδε ἐν τῇ πατρίδι σον. (‘Probably you will quote me this proverb: “Doctor, heal yourself. What we have heard has happened in Capernaum, do here also in your own territory.’ “) Some think that Lk. 4:23b is adequately prepared for by Luke’s summary at Lk. 4:14–15. For example, according to David R. Catchpole, ‘all that is needed to give backing and wholly adequate preparation for v. 23b is to be found in Lk. 4:14 …’ Had Luke omitted the reference to Capernaum one might agree. Luke 4:14–15 clearly indicates both widespread renown for Jesus throughout Galilee and teaching activity on the part of Jesus in Galilaean synagogues. Although Luke has not yet referred to any specific incident in Capernaum, one might surmise that Jesus may have visited Capernaum.

On the other hand, there was no need to refer specifically to Capernaum in Lk. 4:23. The same point could have been made by referring to what Jesus had done elsewhere in Galilee. In addition, as Bock points out,

The reference to Capernaum in Luke 4:23 is raised without any qualifying description. This stands in contrast to 4:31, where Capernaum is described as a city in Galilee, a type of note that usually accompanies the introduction of a city for the first time … This probably indicates that the reference in 4:23 has been relocated ahead of where the first reference to Capernaum existed in the tradition.

Finally, Lk. 4:23 refers to miracles, but so far Luke has recorded no miracle by Jesus. In short, Lk. 4:23 refers to a previous miraculous ministry in Capernaum, for which the summary in Lk. 4:14–15 does not adequately prepare.

Another circumvention of the difficulty presented by the reference to Capernaum in Lk. 4:23 is to understand the verb ἐρεῖτέ in a future sense. For example, according to Conzelmann, ‘The future tense ἐρεῖτέ points to a future rather than a present utterance by those who are being addressed’. In his view, this saying is later fulfilled when in Lk. 8:19–21 members of Jesus’ family wish to see him. By explaining their wish to see Jesus along the lines of Herod’s desire to see Jesus in Lk. 9:9 and 23:8, Conzelmann claimed that ‘they want to take him to Nazareth in order that he may work miracles in what they consider the proper place’. Although Robert Tannehill did not defend Conzelmann’s specific interpretation of ἐρεῖεῖτέ in Lk. 4:23, he nevertheless agreed that it should be understood to refer to some future period: ‘In the light of the fact that there is no clear reference to events at Capernaum prior to Jesus’ coming to Nazareth, it is doubtful that Luke understood iv 23 to refer to such prior events’.

Conzelmann and Tannehill have found little support for their interpretation of ἐρεῖτέ. Consequently, one can hardly resist the conclusion that Lk. 4:23 betrays Luke’s interference with a traditional sequence of events. As Bruce Chilton stated in his analysis of Lk. 4:16–21, “Luke … retains traces of the order he altered, and consciously treats 4:16–21 (which Matthew and Mark do not parallel) as Jesus’ programmatic statement (to which vv. 14, 15 serve as introduction and vv. 22–30 as corollary)’. Or as Bo Reicke wrote in connection with Lk. 4:16–30, ‘Because it presupposes a knowledge of earlier healings in Capernaum (4:23), this pericope must have been moved by Luke to the beginning of the Galilean section’.

In conclusion, both the position of Lk. 4:16–30 within Luke’s narrative and details within this pericope indicate that Luke’s version of Jesus’ rejection at Nazareth is secondary to that represented in Mt. 13:53–58 and Mk 6:1–6. Whether Luke is dependent on either Matthew or Mark (or both) in this pericope is less certain, although the evidence for the influence of Mt. 4:12–17 on Luke’s placement of this episode is significant, even if not overwhelming.

Luke 5:1–11, The Call of the First Disciples

In an article entitled ‘St Luke’s Transpositions’, H.F.D. Sparks noted that some of Luke’s alleged transpositions ‘are capable of being explained just as satisfactorily as “parallels”, or rather “replacements”, derived from a so-called “special source” ‘. Luke 5:1–11 is one of a number of instances for which critics often appeal to a special source or variant tradition. Nevertheless, however one explains the unique content of this pericope, provided a literary relation exists between Luke and at least one of the other two synoptics, Luke clearly chose a different context for his version of the call of the first disciples.107

To begin with, the degree of divergence in order at this point between Luke and the other two synoptics warrants attention. In Lk. 3:1–4:14, Luke’s order is most similar to that of Matthew by virtue of more shared material in the same basic order, but it is also similar to Mark’s. Of course, Luke included special material, for example, Lk. 3:10–14 (John’s Ethical Instruction) and Lk. 3:23–38 (Genealogy of Jesus), and he also diverged from the common sequence by relating John the Baptist’s imprisonment before the baptism of Jesus. At Lk. 4:14, where Jesus returns to Galilee, Luke agrees in relative order with both Matthew and Mark. Matthew 4:12–17, Mk 1:14–15 and Lk. 4:14–15 are all transitional pericopes that set the stage for Jesus’ Galilaean mission. It is at this point that Luke has Jesus proclaim his ‘inaugural sermon’ in Nazareth, which results in rejection and his departure to Capernaum. On the other hand, Matthew and Mark record Jesus’ recruitment of his first four disciples in almost identical terms.

So far, Matthew and Mark agree in relative sequence, but immediately after the call of the first disciples their accounts diverge dramatically: Matthew uses a summary pericope (Mt. 4:23–5:2) to lead into his Sermon on the Mount while Mark has Jesus enter Capernaum to teach in the synagogue on the Sabbath (Mk 1:21–28). Capernaum is then the setting for a series of incidents in which Mark and Luke agree not only in content but in relative order (Mk 1:21–39; Lk. 4:31–44). Only after the Capernaum sequence does Luke recount the call of the first disciples in conjunction with a story about a miraculous catch of fish. A preliminary point, then, is that if one assumes Luke’s familiarity with Matthew, Mark or both (or even a traditional sequence of events reflected in the other two synoptics), his decision to delay his account of Jesus recruiting his first disciples follows immediately upon his decision to locate his version of Jesus’ rejection in Nazareth at the outset of Jesus’ public ministry. There was nothing to stop Luke recounting the call of the first disciples after the Nazareth incident. However, having decided to begin his account of Jesus’ Galilaean mission with the rejection at Nazareth, perhaps it was inevitable that Luke would delay the call of the first disciples until Jesus is shown to have won more positive public acceptance.109 In any case, Lk. 4:16–30 and Lk. 5:1–11 appear to be connected in terms of Luke’s compositional procedure.

With reference to Luke’s account of the call of Jesus’ first disciples, Goulder notes: ‘The unsatisfactoriness of the Marco-Matthaean story has been felt by many expositors since Luke’. He means that Matthew’s and Mark’s call stories suggest that four fishermen were prepared to leave family and livelihood to follow a stranger solely on the strength of his challenge to follow after him. Matthew and Mark emphasize the immediacy of the respective decisions by two pairs of brothers to ‘drop everything’ to follow Jesus. To this point, Matthew has noted Jesus’ ‘retreat’ into Galilee (4:12), his departure from Nazareth to settle in Capernaum (4:13), and the datum that after settling in Capernaum Jesus began to preach and to say, ‘Repent, for the kingdom of the heavens has drawn near’ (4:17). For his part, Mark has noted simply that Jesus went into Galilee preaching the good news of God, which featured a call to repentance because of the nearness of God’s kingly rule (1:14–15). It is true that Mt. 4:17 and Mk 1:14–15 each records Jesus embarking on a preaching ministry, but each reads like an anticipatory summary to be fleshed out in what follows. Their respective call stories therefore recount the first specific incident in Jesus’ public ministry. As a result, the collective decision of the four fishermen appears decidedly rash, even irresponsible.

This is so, of course, only if one expects Matthew and Mark to provide a psychologically satisfying explanation for four fishermen abandoning their livelihoods to follow a stranger. This they were clearly not concerned to offer. The emphasis in Mk 1:14–45 is to illustrate the authority (1:22, 27) and charisma (1:28, 37, 45) of Jesus. Mark ‘explains’ the disciples’ willingness to abandon their work in order to follow Jesus by what he records about Jesus in Mk 1:21–45; their ‘rashness’ becomes more comprehensible in view of Jesus’ charisma and authority, which is amply illustrated in his teaching, healing and exorcisms. In addition, the story of Jesus healing Simon’s mother-in-law might be understood to demonstrate the benefits of the decision to follow Jesus. In short, by the end of Mark 1, Jesus’ first four followers are vindicated, so to speak, in their decision to follow the stranger who called them away from their work and family responsibilities.

Matthew’s sequence at this point is less easy to explain. He mentions Jesus’ ‘disciples’ in the lead-in to his Sermon on the Mount (5:1), so perhaps the call of some followers prior to this was required (even if the fishermen are not described as disciples in Mt. 4:18–22). According to W.D. Davies and Dale C. Allison,

Matthew wishes to place the great sermon at the beginning of Jesus’ public ministry (and thus before the events in Capernaum are recounted). At the same time, since [chs.] 5–7 is teaching for the disciples as well as the crowds, the evangelist must place the call of Peter and the others before 5:1: ‘the disciples came to him’.

In any case, Matthew’s placement of his call story creates more ‘narrative dissonance’ than does Mark’s. The delay caused by the long teaching block in Mt. 5:1–7:29 and the absence of any reference to his newly called followers in Mt. 8:5 (when Jesus enters Capernaum) or Mt. 8:14 (when Jesus enters Peter’s house) makes one wonder whether it would not have served Matthew’s purpose better to delay the call of the first disciples until Jesus’ descent from the mountain in Mt. 8:1. Mark’s call story is abrupt, but it is duly ‘explained’ in subsequent episodes; Matthew’s call story is not only abrupt, but it also fits awkwardly into Matthew’s narrative and structure. Goulder’s comment regarding the ‘unsatisfactoriness of the Marco-Matthaean story’ should perhaps be reserved solely for Matthew’s account.

Despite the danger of providing reasons for Luke’s transposition of his call story to a later context, few are reluctant to supply one. Bultmann, who considered the story of the miraculous catch of fish to have been ‘developed’ from the saying of Jesus about making fishermen into ‘fishers of men’, noted in his discussion of Luke’s editing of traditional materials that ‘he takes the story of the calling of the disciples, told very differently by him and places it later at 5:1–11 because it is psychologically more probable that Jesus had been active for some time before he called any disciples’. Similar, but more explicit, is Fitzmyer’s explanation:

By transposing the scene from its Markan setting, Luke has eliminated the oft-noted implausibility of the Markan story about the call of the four disciples—the first thing that Jesus does in that Gospel after his baptism and desert sojourn. In the Lucan context, Jesus has been seen preaching and healing, and Simon (at least) has witnessed one of his mighty deeds (4:38–39). The preceding Lucan scenes thus provide a psychologically plausible setting for the call of Simon the fisherman.

While possible, there is no way of knowing whether this was Luke’s reason for recounting his call story in a slightly later context. Indeed, there is good reason to be suspicious of such an explanation, as Goulder points out:

Fitzmyer does not offer any evidence for Luke’s supposed interest in psychological plausibility. Such a concern cannot be said to be obvious after a chapter or so in which John is put in prison before baptizing Jesus; in which Jesus preaches a sermon about his anointing with the Spirit which suddenly veers off into the sending of the prophets to the Gentiles; in which the mention of the Gentiles suddenly fills the congregation with murderous rage; and in which Jesus refers to his earlier healings, of which Luke has given no hint.

In addition, since the explanation appealing to psychological plausibility invariably presupposes Luke’s familiarity with Mark’s account, Luke must be seen to have been dissatisfied with Mark’s placement of his call story. If so, however, it is strange that Luke failed to appreciate the dramatic impact of Mark’s order of incidents at the beginning of Jesus’ ministry. In a sense, it would be more understandable to delay the call story if one were following Matthew’s order of incidents. On the other hand, perhaps Luke’s chosen context for his call story was linked to his decision to recount the rejection of Jesus at Nazareth at the outset of his ministry. As noted above, were Luke following the order of Mark or Matthew (or both), his decision to delay the call story follows hard upon his decision to use the Nazareth episode as a frontispiece to Jesus’ mission. Perhaps this partly explains Luke’s placement of his call story; the fishermen’s positive response to Jesus’ call would form an odd sequel to the hostile response to Jesus in Nazareth.

Luke’s later placement of his call story is certainly explicable on the assumption that he altered the Markan or Matthaean sequence. This is not proof that Luke was dependent on either Mark’s or Matthew’s account. Indeed, there is nothing intrinsically implausible about the suggestion that Mark or, indeed, Matthew transposed the call story to the beginning of Jesus’ public ministry. Mark’s version in particular is psychologically satisfying provided one reads it within the context of subsequent events, in which Mark underscores Jesus’ authority and charisma.

However, another feature of Luke’s narrative indicates that his call story is secondary. In the Capernaum sequence preceding Lk. 5:1–11, Luke recounts Jesus entering Simon’s house and healing his mother-in-law in a way that suggests prior acquaintance with Simon. As John Rist noted, ‘Luke causes unnecessary confusion by relating the story of Simon’s mother-in-law, and thus introducing Simon (4:38) before the call of Simon itself (5:1–11)’. Similarly, Goulder’s comment on Lk. 4:38 is that ‘Andrew is dropped, as he is at 5:1 ff; as are James and John, who have not yet been called in Luke. Simon is left without introduction by oversight’. For Drury, Luke’s decision to delay his call story resulted in an unintended ‘slip’:

Jesus heals Simon’s mother-in-law without the reader having any idea of who Simon is or how he is connected to Jesus, as he would have if Mark’s order were intact. He is more careful in the next incident … Jesus is tracked down by the crowd instead of Mark’s ‘Simon and those that were with him’.

According to Harold Riley, ‘At his first reference to persons or places, Luke is careful to say whom or what he is talking about. He introduces his characters, and when he refers to places explains what they were’. If true, the first reference to Simon in Lk. 4:38 is a significant exception. Riley noted this and asked: ‘Was his [Simon’s] name already a household word to the readers?’ Equally, if not more, likely is that Luke was familiar with a sequence of incidents in which the call of the first disciples occurred at the beginning of Jesus’ public ministry, but in choosing to delay the story neglected to introduce Simon at the first mention of his name because he momentarily failed to appreciate that by delaying his call story, Simon would be new to his reader or hearer at this point.

Tuckett has offered another reason why Luke may have delayed his version of the call of the first disciples. Accepting that ‘the later position of the call story does help to reduce the arbitrariness of Peter’s response’, he also noted that the alleged ‘delay’ of Luke’s call story may equally well be understood as a ‘bringing forward’ of the Capernaum sequence in Lk. 4:31–44. Indeed, for Tuckett, ‘Part of the reason for the different order may be a desire by Luke to link the events in Capernaum more closely to the Nazareth story’.

Tuckett claimed that the events in Lk. 4:31–44 are ‘exorcistic’ in character. Of the three pericopes concerned with Jesus’ activity in Capernaum, two specifically refer to Jesus exorcizing demons, but Tuckett noted that Luke records Jesus ‘rebuking’ Simon’s mother-in-law’s fever, which is ‘exorcistic language’. Referring to Acts 10:38 as an example of Luke’s interpretation of the Isaiah 61 text said by Jesus to be fulfilled in Lk. 4:17–21, Tuckett suggested that Luke wished to narrate the incidents in Capernaum immediately following the Nazareth episode to illustrate the liberating power of Jesus. As Tuckett put it,

Acts 10:38 gives what appears to be Luke’s understanding of the ‘liberation’ promise in Isa. 61 // Luke 4 as referring to Jesus’ ‘healing all who were oppressed by the devil’. Luke’s order in Luke 4 may thus be due to his desire to bring out this facet of Jesus’ ministry by bringing forward the events in Capernaum and redacting the material so that the language is predominantly exorcistic. The change in order may thus be due to Luke’s interpretation of the Isa. 61 material and is thus at least as much an anticipation of Mark 1:21–39 as a postponement of Mark 1:16–20.

Tuckett also suggested an alternative reason why Luke may have decided to delay his version of the call story. In his view, not only does the Nazareth pericope in Lk. 4:16–30 prefigure Jewish hostility to the gospel and the church’s mission to the Gentiles, but Lk. 5:1–11 depicts the expansion of Jesus’ mission. Since it is possible to view Lk. 5:1–11 as in some sense prefiguring the church’s expansion, in part because the person on whom Luke’s call story focuses is the one who later initiates the mission to the Gentiles, it is therefore an appropriate beginning to this section of Luke’s Gospel.

For Tuckett and many others, Lk. 5:1–11 is readily explicable on the two-document hypothesis, with allowance made for Luke’s use of special source material. For Goulder, however, Lk. 5:1–11 reveals Luke’s dependence on Matthew as well as Mark. In an article entitled ‘On Putting Q to the Test’, Goulder discussed 12 passages ‘where it appears that Luke agrees with Matthew against Mark in expressions which he, Luke, never uses elsewhere, or to which he shows a marked aversion; or which are unnatural; or in matters of order’. For Goulder, any such agreements between Matthew and Luke against Mark are most simply explained by Luke’ s familiarity with Matthew. The second of his 12 passages is Lk. 5:1–11. Indeed, the focus of his discussion is the agreement in order between Matthew and Luke against Mark at this point. According to Goulder,

The Markan order would be logical in Luke also, for in Luke 4:38 Jesus enters Simon’s house without Simon being mentioned before. Something therefore is drawing the story of the call of the first apostles out of its Markan position. It might be, clearly, that Luke knew the story in a later context in another source, parallel here to Mark, of which the call of the first apostles was a part, but coming after the Capernaum ministry instead of before. But then we have just such a source before us in Matthew.

The weakness of this argument is that Goulder regards Mt. 4:13–17 as a Capernaum ministry, whereas all that Matthew says about Capernaum at this point is that Jesus settled there after leaving Nazareth. As Tuckett noted in response to Goulder, ‘Goulder’s theory requires Luke’s thinking that Matt. 4:13–17 corresponds to the Capernaum ministry of Mk. 1:21–39’. Tuckett regarded this as unlikely and suggested instead: ‘The repositioning of the call story is also conceivable as a change by Luke in view of Luke’s predilection for stressing the value of miracles in bringing people to faith’.134

Goulder has responded to Tuckett’s criticisms, and in the case of Lk. 5:1–11 sees no difficulty in thinking that Luke may have regarded the Capernaum sequence in Mk 1:21–39 as ‘summed up’ in Mt. 4:13–17. In Goulder’s view, there was no need for Luke to note that Jesus was preaching by the lake in Lk. 5:1–2. Certainly Mk 1:16–20 does not suggest Jesus was preaching by the Sea of Galilee before he called the four fishermen to follow him. For Goulder, however, Mt. 4:17–22 can be read as if Jesus was preaching prior to calling his first disciples: ‘Luke could quite naturally have read the two sentences [Mt. 4:17–18] together, and combined proclamation and calling as he has in 5:1–11’.

Goulder is right. Luke could have interpreted Mt. 4:17–18 in this way. However, it is difficult to make a convincing case for this, and Goulder’s explanation of Luke’ s disagreement in order at Lk. 5:1–11 is too tenuous to be convincing. What must be granted is that Goulder has offered possible explanations for Luke’s dependence on Matthew at places where the majority of critics appeals too quickly to hypothetical sources.


My reappraisal of three Lukan pericopes that occur in a different relative order from their functional parallels in Matthew and Mark has discovered nothing to contest the widely held view that Luke’s placement of these pericopes is secondary. In each case, Luke betrays his rearrangement of a traditional order of events that is at least reflected in the Gospels of Matthew and Mark. This is especially so in the case of Lk. 3:19–20 and 4:16–30, but there is also enough evidence to regard Lk. 5:1–11 as a rearrangement of a traditional order of events. In addition, on form-critical grounds, Lk. 5:1–11 appears secondary to Mt. 4:18–22 and Mk 1:16–20. As George Wesley Buchanan has argued, Lk. 5:1–11 looks to be an expanded chreia developed from the ‘purer’ chreia form in Mt. 4:18–19 and Mk 1:16–17. On the basis of the foregoing analysis, then, one can say that Luke is secondary. The phenomenon of order does not support those who argue for Lukan priority.

Beyond this more general conclusion, is it possible to identify which of the other two synoptic gospels Luke may have used? The conventional view that Luke borrowed from Mark remains feasible. On the other hand, as Tuckett has conceded, reasons offered for Luke’s alleged rearrangement of Mark’s pericope order can also be used to explain Luke’s rearrangement of Matthew’s pericope order. Furthermore, Goulder has shown that in relation to Lk. 4:16–30 and 5:1–11, Matthew’s narrative may have influenced Luke’s transposition of these pericopes. On this point, he is supported by advocates of the two-gospel hypothesis. One cannot extrapolate from this, but in the pericopes discussed above Luke’s dependence on Matthew alone is at least as plausible as Luke’s dependence on Mark alone, and in the case of Lk. 4:16–30 it seems more plausible that Luke borrowed from Matthew alone than that he borrowed from Mark alone. On the other hand, Luke’s dependence on Matthew alone is no more plausible than Luke’s dependence on both Matthew and Mark. The data permit no conclusion more refined than this.

Chapter 7

The Disagreements in Order in Matthew 4:23–9:35

In 1984, at the Symposium de interrelatione evangeliorum in Jerusalem, Frans Neirynck remarked, ‘The phenomenon of order, and particularly the relative order of Mt. 4:23–13:58/Mk 1:21–6:13, is a major issue in the discussion of the synoptic problem’. Disagreements in the order of triple-tradition material between the Gospels of Matthew, Mark and Luke have been at the heart of synoptic source criticism since its scientific inception with Johann Jakob Griesbach. Later, the credibility of the Markan hypothesis was based to a significant extent on critics’ ability to explain Matthew’s disagreements with Mark’s order as editorial modifications of Mark’s narrative sequence. For example, in an essay published in Studies in the Synoptic Problem, Sir John Caesar Hawkins noted that Matthew 8–13 and Lk. 9:51–18:14 are two exceptions to the general rule that Matthew and Luke do not ‘desert’ Mark’s order and arrangement. He then wrote:

On the one hand, it can hardly be doubted that in Mt. 8–13 the compiler had our Mark, or its general equivalent, before him … In those chapters of Matthew therefore, the main task of students of the Synoptic Problem is to discover the reasons which induced Matthew (meaning the compiler of the First Gospel) here, and here only, to break up his Markan source, and to rearrange it among other materials, instead of merely inserting those materials into it as it stood.

This chapter is concerned to evaluate explanations of Matthew’s discrepancies in order vis-à-vis the more similar relative orders shared by Mark’s and Luke’s Gospels—at least in their respective accounts of Jesus’ Galilaean ministry—on the theory of Markan priority.

On my arrangement of parallel pericopes, Matthew’s first series of disagreements with the common order of the triple tradition comprises an introductory summary (4:23–5:2), the Sermon on the Mount (5:3–7:27), a transition (7:28–8:1), a pericope concerned with the healing of a leper following Jesus’ descent from the mountain on which the Sermon was spoken (8:2–4), and a pericope concerned with the healing of a centurion’s servant (or son) after Jesus’ return to Capernaum (8:5–13). Restricting attention to the triple tradition, the first two pericopes to be discussed are Mt. 4:23–5:2 and Mt. 8:1–4.

Matthew 4:23–5:2, Matthew’s Introduction to the Sermon on the Mount

Matthew 4:23–5:2 occurs at an important juncture in Matthew’s Gospel because it is at this point that Jesus embarks upon his public ministry. At a formal level, one could argue that Jesus begins his public ministry at Mt. 4:12–13, where he is said to have gone into Galilee, left Nazareth, and settled in Capernaum; or at Mt. 4:17, which relates that he began to proclaim his message of repentance; or even at Mt. 4:18–22, where he is reported to have begun recruiting followers. However, only in the summary at Mt. 4:23–5:2 does Jesus actually begin to encounter those to whom his mission is directed. Possibly no strict divisions should be drawn within this larger segment (Mt. 4:12–5:2), for as Ulrich Luz has said, ‘Matthew does not seem to value delimitations. On the contrary, again and again we find transitional verses or pericopes which because of their relationships to what precedes and what follows can be considered only as transitional pericopes’.

In any case, Mt. 4:23–5:2 has two main functions within Matthew’s Gospel. First, it is widely recognized that Mt. 4:23 is repeated or ‘echoed’ at Mt. 9:35 (and again, though less conspicuously, at Mt. 11:1), thereby enclosing the intervening material. The inclusio formed by Mt. 4:23 and 9:35 indicates that the Sermon on the Mount in Matthew 5–7 and the cycle of miracle stories in Matthew 8–9 belong together in some sense. As Christoph Burger noted, ‘Die gemeinsamen Rahmenverse Mt. 4:23 und 9:35 verbinden die Bergpredigt und die beiden folgenden Kapitel zu einem geschlossenen Block. Die Kapitel 8 und 9 bilden in dieser Komposition unverkennbar ein Gegenstück zu den Kapiteln 5–7’. Indeed, Mt. 4:23 prefigures what Jesus does in Matthew 5–9 by first speaking of Jesus teaching and proclaiming the good news of the kingdom and then referring to his healing activity.

Second, the remainder of the summary in Mt. 4:24–5:2 provides both a general and an immediate setting for what follows. Matthew 4:24–25 expands upon Mt. 4:23 and sets the stage for Jesus’ teaching and healing ministry in Matthew 5–9, whereas Mt. 5:1–2 provides the immediate setting for the Sermon on the Mount by identifying where and to whom it was addressed.

Commenting on this summary in their ICC commentary on the Gospel of Matthew, W.D. Davies and Dale C. Allison assert:

4:23–5:2 is not particularly helpful for solving the synoptic problem. As the commentary shows, the hypothesis that Matthew’s text is based upon Mark and Q has nothing against it. At the same time, our passage and its parallels do not supply any reason to reject the Griesbach hypothesis or any other serious theory of synoptic relationships.

However, at the 1984 Jerusalem Symposium, Neirynck’s textual discussion supporting the two-source hypothesis focused on ‘Matthew 4:23–5:2 and the Matthean Composition of 4:23–11:1’. His opening statement explained why: ‘Mt. 4:23–5:2 is one of the key passages in the study of the synoptic problem’. After noting how the disruption at this point of an established sequential parallelism between Matthew’s and Mark’s Gospels had been important for Griesbach’s view of Mark’s compositional procedure, he summed up the function of this pericope when seen from the perspective of the two-source theory: ‘Mt. 4:23–5:2 is Matthew’s preparation for the insertion of the Sermon (from his second source) in the framework of Mark, and this introduction to the Sermon should be studied in connection with the Matthean editorial composition of Mt. 4:23–11:1’.

Neirynck’s analysis of Mt. 4:23–5:2 progressed on the basis of a threefold subdivision of the pericope: 4:23–24a, 4:24b–25, 5:1–2. Regarding Mt. 4:23–24a, he noted that while Mk 1:39 is generally regarded as the primary parallel of Mt. 4:23, one should not ignore Mk 6:6b, Mark’s parallel to Mt. 9:35, which is largely a repetition of Mt. 4:23. According to Neirynck, ‘Mt. 9:35 precedes 10:1, 5a like Mk 6:6b precedes 6:7, and two differences between Mt. 4:23 and Mk 1:39 are probably due to the influence of this Markan parallel to Mt. 9:35 … The two differences are the imperfect verb περιῆγεν and the participle διδάσκων, each of which appears in Mt. 4:23, Mt. 9:35 and Mk 6:6. Because Matthew twice uses περιάγειν with the accusative (9:35; 23:15), Neirynck averred that ‘the construction with ἐν … in 4:23 looks like a conflation of Mk 1:39 and 6:6b’. But this is too speculative. That Matthew twice used περιάγειν with the accusative is no grounds for thinking it was the influence of Markan texts that caused him to use περιάγειν with the dative. If Matthew’s compositional procedure included conflating passages from disparate sections of Mark’s Gospel, as many think, surely it would have been easier to write καὶ περιῆγεν εἰς ὅλην τήν Γαλιλαίαν διδάσκων.

Neirynck rightly regarded the reference to Jesus’ preaching in Mk 1:39 as ‘an echo of the first summary in 1:14–15’, but he also contended that ‘the complement which is added to κηρύσσων in Mt. 4:23, to τὸ εὑαγγέλιον τῆς βασιλείας, conveniently summarizes Mk 1:14b–15′. Apart from the fact that on the two-source theory Matthew had already summarized Mk 1:14b–15 at Mt. 4:17, there is the view of many that Mark’s use of to τὸ εὑαγγέλιον reflects a secondary development in the gospel tradition. Matthew’s use of this term invariably signifies Jesus’ own message about the advent of the kingdom of heaven, whereas Mark’s use of the term is nearer to the later Christian message about the ‘benefits’ of Jesus’ death and resurrection. According to William R. Farmer,

This emphasis upon the absolute use of ‘the gospel’ or the use of the expression ‘the gospel of God’ is very Pauline; it appears to be a Paulinism that has been taken over by the author of Mark. Since the whole church eventually came to adopt this Pauline usage, there is no reason to think that Matthew or Luke would have omitted this feature of the developing tradition. It follows that the use of this expression in Mark and its absence in Matthew and Luke indicates that Mark is secondary to Matthew and Luke, and that Mark was written in a church which had a great regard for Paul and his gospel.

Harold Riley also considered Mark’s use of this term to reflect a later development:

The word ‘gospel’ is used absolutely once in Acts (15:7, ‘the word of the gospel’ on the lips of Peter; the phrase ‘the gospel of the grace of God’ is used in the speech of Paul at Ephesus, Ac 20:24). Elsewhere in the New Testament the word is used absolutely in the letters of Paul. Its use in the church would be a natural development, and certainly churches that had known Paul’s preaching, or knew one or more of his letters, would be familiar with it. But the fact that Matthew and Luke do not use the word in this way in their Gospels suggests they represent an earlier usage than Mark.

Even advocates of Markan priority concede that to τὸ εὑαγγέλιον in Mark’s Gospel is secondary. Regarding its use in Mk 1:15, Rudolf Bultmann remarked, ‘This is a quite secondary formulation made under the influence of a specifically Christian terminology (πιστεύετε ἐν τῷ εὑαγγελίῳ) which might very well derive from Mark himself’. More recently, Helmut Koester argued that in at least four instances in which Mark used the term εὑαγγέλιον (Mk 8:35; 10:29; 13:10; 14:9), his text is a secondary development. For Koester, however, this is not so much evidence of Mark’s posteriority as of what he described as different stages in the development of Mark’s Gospel:

It seems that even a new Ur-Marcus hypothesis will not illuminate the problem of Mark, nor will the assumption of Matthean priority. Rather, a solution may be sought in analogy to more recent studies of the Gospel of John which have recognized several stages of rewriting in the text of this Gospel, thus testifying to the instability of such texts as they were used by early Christian communities.

In reply to Koester’s argument, David B. Peabody criticized his decision to divide the evidence in Mark’s Gospel on the basis of parallels (or lack thereof) in the Gospels of Matthew and Luke. When viewed as a whole, he argued, it is clear that Mark’s use of τὸ εὑαγγέλιον is consistent throughout his Gospel, irrespective of whether parallels exist in the other two gospels. For Peabody, ‘This uniform usage is most easily explained if one hand is responsible for all of the occurrences of “the gospel” within the text of Mark. In this case Koester’s theory of multiple redactions of the Markan Gospel has obscured the uniformity of the data’. After examining Matthew’s use of τὸ εὑαγγέλιον, he concluded that Mark’s dependence on Matthew accounts for the evidence somewhat better than vice versa:

The Griesbach hypothesis has no difficulty with the fact that the usage in Mark when a parallel in Matthew or Luke is present is no different from a usage in Mark when no parallel is present. On Koester’s reconstruction the agreement of the redactor of Ur-Marcus with the Secret Gospel redactor on this usage is problematic. It is better understood as the work of a single redactor.

Peabody was careful not to claim too much on the basis of Matthew’s and Mark’s respective uses of the term ‘gospel’. However, given that it is widely accepted that τὸ εὑαγγέλιον is used anachronistically at Mk 1:14–15, it was disingenuous for Neirynck to suggest that Matthew summarized Mk 1:14–15 at Mt. 4:23.

Neirynck’s argument also relied on other features of Mk 1:39. In his view, this verse is a summary that extends what Jesus did in Capernaum throughout Galilee. In other words, Mark’s account suggests that what Jesus is reported to have done in Mk 1:21–28 was typical of what he did throughout Galilee. Based on this accurate insight he made an adventurous suggestion:

The Markan parallel which would be expected in Mt. 4:23 after Mt. 4:18–22/Mk 1:16–20 is that story of Jesus’ action in the synagogue of Capernaum. But the arrival at Capernaum (Mk 1:21a) is anticipated in Mt. 4:13 … and Matthew apparently has no parallel to the individual story of Mk 1:23–27. Or should we say that, by anticipating here the generalization of Mk 1:39, he created a substitute for that story? There is a firm indication in that sense. After 4:23 Matthew continues with the conclusion of that story …

Working with Aland’s Synopsis, one is not given the opportunity to see the similarities between Mt. 4:24a and Mk 1:28. Huck-Greeven at least displays Mt. 4:24a as a potential parallel of Mk 1:28, as does Boismard-Lamouille’s Synopsis. Matthew uses the slightly different compound ἀπέρχομαι rather than ἐξέρχοπαι, and he speaks of the report about Jesus spreading throughout Syria rather than the surrounding region of Galilee, but otherwise his text parallels Mark’s closely. Certainly omitting εὑθὺς πανταχοῦ from Mk 1:28 would not have been difficult, as Neirynck pointed out. What is more difficult, however, is to envisage Mt. 4:23 as in some sense Matthew’s ‘substitution’ for Mk 1:23–27. If this were so, surely Matthew would have at least noted Jesus’ exorcism of demons among his other deeds because Mk 1:23–27 recounts an exorcism.

Neirynck had an answer to this objection. He showed that Matthew occasionally ‘suppresses’ references to demon exorcism, for example, at Mt. 17:16 (cf. Mk 9:18), 12:22 (cf. Lk. 11:14) and 15:28 (cf. Mk 7:30). Granted that Matthew emphasized Jesus’ healing activity at the expense of exorcisms, even if only marginally, this does not make more credible Neirynck’s conjecture that Mt. 4:23 is a ‘substitute’ for Mk 1:23–27. Nor do his comments on Matthew’s phrases, πᾶσαν νόσον and πᾶσαν μαλακίαν. In Mt. 4:23, 9:35 and 10:1 these two phrases are used together, preceded by καὶ θεραπεύειν. As Dennis Tevis has noted, ‘This phrase consists of seven elements which occur together in Matt 4:23, 9:35, and 10:1. These elements do not occur together elsewhere in the New Testament’. Both Tevis and Neirynck noted that the term μαλακία occurs nowhere else in the entire New Testament. Neirynck also noted that the word νόσος occurs five times in Matthew but only once in Mark’s Gospel, at Mk 1:34. In his view, the appearance of νόσοις at Mt. 4:24 is the result of borrowing Mark’s only instance of the term at Mk 1:34. This is difficult to credit. If source-critical significance can be attached to this word, which is doubtful, the fact that it occurs in Mt. 4:23 as part of a characteristically Matthaean phrase makes it more likely that its presence at Mk 1:34 is due to its presence in Matthew’s, or better, Luke’s text.

Regarding Mt. 4:23–24a Neirynck concluded:

the phrase of Mt. 4:23 is more directly inspired by Mk 1:39, but, as it is placed just before Mt. 4:24a (= Mk 1:28) and in continuation of Mt. 4:18–22/Mk 1:16–20, Mt. 4:23b can be seen as the Matthean parallel to Mk 1:21, and Mt. 4:23d θεραπεύων … can be considered the Matthean substitute for the (exorcism-)healing in the synagogue (Mk 1:23–27).

Assuming Neirynck’s conception of Matthew’s compositional procedure, how is one to imagine Matthew writing up this brief summary? He has only just recounted the call of the first four disciples, based on Mk 1:16–20. His rewriting of Mark at this point is negligible. After transcribing the final words of Mt. 4:22 he perhaps picks up Mark’s Gospel again and resumes reading at Mk 1:21, where Mark has Jesus enter Capernaum and teach in the synagogue on the Sabbath. Matthew has already determined that he wishes to place a block of Jesus’ teaching in a prominent position near the beginning of Jesus’ public ministry. Mark 1:21 records the first instance of Jesus teaching in Mark’s Gospel. It is especially felicitous that Mk 1:22, which records the crowd’s response to Jesus’ teaching, could not have been better designed to serve as a conclusion to Matthew’s block of teaching material, now known as the Sermon on the Mount. He therefore ignores the details in Mk 1:21 and sets about writing a suitable introduction to his Sermon on the Mount.

Thus far everything predicated of Matthew is simple and well within the compositional conventions of his time. But, following Neirynck, one must now imagine Matthew reading through Mk 1:21–39 looking for motifs and phrases with which to construct the first part of his summary introduction. He chooses Mk 1:39 as a suitable summary statement on which to build, but when writing Mt. 4:23a he is also influenced by the wording of Mk 6:6b. This implies either that he was very familiar with Mark’s account (which Neirynck considers likely) or that he browsed well ahead in his version of Mark’s Gospel until he came to this verse. In view of contemporary compositional conventions, the latter is unlikely. Given that περιάγω occurs three times in Matthew’s Gospel (4:23; 9:35; 23:15) but only once in Mark’s, it is more likely that Matthew favoured this word independently of Mark’s usage; indeed, one might even consider whether its occurrence at Mk 6:6b was influenced by its presence at Mt. 4:23 and 9:35. Neirynck did not consider this alternative, however, because his purpose was to illustrate the two-source theory.

As for the remainder of Mt. 4:23 and the beginning of Mt. 4:24, Neirynck proposed that this part of Matthew’s summary was influenced by the following passages: Mk 1:14–15, which Matthew ‘conveniently summarizes’ by adding to κηρύσσων the phrase τὸ εὑαγγέλιον τῆς βασιλείας; Mk 1:23–27, which is accounted for by Matthew’s ‘substitute’ reference to Jesus healing all diseases and sicknesses among the people; and Mk 1:28. Neirynck’s source-critical analysis of the one-and-a-half verses in Mt. 4:23–24a envisages Matthew borrowing from, or at least being ‘influenced by’, passages from widely disparate sections of Mark’s Gospel. Such a procedure, though possible, does not readily fit recognized conventions of composition in the first century.

Concerning Mt. 4:24b–25, Neirynck began by pointing out how details from Mk 1:32–34 could have been incorporated into Mt. 4:24b. The verbal parallels between Mt. 4:24b and Mk 1:32, 34 are striking, perhaps even more striking than the verbal closeness between this same Markan pericope and Mt. 8:16, which is its contextual parallel. In Neirynck’s view, Matthew incorporated significant elements from Mk 1:32, 34 at both Mt. 4:24b (immediately before his Sermon) and Mt. 8:16 (four pericopes after the Sermon).

As for Mt. 4:25, Neirynck regarded this verse as completely dependent on Mk 3:7–8. Again, in brief compass Matthew must be seen to have conflated details from two disparate pericopes, which involved moving both backward (from Mk 1:39) and forward past six pericopes. Did first-century writers do this kind of thing, even when writing transitional passages? Or are we to imagine that Matthew was so familiar with Mark’s Gospel that these combinations of passages occurred to him without having to move backward and forward in his scroll of Mark?

Furthermore, if Matthew borrowed from Mk 3:7–12 when writing Mt. 4:25, he made further use of this same pericope when writing Mt. 12:15–16. This is similar to his alleged ‘double use’ of Mk 1:32–34 at Mt. 4:24b and 8:16. On Neirynck’s view, Matthew was not averse to ‘recycling’ previously used Markan sections.

With regard to Mt. 5:1–2, Neirynck wrote, ‘In Mt. 5:1, after the parallel to Mk 3: (10a) 7–8 in the preceding verse 4:25, Matthew continues with Mk 3:13, but not without adding a connecting formula …’ The ‘connecting formula’ is the phrase ἰδὼν δὲ τοὺς ὄχλους. Neirynck also noted that the reference to Jesus’ disciples in Mt. 5:1 makes it difficult to determine who comprised the audience for Jesus’ Sermon, the disciples alone or the disciples and the crowds, who are present at the end of the Sermon and react with amazement (Mt. 7:28). In his view, this ambiguity can be explained as follows: ‘This is the first mention of “his disciples” in Matthew and the only reference to them in Mt. 4:23–8:17. Since in Matthew “his disciples” are the twelve disciples (cf. 10:1), this precision added to the text of Mark may be inspired by the context of Mk 3:13 (v. 14 καὶ ἐποίησεν δώδεκα)’.

Given Neirynck’s conception of Matthew’s redactional procedure, it is certainly possible to imagine Matthew inadvertently introducing ambiguity into his narrative because of the influence of Mk 3:14. But what does this imply? Neirynck imagined that after writing Mt. 4:25 around various details from Mk 3:7–10, Matthew then skipped over the next two sentences in Mk 3:11–12 and used the opening words of Mk 3:13 in Mt. 5:1 to provide the immediate setting for his Sermon on the Mount. In doing so, however, he was ‘inspired’ by the succeeding verse—Mk 3:14, which speaks of Jesus ‘making’ twelve—to write that ‘his disciples’ came to Jesus, thereby introducing ambiguity into his otherwise carefully crafted narrative. In other words, Matthew turns out not to be the complete master of his source materials. The image of Matthew that emerges is not unlike that held by G.M. Styler, who, with reference to his argument for Matthew’s dependence on Mark, noted, ‘My own argument … makes Matthew capable of a measure of pedantry, and also of errors and illogicalities in some of his sequences’.

Even if one grants that Matthew’s Gospel is not free from ambiguity, or even ‘errors and illogicalities’, is Neirynck right that by οἱ μαθηταὶ αὑτοῦ Matthew always meant the twelve disciples? Matthew 5:1 may be the only reference to οἱ μαθηταὶ αὑτοῦ within Mt. 4:23–8:17, but Matthew refers to ‘his disciples’ at least six more times before Mt. 10:1, at which point they are referred to as ‘his twelve disciples’.34 It is by no means certain that prior to Mt. 10:1 Matthew envisaged each reference to ‘his disciples’ to imply the twelve disciples. Neirynck’s conjecture that the reference to ‘his disciples’ in Mt. 5:1 was influenced by the reference to the Twelve in Mk 3:14 is tenuous. This does not mean that the reference to the disciples in Mt. 5:1 denotes only the four fishermen called in Mt. 4:18–22. Given that Mt. 4:23–25 is an anticipatory or encapsulatory summary of Jesus’ Galilaean ministry, the reference to ‘his disciples’ in Mt. 5:1 may well denote a group representative of those whom Jesus commissioned to extend his own healing ministry (see Mt. 10:1–8). There is no need to infer the influence of Mk 3:13.

Building on his less than convincing discussion of Mt. 4:23–5:1, Neirynck made a further speculative and somewhat pedantic suggestion:

Since the phrase διδάσκων ἐν ταῖς συναγωγαῖς αὑτῶν in Mt. 4:23b (before 4:24a=Mk 1:28) has drawn our attention to Mk 1:21, it is to this first mention of ἐδίδασκεν in Mark that we turn now at the end of Matthew’s introduction to the Sermon. The text of Mk 1:22 is used by Matthew, in combination with Mt. 7:28a, as the conclusion of the Sermon on the Mount … The rather strict correspondence between Mt. 4:25–5:2 and 7:28–8:1 indicates that the parallel of Mt. 7:28–29/Mk 1:22 can be extended to Mt. 5:2/Mk 1:21b (ἐδίδασκεν).

Once again, Neirynck revealed his penchant for finding a Markan parallel for every possible word in Matthew’s composition. Why ἐδίδασκεν in Mt. 5:2, rather than διδάσκων in Mt. 4:23, should be the parallel to Mk 1:21 is difficult to surmise, apart from Neirynck’s determination to regard Mt. 4:23 as ‘influenced’ by Mk 6:6b. Indeed, why must either occurrence at Mt. 4:23 or 5:2 be a parallel to Mk 1:21? On the Markan hypothesis, is it not as likely that the entire Sermon on the Mount is Matthew’s ‘parallel’ to Mark’s reference to Jesus teaching in Mk 1:21 ? At one level, an ability to detect verbal parallels is commendable, but the image of Matthew that emerges from Neirynck’s discussion is of a mechanical, pedestrian redactor. His introduction to this study concluded with the statement: ‘The originality of Matthew’s composition can first of all be demonstrated in his use of Mark’. However, his analysis of Matthew’s use of Mark all but robs Matthew of any originality. As David L. Dungan remarked in his response to Neirynck’s discussion,

We find little … in the way of evidence from literary precedents or of the church’s historical setting or of the author’s theological motivations, to balance or justify the complicated manipulation of words and phrases Neirynck alleges the author of Matthew to have carried out. As a result, this textual discussion is perceived by us, at least, as an arid mechanical tour de force having little human feeling, historical validity or persuasive power.

In this connection, Downing’s response to Goulder’s conception of Luke’s literary achievement applies also to Neirynck’s view of Matthew:

Goulder is not alone … in supposing that most if not every word retained in common, and echoes and also divergences among linked authors are to be interpreted as intentionally significant. But this kind of supposition is badly out of tune with first-century literary composition, which at its most sophisticated had detailed concerns only with syntax and vocabulary, however open it may have been to un- or half-conscious resonances.

As part of his discussion, Neirynck also considered the composition of Mt. 4:23–8:17. On his source theory, Matthew must have rearranged the sequence of incidents in Mk 1:21–45 to a considerable extent, and on a number of occasions he must also have used the same material in different contexts, for example, Mk 1:32, 34 and 3:7–12. In Neirynck’s view, these redactional decisions on Matthew’s part are consistent with his compositional and structural plan:

Almost all commentators consider 8:1–17 a special section within Mt. 8–9. The study of Matthew’s sources does corroborate this position. From 8:18 on, the Markan material comes predominantly from Mk 4:35–5:20, 21–43, anticipated before and after Mt. 9:1–17=Mk 2:1–22 … The Markan material in Mt. 8:2–4, 14–16 is derived from Mk 1, in continuation of the parallels to Mk 1 before and after the Sermon. Source-critically, the material of 8:1–17 holds together with Mt. 5–7 and the doublets of 4:24b and 8:16 form an inclusio around this larger section.

In other words, given that Mt. 8:1–17 appears to be a distinct subsection of Mt. 8:1–9:34, and that the similarities in expression between Mt. 4:24b and Mt. 8:16 might be said to enclose the intervening material, Matthew’s alleged reliance on materials from Mk 1:21–45 within the larger section Mt. 4:23–8:17 ‘corroborates’ Matthew’s dependence on Mark. Were Neirynck’s claim strictly true, it would be a valid consideration in favour of Matthew’s reliance on Mark because Matthew could be seen to have made significant modifications—including transpositions, interpolations and duplications—within a relatively discrete section of Mark’s Gospel, namely, Mk 1:21–45. However, what of Matthew’s foray into Mk 3:7–12 when writing Mt. 4:25 and the alleged ‘influence’ of Mk 6:6b on Mt. 4:23? The apparent simplicity—hence credibility—of Matthew’s composition of Mt. 4:23–8:17 from Markan materials is contradicted by Neirynck’s more detailed source—critical analysis. As a result, even if the ‘Markan’ materials in Mt. 8:2–17 were incorporated from Mk 1:21–45, which is obviously the case on the Markan hypothesis, this is not evidence of a redactional policy on Matthew’s part because Mk 1:21–45 was not the sole source of Matthew’s Markan parallels ‘before and after the Sermon’.

Nevertheless, while Neirynck’s discussion of Mt. 4:23–8:17 oversimplified the data, it was suggestive at some points. Had he considered Mark’s compositional procedure on the theory that he was working with Matthew’s and Luke’s Gospels, it would have been more valuable.

Neirynck also suggested that the inversion in Mt. 8:2–17 of the relative sequence of Mk 1:29–34 and Mk 1:40–45 is another indication of Matthew’s dependence on Mark’s Gospel, as well as Q. He pointed out that following the healing of the leper in Mk 1:40–45, Mark has Jesus return to Capernaum, at which time he heals a paralytic (Mk 2:1–12). Matthew’s account is similar to Mark’s in that following the healing of the leper in Mt. 8:2–4, Jesus enters Capernaum and heals a centurion’s paralysed servant (8:5–13). Neirynck added that πάλιν in Mk 2:1 refers back to Mk 1:21, thereby making Jesus’ entry into Capernaum at Mt. 8:5 a parallel to both Mk 1:21 and 2:1. In Neirynck’s words,

Mk 2:1, Jesus’ second coming to Capernaum (πάλιν), refers back to 1:21, followed by 1:29 (Jesus in the house of Peter). In Matthew, this last episode is at 8:14 (καὶ ἐλθὼν ὁ Ἰησοῦς εἰς τὴν οἰκίαν Πέτρου) and the preceding entry into Capernaum (cf. Mk 1:21) is that of Mt. 8:5. Thus, the context before and after Mt. 8:5–13 seems to suggest that the Q motif of Jesus’ entry into Capernaum has been associated with both Mk 1:21 and 2:1:








8:5 (–13) Q





Why this should be so is not obvious. After all, Mt. 9:1 has Jesus returning to Capernaum (‘his own town’; cf. Mt. 4:13), followed immediately by the parallel to Mark’s healing of the paralytic in Mk 2:1–12. On Neirynck’s hypothesis, an equally feasible explanation of Matthew’s redaction is that, following his Sermon on the Mount, he transposed the healing of the leper from immediately after to immediately before Jesus’ ‘day at Capernaum’ (Mk 1:21–39). Matthew 8:5–13 might then be seen as Matthew’s substitute for Mk 1:23–27, for each involves giving orders that result in wonder—working and each highlights the issue of authority. Having made major modifications to his source at the first mention of Capernaum (i.e. Mk 1:21), it was perhaps natural to do so again immediately before Mark has Jesus return to Capernaum ‘some days later’ (2:1).

In conclusion, while Neirynck’s detailed source-critical argumentation is more satisfactory than many similar claims about the composition of Mt. 4:23–5:2, his conception of Matthew’s convoluted compositional procedure is anachronistic and therefore lacks plausibility. According to his view of the composition of this transitional summary, Matthew borrowed from the following passages of the Gospel of Mark in this order: Mk 1:39, 6:6, 1:28, 1:32–34, 3:7–8, 3:13, 1:21. Because he envisaged Matthew writing Mt. 4:23–5:2 when he came to Mk 1:21–22, all but the final verse are anticipations of words and phrases that occur in a number of later Markan contexts, two of which are far removed from the point in Mark’s narrative where Matthew allegedly chose to disrupt the Markan sequence so as to lead into his Sermon on the Mount and the subsequent cycle of miracle stories. Neirynck’s analysis would be more credible if Matthew’s alleged anticipations were not from so many and, in some cases, such far-flung Markan contexts.

Excursus on Matthew 8:1–9:34

Each of the remaining pericopes discussed in this chapter falls within Mt. 8:1–9:34, which follows the Sermon on the Mount and is dominated by a series of ten miracles. That Matthew intended this section to illustrate the wonder-working activity of Jesus is supported by the fact that there are as many miracles in Matthew 8–9 as in the rest of his Gospel. As the Sermon on the Mount in Matthew 5–7 illustrates Jesus’ authority in word, so the series of miracles in Matthew 8–9 illustrates his authority in deed. As Heinz Joachim Held noted,

The similarly-worded verses in Matt. 4:23 and Matt. 9:35 show by their contents (summary account of the activity of Jesus in word and deed) and their position (in the one case before the Sermon on the Mount and in the other immediately after the chapters containing the miracles), that Matthew’ s purpose in the chapters enclosed by these verses is to portray the double office of Christ: his teaching and his healing activity.

Similarly, Terence L. Donaldson maintained that ‘by means of the inclusio of 4:23 and 9:35 the words of Jesus in chs. 5–7 stand alongside the deeds of Jesus in chs. 8–9 as a two-panel depiction of Jesus’ ministry’.

Roughly half of Matthew’s disagreements with Mark’s order occur in Matthew 8–9. At least since Heinrich Julius Holtzmann’s Die Synoptiker, critics have appealed to Matthew’s desire to collect together miracles illustrative of Jesus’ wonder-working power to explain his disagreements with Mark’s order in this section of his Gospel. Holtzmann was satisfied that this editorial concern was in itself sufficient to explain the disagreements with Mark’s order in Matthew 8–9. At one level, the desire to recount a representative—and relatively extensive—number of miracles so as to match the impressive picture of Jesus the teacher in Matthew 5–7 would account for order changes in Matthew’s source material if that source (or sources) did not have sufficient miracles recorded one after the other. In general terms, therefore, the editorial concern to form a miracle collection explains why Matthew disagrees with Mark’s order in this block of material, even if it does not account for each specific disagreement. As Bultmann wrote,

Admittedly Matthew undertook some transpositions in Mark’s outline, of which the most important is that in 8:1–9:34 he gives us a collection of miracle stories and in doing so brings together from Chaps. 1, 2, 4 and 5 of Mark stories that are separated there and in a different order.

Nevertheless, were Matthew primarily dependent on Mark for his individual miracle stories, he could have constructed his miracle section without rearranging Mark’s order so radically. As B.W. Bacon pointed out in his Studies in Matthew, ‘Had he [Matthew] given any consideration to the order of Mk he could easily have made his extracts from Mk 4:35–5:20 in 8:23–34, from Mk 2:1–22 in 9:1–17 and from Mk 5:21–43, conform much more nearly to the sequence of the Gospel from which he took them’. As a result, most critics consider it necessary, or at least desirable, to explain individual disagreements in order between Matthew and Mark in these respective sections of their Gospels.

Matthew 8:1–9:34 has been the focus of numerous studies. At the end of the nineteenth century, W.C. Allen published ‘The Dependence of St Matthew 1–13 upon St Mark’, in which he aimed to show that ‘the difference in the order of events between the two Gospels can only be accounted for on the hypothesis that Matthew has rearranged Mark’s order to suit the scheme upon which he has built up the first half of his Gospel …’ In response to Allen’s article, J.C. Hawkins published a two-part study on ‘The Arrangement of Materials in St Matthew 8–9’,52 in which he supplemented and in some respects sought to correct Allen’s arguments for Matthew’s dependence on Mark. These were influential studies that strengthened the consensus concerning Matthew’s dependence on the Gospel of Mark. Certainly Bacon’s study of 1920, ‘Editorial Arrangement in Matthew 8–9’, assumed Matthew’s dependence on Mark and was more concerned with the principles governing the composition of this section of Matthew’s Gospel. He considered that Mt. 8:1–9:34 comprised three subsections with different themes: the first three miracles function as ‘a testimony to Israel’; the second group of three miracles encourages wholehearted faith on the part of those considering the cost of Christian discipleship; and the third subgroup of miracles demonstrates that Jesus’ miracles elicited both faith and unbelief.

Following Bacon, more recent studies have focused on the structure, function and themes governing the composition of Mt. 8:1–9:34. Particularly influential has been Heinz Joachim Held’s form-critical and redactional analysis, ‘Matthäus als Interpret der Wundergeschichten’, which was published in 1960 alongside studies by Gunther Bornkamm and Gerhard Barth. Like Bacon, Held argued that Mt. 8:1–9:34 is best subdivided into three subsections—8:2–17, 8:18–9:17, 9:18–31, each controlled by a different theme, namely, Christology, discipleship and faith, respectively. As for Mt. 9:32–34, Held maintained that this brief pericope rounded off the entire section by illustrating varying responses to Jesus’ miracles.

K. Gatzweiler’s study of Matthew’s miracle stories took Held’s landmark work as its point of departure. Indeed, his study was largely a summary and review of Held’s work. As he remarked, ‘On ne saurait aujourd’hui aborder le problème des récits de miracles chez Mt. sans passer par Held’.58 As part of his study, Gatzweiler attended to the miracles in Mt. 8:1–9:34. Like Held, he perceived within this section three groups of three miracles each (8:1–17; 8:23–9:8; 9:18–34) interrupted by two series of sayings (8:18–22; 9:9–17):

Le premier groupe de miracles ne comporte que des guérisons qui sont interprétées par le verset final comme réalisation des prophiéties sur le serviteur de Jahweh (cfr v. 17). Dans le dernier groupe de miracles, le thème de la foi paraît particulièrement mis en évidence, comme il ressort des logia de Jésus aux vv. 22 et 29. Dans le deuxième groupe de miracles, les disciples occupent une place prépondérate. Jésus ne se révèle pas seulement comme thaumaturge, mais aussi comme Seigneur de la communauté.

Gatzweiler went on to show that in certain respects the larger ‘ensemble’ of Mt. 4:23–9:35 prepares for the mission discourse, in which Jesus gives his disciples authority to do what he has been doing (cf. Mt. 10:1, 7–8a). This is an aspect of Matthew’s christological interpretation of the miracle stories, through which Jesus is depicted as one who fulfils Old Testament prophecies, as the Servant of God who acts in power, as the Lord and helper of his community, and as one who calls his disciples to participate in his power and mission:

Au début de son évangile, Mt. nous présente donc en Jésus le messie de la parole et le messie de l’action. Il nous présente Jésus comme celui qui charge ses disciples de poursuivre sa mission. Jésus n’est pas seulement le Christ autorisé par Dieu en paroles et en actions, il est aussi le Seigneur qui accorde à ses disciples une participation à son pouvoir et à son autorité. En résumé, nous pouvons dire que Mt. interprète les miracles au service de sa christologie. Jésus se manifeste par eux comme le réalisateur des prophéties vétérotestamentaires, comme le serviteur de Dieu qui agit en puissance, comme le Seigneur et l’assistant de la communauté, comme celui qui appelle ses disciples à participer à son pouvoir et à sa mission.

Like Gatzweiler, B.F. Drewes opened his 1971 study, ‘The Composition of Matthew 8–9’, by noting the importance of Held’s work. Acknowledging that Held had provided ‘clear starting-points’, Drewes thought ‘some aspects of the structure [of Mt. 8:1–9:34] would merit a deeper study’. He first pointed out that in three key places Matthew focuses attention on the demonized, namely, in the summary concluding the first triad of miracles (8:16–17), in the central story of the second group of miracles (8:28–34), which Drewes regarded as ‘the core of the whole composition’, and in the conclusion to the third group of miracles (9:32–34), which is also the conclusion to the whole section.

Drewes then argued that there is ‘movement’ within Matthew 8–9 from relative privacy in Mt. 8:14–17 (in Peter’s house) to public proclamation in the final group of miracles (Mt. 9:26, 31). In his view, this explains why Matthew did not use Mark’s ending to the story of the cleansing of the leper (Mk 1:45) in his parallel account, but reserved it for the final miracle in this larger block, Mt. 9:27–31. He also maintained that Matthew enclosed this block of miracle stories within two miracle stories that contain a command by Jesus not to speak (Mt. 8:4a; 9:30b). Finally, he urged that the pericope about the Gadarene demoniacs is not only the centre of Mt. 8:18–9:17 (the middle triad of miracle stories preceded by two conversations regarding discipleship, Mt. 8:18–22, and followed by two disputes, Mt. 9:13–17), but also the centre of the entire block of miracle stories in Matthew 8–9. For Drewes, ‘The function of 8:28–34 is thus first of all to show within the whole composition the central place of the casting out of demons in the healing activity of Jesus …’

Drewes also pointed out that Mt. 4:23–9:35 is enclosed within the calling of the disciples in Mt. 4:18–22 and the sending out of the disciples in Mt. 10:5–8. Together with the concentric structure and content of Matthew 8–9, this compositional detail reveals ‘the close link between the Messiah and his community’.

Also building on Held’s work, William G. Thompson’s ‘Reflections on the Composition of Mt. 8:1–9:34’ focused on the arrangement of this section of Matthew’s Gospel rather than on Matthew’s modification of source material. While cognizant of the value of source criticism for redaction criticism, and appreciative of previous redaction-critical work on Matthew’s Gospel, Thompson concentrated on Matthew’s text in its own right, seeking to discern the ‘logic’ of his composition from within, so to speak, rather than by comparison with synoptic parallels. Contrasting his approach to that of Bornkamm and Held, he remarked,

I suggest that broader questions about the over-all composition of Mt. 8:1–9:34 demand another approach which can be called ‘vertical’. For it is well known that Matthew differs from Mark in the selection and arrangement of his material, and the somewhat acrobatic attempts to explain his editorial activity from the viewpoint of Mark leave several issues open for further discussion. What is needed is to study Matthew in terms of Matthew, as we tend to study John in terms of John. Consequently, I wish to by-pass synoptic comparison and exegetical detail—not neglecting them but building on them—and concentrate on Matthew’s composition.

Two years after Thompson’s study, Christoph Burger published a study entitled ‘Jesu Taten nach Matthäus 8 und 9’. Whereas Held had argued that the cycle of miracle stories in Matthew 8–9 comprised three main subsections concerned with Christology (8:2–17), discipleship (8:18–9:17), and faith (9:18–31), Burger argued that Matthew 8–9 comprised four subsections concerned with Christology (8:1–17), discipleship (8:18–34), issues regarding the relation between Jesus (and his followers) and Israel (9:1–17), and faith (9:18–34). Beyond this refinement of Held’s analysis, Burger also urged that the overarching theme of this section of Matthew’s Gospel is the church of Jesus Christ: ‘Das Gesamtthema dieser Komposition ist die Kirche Jesu Christi. Matthäus schildert das Auftreten Jesu in einer Weise, daß darin Wesen und Aufgabe der Kirche im voraus abgebildet sind’. Indeed, for Burger, Matthew 8–9 provides what he called the ‘foundation legend’ (die Gründungslegende) of the church, by which he meant that Matthew ‘read back’ his understanding of the church into his depiction of Jesus’ life. Once one appreciates Matthew’s ‘ecclesiological arrangement’, one perceives that Matthew intended his portrayal of the deeds of Jesus in Matthew 8–9 to serve as a counterpart to his Sermon on the Mount. As Burger put it,

Matthäus hat sein Verständnis der Kirche in die Darstellung des Lebens Jesu zurückgetragen und hier entfaltet. Etwas überspitzt läßt sich behaupten: Die Kapitel 8 und 9 seines Evangeliums bieten den ἱερὸς λόγος, die Gründungslegende der christlichen Kirche. Erkennt man diese ekklesiologische Ausrichtung der Komposition, stellt sie in der Tat ein Gegenstück zur Bergpredigt dar.

In 1974, Gerd Theissen published his Habilitationsschrift on The Miracle Stories of the Early Christian Tradition, in which he briefly considered various explanations for ‘the particularly striking collection of miracles in between the Sermon on the Mount and the sending out of the Twelve (Chs 8–9)’. Theissen referred first to Matthew’s alleged desire to balance his depiction of Jesus as Messiah of the Word in Matthew 5–7 with a presentation of Jesus as Messiah of action in Matthew 8–9. He noted, however, that this section includes not only miracles but also the ‘apophthegms’ in Mt. 8:18–22, 9:9–13 and 9:14–17, which in his view were either inserted or retained for theological reasons. Another explanation for the collection of miracles in Matthew 8–9 is that they were assembled together at this early stage in Jesus’ public ministry to prepare for Jesus’ reply to John the Baptist’s question in Mt. 11:2–6. This explanation, for which there were hints from as early as Karl Lachmann’s 1835 study on the order of the stories in the synoptic gospels, was not fully satisfactory for Theissen because it does not account for all of Matthew’s alleged rearrangements, for example, the stilling of the storm and the exorcism of the Gadarene demoniacs in Mt. 8:23–34.

Finally, Theissen offered an unlikely explanation of his own:

The reflexive quotation in [Mt.] 4:15f. mentions the sea, Transjordan and Galilee of the Gentiles. In Chs 8–9 we find precisely a journey across the ‘sea’, activity beyond the Jordan (8:28–34) and a Gentile (8:5ff.). The anticipation of the journey on the lake might be explained by Matthew’s desire to make the prophecy of Is 8:23–9:1 come literally true.

This ‘geographical consideration’ was apparently meant to explain what the previous explanation failed to explain, namely, the storm-stilling and the exorcism in Mt. 8:23–34. However, it ignores the explicit statement in Mt. 4:13–14 that the prophecy of Isa. 8:23–9:1 was fulfilled when Jesus settled in Capernaum-by-the-lake.

In 1978, Jack Dean Kingsbury published his ‘Observations on the “Miracle Chapters” of Matthew 8–9’, in which he explored the arrangement, Christology and context of Mt. 8:1–9:34, together with what he described as the ‘paradigmatic function’ of the pericopes within Matthew 8–9. Commenting on previous research, he noted that three major attempts had recently been made to analyse this section of Matthew’s Gospel: the ‘classic study’ by Held, who built upon the observations of Julius Schniewind; the 1971 study by Thompson, who built upon Held; and the 1973 study by Burger, who regarded his work as a corrective to Held’s position.

While Kingsbury agreed with Burger that there are four main subsections in Mt. 8:1–9:34, he disputed Burger’s thesis that the overarching theme of this section is ecclesiology rather than Christology. In line with arguments elaborated in his Matthew: Structure, Christology, Kingdom, Kingsbury argued that Matthew 8–9 is dominated by a ‘Son of God’ Christology. He also maintained that the respective pericopes in Matthew 8–9 provide ‘extended commentary’ on the summary in Mt. 4:24–25 The portrait of Jesus that emerged from his analysis of Matthew 8–9 is as follows:

in the subsection 8:1–17 Jesus reveals himself, through his healing activity in and around Capernaum, to be the one who, in fulfillment of Old Testament prophecy, ‘takes’ his people’s sicknesses and ‘bears’ their diseases. In the subsection 8:18–34 Jesus makes known, through his interaction with his disciples and others, the cost and commitment of discipleship. In the subsection 9:1–17 Jesus, in Capernaum, engages in debate with the leaders of the Jews and with the disciples of John, showing thereby that he and his followers constitute a group that stands apart from contemporary Israel. Finally, in the subsection 9:18–34 Jesus performs additional healings in Capernaum, and, through his encounter with suppliants, the essence of faith is disclosed.

Turning to the ‘paradigmatic function’ of the pericopes that make up Mt. 8:1–9:34, Kingsbury maintained that these miracle stories served as reminders of the cost of discipleship and as ‘paraenetic paradigms’ or stories that encouraged community members to continue to rely on the exalted Son of God in faith during times of hardship. Then, in a concluding paragraph, he wrote:

An analysis of chaps. 8–9 reveals that Matthew has appropriated the contents of Mark and Q but rearranged them to form four subsections that treat, in turn, such matters as christology (8:1–17), discipleship (8:18–34), questions pertaining to the separation of Jesus and his followers from Israel (9:1–17), and faith (9:18–34).

However, no source-critical conclusions emerge from Kingsbury’s analysis of Matthew 8–9. A defensible recasting of his concluding statement would read: ‘An analysis of chaps. 8–9 reveals that if Matthew appropriated the contents from Mark and Q, he rearranged them to form four subsections …’ Kingsbury’s only source-critical remarks appear in an early paragraph as part of his consideration of the arrangement of materials in Matthew 8–9, beginning with this statement, ‘If one presupposes the two-source hypothesis, all of the pericopes except three stem from Mark’.

In 1979, the year after Kingsbury’s study, John Paul Heil examined ‘Significant Aspects of the Healing Miracles in Matthew’. Like Held and Gatzweiler, Heil did not restrict himself to the miracle stories in Matthew 8–9. Against Held and others, however, he argued that in redacting the miracle stories—which he accepted as having been taken from Mark and Q—Matthew did not de-emphasize the miraculous element to the point that they almost cease to be miracle stories. Heil also contested the value of grouping Matthew’s miracle stories around particular themes such as Christology, discipleship and faith, arguing that such an approach fails to appreciate the ‘multi-dimensional character and theological richness of a miracle story’. He did not maintain that themes such as Christology and faith are absent from such miracle stories as the healing of the leper in Mt. 8:1–4, but rather that to emphasize one theme at the expense of others is reductionistic: ‘The miracle has many dimensions. Each dimension plays its part in the drama of the narrative. To emphasize one theme as the “goal” or “highpoint” at the expense of the others detracts from the theological richness of the event’.

Heil not only criticized one-sided interpretations that seek to press miracle stories into the service of one theme; he also maintained that one theme or dimension often neglected by critics when considering the miracle stories is what he termed ‘salvation experience’. According to Heil,

In no other type of synoptic gospel material do people receive the gifts of salvation in a way that visibly improves their personal lives. In this sense the uniqueness of the healing of the leper as well as all healings and exorcisms belongs to the theme of soteriology.

Heil then proceeded to consider both (a) the relation between Matthew’s soteriological healing/exorcism stories and his account of Jesus’ passion, death and resurrection and (b) the significance of the miracle stories for Matthew’s community, including those who performed healings and exorcisms and those who experienced the benefit of healing. His concluding remarks aptly encapsulate his conception of the miracle stories in Matthew’s Gospel:

Matthew has fully integrated the healing miracles of Jesus into the plan of his gospel. In so doing he has underlined their abiding significance: As faith experiences they are christological in that they reveal Jesus as the Christ who brings about the messianic end-time in accord with scriptural fulfillment. They are soteriological in that people personally experience Jesus as savior in their lives. They are ecclesiological in extending their meaning beyond the time of the earthly Jesus to the church of Matthew.

Heil’s study is an important corrective to Held’s redaction-critical evaluation of Matthew’s healing miracles and exorcisms. However, while his critique of the tendency to interpret Matthew’s miracle stories in terms of single themes is both valid and enlightening, he did not consider whether prominent themes in pericope blocks might not disclose the logic of Matthew’s compositional arrangement. He was right to protest against interpretations of individual miracle stories that highlight one theme at the expense of others. But if one theme recurs in an obvious way throughout a series of contiguous miracle stories, that theme may well explain the arrangement of material in that larger section.

In the 1987 Festschrift for E. Earle Ellis, Ulrich Luz re-examined ‘Die Wundergeschichten von Mt. 8–9’. He began by making some general source-critical and compositional observations:

Die Sammlung matthäischer Wundergeschichten in Kap. 8–9 scheint ein untypisches Beispiel matthäischer Interpretation der Tradition. Matthäus, der sonst so konservative Evangelist, hat sich hier aussergewöhnliche Freiheiten im Umgang mit seinen Quellen erlaubt. Statt, wie sonst üblich, der Abfolge der Markuserzählung zu folgen, hat er zwei verschiedene Abschnitte seiner Markusquelle ineinander verwoben (Mk 1:29–2:22; 4:35–5:43). Deren relative Reihenfolge ist nur im allgemeinen unangetastet geblieben; mindestens zwei auffällige Umstellungen sind festzustellen.

After reviewing research on Matthew’s miracle stories following Held’s monograph, Luz identified the basis for his own interpretation of Matthew 8–9: ‘Es geht Matthäus keineswegs um eine blosse Sammlung von Wundergeschichten, die beispielhaft die Taten des Messias oder gar verschiedene Aspekte seiner Lehre und des christlichen Glaubens erläutern, sondern es geht ihm um eine zusammenhängende Geschichte‘,

For Luz, Mt. 8:1–9:34 forms the beginning of the second major part of Matthew’s Gospel. As he put it, if one understands a narrative account as a sequence of three main elements—orientation, complication and resolution—Matthew 1–7 provides an orientation, and the miracle stories in Matthew 8–9 comprise the beginning of the complication. In addition, like Drewes, he pointed out that Matthew’s conclusion to his story of the healing of two blind men (Mt. 9:30–31) has close affinities with Mark’s conclusion to the story of the healing of the leper (Mk 1:44–45). In Luz’s view, ‘Man könnte also sagen: Matthäus braucht die markinische Geschichte 1:40–45 als Rahmen für seinen ganzen Wunderzyklus Mt. 8:1–9:34’.

Luz’s understanding of Matthew 8–9 resembled that of Burger:

Matthäus erzählt in Kap. 8–9 die Entstehung der Jüngergemeinde aus Israel und die damit verbundene Spaltung in Israel. Die These Burgers, dass es in Mt. 8–9 um ‘die Gründungslegende der christlichen Kirche’ gehe, ist also m.E. nicht ‘etwas überspitzt’, sondern richtig. Nur muss man deutlicher als Burger sagen: Es geht nicht—statisch—um eine Zurückprojektion der Wirklichkeit der Kirche ins Leben Jesu und nicht um eine Darstellung der Kirche durch das sprachliche Medium eines Berichts über Jesus, sondern es geht um eine Erzählung, die Erzählung nämlich, wie durch die Wunder des Messias in Israel Gemeinde entstand.

To conclude this overview of research on Mt. 8:1–9:34, volume II of the new ICC commentary on the Gospel of Matthew by W.D. Davies and Dale C. Allison opens with an excursus on the structure, Christology and function of this section of Matthew’s Gospel.98 Regarding the structure of Mt. 8:1–9:34, they reconsider many of the proposals surveyed above, but most significant are their criticisms of Held, Thompson, Burger and Kingsbury. While agreeing with Held and Thompson that a threefold subdivision of Mt. 8:1–9:34 ‘probably points us in the right direction’, they regard as unsustainable any attempt to discern a thematic unity in any of the proposed subsections. By contrast, their own structural analysis is based on a purely formal consideration, namely, Matthew’s documented penchant for triadic arrangement, which is represented in Matthew 8–9 by three clearly demarcated groups of three miracle stories. According to Davies and Allison,

The point to stress is that the key to unlocking the structure cannot be found in topical interests (Christology, discipleship, faith). Not that there are no thematic threads. 8:14f. (the formula quotation) follows nicely after three healing stories, and 8:18–22 (on discipleship) is neatly placed into a context where Jesus’ disciples literally follow him around. Further, in 9:9–17, the very first verse (the call of Levi [sic], 9:9) appropriately picks up once again the theme of discipleship, and 9:10–13 (on toll collectors and sinners) gains meaning from the low social status of most of those healed in Mt. 8–9. But the arrangement of the entire section is dictated by a formal consideration, the triad.

As for the Christology of this section of Matthew’s Gospel, Davies and Allison resist the tendency to make one title dominant, as Kingsbury had. In their view, ‘the concatenation of various christological appellations in our two chapters should not be muted …’

Regarding the function of Matthew 8–9 within the narrative structure of Matthew’s Gospel, Davies and Allison make two major points. First, together with the Sermon on the Mount, which it complements, the cycle of miracle stories prepares for the missionary discourse in Matthew 10. ‘The many parallels between what Jesus has already said and done and what the disciples will say and do demand that one function of the miracle chapters is to set up an example: like master, like disciple (cf. 10:24f.)’. Second, in the first half of Matthew’s Gospel there is a particular focus on Israel. Throughout Matthew 5–10 Jesus challenges God’s people, Israel; on the other hand, Matthew 11–12 is dominated by Israel’s negative response to Jesus. In short, ‘Mt. 5–7 (the words of Jesus), 8–9 (the deeds of Jesus), and 10 (the words and deeds of the disciples) largely record an overture, chapters 11 and 12 a response’.

Matthew 8:1–4, Jesus Heals a Leper

Before analysing Matthew’s account of Jesus healing a leper, it is worth recalling the ambiguity of determining which of Matthew’s pericopes are ‘out of order’ relative to Mark’s and Luke’s sequence of events. When constructing a synopsis or table of synoptic parallels, deciding to place Matthew’s Sermon on the Mount after Mk 1:20, as Farmer et al. suggest, or between Mk 1:21 and 1:22, as Neirynck suggests, does not determine which of the three or four pericopes after the Sermon follow the same relative order as their parallels in Mark and Luke. I have placed Mt. 8:14–17 in parallel order with Mk 1:29–34 and Lk. 4:38–41, thereby making Mt. 8:1–4 disagree in order against Mark and Luke. An equally sustainable decision, however, would be to place Mt. 8:1–4 opposite Mk 1:40–45, thereby making Mt. 8:14–17 disagree in order against Mark and Luke. Tuckett has said that Neirynck’s scheme of parallelization ‘implies that there is no alteration of the relative order of Mark 1:29–34; rather, the pericope whose order is changed is that of the healing of the leper (Matt. 8:1–4 // Mark 1:40–45)’. But Neirynck’s schema implies no such thing; either Mt. 8:1–4 or Mt. 8:14–17 must be displayed as a disagreement in order relative to Mark’s and Luke’s orders, but there is no objective criterion for determining which should be so displayed. Tuckett preferred to regard Mt. 8:1–4, rather than Mt. 8:14–17, as a transposition precisely because it enabled him to explain why Matthew rearranged this pericope by bringing it forward.

Matthew 8:1 is unique and is commonly regarded as an editorial transition between Mt. 7:28–29 and Mt. 8:2–4. Equally plausible, however, is that Mt. 8:1 belongs with 7:28–29, as Hawkins maintained:

these three verses combine to tell us the whole immediate result of the Sermon on the Mount, namely, that οἱ ὄχλοι were astonished at the teaching of Jesus, and that, consequently, when He was come down from the mountain those multitudes or a large portion of them—ὄχλοι πολλοί—followed Him. So ends that incident, and a completely fresh one, quite disconnected with what had gone before, begins in 8:2.

While perhaps not crucial, whether one regards Mt. 8:1 as Matthew’s introduction to the story in Mt. 8:2–4 or as part of his conclusion to the Sermon on the Mount will influence one’s response to an argument occasionally put forward as evidence of Markan priority, namely, that Jesus’ command to silence in Mt. 8:4 conflicts with the setting provided for the story in Mt. 8:1, in which Matthew notes the presence of crowds, and is only present because he found it in his Markan source.

The many crowds (ὄχλοι πολλοί) of Mt. 8:1 might seem to be those mentioned in Mt. 4:25, 5:1 and 7:28. However, that καταβάντος in Mt. 8:1 is aorist implies that crowds followed Jesus after he had descended from the mountain. In addition, the most natural reading of Mt. 5:1 is that Jesus ascended the mountain to escape the crowds, so that his sermon is addressed primarily to his disciples. As Warren Carter pointed out in his discussion of ‘The Crowds in Matthew’s Gospel’:

The focus on the crowds at the beginning of 5:1 a gives way in 5:1b with the use of a third person singular verb (ἀνέβη) to describe Jesus going up the mountain. The singular verb suggests that he goes alone without the crowds. The focus again shifts in 5:1c: once he is seated, the disciples, identified closely with Jesus as his disciples (οἱ μαθηταὶ αὑτοῦ), come to Jesus (5:1d) and he teaches ‘them’ (5:2). The effect of the narrative at this point is to differentiate the crowds from the disciples, to ally the disciples with Jesus, and to distance the crowds from Jesus.

This reading is supported not only by Mt. 8:18, where the sight of crowds provokes Jesus’ command to cross to the other side of Lake Galilee accompanied only by his disciples (Mt. 8:23), but also by details in the parables discourse in Mt. 13:1–53. At Mt. 13:1–3 Jesus explicitly addresses many crowds (ὄχλοι πολλοί), thereby demonstrating that Matthew could make it clear when Jesus addressed the crowds. There is also the statement in Mt. 13:34, remarkable by contrast with Mt. 7:28–29, that without a parable Jesus told the crowds nothing. On this reading, the reference in Mt. 7:28–29 to the amazement of the crowds who heard Jesus’ authoritative teaching is anomalous, which might be taken to suggest Matthew’s incorporation of material that was not sufficiently edited to fit unobtrusively into his narrative. However, while Matthew shares with both Mark and Luke the phrase ἐξεπλήσσοντο ἐπὶ τῇ διδαχῇ αὑτοῦ, in using it only Matthew refers to οἱ ὄχλοι. Had all three synoptists referred to οἱ ὄχλοι one might have appealed to an adapted version of one of E.D. Burton’s six criteria for determining ‘features of a secondary character’ within documents that stand in literary relation to one other. Burton’s second criterion was the ‘insertion by one writer of material not in the other, and clearly interrupting the course of thought or symmetry of plan in the other’. Had Mark used οἱ ὄχλοι at Mk 1:22, one could argue that in taking over Mark’s wording, Matthew incorporated material from Mark that muddied his own course of thought. As it stands, however, Matthew seems to have incorporated a phrase unique to himself that appears to conflict with his own course of thought!

Matthew 8:2–4 recounts the ‘cleansing’ of a leper in Matthew’s stark style. (Excluding the 10 words in Mt. 8:1, Matthew uses 52 words, Mark uses 99 words and Luke uses 98 words.) When recording direct speech, and in his description of Jesus reaching out to touch the leper (Mt. 8:3a), Matthew’s wording is close to that of both Mark and Luke. Only Mk 1:41a notes Jesus’ deep emotion (σπλαγχνισθεὶς; v.l. ὀργισθεὶς). Matthew’s description of the cure is closer to Mark’s by using the same word ἐκαθαρίσθη, but in Matthew the leprosy rather than the leper is ‘cleansed’ (cf. Mk 1:42b). Mark and Luke each report that the leprosy left the man (ἀπῆλθεν ἀπʼ αὑτοῦ), which suffices for Luke, but Mark also notes that the man was ‘cleansed’, an example of his penchant for ‘duplicate expressions’, one of which he shares with Matthew, the other with Luke. Such details are emphasized by advocates of the two-gospel hypothesis. For example, commenting on Mk 1:40–45 in their ‘Narrative Outline of the Markan Composition According to the Two Gospel Hypothesis’, Farmer et al. assert:

This story contains classic evidence of Mk’s procedure in conflating his sources Mt. and Lk. Words and phrases are woven together into a number of ‘duplicate expressions’. In the process Mk adds or omits numerous tiny details thus creating the familiar pattern of ‘minor agreements’ (both positive and negative) against Mk; note especially the famous ‘moved with pity’ in vs. 41a. Basically, he follows Lk’s form and works in Mt’s details.

Conversely, Davies and Allison represent the majority: ‘For our part we find it easier to tackle the minor agreements than what would have to be considered from the Griesbach perspective Markan additions’.

While the minor agreements relate more directly to the question of the relation between Matthew and Luke than to the relation between Matthew and Mark, they nevertheless form part of the larger picture of the relations between all three gospels. Responding to Streeter’s treatment of those minor agreements that he tried to explain by recourse to textual corruption, Farmer’s analysis of Mk 1:40–42 and parallels not only calls attention to the positive and negative agreements between Matthew and Luke against Mark, but also underscores the extensive verbatim agreement between Mt. 8:2b–3 and Lk. 5:12c–13. He asked:

Is the reader to think that Matthew and Luke in complete independence omitted αὑτῷ ὅτι, added κύριε, omitted σπλαγχνισθεὶς, transposed the word order of αὑτοῦ ἥψατο, substituted a participle for Mark’s historic present καὶ λέγει, omitted αὑτῷ, and then substituted ωὑθέως for ωὑθὺς, all in the compass of three verses? It would be far simpler to account for this series of agreements and the eighteen word consecutive verbatim agreement on the hypothesis that Luke copied Matthew.

Few critics have responded positively to Farmer’s argument, but one who accepts a literary connection between Matthew and Luke (while still holding to Markan priority) is Robert H. Gundry. Commenting on Mt. 8:2, Gundry wrote, ‘The emphatic ἰδού (not in Mark) typifies Matthew’s style … Yet Luke has it, too. In conjunction with other agreements between Matthew and Luke against Mark-agreements that sometimes are typical of Matthew but not Luke—this one suggests that Luke knew Matthew as well as Mark’. In this connection, it is also worth recalling that Bultmann regarded Mk 1:43 as a Markan addition, although it is not entirely clear whether he thought this was evidence of a Deutero-Mark or a Markan addition to the original form of the story that was in turn omitted by both Matthew and Luke. E.P. Sanders and Margaret Davies also regard Mk 1:45 and Lk. 5:15–16 to be later additions. When compared with Mt. 8:2–4, Mk 1:40–45 by no means provides unambiguous evidence of Markan priority.

A distinctive note in Mt. 8:2 is that the leper prostrated himself before Jesus (προσεκύνει αὑτῷ) before requesting cleansing. A similar construction occurs at Mt. 9:18 when a ruler prostrates himself in front of Jesus before requesting Jesus’ resuscitating touch for his recently deceased daughter. Tevis identified προσκυνέω, often preceded by a participle and always followed by a pronoun in the dative, as a linguistic construction likely to come from a single writer, possibly from a redactor. The verb appealed to Matthew; it occurs 13 times in Matthew, but only twice in Mark, 3 times in Luke and 4 times in Acts.125

Linked to Matthew’s use of προσκυνέω here is προσέρχομαι. Of the 86 occurrences of this word in the New Testament, 51 are in Matthew’s Gospel. (Mark uses the term 5 times, Luke—Acts 20 times, 10 times in each book, and John once.) Kingsbury has noted the ‘cultic connotation’ προσέρχομαι has for Matthew, especially when coupled, as it is in Mt. 8:2–4, with προσκυνέω, which already has ‘cultic colouring’.

Like Mark and Luke, Mt. 8:4a records Jesus commanding the ‘cleansed’ leper not to tell anyone, presumably about his healing by Jesus’ touch and command. However, some regard this detail to be dissonant because Mt. 8:1 notes the presence of many crowds. For example, according to Samuel T. Lachs,

The author-editor was apparently unaware of the problem which has resulted by placing this first miracle story here. The connecting literary link between the Sermon and the miracle stories is v. 1, which related that ‘great crowds followed him’. In the leper story, however, Jesus tells the leper not to tell it to anyone, an act witnessed by the great crowds.

On the basis of this observation, one might infer that Matthew’s version of this story is secondary because he incorporated source material—namely, the command to secrecy—that did not fit neatly into his narrative framework. However, as noted above, judgment on this depends on whether one reads Mt. 8:1 as Matthew’s introduction to this pericope or as part of his conclusion to the Sermon on the Mount.

On the Markan hypothesis, Mt. 8:1–4 is Matthew’s second transposition of Markan material. The first, Mt. 4:23–5:2, was constructed around Mk 1:39, Mark’s summary statement concluding Jesus’ ‘day at Capernaum’ (1:21–39). Matthew 8:1–4 is Matthew’s version of the incident that immediately follows Mk 1:39, the healing of a leper in Mk 1:40–45. If Matthew did use Mark’s Gospel, one can say he used Mk 1:39 and 1:40–45 as the ‘frame’ for his Sermon on the Mount. Neirynck may be correct that Mk 1:21–22 is the Markan context for Matthew’s Sermon, but Mk 1:39 and 1:40–45 provided the literary frame for it. In other words, having decided to incorporate his Sermon at the earliest possible point in the Markan outline, Matthew did so at the first recorded instance of Jesus teaching in Mark’s Gospel—Mk 1:21. Mark 1:22 provided a suitable summary of people’s response to Jesus’ teaching, so Matthew incorporated it at the conclusion of his Sermon (Mt. 7:28–29), albeit with the puzzling addition of ‘crowds’. Having read ahead in Mark’s account, however, Matthew chose to build his introduction to the Sermon around Mk 1:39. After completing his Sermon and recording the crowd’s response, based on Mk 1:22, he continued with the episode that immediately follows Mk 1:39, namely, Mk 1:40–45. From the vantage point of the Markan hypothesis, Mt. 8:1–4 may be regarded as part and parcel of Matthew’s initial decision to interrupt the Markan sequence following Mk 1:20 to incorporate his Sermon on the Mount. Seen from this perspective, Mt. 8:1–4 is not really an additional transposition on Matthew’s part.

In this connection, it is worth recalling the view of Drewes and Luz that Matthew used Mk 1:40–45 as the literary frame for the entire block of miracle stories in Matthew 8–9 by omitting the substance of Mk 1:45 at Mt. 8:1–4 and incorporating it at Mt. 9:31. If this view is correct, Mk 1:39–45 was pivotal for the composition of Matthew’s ‘two-panel depiction’ of Jesus the teacher and wonder-worker. He used the seam between Mk 1:39 and 1:40 as the framework for his sermon in Matthew 5–7 and the seam between Mk 1:44 and 1:45 as the framework for his collection of miracle stories in Matthew 8–9.

However, more elaborate explanations of Matthew’s compositional procedure at this point have been offered. In his initial discussion of Matthew’s rearrangement of Mark’s narrative order, Allen maintained that Matthew’s transposition of Mk 1:40–45 and his subsequent inclusion of the non-Markan story of Jesus healing a centurion’s servant in Capernaum were part of a larger plan to begin this section on the healing ministry of Jesus (Mt. 8:1–9:34) with three healings of typical diseases: leprosy, paralysis and fever. Matthew then concluded this subsection of three healing miracles both by altering the πολλοὺς of Mk 1:34 to πάντας in Mt. 8:16b and by appending an Old Testament quotation in Mt. 8:17 to show that Jesus’ healing ministry fulfilled an Isaianic prophecy.

In response to Allen’s discussion of Mt. 8:1–4, Hawkins supplemented his argument:

I have no doubt that it was mainly, even if it was not exclusively, the subject-matter of this miracle which disposed Matthew to give it the place of honour [as the first miracle story]. For it would have a unique interest for him, and for the Jewish-Christians, whose habits of thought and whose needs he seems to have primarily regarded, both because of the prominence given to leprosy in Lv 13–14 and elsewhere in the Old Testament, and because of the illustration of the respectful attitude of Jesus towards the Mosaic law (as in Mt. 5:17–19) which is supplied by the reference to the priesthood.

Allen’s subsequent discussion of Mt. 8:1–4 in his ICC commentary betrays the influence of Hawkins’s remarks. After reiterating his earlier observation that Matthew probably wished to begin his cycle of miracle stories by recounting ‘three incidents of healing of typical diseases—leprosy, paralysis, fever’, Allen observed:

The incident of the leper, which in Mk seems to have no expressed details of time or place, is therefore substituted for that of the demoniac [i.e. Mk 1:23–28], and becomes the first miracle (8:1–4). The fact that this incident illustrates Christ’s attitude towards legal ceremonies may have co-operated in influencing the editor to place it immediately after the Sermon on the Mount.

Building on the work of both Allen and Hawkins, Bacon concurred that it was reasonable to maintain that Matthew rearranged the order of Mk 1:40–45 so as to begin his cycle of miracle stories with an episode in which Jesus acts in accordance with Jewish legal requirements. Despite his apparent agreement, however, he also warned against reading too much into Jesus’ instruction in Mt. 8:4.

It is true that the direction of verse 4 to ‘offer the gift that Moses commanded’, is thoroughly in line with the evangelist’s conservative attitude on matters affecting the Law (v. 17–20), and the loose chronological and geographical setting of this incident in Mark would permit an easy transfer to the position Matthew has given it. These reasons are often alleged (and not without force) as showing why this incident was made to stand first in the group, itself following immediately upon the Sermon. But the only motive which receives explicit support from the evangelist’s own words is the peculiarity of this miracle in directing a report to the temple authorities ‘for a testimony unto them’.

More recently, Tuckett reiterated the argument that the reference to the priesthood in Mt. 8:4 explains Matthew’s transposition of Mk 1:40–45. His discussion is instructive and merits close attention. He began by stating, ‘It is incumbent upon the defender of any source hypothesis to account for the changes which the secondary writers (on the hypothesis) must have made to their source(s)’. With respect to the disagreements in order between Matthew and Mark following Mt. 4:18–22 and Mk 1:16–20, Tuckett finds ‘too simplistic’ the broad explanation that Matthew wanted to group together a cycle of miracle stories, either to balance the impression of Jesus the teacher in Matthew 5–7 with a similarly extensive section on Jesus the wonder-worker or to set the stage for Jesus’ statement in Mt. 11:5. Referring to this kind of ‘blanket statement’, he remarked:

Not only does it make Matt. 8–9 too monochrome … but it also fails to account for the detailed changes in order which Matthew made (assuming MP) within this section. (E.g. within chs. 8–9, the relative order of the Markan pericopes used is not maintained.) A more precise set of reasons is thus necessary to make MP more credible here.

Tuckett proceeded to discuss the ramifications of two different arrangements of pericopes for explaining Matthew’s alterations to Mark’s order. On Huck’s arrangement, Mt. 8:1–4 and Mk 1:40–45 are parallels in order, which suggests that Matthew chose to insert his Sermon on the Mount into the Markan outline after Mk 1:39 (cf. Mt. 4:23), and continued with Mk 1:40–45 after concluding the Sermon. He then backtracked to incorporate Mk 1:29–34 after recounting the non-Markan story of the healing of the centurion’s servant in Capernaum (Mt. 8:5–13). On this arrangement, reasons must be given for Matthew’s transposition of Mk 1:29–34, not Mk 1:40–45.

However, Tuckett preferred Neirynck’s arrangement of pericopes, which suggests that Matthew rearranged Mk 1:40–45, not Mk 1:29–34. Is there an objective reason why he preferred Neirynck’s to Huck’s scheme of parallelization? All he said was, ‘By using Neirynck’s model a relatively simple explanation of the change in order now seems available, assuming MP’. In short, he preferred Neirynck’s schema because it makes Matthew’s rearrangement of Mark’s order easier to explain at this point. He would no doubt have admitted that to explain Matthew’s transposition of Mk 1:40–45 does not prove Matthew’s dependence on Mark, and that one should probably also try to explain Matthew’s rearrangement of Mk 1:29–34, assuming Huck’s arrangement of pericopes. However, he did not do so. Instead, he focused solely on potential reasons for Matthew’s rearrangement of Mk 1:40–45.

Following Held and others, Tuckett wrote:

The story of the leper shows Jesus fulfilling the requirements of the Old Testament Law by commanding the healed leper to go to the priests and offer the appropriate sacrifices. Indeed this forms the climax of the story in Matthew. (In Mark the climax is the wide fame which Jesus enjoyed [1:45], but this does not appear in Matthew: on MP this change is thus due to Matthew’s redaction [MattR], but in any case this feature is identifiable as the climax in Matthew’s story without reference to the Markan parallel.) This now fits in extremely well with Matthew’s overall tendency, clearly evidenced throughout his gospel, to portray Jesus as the fulfilment of, and in a line of positive continuity with, the Old Testament dispensation. (One can say this on almost any source hypothesis.) Matt. 5:17 (whose final form may owe a lot to MattR) sums this up well. It is thus quite in keeping with Matthew’s overall interests to have re-ordered Mark’s account so that the first miracle after the Sermon on the Mount (where 5:17 acts as a kind of ‘title’) shows Jesus as maintaining a thoroughly positive attitude to the Old Testament.

In response to Tuckett, it is worth asking whether Matthew’s alleged transposition of Mk 1:40–45 really does ‘fit in’ with Matthew’s redactional tendencies? Gundry and F.W. Beare141 represent those who think that Matthew’s placement of the healing of the leper was motivated by a desire to illustrate the fulfilment motif sounded at Mt. 5:17. Leaving aside the question of how Mt. 5:17 serves as a ‘title’, in Mt. 5:17 Jesus not only refers to the Torah but also to the Prophets. In other words, more than the Mosaic Law is in view at Mt. 5:17. A more important question is whether the command of Jesus in Mt. 8:4 really illustrates what he says about his relation to the Law and the Prophets in Mt. 5:17. In other words, is what Jesus commands the cleansed leper to do in Mt. 8:4—act in conformity with Mosaic regulations for the social reintegration of a healed leper—really an example of Jesus fulfilling the Law?

Without attempting to resolve the crux of how to interpret Mt. 5:17, which needs to be understood within the larger context of Mt. 5:17–48, one can at least say that for Matthew ‘fulfilling’ the Law was not simply a matter of acting in conformity with it. Davies and Allison adopt the conservative position that while Mt. 5:17–48 shows Jesus going beyond the Mosaic Law, it does not present Jesus as being in conflict with the Law. Nevertheless, they acknowledge that ‘πληρόω must at least be consistent with a transcending of the Mosaic law’. If the teaching of Jesus in Mt. 5:21–48 illustrates how he fulfils rather than abolishes the Law, which must surely be the case, Mt. 8:4 simply records an occasion when he advised someone to conform to Jewish legal requirements. Only by interpreting Matthew’s conception of fulfilment to mean obedience can one maintain that Mt. 8:4 illustrates the saying of Jesus at Mt. 5:17.

Given the most likely original socio-historical setting of this incident, the record of Jesus’ instruction to act in accordance with Levitical instruction may be little more than historical reminiscence. Like Gundry and Beare, Kingsbury maintained that ‘because Jesus commands compliance with the Mosaic law, he stands out in this story as one whose coming means, not the abolition, but the fulfillment of the law (cf. 5:17)’. Since Jesus’ command to offer what was required by Mosaic Law appears in all three synoptic accounts, however, one cannot place too much emphasis on its significance for Matthew. Tuckett maintained that the climax of Matthew’s account is Jesus’ command to the cleansed leper to present himself to the priest with the gift prescribed by Mosaic Law. But surely the climax occurs in Mt. 8:3 when Jesus responds positively to the request for cleansing, with the result that the leprosy is ‘cleansed’ immediately. The theme of Jesus’ authority, with which the Sermon on the Mount ends (Mt. 7:28–29), is carried over into the cycle of miracle stories. Just as Jesus’ words were authoritative in a way that those of the teachers of the Law were not, so his word also commanded authority over disease and demons (8:3, 13, 16). As John P. Meier noted in connection with the twofold depiction of Jesus as teacher and miracle-worker in Matthew 5–9, ‘Matthew, vis-à-vis Mark, stresses the word of Jesus in the miracle stories. This powerful word of Jesus, which possesses divine exousia (power, authority), is what binds together the sermon on the mount and the catena of nine miracle stories’.

Similarly, Thompson argued that within the subsection Mt. 8:1–17, Matthew emphasizes the words and actions of Jesus, particularly those that illustrate the fulfilment quotation in Mt. 8:17. Regarding Mt. 8:1–4, for example, he wrote:

Matthew emphasizes the power of Jesus’ word by repeating the same keyword [καθαρίζω] in the leper’s request … in Jesus’ response … and in his own brief description of the cure … He has also framed the miracle with a verbal inclusio: [λεπρὸς (vs. 2) ἡ λέπρα (vs. 3)]. In this way the evangelist gives a concrete example of how Jesus ‘took our infirmities and bore our diseases’ (8:17).

Heil also criticized the view of Held—subsequently seized upon by Tuckett—that the command of Jesus in Mt. 8:4 is ‘the entire goal of this story’. In his view, while the story of the healing of the leper is multidimensional and should not be understood to illustrate one particular theme—whether Christology, ecclesiology or faith—a unique dimension of this as well as other healing stories is soteriology. Using Matthew’s story of the healing of the leper as an example, he asked:

What is the unique aspect here which is not found in other gospel material in precisely the same way? Is it not the fact that the leper visibly experiences in his own life the miraculous saving power of Jesus? In no other type of synoptic gospel material do people receive the gifts of salvation in a way that visibly improves their personal lives … The healing of the leper can thus be called a salvation experience. It is soteriological in that the leper experiences salvation in the form of divine healing of his disease plus its social consequences.

In light of these observations, surely it is more defensible to argue that the focus of Matthew’s version of this story is the healing authority of Jesus—in fulfilment of the prophecy deliberately positioned by Matthew at the conclusion of his first triad of miracle stories—than to maintain that Mt. 8:4 forms the climax of this episode and thereby illustrates the fulfilment theme of Mt. 5:17.

Even if one accepts Tuckett’s evaluation of Matthew’s placement of the healing of the leper within his narrative, this is not evidence of Matthew’s dependence on Mark’s Gospel, a point Tuckett implicitly acknowledged in his parenthetical remark, ‘One can say this on almost any source hypothesis’. All that one must concede is that if Matthew borrowed from Mark, his redactional decision to transpose Mk 1:40–45 so that the healing of the leper occurred immediately after the Sermon on the Mount was reasonable and consistent, even if not for precisely the same reasons given by Tuckett. To make this concession, however, does not rule out the possibility that another explanation for the disagreement in order between Matthew and Mark might also be both reasonable and consistent.

Why not explore another perspective by asking whether the different emphasis in Mark’s account of the healing of the leper is suggestive of Mark’s posteriority? As I have pointed out elsewhere:

it is conceivable that Mark altered the purpose of this pericope because he was unconcerned about the fulfilment of Jewish legal requirements and rearranged its relative order to make this the first specific incident of Jesus’ preaching tour through the cities of Galilee (Mk 1:39) and to build on the fame that Jesus experienced as a result of his teaching and healing activity in Capernaum (Mk 1:21–34).

That such an alternative perspective is worthy of serious consideration is supported by the following observations. First, one can appeal to Bultmann’s form-critical judgment that Mk 1:43 is not original but is rather a Markan addition. Coupled with this, Mk 1:45, which Tuckett considered to be the climax of Mark’s account, appears to be a secondary expansion. After pointing out that Bultmann regarded Mk 1:43 as a Markan addition, Sanders and Davies comment: ‘We might also suppose that the evangelistic conclusions of Mark and Luke (the healed man testified to Jesus’ deed) are later additions’. This is not evidence of Mark’s posteriority any more than Tuckett’s comments about Matthew’s version of the story constitute evidence of Mark’s priority. However, such observations are consistent with Mark’s dependence on, and alterations to, Matthew’s account.

In addition, as C.H. Cave noted in his study, ‘The Leper: Mark 1:40–45’, there is an ‘apparent lack of connection with its context in Mark’. By this he meant that within Mark’s narrative, Mk 1:40–45 falls between two clearly delineated sections, Mk 1:21–39 (the day at Capernaum) and Mk 2:1–3:6 (Mark’s cycle of conflict stories). However, as Joanna Dewey has shown, Mk 1:45 and Mk 3:7–8 form a ‘frame’ around Mk 2:1–3:6. Cave also noted the view of commentators like D.E. Nineham that because Jesus commands the leper to comply with the requirements of Jewish law, this may have influenced Mark to place this incident immediately before his controversy section (Mk 2:1–3:6). Robert A. Guelich agreed that ‘the pericope’s specific content involving Jesus’ ministry and the Mosaic Law (1:44) helps set the thematic stage for the following conflict narratives in 2:1–3:6 involving issues of the Law’. As a result, the argument that Matthew placed his account of the healing of a leper immediately after the Sermon on the Mount to illustrate Jesus’ regard for the Mosaic Law is no more convincing than the argument that for similar reasons Mark chose to place his version of this story immediately before his cycle of conflict stories in Mk 2:1–3:6. In each case, a reasonable explanation can be given for the placement of this pericope within its respective literary settings, but it is tenuous to argue from this that one or the other arrangement is secondary.

Other evidence within Mk 1:40–45 that is at least consistent with—even if not unambiguous evidence of—Mark’s posteriority is the redundancy of Mk 1:42, in which part of Mark’s expression parallels Lk. 5:13 while the other parallels Mt. 8:3. According to J. Bernard Orchard, ‘Though it is possible that Matt and Luke each adopted a separate half of Mark’s sentence, not knowing that the other evangelist had chosen the other part, it is a lot easier to see Mark combining the phrases of both Matt and Luke’. He is supported in this instance by Marie-Émile Boismard, who with reference to the healing of the leper wrote, ‘Luc et Marc dépendent en grande partie de Matthieu, et Marc pourrait avoir fusionné Matthieu et Luc’.158

Given the inconclusiveness of source-critical judgments based on either the relative order of the story of the healing of the leper within the synoptic gospels or the verbal agreements and disagreements within their respective accounts of this healing miracle, it seems prudent to say, with Davies and Allison:

It must be conceded … that the implications of Mk 1:40–45 par. for the synoptic problem may be perceived as ambiguous. The signs can be said to point in two different directions. Thus the passage, instead of contributing to the solution of the riddle of synoptic relationships, simply typifies the problem.

Matthew 8:18–34, Jesus Calms a Storm and Exorcizes Two Demons

In most synopses Mt. 8:18–34 is subdivided into three pericopes: 8:18–22 (On Following Jesus), 8:23–27 (Stilling the Storm), and 8:28–34 (Two Gadarene Demoniacs). This is a reasonable subdivision of the text, but the narrative coherence of this section as a whole should not be ignored. On the Markan hypothesis, Matthew evidently respected the integrity of Mk 4:35–5:20 and transposed it as a unified section, even though he inserted sayings on discipleship after modifying Mk 4:35 to suit a different context. On the other hand, he did not transpose Mk 5:21–43 along with Mk 4:35–5:20, despite the strong narrative continuity within Mk 4:35–5:43 as a whole (cf. Lk. 8:22–56). In both Mark’s and Luke’s accounts, the stories of the storm-stilling and the Gerasene demoniac, followed by the story of the haemorrhaging woman enclosed within the story of the resuscitation of Jairus’s twelve-year-old daughter, occur after their respective parable collections.162 Together, these four incidents form a unity within Mark’s and Luke’s narratives and are held together by the first return lake-crossing indicated in Mk 4:35–36 // Lk. 8:22, Mk 5:1 // Lk. 8:26 and Mk 5:21//Lk. 8:40 (cf. Lk. 8:37c).

While the same four incidents occur within Mt. 8:1–9:34, the first two and last two episodes are separated by three pericopes in Mt. 9:1–17, all of which are paralleled in order by Mk 2:1–22 and Lk. 5:17–39. On the Markan hypothesis, Matthew must be seen to have brought forward Mk 4:35–5:43 into an earlier context—certainly well before his own parable collection—but to have used Mk 4:35–5:20 before incorporating any of the controversy pericopes in Mk 2:1–3:6 and to have delayed using Mk 5:21–43 until after recording the substance of Mk 2:1–22. Before concluding Mark’s controversy section, Matthew must also have inserted two further miracles in Mt. 9:27–34, the inclusive summary of Mt. 9:35–38 (cf. Mt. 4:23) and the whole of chs 10 and 11 in which there are three further ‘anticipations’ of Markan pericopes. This section evaluates reasons for Matthew’s alleged anticipation of Mk 4:35–5:20; the following section reviews reasons offered for Matthew’s separate transposition of Mk 5:21–43.

On the Markan hypothesis, Matthew incorporated Mk 4:35–5:20 after relating three miracles in Mt. 8:1–15 and adapting Mk 1:32–34 to demonstrate that Jesus’ healings fulfilled Old Testament prophecy. To explain Matthew’s anticipation of Mk 4:35–5:20, Allen envisaged the following compositional scenario: Matthew omits Mk 1:35–39 because it is ‘irrelevant to his purpose’ of illustrating Jesus’ healing work; having already anticipated Mk 1:40–45 at Mt. 8:2–4, he now bypasses it; because Mk 2:1 indicates that the healing of the paralytic occurred on a subsequent visit to Capernaum, he chooses not to use Mk 2:1–12 immediately after Jesus’ healing of Peter’s mother-in-law in Mk 1:29–31 so as not to ‘confuse two visits’ to Capernaum; he postpones Mk 2:13–22 because of its close connection to Mk 2:1–12 and also because ‘it is not relevant to Matthew’s immediate purpose’; he ‘reserves’ Mk 2:23–3:6 for a controversy section; he bypasses Mk 3:7–35 because it ‘affords no illustration of healing’; and he ‘reserves’ Mk 4:1–34 for his own parable section. ‘This brings him to [Mk] 4:35–5:20. At 4:35 Christ is wearied with His ministry, and surrounded by a multitude. Matthew adapts this situation to what he has just related (8:18), inserts 8:19–22, and then takes over Mk 4:35–5:20, with considerable omissions, Mt. 8:23–34’.

Two aspects of Allen’s description of Matthew’s redaction are unconvincing. The first is Matthew’s alleged concern not to confuse Jesus’ two visits to Capernaum in Mk 1:21 and Mk 2:1. This is inconsistent with the kind of radical alterations Matthew must already have made to his Markan framework. In addition, while Matthew retained the relative sequence of Mk 2:1–22 at Mt. 9:1–17, it is not necessary to suppose that Matthew thought Mk 2:1–12 and 2:13–22 to be inseparable, as Allen suggested. The connection between Mk 4:35–5:20 and 5:21–43 is closer than that between Mk 2:1–12 and 2:13–22, yet on Allen’s theory Matthew separated Mk 4:35–5:20 from 5:21–43. In this connection, Allen’s note that Mk 2:13–22 was not relevant to Matthew’s immediate purpose is contradicted by Matthew’s inclusion of this material shortly thereafter, while still composing his section illustrative of Jesus’ miraculous deeds. Allen later noted that Matthew incorporated Mk 2:13–22 after his second triplet of miracles ‘simply because it was connected in Mark with the preceding, and in spite of the fact that it interrupts his illustrations of Christ’s miracles’. The connection between Mk 2:1–12 and 2:13–22 is not so close that Matthew must have felt bound to retain it. Nor can one avoid thinking that Mt. 9:9–17 was included where it was for some more significant reason than that it occurred in the same relative sequence in Matthew’s source material.

In part, a similar concern prompted Hawkins to respond to Allen’s study. Hawkins disagreed that there was a close connection between Mk 2:1–12 and 2:13–22, although he suggested that the connection between Mk 2:13–17 and 2:18–22 was probably too close to be broken. In his view, a more satisfactory reason for Matthew’s inclusion of Mk 2:13–17 at Mt. 9:9–13 was his concern to prepare for the next major section of his Gospel—the mission of the Twelve: ‘Therefore he includes in the division now before us [Mt. 8:1–9:34] any details he can find as to the calling of those who were to become apostles, or at least members of that body of disciples out of whom the apostles were to be chosen’.

This same purpose, according to Hawkins, also explains the presence of Mt. 8:18–22, which on the Markan hypothesis was not taken over from Mark but deliberately inserted into a section largely concerned to illustrate Jesus’ authority in deed. As Hawkins explained,

even if neither of these particular men [in Mt. 8:19, 21] was thought by Matthew to have become an apostle, he might well think that the records of both of them would serve to show how Jesus tried and sifted each member of that whole band of disciples out of whom the Twelve were to be selected.

Certainly Hawkins’s explanation for the presence of Mt. 8:18–22 and 9:9–13 is more convincing than Allen’s. Indeed, his view that the theme of discipleship is prominent in sections of Matthew 8–9, and is therefore an important clue to the arrangement of this section of Matthew’s Gospel, is now widely accepted.

Thus far Hawkins had provided perspicacious observations about Matthew’s incorporation of material relating to discipleship within a section widely acknowledged to be concerned to depict Jesus’ wonder-working authority, but he had yet to explain Matthew’s alleged transposition of Mk 4:35–5:20. Unlike Allen, he doubted whether Matthew was concerned about confusing two visits to Capernaum and thus postponed Mk 2:1–12 by bringing forward Mk 4:35–5:20. Instead, he offered three reasons why Matthew transposed this material: first, reiterating an idea proposed by J. Vernon Bartlet, he suggested that Matthew’s use of Mk 1:32–34, which mentions ‘evening’ and the gathering of a crowd, may have reminded him of Mk 4:35–36, which concludes Mark’s parables discourse and also refers to ‘evening’; second, he suggested that Matthew may have wanted to keep well apart the two healings of paralysis in Mt. 8:5–13 and 9:1–8; and third, he recommended Allen’s suggestion that after the first triad of miracles in Mt. 8:1–17, Matthew arranged a second ‘ascending triad’ of miracles to illustrate Jesus’ authority over natural forces (8:23–27), demonic forces (8:28–34) and spiritual forces (9:1–8).

Concerning Hawkins’s first suggestion, it is noteworthy that Matthew relates in immediate sequence two incidents that in Mark’s Gospel occur ‘when evening had come’ (Mk 1:32–34; 4:35–41). However, it is not entirely clear whether Mt. 8:18–27 is meant to have occurred at night. Perhaps the detail in Mt. 8:24–25 that Jesus was asleep and had to be awakened suggests a nocturnal lake-crossing, as in Mk 4:35–41. But Mt. 8:16 may provide the detail that explains this. As a result, it is difficult to attach too much significance to the temporal coincidence to which Hawkins referred. The ‘crowd’ of Mt. 8:18 is less easily explained. It might be the gathered company implied by Mt. 8:16 or an adaptation of Mk 4:35–36. On the other hand, it might also be Matthew’s way of beginning a new pericope.

If, as Hawkins suggested, Matthew postponed the story of the paralytic in Mk 2:1–12 because he had only just recounted the healing of the centurion’s paralysed servant, this implies that his incorporation of material from a non-Markan source caused him to rearrange the order of material in the source that provided him with his narrative framework. Would it not have been easier to incorporate Mk 2:1–12 along with Mk 1:40–45 and to reserve the story of the healing of the centurion’s servant until later in this section of his Gospel?

As for the final suggestion, first articulated by Allen, Hawkins had already expressed the reservation that while the sense of ascendancy in the second triad of miracles in Mt. 8:23–9:8 is remarkable, ‘the only question is whether it is not too subtle and (in no bad sense of the word) too artificial to have been designed by any of the Synoptists …’ Furthermore, as he also noted, if Matthew was responsible for the ‘design’ seemingly present in Mt. 8:23–9:8, the final sequence of miracles in Mt. 9:18–34 is, by comparison, anti-climactic. Hawkins might also have noted that in the same way that the material in Mt. 8:18–22 and 9:9–17 prepares for the following section on the mission of the Twelve, so the miracles of Mt. 9:27–34 anticipate themes that are developed further in succeeding sections (Mt. 10:25; 12:22–32).

In his article on the editorial arrangement of Matthew 8–9, Bacon acknowledged that with respect to the motives governing the arrangement of this section, ‘nothing thus far published is more penetrating than the analysis of Dr. W.C. Allen …’ However, he was unconvinced by the specific motives that Allen had attributed to Matthew:

Thus there is abundant room to doubt the desire imputed by Dr. Allen to the editor to present ‘three healings of disease’ followed by ‘three miracles which illustrate Christ’s power over natural forces, over the hostility of demons, and in the spiritual sphere’. We may well be even more sceptical of Matthew’s alleged desire to conclude with ‘a third series illustrating Christ’s power to restore life, sight and speech’. Motives of this kind will hardly explain the evangelist’s destruction in these two chapters of that Markan order which in all the latter part of his Gospel he follows so religiously. The motive supposed to control in the formation of sub-group b (8:18–9:8) has a very modern ring, and that of sub-group c (9:9–34) should surely not produce the anti-climactic order ‘life, sight and speech’. Rearrangement from such a motive would surely result in the order: speech, sight and life.

Concerning Matthew’s alleged rearrangement of Mk 4:35–5:20, Bacon cited Allen’s explanation for Matthew’s transposition ‘as an example of perplexity taking refuge in bare statement of fact which calls for explanation …’ because Allen offered no reason for Matthew’s incorporation of non-miraculous non-Markan material (Mt. 8:19–22) along with his rearrangement of Mk 4:35–5:20. Like Hawkins, but without reference to Hawkins’s discussion, Bacon urged that this subsection of Matthew 8–9 is concerned not only to illustrate Jesus’ wonder-working power, but also to bolster faith among Christian disciples.

A decade later, in Studies in Matthew, Bacon continued to hold that the second sub-group of miracles in Mt. 8:23–9:8, prefaced by Mt. 8:18–22, was intended to build up the faith and commitment of those contemplating the cost of Christian discipleship. However, he seems to have modified his earlier position in favour of Allen’s view that this triad of miracles is arranged in order of ascending importance:

We have defined as the probable purpose of the section Mt. 8:18–9:8, to encourage the applicant ready to undertake the ‘gospeller’s’ work undeterred by a life of homeless wandering, detached from close family ties, by assuring him that ‘authority over unclean spirits’ and ‘over all the power of the Evil One’ is committed to him if only he does not ‘doubt in his heart’. If the compiler’s aim was indeed, as we have surmised, to send forth not only exorcisers and healers but ‘men’ conscious of a divine authorization to declare to the penitent the forgiveness of their sins, then no selection could have been better among the accounts of Jesus’ ministry of faith than just the series: Command of Stormy Winds, verses 23–27; Command of Hosts of Demons, verses 28–34; Authority to Forgive Sins, 9:1–8. All three examples are naturally taken from Mark, but chronological order is totally disregarded, the examples being given in climactic sequence; for ‘authority to forgive sins’ can come from God alone.

Despite his earlier objections to Allen’s view that the miracle stories in Mt. 8:18–9:8 are arranged in climactic sequence, Bacon did not explain his change of mind. More important, however, were his observations on the pericopes that make up this second triad of miracle stories. For example, regarding the storm-stilling pericope, he pointed out that in Matthew’s account Jesus rebukes the disciples before rebuking the storm. Although he drew no source-critical conclusions from this observation, more recent scholars such as Bornkamm and Held have made much of this alleged transposition as part of Matthew’s redaction of Mark’s storm-stilling pericope.

As for Mt. 8:28–34, Bacon asserted that one proof of the priority of Mark is that ‘the Gerasene (Mt. “Gadarene”) Demoniac story is so greatly condensed that the reader fails to see why a whole “herd” of swine are needed to afford lodgment for only two demons, for the “legion” feature has disappeared’. Bultmann made the same point to illustrate Matthew’s occasionally less than apt abbreviation of Mark’s stories: ‘it is unintelligible why the spirits from the two demoniacs should pass into the whole herd of swine, because Matthew has omitted the saying about Legion (Mk. 5:9f.)’.

In 1948, Günther Bornkamm helped to launch the redaktions-geschichtlich approach to the gospels by publishing ‘Die Sturmstillung im Matthäusevangelium’. Assuming rather than explicitly arguing for Matthew’s dependence on Mark, Bornkamm claimed that Matthew reinterpreted this miracle story by removing it from its Markan context and inserting it into his series of miracle stories in Matthew 8–9. More specifically, he placed this story immediately after the sayings on discipleship in Mt. 8:19–22 and altered the details of the story to emphasize the motif of discipleship. In Bornkamm’s view,

Matthew is not only a handeron of the narrative, but also its oldest exegete, and in fact the first to interpret the journey of the disciples with Jesus in the storm and the stilling of the storm with reference to discipleship, and that means with reference to the little ship of the Church.

In other words, Matthew’s changes to Mark’s version—both by rearrangement and editorial modification—illustrate what it means to be a follower of Jesus.

Among Matthew’s alleged alterations is the internal transposition of Jesus’ rebuke to the disciples and the storm-stilling miracle itself. Whereas the miracle comes before Jesus’ reproach to his disciples in Mark’s and Luke’s accounts, Matthew reverses the sequence. ‘Before the elements are brought to silence, thus in the midst a mortal threat, the word of Jesus goes forth to the disciples and puts them to shame for their little faith’.

Bornkamm also noted that in contrast to Mark’s account, Mt. 8:27 refers to unidentified people who by means of their astonished question confirm what has occurred. Appealing to the form-critical investigations of Martin Dibelius, he identified Mt. 8:27 as a ‘choral ending’ typical of the style of paradigmatic narratives.

Building on Bornkamm’s work, Held noted that Mt. 8:19–22 is one of four occasions (cf. Mt. 8:11–12; 14:28–31; 15:22–24) in which Matthew added to, rather than subtracting from, a Markan miracle story. For Held, such ‘expansions’ are ‘a means of interpretation’. As he explained, ‘If the stilling of the storm is interpreted by Matthew as a paradigm of discipleship, then the placing of the two conversations about discipleship immediately before it can be nothing other than a means of this interpretation’. On this view, it is worth noting that in Mt. 8:18–27, Matthew not only anticipated Mark’s storm-stilling pericope but also his Q ‘expansion’. If current opinion that the order of Q is most faithfully reflected in Luke’s Gospel is correct, Matthew had to search a long way forward for Mt. 8:19–22 (cf. Lk. 9:57–60).

For Held, Mark’s storm-stilling account is a ‘primitive miracle story in the authentic style’ because ‘the typical marks of a miracle story come to the fore in it’. In his discussion of the redactional changes Matthew allegedly made to Mark’s ‘primitive miracle story’, he focused on two principal features: (1) the differences between Matthew’s and Mark’s introductory verses (Mt. 8:18–23; Mk 4:35–36); and (2) the inversion between the words of Jesus and the miracle itself. In Mt. 8:18–23 Matthew not only interposed the sayings on discipleship between the connecting verse, Mt. 8:18, and the miracle story itself in Mt. 8:23–27, but he also linked these sayings with the following pericope by catchwords. As for the inversion within the storm-stilling pericope itself, Held remarked:

Mark places the nature miracle of the stilling of the storm in the centre and the words addressed to the disciples are an appendage. By transposing the scene Matthew has created a conversation between the disciples and Jesus and placed this in the centre, so that now the stilling of the storm looks like an appendage. In this way it is no longer Jesus and the elements that constitute the theme of the narrative but Jesus and his disciples who are in peril. The miracle story becomes a story about the disciples, so to speak.

Building on Bornkamm’s and Held’s view that Mt. 8:18–27 is primarily concerned with the theme of discipleship, Tuckett noted that Matthew’s transposition of Mk 4:35–41 to follow the words of Jesus on discipleship is at least intelligible on the theory of Mark’s priority. He also suggested that the ending to the story in Mt. 8:28–34—the rejection of Jesus by compatriots of the two demoniacs—illustrates one aspect of Matthew’s theme of discipleship. For Tuckett,

The story now acts in Matthew as an illustration of the saying in 8:20 about the Son of man having ‘nowhere to lay his head’, and thus continues the theme of the costly nature of discipleship, following the one who finds no welcome with men. The transfer of this story which Matthew must have made, assuming MP, would be coherent with the dominant themes which seem to be emerging in this section of Matthew’s gospel.

In short, Tuckett contended that on the Markan hypothesis, Matthew’s transposition of Markan materials can be understood to serve thematic designs. The cogency of this explanation should not be underestimated, provided one is confident about Matthew’s primary thematic concern(s) in this section of his Gospel.

However, the view advanced by Bornkamm and Held that Matthew redacted Mk 4:35–41 to emphasize the theme of discipleship has not gone unchallenged. According to Paul F. Feiler, the wider setting of Mt. 8:23–27 (i.e. the series of miracles in Matthew 8–9), the immediate context (Mt. 8:18–22), Matthew’s redaction and parallels from Old Testament and Rabbinic sources (especially Jon. 1:3–16 and Ps. 107:23–32) indicate that Matthew’s primary theme in this pericope is not discipleship but Christology. As he observed,

The expression, ‘This was done in order that what was written in the prophets might be fulfilled’, is unique to Matthew. It focuses our attention on the fact that Matthew consistently interprets Jesus on the basis of Old Testament messianic expectations. In the light of such a methodology, the stilling of the storm can be understood as a portrayal of Jesus. As Jonah’s act was in the interest of Israel [according to Rabbinic understanding] and as Yahweh delivered those tossed by the currents of the deep, so Jesus is the Deliverer who saves those overwhelmed by the chaos and afflictions of life.

It is unnecessary to resolve the question whether Christology or discipleship was foremost in Matthew’s mind in this pericope because it is unlikely that he had a single purpose in recounting the story of Jesus stilling a storm. One must therefore be cautious about source-critical judgments based on overly restrictive understandings of Matthew’s thematic concerns in this pericope.

Returning to basics, the main differences between the three synoptic accounts of the stilling of the storm are as follows. First, each synoptist introduces the story differently, in part because in each case the episode preceding it is different. Matthew has Jesus respond to the sight of a crowd (ὄχλον) by giving the command (κελεύω, but to whom is not specified) to go to the other side of the lake. Before getting into a boat, however, he is confronted by a scribe who asks to follow him and another follower who asks permission first to bury his father. After responding to each in turn, Jesus boards a boat and is followed by his disciples (ἠκολούθησαν αὑτῷ οἱ μαθηταὶ αὑτοῦ). Mark has Jesus express his desire to go to the other side of the lake at the end of the day on which he taught a large crowd from a boat (cf. Mk 4:1). Leaving the crowd, the disciples (cf. Mk 4:34) take Jesus ‘as he was’ aboard a boat (παραλαμβάνουσιν αὑτὸν ὡς ἦν ἐν τῷ πλοίῳ). Luke has Jesus and his disciples get into a boat ‘on one of those days’ (ἐν μιᾷ τῶν ἡμερῶν), after which Jesus instructs his disciples to cross to the other side of the lake. They set sail and while they sail Jesus falls asleep.

Especially noteworthy in these respective introductions are the verbal agreements between Matthew and Luke against Mark in Mt. 8:23 and Lk. 8:22b and the uniqueness of Mk 4:36, both in expression and content. The subject matter—Jesus boards a boat with his disciples—allows for coincidental agreement between Mt. 8:23 and Lk. 8:22b. Mark’s detail that Jesus was taken aboard ‘as he was’ is curious, as is the seemingly superfluous reference to other boats in Mk 4:36b.

From a source-critical viewpoint, the reference to a ‘crowd’ (ὄχλον) in Mt. 8:18 is also enigmatic. As Davies and Allison ask, ‘Are we to imagine that the great crowds surrounding Peter’s house have become troublesome and that Jesus is seeking repose? Or is Jesus’ movement a challenge to the crowds to follow him?’195 According to Gundry, the latter is correct; indeed, in his view, the crowd consists of disciples. But this is unlikely. As in Mt. 5:1–2, Matthew differentiates the crowds from Jesus’ disciples (cf. Mt. 8:23).

The standard view is that ὄχλον at Mt. 8:18 was influenced by the presence of the term at Mk 4:36a. On the Markan hypothesis, this is reasonable. But it is not the only explanation for Matthew’s choice of terminology here. The phrase ἰδὼν δὲ ὁΊ ὄχλον (v.l. πολλοὺς ὄχλους) περὶ αὑτὸν ἐκέλευσεν … has marked affinities with the redactional phrase ἰδὼν δὲ τοὺς ὄχλους+verb used at Mt. 5:1 and 9:36. As a result, one may say that just as Matthew chose to make a transition to a new episode at Mt. 5:1 by having Jesus respond to gathered crowds, so too Mt. 8:18 marks a transition to a new episode by having Jesus respond to the crowd implied by Mt. 8:16–17. Matthew’s terminology is explicable without recourse to dependence on Mark.

Second, whereas Mark and Luke speak of a wind-storm (λαῖλαψ ἀνέμου), Matthew uses the phrase σεισμὸς μέγας … ἐν τῇ θαλάσσῃ). While ὁ σεισμὸς may mean a sea-storm, this is the same term used to denote an earthquake at Mt. 24:7; 27:54; 28:2; Mk 13:8; Lk. 21:11; and Acts 16:26. Gundry understands Matthew’s use of the term σεισμὸς in Mt. 8:24 to parallel its usage in Mt. 28:2, where it accompanies Jesus’ resurrection. In his view,

In Matthew … the storm does not pose a threat to the disciples. Rather, in correspondence with 28:2 it is a sign of Jesus’ majesty. The great shaking in the sea previews the majesty of Jesus in his resurrection, at which a great shaking again occurs and upon which he claims ‘all authority in heaven and on earth’ (28:18). The similarity of Matthew’s peculiar description of the storm with the phraseology of Jonah 1:4 (see esp. the Hebrew text) combines with his distinctive comparison between Jonah’s experience and Jesus’ rising from the dead (12:40) to confirm the symbolism of the great shaking as a preview of the risen Jesus’ authority.

While the language of Mt. 28:2a mirrors Mt. 8:24a closely, and while two of the three later occurrences of the term σεισμὸς appear at decisive points in Matthew’s narrative—namely, the crucifixion and resurrection of Jesus—one wonders whether Gundry does not read too much into the use of σεισμὸς in Mt. 8:24. At a more mundane level, Hagner comments that ‘the introduction of an earthquake here stands in some tension with the later reference, borrowed from Mark, to “the wind”, usually associated with storms rather than earthquakes’, ‘but he then notes that ‘evidence exists in Greek literature … that earthquakes were at times associated with winds and heavy seas (cf. BAGD, 746)’. If, as O. Lamar Cope and Davies and Allison affirm,202 the synoptists were aware of the close parallels between the storm-stilling pericope and Jonah 1, Matthew may well have expected his audience to assume that the quake in the sea was accompanied by a tempest (cf. Jon. 1:4).

Third, only Mark notes that the boat was filling with water and that Jesus was sleeping on a cushion in the boat’s stern. Once appealed to as evidence of Mark’s primitive character, such graphic details are now generally recognized to be stylistic traits that have little or no bearing on the issue of priority. Commenting on this pericope, John Rist observed that ‘there is no denying that Mark’s is the most graphic version. But “most graphic” does not equal “necessarily earlier” ‘.204

Fourth, all three synoptists put a different form of address on the lips of the disciples when they awaken Jesus. Matthew uses κύριε, Mark διδάσκαλε and Luke ἐπιστάτα (twice). Only Matthew has the disciples cry out to be rescued (σῶσον), whereas only Mark has them inform Jesus that they are sinking by means of the accusing question, οὑ μέλει σοι ὅτι ἀπολλύμεθὰ;

Fifth, alone among the synoptists, Matthew has Jesus reprove the disciples before rebuking (ἐπιτιμάω) the wind and water. Bornkamm and Held probably read too much into this variation. Certainly Held went too far in claiming that by means of this internal transposition—if transposition it was—Matthew makes the storm-stilling itself look like an ‘appendage’. Indeed, it is possible to read Matthew’s story as heightening the miraculous power of Jesus by delaying the miracle, thereby increasing the tension within the story. Only after speaking to his disciples does Jesus rise from where he was sleeping: τότε ἐγερθεὶς … Confronted by imminent, life-threatening danger, Jesus takes time to chide his disciples.

Theissen offered yet another perspective on the inverse order of miracle and reproach within Mt. 8:23–27. In his view, ‘motifs of secrecy and incomprehension which indicate a revelation to come’ in Mark’s narrative are absent from Matthew’s version of the same incidents. ‘In Mark, for example, the reproach after the stilling of the storm provokes the question when the disciples will have faith. In Matthew (8:23–27) the reproach is answered by the miracle. The story is complete in itself,. Theissen’s remark that in Matthew’s account ‘the reproach is answered by the miracle’ supports the view that Matthew did not intend to play down the miraculous power of Jesus; rather, the inappropriate fear of the disciples serves to emphasize Jesus’ mastery over life-threatening elements.

Like Bornkamm and Held, Theissen assumed Matthew’s dependence on Mark, but his observations on this pericope do not reinforce Mark’s priority. Indeed, after mentioning Mt. 14:33 and 9:25 as further examples of a similar absence of signs that ‘the miracle stories are about truth not yet disclosed’, he then observed: ‘While Matthew has generally strengthened the structural linking of individual pericopae, his treatment of the miracle stories makes them seem relatively self-contained narratives. They have an episodic character. Matthew is “even more than Mark a book of pericopae” (K.L. Schmidt)’.

While not dogmatic about Matthew’s dependence on Mark, B.M.F. van Iersel and A.J.M. Linmans thought there was ‘sufficient reason’ to regard Mt. 8:18–27 as dependent on Mk 4:35–41: ‘The main indication is indeed that vss. 26–27 presuppose that in the preceding verses mention has been made of the wind, which is not the case in Mt., but indeed in Mk (4:37 cf. Lk. 8:23)’. Davies and Allison also regard the detail that in Mt. 8:26 Jesus rebukes the previously unmentioned wind as one of three inconcinnities within Mt. 8:23–9:17 that are ‘readily explained as due to over-zealous abbreviation of Mark’. This judgment probably implies too literal a reading of Matthew’s narrative. What early reader or hearer of Matthew’s version of this episode would not have associated strong winds with a ‘quake in the sea’ and a boat being hidden by waves (Mt. 8:24)? If, as suggested above, Matthew was aware of the parallels between this pericope and Jonah 1, there is no need to posit ‘over-zealous abbreviation’ of Mark’s account to explain Matthew’s failure to mention the wind in Mt. 8:24. If he assimilated Mt. 8:24a to Jon. 1:4, which Davies and Allison accept as a possibility, that would readily explain his failure to mention the wind. The wind rebuked by Jesus in Mt. 8:26 may not have been mentioned before, but it was surely alluded to. As a result, little source-critical weight can be attached to this alleged ‘inconcinnity’.

If the storm-stilling pericope provides little evidence of Matthew’s familiarity with the text of Mark’s Gospel, what about Matthew’s account of the Gadarene demoniacs? Apart from Matthew’s conciseness, two details in his version of the story are particularly curious: (1) Matthew sets his account in the region of the Gadarenes (Mt. 8:28) rather than the region of the Gerasenes (cf. Mk 5:1; Lk. 8:26); and (2) Matthew has two demoniacs rather than one.

Textual evidence suggests that Γαδαρηνῶν is likely original at Mt. 8:28 and Γερασηνῶν at Mk 5:1 and Lk. 8:26. It is sometimes suggested that Matthew altered Mark’s ‘region of the Gerasenes’ to ‘region of the Gadarenes’ because Gerasa was too far inland to provide a suitable setting for this story.212 But textual uncertainties preclude source-critical confidence on this issue.

As for Matthew’s two demoniacs, perhaps the simplest suggestion is the view of Theissen that ‘Matthew frequently doubles the suppliants (exception: 15:22) simply on the ground that two shout louder than one. At least he doubles only where the motif of the cry of help is also present’. Bultmann noted that numbers, especially the number two, play an important role in popular story-telling. Despite listing numerous examples of the presence of two characters, often simply ‘nominal figures’, in classical literature as well as in Jewish-Christian tradition, he claimed that the presence of two demoniacs in Mt. 8:28–34 is undoubtedly a secondary development of the Markan version of the story.

There is a tradition of ascribing to Matthew the change from one to two demoniacs as a ‘compensating’ device. Hawkins suggested that Matthew’s two demoniacs in Mt. 8:28 were the result of his ‘conflation’ of the two exorcism stories in Mk 1:23–28 and 5:1–20, thereby helping to explain Matthew’s omission of Mk 1:23–28. This suggestion then found its way into Allen’s ICC commentary:

In view of the brevity of Mt. as compared with Mk in this section and the following, and to a less extent in the preceding one, it seems not improbable that when the editor came to Mk 1:45 and was proposing to pass on to Mk 4:35–5:20, he did not unroll Mk’s Gospel to these verses, but summarised them from memory, perhaps purposely shortening them. If that was the course adopted, δύο may be a slip of the memory; but it should be borne in mind that, having omitted a previous history of a demoniac, he may purposely have duplicated here by way of compensation. Cf. 20:30, where he has two blind men and Mk has one, with the fact that he had previously omitted a history of a blind man, Mk 8:22–26.

Gundry followed Allen in this respect, claiming that one of Matthew’s compositional habits was to compensate for omitted material by incorporating elsewhere fragments of the omitted material. In his view, such is the case in Mt. 8:28, although he also thought Matthew had a theological reason for adding another demoniac:

Matthew omitted the story of the man with an unclean spirit in the synagogue at Capernaum (Mark 1:21–28; Luke 4:31–37). Here he compensates for the omission by associating that man with the demoniac on the other side of the Sea of Galilee. The similarity of the two demoniacs’ responses to Jesus made the association natural (see Mark 1:24 with Mark 5:7). Cf. Matthew’s adding a second blind man in 9:27–31; 20:29–34 because of the omission of the blind man in Mark 8:22–26. The compensatory doubling may have something to do with Matthew’s distinctive interest in establishing every matter by two or three witnesses (18:16). The coming reference to Jesus as God’s Son would benefit from a double witness (v 29).

According to Hagner, ‘There can be little doubt that Matthew has introduced two demoniacs into Mark’s (and Luke’s) story of a single demoniac. It is easier to explain why Matthew would have increased the number than why Mark would have reduced it’. He accepted as possible that Matthew here compensated for earlier omitting Mk 1:23–28, but gave more weight to the suggestion that Matthew preferred stories about pairs of healed individuals because of the importance Jewish tradition attached to multiple witnesses. If this latter suggestion is cogent, however, it stands in tension with Hagner’s confident assertion that Matthew ‘introduced two demoniacs into Mark’s (and Luke’s) story of a single demoniac’. Among the synoptists, such a concern best fits Matthew’s Jewish-Christian mind-set (cf. Mt. 18:16). Add to this that Matthew was fond of ‘echoes’, repetition, duplication and what Ernst von Dobschütz described as ‘the schematic use of numbers’, and one is less inclined to say that where Matthew has two demoniacs, two blind men or two donkeys, he must have modified Mark. Were we certain that Matthew used Mark’s Gospel, these instances would confirm his tendency to alter traditional materials by creating doublets and doubling characters. But in the absence of certainty about the relation between Matthew and Mark, nothing can be concluded from the simple fact that in four pericopes Matthew has two characters where his synoptic counterparts have only one.

Regarding various suggestions as to why Matthew ‘changed’ one man with an evil spirit into two demoniacs, Davies and Allison comment that while none is impossible, none is ‘obviously probable’. From the traditional Augustinian perspective, Norman Walker asserted:

The ‘two’ Gadarene demoniacs (8:28), the ‘two’ blind beggars at Jericho (20:30) and the ‘two’ donkeys of Palm Sunday (21:1–6) recorded by Matthew, as against one in each case by Mark, point … to the improbability that Matthew was copying Mark. The Matthaean text with ‘two’ is more likely to have been the original rather than the Markan without ‘two’.

Walker did not argue his case, but he has been followed by John Wenham, who glossed Walker’s statement by saying:

There could well be something in this assertion, for in all three instances Matthew gives the shorter and more restrained account, altogether belying the customary notion that Matthew is heightening the miraculous. (One extra individual in the content of Jesus’ myriad cures only heightens the miraculous in a most trivial sense.)

Different again is the view of Rist, who argued that Matthew’s and Mark’s accounts of this incident are independent:

Some say that Matthew’s two demoniacs at Gadara are a conflation of the Markan Gerasene demoniac with the demoniac from Capernaum. But if so, someone (Matthew) has deliberately (and for no very obvious theological reason) scrapped the Capernaum story and deliberately botched the Gerasene. And if Mark knew Matthew, the same difficulties arise. Mark has deliberately separated the two demoniacs and landed one of them in Capernaum. Of course the solution may be that either Mark knew (or supposed) that Matthew was mistaken, or vice versa. But it may be more plausible to suppose that Matthew does not follow Mark (or know Mark?) on these matters, nor Mark Matthew.

Such diversity of opinion precludes confidence about the source-critical significance of the two demoniacs in Matthew’s account, as opposed to one in Mark’s and Luke’s accounts. Hagner’s confident assertion, ‘There can be little doubt that Matthew has introduced two demoniacs into Mark’s (and Luke’s) story of a single demoniac’, is unwarranted. In fact, Matthew’s disagreement with Mark and Luke in this respect is not obvious evidence of priority or posteriority.

It has been noted that Bacon and Bultmann considered Matthew’s references to the herd of pigs in Mt. 8:30–32 to betray his familiarity with Mark’s account because he fails to mention the large number of demons involved (cf. ‘Legion’ in Mk 5:9). One might have expected Davies and Allison to appeal to this datum as an ‘inconcinnity’ due to Matthew’s severe abbreviation of Mark’s text, but they do not do so. Nor do other recent commentators. Rist referred to the absence of the name ‘Legion’ in Matthew’s account, maintaining that if Matthew abridged Mark’s much longer version of the story, he did it poorly. As a result, he maintained that a more likely explanation is that Matthew did not have Mark’s account to abridge.227 If Rist is correct, the difficulty noted by Bacon and Bultmann remains unexplained—except perhaps by Matthew’s inability to relate an internally consistent and coherent narrative—or is not a genuine difficulty. The latter is more likely.

In conclusion, one is forced to say that although it is possible to interpret Mt. 8:18–34 as a redacted version of Mk 4:35–5:20, neither the different context nor details within these parallel pericopes requires this redaction-critical interpretation. One must also say that redactional analyses based on allegedly different thematic concerns within these parallel pericopes are especially tenuous. A further difficulty for the Markan hypothesis is that if Matthew borrowed from Mark, he chose the more difficult compositional procedure of interposing material from a much earlier Markan context (Mk 2:1–22) between his anticipation of Mk 4:35–5:20 and his later anticipation of Mk 5:21–43. Markan priorists need to provide a satisfactory explanation for Matthew’s failure to bring forward Mk 4:35–5:43 as a single block.

Matthew 9:18–26, Jesus Heals a Woman and Resuscitates a Girl

It has been noted that in Mark’s and Luke’s Gospels, this composite or intercalated perieope immediately follows the episodes of the stilling of the storm and the Gerasene demoniac (Mk 4:35–5:20; Lk. 8:22–39). By contrast, Matthew interposed three pericopes between his account of the two Gadarene demoniacs and this perieope. In addition, Mt. 9:18–26 immediately precedes the final two miracle stories in the cycle of miracle stories in Matthew 8–9, one of which, Mt. 9:27–31, is a doublet of Mt. 20:29–34 (cf. Mk 10:46–52), while the other, Mt. 9:32–34, is a doublet of Mt. 12:22–24 (cf. Mk 3:22).

On the Markan hypothesis, Matthew’s incorporation of Mk 5:21–43 follows two relatively drastic redactional decisions. After choosing to ‘anticipate’ the substance of Mk 4:35–5:20 at Mt. 8:18–34, Matthew then ‘backtracked’ to retrieve the substance of Mk 2:1–22 at Mt. 9:1–17. This completed two subsections of the cycle of miracle stories in Matthew 8–9, as illustrated below:


Jesus Heals a Leper


Jesus Heals a Centurion’s Servant


Jesus Heals Simon’ s Mother-in-law


Exorcisms and Healings Fulfil Isaiah’s Prophecy


Sayings on Discipleship and Jesus Calms a Storm


Jesus Exorcizes Two Gadarene Demoniacs


Jesus Forgives and Heals a Paralytic


Jesus Calls Matthew and Responds to Questions

To begin his third subsection, Matthew must have bypassed once again a large section of Mark’s Gospel, Mk 2:23–4:34. This implies that within a relatively brief, discrete section of his Gospel, he borrowed from two disparate sections of Mark’s Gospel (Mk 4:35–5:43 and 1:29–2:22). More importantly, his manner of borrowing from these two Markan sections was not ‘economical’. In light of the first-century convention of using relatively simple compositional procedures, it is difficult to explain why Matthew did not incorporate Mk 4:35–5:43 as a single block.

In ‘The Dependence of St Matthew 1–13. upon St Mark’, Allen urged that Matthew altered the opening verse of Mk 5:21–43 to suit his new context and subsequently added two further miracles so as to end his cycle of miracle stories with ‘a third triplet of miracles illustrative of Christ’s power to restore life, sight, and hearing’. In a ‘Note on Some Changes Necessarily Made in Mark’s Narrative by Matthew when he Changed their Position’, he observed that whereas Mk 5:21 indicates that Jairus comes to Jesus while he is by the lake and surrounded by a crowd, Matthew places this incident immediately after Jesus’ response to the question about fasting posed by the disciples of John. As a result, Matthew was forced to alter Mk 5:21 to read ταῦτα αὑτοῦ λαλοῦντος αὑτοῖς …

The cogency of Allen’s explanation rests largely on his contention that Mt. 9:18–26 begins the third and final triplet of miracles in Matthew 8–9. However, this view required him to count the healing of the haemorrhaging woman and the resuscitation of a girl as a single miracle. In reply to Allen, Hawkins argued that there are ten miracles in Matthew 8–9, not nine as Allen thought. Furthermore, while impressed by the seemingly intentional arrangement of Allen’s first two triplets of miracles in Mt. 8:1–15 and 8:23–9:8, Hawkins considered that Allen’s explanation failed with respect to the miracles in Mt. 9:18–34:

Not only … are there four miracles remaining to be reckoned; but if we get rid of that difficulty by ignoring the healing of the issue of blood, the three miracles which remain are hardly such as could be congruously grouped together as a ‘triplet’ (see 9:18–34). It is suggested that they are chosen as ‘illustrative of Christ’s power to restore life, sight, and hearing …’ But they are miracles of such very different degrees of importance that the idea of restoration seems quite insufficient to distinguish them as a special class. And when we look at the order of the three items in this supposed class, there is in the descent from ‘life’ to ‘sight’ and ‘hearing’ an anti-climax which we cannot easily attribute to the compiler who ex hypothesi arranged so skilfully the ascending scale of miracles in the second triplet.

Allen was not convinced by Hawkins’s objections, as may be seen in his discussion of Mt. 9:18–34 in his ICC commentary. Although he accepted that there are, strictly speaking, ten miracles in Mt. 8:1–9:34, he held to his view that Matthew regarded Mt. 9:18–26 as one incident:

It is true that as a matter of fact there are ten miracles in 8:1–9:34, but 9:18–26 contains a miracle within another, and may be counted as one. And the fact that there are two previous series of three miracles, suggests that the editor reckoned this last series as three, not four.

Unfortunately, Allen did not respond to Hawkins’s comment that the miracles in Mt. 9:18–34 differed so much in importance that they failed to illustrate a single theme such as restoration. Nor did he take up Hawkins’s challenge about the ‘anti-climactic’ sequence of this final group of miracles when compared with the ascending scale of importance in the middle triplet of miracles (Mt. 8:23–9:8).

Like other miracle stories in this section of Matthew’s Gospel, Mt. 9:18–26 is much briefer than its Markan and Lukan parallels. The brevity of Matthew’s version has sometimes been claimed to reveal his dependence on Mark; at least his alleged omissions have been thought to be best explained by his alterations to Mark’s account. For example, according to Allen,

This section is much longer in Mk than in Mt. In part, this is due to the difference of situation in the two Gospels. In Mk. Jairus comes to Christ when He is by the lake side, and surrounded by a multitude (v. 21). But when Mt. transfers the incident to 9:18, Christ is in a house discoursing to the disciples of John. Consequently he has to omit Mk vv. 30–33, which could not have taken place in a house.

The word τότε at Mt. 9:14 does not necessarily indicate that Mt. 9:14–17 shares the same setting as Mt. 9:10–13, but the phrase καὶ ἐγερθεὶς ὁ Ἰησοῦς in Mt. 9:19 makes sense only in light of what is said of Jesus in Mt. 9:10, Καὶ ἐγένετο αὑτοῦ ἀνακειμενου ἐν τῇ οἰκίᾳ …

Equally, if not more, plausible is Allen’s other suggestion:

The shortening may also be due to the method adopted by the compiler, who, instead of unrolling his copy of Mk from 2:22–5:20, may have summarised 5:20–43 from memory, purposely shortening … It is certainly noticeable that the sections in which Mt. is considerably shorter than Mk, viz. Mk 4:35–41, 5:1–20. 21–43, are just those to obtain which the editor must be supposed to have unrolled his copy of Mk if he wished to see them before him.

In other words, Matthew’s parallels to the pericopes in Mk 4:35–5:43, all of which are ‘anticipations’, are briefer at least partly because instead of rolling forward to these passages in his copy of Mark, Matthew rewrote them from memory. This is certainly possible and might help to explain why Matthew did not relate Mk 5:21–43 along with Mk 4:32–5:20. If he retold these incidents from memory, he may not have recalled that they follow one another in strict sequence in Mark’s narrative.

Another argument based on the comparative brevity of Matthew’s account is Bacon’s:

Condensation of the story of the Healing of the Bloody Flux and Raising of Jairus’ Daughter (Mk 5:21–43) has reduced its compass by nearly two-thirds while enhancing the wonder beyond the psychologically credible (another proof of Markan priority). Thus Jairus evinces from the outset an incredible degree of faith by coming to ask from Jesus an unheard-of miracle: ‘My daughter has just died; but come and lay thine hand upon her and she shall live’. In the interwoven Healing of the Bloody Flux similar condensation leaves only the moral of the story, ‘Courage, daughter, thy faith hath saved thee’.

Apart from the danger of arguing on the basis of what may or may not have been ‘psychologically credible’ to Matthew and his audience, this argument is clearly reversible. Indeed, if psychological credibility were a legitimate criterion for determining literary priority, Matthew’s version is the ‘harder reading’ and therefore more likely to be prior, with Mark’s version modified to make the story more psychologically credible. Bacon’s inference is worthless.

Bultmann did not discuss Matthew’s version of this pair of intercalated miracle stories, but in his discussion of Mk 5:21–43 he raised three pertinent issues. First, he considered that the interweaving of these two miracle stories was pre-Markan. Second, he maintained that the original motive for combining these two miracle stories was to create an interval between the father’s request for healing for his nearly dead daughter and the subsequent message that his daughter had since died. Third, he regarded the name Jairus in Mark’s text as secondary: ‘Since the name is not to be found in D and is lacking in Matthew it must be regarded as in analogous instances … as secondary. It has been introduced from Luke’.238

Following Bultmann, Vincent Taylor suggested that ὀνόματι Ἰάϊρος was an early scribal addition, but Metzger found this conjecture unconvincing:

from a text-critical point of view it is more probable that the name Jairus was accidentally dropped during the transmission of part of the Western text (represented by one Greek manuscript and several Old Latin witnesses) than that it was added, at the same point in the narrative, in all the other textual groups.

Metzger also asserted that since Matthew severely ‘condensed’ Mark’s account, there is little difficulty in thinking that Matthew omitted the name Jairus along with his other omissions.

As for Bultmann’s two other points concerning Mk 5:21–43, later critics have concurred with his judgment regarding the motive for interrupting the story of the ruler’s daughter with the story of the healing of a haemorrhaging woman, but have generally disagreed with his view that the combination of these two miracle stories was pre-Markan. As Theissen noted, ‘The encapsulation of one miracle story inside the other is most probably the work of Mark’. If Theissen rather than Bultmann is correct, this has potential source-critical significance, for if ‘encapsulations are frequent in Mark’242 but relatively infrequent in Matthew, one can argue that Mt. 9:18–26 retains a Markan compositional technique because in this case there was no good reason not to do so, thereby revealing Matthew’s dependence on Mark’s version of these intercalated miracle stories.

Before evaluating this argument, Held’s discussion of this pericope warrants consideration. For Held, Mt. 9:18–26—particularly 9:20–22—illustrates Matthew’s abbreviation of Mark’s miracle stories for the purpose of interpretation. After noting Matthew’s ‘far-reaching abbreviation’ of the stories of the two Gadarene demoniacs and the raising of Jairus’s daughter, he then wrote:

Above all, the healing of the woman with the haemorrhage … gives the impression that Matthew has substituted a concise and clear report for one which in Mark is obscured by detail. Everything that is accidental is omitted, only the main line is worked out. In actual fact one can hardly speak of a real story any more. We have here no variegated picture, bewildering by reason of its diverse scenes, so to speak, but a simple outline drawing … Certainly it is not a matter of carelessness in this instance. It looks very much more as though the abbreviating is done in the interest of concentration on what is essential and must consequently be regarded as a means of interpretation.

In Held’s view, Matthew’s abbreviation of Mk 5:21–43 serves to underscore Jesus’ saying about ‘saving faith’ in Mt. 9:22, which is the key not only to the story of the haemorrhaging woman but also to the story about the ruler’s daughter. This is confirmed by the sequel to this pericope, Mt. 9:27–31, in which Jesus says to two blind men who express their confidence in his ability to give them sight, κατὰ τὴν πίστιν ὑμῶν γενηθήτω ὑμῖν. According to Held,

This statement of Jesus describes exactly the content of both miracle stories told in Matt. 9:18–26. Both people, the ἄρχων and the woman with the haemorrhage, are rewarded according to their faith. Thus it is seen not only that all three miracle stories are closely bound together with regard to content but also that the stories of the woman with the haemorrhage and of the restoration to life must, in fact, be understood as examples to illustrate genuine miracle-faith.

These observations are certainly consistent with the Markan hypothesis. Tuckett, who relied on Held’s and others’ analysis of the structure and major themes of Matthew 8–9, emphasized the consistency of Matthew’s redaction—and transposition—of Mk 5:21–43 if, in fact, the final subsection of Matthew 8–9 is concerned primarily with faith as the appropriate response to Jesus: ‘If … this whole section 9:18–34 is intended to illustrate the proper response to Jesus as being one of faith, there is little difficulty in seeing Matthew’s redaction of, and transposition of, Mark 5:21–43 in Matt. 9:18–26 as perfectly consistent with this’.

However, redactional consistency does not demonstrate Matthew’s dependence on Mark, particularly in the absence of discussion about the consistency of Matthew’s composition assuming some other source theory. Held himself posed the question whether Matthew’s brevity vis-à-vis Mark does not suggest Matthew’s use of an independent tradition, especially in relation to the story of the haemorrhaging woman. However, he presented a number of considerations that in his view made it improbable that Matthew is independent of Mark:

The placing of the miracle story within the narrative of the raising of Jairus’ daughter is only intelligible in connexion with Mark’s account of this story since it there brings out the halt in the journey, during which the child which is sick unto death does, in fact, die. The behaviour of the woman, who comes to Jesus ‘from behind’ and desires to touch ‘only’ his garment, is only intelligible in the situation which Mark depicts, namely that Jesus is hemmed in by a great crowd. The same applies to the remark that Jesus turns round and sees the woman. Finally Matthew follows the Markan account both with regard to order and the choice of words. Dependence on Mark is seen in the following passages:

Matt. 9:20a αἱμορροοῦσα.

Mark 5:25 ἐν ῥύσει αἵματος.

Matt. 9:20b προσελθοῦσα.

Mark 5:27 ἐλθοῦσα.

Matt. 9:22 στραφείς.

Mark 5:30 ἐπιστραφείς.

Matt. 9:22 ἰδὼν αὑτήν.

Mark 5:32 ἰδεῖν τήν τοῦτο ποιήσασαν.

Thus here also one must firmly retain the dependence of Matthew upon Mark and also at the same time ascribe the formal arrangement of this miracle story to Matthew’s conscious fashioning.

Here Held offered three principal reasons for Matthew’s dependence on Mark: (1) the intercalation of the two miracle stories is intelligible only in Mark’s account because it is only in Mark’s version of these stories that the interruption of the story about Jairus’ daughter serves any purpose; (2) the woman’s approach to Jesus ‘from behind’ and his ‘turning around’ to see her is intelligible only in Mark’s account because the presence of a crowd is integral to the story; and (3) the common order and use of similar vocabulary. These may be evaluated in reverse order.

First, the only potentially significant aspect of the similar order between Matthew and Mark is that the two miracle stories are intercalated in the same way, which is discussed below. As for the similar vocabulary, Held’s first two partial parallels do not indicate dependence one way or the other. They only indicate Matthaean dependence to one who already holds to Markan priority. The other two partial parallels relate to Held’s second reason for postulating Matthew’s dependence on Mark and should therefore be discussed in that connection.

As Allen pointed out, the fact that in Matthew this pericope follows a different episode than in Mark results in the absence of a crowd, which is prominent in both Mark’s and Luke’s parallel pericopes. Held’s and others’ reasoning is that the haemorrhaging woman could only approach Jesus unnoticed if he were surrounded by the crowd mentioned in Mk 5:27. In addition, Jesus would have to turn around to see her only if she were obscured by the crowd. However, Jesus turns around in both Matthew and Mark-but not in Luke—because the woman came and touched his garment from behind (ὄπισθεν). In this instance, Matthew’s account is as internally consistent as those of his synoptic counterparts.

As for the view that the woman could approach Jesus inconspicuously only if he were surrounded by a crowd, it is worth noting that whereas Mark and Luke note the presence of a crowd (Mk 5:21, 24; Lk. 8:40, 42), only Matthew states that Jesus’ disciples followed Jesus and the ruler (Mt. 9:19). While the disciples do not constitute a crowd, for Matthew their presence may have provided enough ‘cover’ for the woman to think that she could approach Jesus without being noticed. On the other hand, this may be reading Matthew’s account through Markan lenses. Perhaps inconspicuousness was not the issue for Matthew. As Schweizer comments, ‘In Mark the woman’s approach from behind Jesus is occasioned by the crowd’s surrounding him; in the briefer version of Matthew the shyness of the woman, who dares do no more than touch the tassels from behind (cf. 14:36), receives greater emphasis’. In any case, a close reading of Matthew does not support Held’s claim that these details are ‘only intelligible in the situation which Mark depicts, namely that Jesus is hemmed in by a great crowd’.

What of Held’s view that the intercalation of the story of the haemorrhaging woman within the story of Jesus’ resuscitation of a dead girl is ‘only intelligible in connexion with Mark’s account of this story since it there brings out the halt in the journey, during which the child which is sick unto death does, in fact, die’? This view, which echoes Bultmann, is shared by Schweizer:

In contrast to the Markan account, the official’s daughter is dead even before her father comes to make his request; this emphasizes the miracle of his faith; according to Luke 8:42 she is at the point of death. The interpolation of the healing of the woman with a hemorrhage makes sense only in Mark, because the child dies while this is taking place; this probably represents the earlier form.

‘Intercalation’, ‘encapsulation’, ‘interpolation’, ‘framing’, ‘sandwiching’ or, more simply, the insertion of one pericope within another is now generally considered to be a characteristically Markan literary technique. Although Ernst von Dobschütz’s 1928 study, ‘Zur Erzählerkunst des Markus’,253 is often cited as the earliest to identify intercalation as characteristic of Mark’s narrative style, Geert Van Oyen asserts that ‘in modern exegesis, the interpretation of Markan intercalation started with B. Weiss (1872)’. Some twenty passages have been suggested as illustrating Mark’s technique of intercalation, but six Markan pericopes provide the clearest examples of this phenomenon: Mk 3:20–35; 5:21–43; 6:7–30; 11:12–21; 14:1–11; 14:53–72. Of these, only the stories of Jairus’s daughter and the haemorrhaging woman are intercalated in both Matthew and Luke; in addition, Matthew’s parallels to Mk 14:1–11 and 14:53–72 are also intercalations.

James R. Edwards’s discussion of Mk 5:21–43 provides a useful starting point. After pointing out stylistic differences between the two stories in Mark’s account, he also noted numerous differences between Jairus and the haemorrhaging woman, particularly in terms of their social standing. In his view, the woman displays greater faith than Jairus. As a result,

The insertion of the woman with the hemorrhage into the Jairus story is thus not an editorial strategem whose primary purpose is to create suspense or ‘to give time for the situation in the main incident to develop’. The woman’s faith forms the center of the sandwich and is the key to its interpretation. Through her Mark shows how faith in Jesus can transform fear and despair into hope and salvation.

Edwards makes too much of the contrast between the faith of Jairus and that of the haemorrhaging woman. There is no indication that Jairus’s faith faltered, but the conclusion that ‘the woman’s faith forms the center of the sandwich and is the key to its interpretation’ is not thereby invalidated.

Edwards’s interpretation of Mk 5:21–43 is close to Held’s and Thompson’s interpretation of Mt. 9:18–26. This suggests that both versions of the story make the same essential point, even if, as Vernon K. Robbins has argued, the rhetorical composition of each account is different. According to Robbins, while the degree of verbal agreement between the three accounts of this intercalated pericope suggests some kind of literary relation between them, or at least dependence on a common source, that relation is probably ‘rhetorical’ rather than ‘scribal’. In other words, ‘Each version contains adoption, adaptation, and composition which exhibits the use of traditional material in a manner congenial to the rhetorical style and purposes of its author’.258 Thus,

Matthew has written an abbreviated version … and Mark has written an expanded version … Unless an interpreter has determined by other means that one has used the other as a source, there is no basis to know which writer may have used the other. From the data in this particular unit, Matthew could have abbreviated the Markan version, Mark could have expanded the Matthean version, or both Matthew and Mark could have composed from a common source.

This view is supported by Sanders and Davies, who, after noting various differences between Mt. 9:18–26 and Mk 5:21–43, comment: ‘The observation of these differences, we think, does not help us solve the synoptic problem. We learn instead about different interests on the part of the authors and, perhaps, their communities’.

But what of the argument that Matthew’s intercalation of these two miracle stories reveals his unconscious borrowing of a Markan compositional technique? If no specific detail within this intercalated pericope demonstrates Matthew’s dependence on Mark, does his use of the technique of intercalation betray dependence on Mark? While there is not unanimity that each example of intercalation in Mark is the result of his own composition, the current consensus is that intercalation is usually the result of Markan redaction. Tom Shepherd, whose doctoral research focused on the definition and function of Markan intercalation,262 has said: ‘All of my research leads me to the conclusion that intercalation in Mark is an intentional, planned storytelling device. As such it illustrates, or reveals, the Evangelist’s hand’.

On the basis of this conviction, Shepherd argued that if Matthew and Luke borrowed from Mark, Mark’s technique of intercalation ought to show up in their respective compositions, or at least that clear reasons should be forthcoming for their redactional deviations from Mark’s wording and theological emphases where he intercalated materials. In his view,

It follows that if Matthew and Luke utilized Mark’s manuscript as a primary source for their Gospels, intercalation ought to be reflected in their work, or, on the other hand, a clear redaktionsgeschichtlich reason ought to be explicable for their shift away from both the Markan wording and theological point. To postulate that a shift from the Markan wording is due to the other Evangelists merely not recognizing this Markan storytelling technique, is really to abdicate the redaction critical endeavor at a key point. Intercalation is not merely a matter of a few words added here or there, it is a pattern, a design for telling a story to make a theological point. For the other Evangelists to leave out intercalation without a traceable reason is tantamount to admitting that they did not have the Markan text before them when writing.

Before examining two specific examples of Markan intercalation along with their synoptic parallels, one of which is Mk 5:21–43, Shepherd made some general observations about the presence of intercalation in the Gospels of Matthew and Luke. He pointed out that of the six clearly intercalated pericopes in Mark’s Gospel, only one intercalation is paralleled in Luke and three in Matthew. In his view, the fact that only one example of Markan intercalation is paralleled in Luke’s Gospel tells against Luke’s dependence on Mark as a primary source. On the other hand, that three of Mark’s six intercalations are paralleled in Matthew indicates a closer relation between Matthew and Mark than between Luke and Mark. Nevertheless, as Shepherd argued, in each case where Mark’s intercalations are paralleled as intercalations in Matthew, either the elements that combine to form what he described as ‘dramatized irony’ is absent or its force is effaced.

Shepherd explained that intercalation within Mark’s Gospel is not only a technique of arrangement but also an integral aspect of story plot. As a result,

The dramatized ironies which intercalation sets up are communications between the implied author and the implied reader. In Mark 5:21–43 the dramatized irony of the intercalation centers on the life and death situation of Jairus’ daughter and the healing of the woman with the hemorrhage. The irony is that the little girl dies even as Jesus is healing (rather, being a passive agent of healing!) the woman.

Shepherd’s discussion of Mk 5:21–43 shows how various details in Mark’s account serve to emphasize the irony that the daughter (Mk 5:23) whom Jesus intends to heal dies while another ‘daughter’ (Mk 5:34) receives healing; indeed, one learns of the death of Jairus’s daughter while Jesus is still speaking words of healing and assurance to the haemorrhaging woman (Mk 5:35).

In Mt. 9:18–26, however, three features accentuate the absence of ‘dramatized irony’, as compared with Mk 5:21–43. First, the synagogue official’s daughter is dead from the outset; there is therefore no suspense or anxiety over whether Jesus will reach the girl before she dies. Second, Mark’s distinctive use of ‘crossover’ key terms such as σῴζω, φόβος/φοβέομαι and πίστις/πιστεύω either do not appear or do so in connection with only one main character. And third, whereas in Mark’s account the Jairus story moves from open to more closed settings in contrast to the haemorrhaging woman’s story moving from secrecy to openness, no such contrast of settings appears in Matthew. For Shepherd, these contrasts in ‘story patterns’ suggest that Matthew did not use Mark’s account of Jairus’s daughter and the haemorrhaging woman.

Shepherd also discussed Mk 14:53–72 and parallels, which confirmed his view that intercalation was an intentional technique used by Mark to create ‘dramatized irony’ between major characters and their actions. By contrast, this technique appears rarely in Matthew’s or Luke’s Gospel; even when it does, it does not perform the same function. He therefore concluded:

The Two Document Hypothesis contends that Matthew and Luke utilized Mark as a primary source. A lack of the appearance of Markan intercalation with its specialized function [dramatized irony] in either of these Gospels, without an adequate explanation for its absence, produces strain on the adequacy of the Two Document Hypothesis as a solution to the Synoptic Problem.

The conclusion of one who has examined the phenomenon of intercalation as carefully as Shepherd has cannot be dismissed lightly. Even if his conclusion were largely intuitive, it would warrant serious consideration. Nevertheless, there are two ways in which his conclusion might be rebutted. First, a key clause in Shepherd’s concluding remarks is ‘without an adequate explanation for its absence’, that is, the absence of intercalation with its specialized function of dramatized irony in Matthew’s and Luke’s Gospels. To counter Shepherd’s conclusion, advocates of the Markan hypothesis have only to provide an appropriate explanation for the absence of ‘dramatized irony’ in Matthew’s and Luke’s intercalations. Few critics would balk at suggesting that Matthew was willing to sacrifice dramatized irony, provided he was aware of it, to heighten Jesus’ authority and to emphasize further the importance of faith in Jesus.

Second, one might turn Shepherd’s conclusion on its head, arguing instead that the presence in Matthew’s Gospel of a Markan literary feature—but without the specialized function it has for Mark-indicates Matthew’s dependence on Mark. It is not difficult to imagine Matthew taking over the substance of a Markan intercalation without perceiving its specialized function, especially if he did not study Mark’s version of a story before incorporating it into his own narrative. One must keep in mind that Matthew probably relied heavily on his memory of particular incidents, whether those incidents were first written down in Mark’s Gospel or came from oral tradition.

In connection with the story of Jesus and the haemorrhaging woman, one should not overlook the verbal agreements between Mt. 9:20b and Lk. 8:44a against Mk 5:27b: The most important is the detail in Mt. 9:20 and Lk. 8:44 that the woman touched the ‘tassel’ (τοῦ κρασπέδου) of Jesus’ outer garment. In an ingenious study, J.T. Cummings argued that Mk 3:9–10 contains ‘verbal anticipations’ of Mk 5:24–34, whereas Mk 6:56 serves as a ‘sequel’. According to Cummings, τοῦ κρασπέδου was present in the pre-Markan tradition, but Mark shifted it to create a ‘tripartite ascending climax’ detectable in Mk 3:9–10, 5:24–34 and 6:56. Subsequently, Matthew and Luke noted τοῦ κρασπέδου in Mk 6:56 and incorporated this detail into their respective accounts of the haemorrhaging woman, thereby restoring an original detail into their secondary accounts! If so, Matthew borrowed this Markan detail twice, once at Mt. 9:20, where it is an anticipation, and once in his parallel to Mk 6:56 (cf. Mt. 14:36). Cummings is followed, albeit without acknowledgment, by Davies and Allison:

The Markan text [i.e. Mk 6:56] appears to function as a climax. In Mk 3:7–12 people touch Jesus. In Mk 5:21–34 a woman touches his garment. In Mk 6:53–56 people touch only the tassel of his garment. Because both Matthew and Luke have ‘the tassel of his garment’ in their parallels to Mk 5:21–34, neither has followed the Markan scheme. Perhaps they did not notice the clever arrangement.

Why these three Markan texts are thought to progress to a ‘climax’ is difficult to surmise. It is certainly understandable, if this is an example of ‘clever arrangement’ on Mark’s part, why a later redactor would fail to notice it!

Of the five occurrences of the term τὸ κράσπεδον in the New Testament, it occurs three times in Matthew’s Gospel (9:20; 14:36; 23:5) as opposed to one occurrence each in Mark’s and Luke’s Gospels. As a result, it is unlikely that its presence at Mt. 9:20 was influenced by Mk 6:56. Indeed, given that the mention of the tassel in Mt. 9:20 may reveal Matthew’s concern to depict Jesus as faithful to the Torah (cf. Num. 15:37–41; Deut. 22:12), one might argue from the two-gospel perspective that Mark omitted the term at Mk 5:27 because he did not share Matthew’s concern to portray Jesus as faithful to the Torah.

Finally, it needs to be reiterated that the disagreement between Matthew and Mark with respect to the relative order and context of this intercalated pericope is not obviously explained by Matthew’s alleged rearrangement of Mark’s order of incidents. As Rist opined in connection with this pericope:

There seems no particular reason for Matthew’s position. He either put Jairus where he did because, rightly or wrongly, he thought that this was the correct chronological sequence, or because some tradition had it there; it is hard to see why he should have put it where he has on grounds of appropriateness of context. The context he uses is not, admittedly, inappropriate, but neither is it particularly appropriate. In other words if Matthew were following Mark, it is hard to see why he has changed Mark’s order in this case; and if Mark were supposed to be following Matthew, the same point could be made.

Redaction critics can explain Matthew’s redaction of individual details in Mt. 9:18–26 and parallels, but they have not provided a convincing reason for Matthew’s transposition of this episode, especially his decision, ex hypothesi, not to transpose Mk 4:35–5:43 as a single block. On the other hand, if the intercalation of one pericope within another is a Markan literary technique, the presence of this technique in Mt. 9:18–26 (and elsewhere) may indicate Matthew’s dependence on Mark.


At one level, to explain Matthew’s disagreements in order as transpositions of Markan pericopes provides a relatively coherent explanation of the narrative order of Matthew’s Gospel. According to Wenham, ‘Supporters of Markan priority have found it difficult to account satisfactorily for Matthew’s changing of the order of Mark’. But this is inaccurate. More representative of the general consensus is the view of Graham N. Stanton, who accepts that the disagreements in order between Matthew and Mark are a prima facie challenge to the Markan hypothesis. Appealing to studies by Neirynck and Tuckett, however, he asserts:

Within the section 4:23–11:1 Matthew’s liberty of order is only relative, for Mark’s order can still frequently be traced; where the Markan order is changed by Matthew, he can be shown to have been inspired by his sources. In short, on the hypothesis of Matthean use of Mark, the evangelist’s changes to Mark’s order are not arbitrary but consistent and coherent.

The widespread acceptance of Markan priority during the twentieth century is due in no small part to the ability of scholars to explain Matthew’s disagreements in order as rearrangements of Mark’s sequence to suit his own narrative plan. However, when one examines the relevant pericopes (or pericope blocks) in detail, it is not such a simple matter to explain Matthew’s disagreements in order as rearrangements of Mark’s narrative sequence.

Perhaps the most important consideration against viewing Matthew’s disagreements in order as rearrangements of Mark’s order is the redactional acrobatics required to effect a number of these transpositions, for example, the alleged conflation of materials from widely disparate sections of Mark’s Gospel when composing Mt. 4:23–5:2. Not only would this be cumbersome, but in view of the compositional conventions of the time, to attribute such a redactional procedure to any of the gospel writers is probably anachronistic. On the other hand, close familiarity with the text of an earlier gospel(s) might lead a secondary redactor to conflate words and phrases from different sections of his or her source(s) subconsciously or to include from memory an apposite word or phrase into his or her text from an earlier or later context of the source(s) at hand. But in such cases one cannot expect source-critical certainty.

Another consideration that does not sit entirely comfortably with the Markan hypothesis is that while Matthew clearly arranged the earlier part of his Gospel on thematic—rather than chronological—grounds, it is within these early sections that he allegedly made the most radical alterations to his narrative source. One can argue that if Mark provided Matthew with the narrative framework for his own composition, Matthew was forced to rearrange Mark’s narrative sequence early on to achieve his more thematic or topical arrangement. But this begs the question.280 In addition, the argument is reversible. If one suspends judgment on the question of priority between Matthew and Mark and asks whether the organization of material in the first half of Matthew’s Gospel is better accounted for by Matthew’s rearrangement of Markan materials or by free composition—without denying the use of oral and written traditions—the latter seems equally likely, if not more so.

Chapter 8

Is Mark’s Gospel Posterior?
The Evidence of Pericope Order

In this chapter, I examine three passages in Mark’s Gospel to ascertain whether they can be understood as posterior to and dependent on parallel passages in the Gospels of Matthew and Luke. My approach in this chapter differs from that of the previous two chapters because, on the two-gospel hypothesis, Mark’s compositional procedure differs markedly from that of either Matthew or Luke on the Markan hypothesis, according to which each rearranged some of Mark’s pericopes. On the two-gospel hypothesis, one’s task is not to discern why Mark transposed certain pericopes in either Matthew’s or Luke’s Gospel, but to explain at certain junctures in Mark’s narrative his alternations between Matthew’s and Luke’s Gospels, with their distinctive relative orders.

In the same way that order differences determined which of Luke’s and Matthew’s pericopes were discussed in the previous two chapters, here Mark’s alleged transitions between the Gospels of Matthew and Luke determine which pericopes are discussed, namely, Mk 1:21–22, Mk 3:7–12 and parallels, and Mk 6:1–13 and parallels. On the two-gospel hypothesis, at each of these points Mark alternated from one principal source to the other, and in some instances changed from one source to the other more than once. If the phenomenon of order is to be regarded as evidence of Mark’s dependence on Matthew and Luke, Mark’s compositional procedure in these (and other) pericopes must be explained in a credible way. As C.S. Mann recognized, when commenting on Mark’s method of composition, ‘Since the present writer is convinced of the persuasive character of Griesbach’s hypothesis, he must indicate the manner in which the evangelist ordered his material. Nothing is more important than to discover why Mark chose his material in the fashion he did’.

An important aspect of analysing these Markan pericopes for evidence of dependence on Matthew and Luke will be to examine explanations of Mark’s compositional procedure proffered by those who either argue for Markan posteriority or assume it, including William R. Farmer, the two-gospel team of Farmer, David L. Dungan, Allan J. McNicol, David B. Peabody and Philip L. Shuler, the British Benedictine J. Bernard Orchard, and two recent commentators on Mark’s Gospel, C.S. Mann and Harold Riley. When appropriate, Johann Jakob Griesbach will also be consulted.4 Advocates of Markan priority will naturally be brought into the discussion whenever their viewpoints are pertinent. Before turning to the relevant texts, however, it is instructive to review how, after a period during which Mark’s priority was almost unquestioned, reconsideration of whether Mark might be the latest of the synoptic gospels is now possible.

Recent History of the Two-Gospel Hypothesis: An Overview

In The Synoptic Problem: A Critical Analysis, William Farmer’s critique of the two-document hypothesis restated and expanded upon earlier criticisms by H.G. Jameson, J.F. Springer, John Chapman, B.C. Butler and A.M. Farrer. Apart from the importance of his historical critique, however, the constructive part of his book was a new departure in at least two ways. The first was that the theory of synoptic relations he preferred to the two-source theory was not that of any twentieth-century critic upon whose work he built. Instead of adopting the traditional Augustinian hypothesis advocated by Jameson, Springer, Chapman and Butler, or retaining Markan priority but dispensing with Q, as Farrer had, Farmer resuscitated Griesbach’s theory, in which Mark is the latest of the synoptic gospels and dependent upon the earlier two.

The second noteworthy feature of his book was the systematic way in which he presented his hypothesis. After reviewing the various arguments that had led many to regard the two-source theory as an ‘assured result’ of New Testament criticism, Farmer devoted two chapters to his own theory. His ‘New Introduction to the [Synoptic] Problem’ is carefully argued, beginning with reasons for regarding the first three gospels as interrelated in a literary way and then setting out, step by step, a ‘web of evidence’ for Mark’s dependence on the Gospels of Matthew and Luke. In a final chapter entitled ‘Notes for a History of the Redaction of Synoptic Tradition in Mark’, Farmer supplemented the more formal argumentation in his ‘New Introduction’ with specific compositional suggestions about Mark’s literary procedure. Although many remain unconvinced by Farmer’ s argumentation, his hypothesis was nevertheless presented clearly.

The publication of The Synoptic Problem in 1964 rekindled debate on the synoptic problem, and Farmer gradually gained support from other scholars. As E.P. Sanders noted in 1975,

It is a rare book that changes anyone’s mind on a subject about which he has already formed an opinion, and in that sense The Synoptic Problem is rare. The appearance of the book was followed by a spate of dissertations and articles which challenged some aspect or other of the traditional two-source hypothesis. Perhaps not all of these were immediately triggered by Farmer’s book, but his work certainly paved the way for them. This prompt, positive reaction is also rarely seen, and it elevates the appearance of The Synoptic Problem to the status of an academic event.

Farmer was also instrumental, whether directly or indirectly, in organizing numerous conferences concerned to re-evaluate critical opinion on the gospels and their interrelations. The following survey outlines important developments in the revival of Griesbach’s hypothesis.

In 1970, at the Pittsburgh Festival on the Gospels, David Dungan emerged as a daunting critic of the two-document hypothesis and a staunch advocate of Griesbach’s hypothesis. His three-part paper (a) offered a caustic critique of traditional arguments for Markan priority and Q, (b) reviewed other hypotheses and (c) replied to major criticisms of Griesbach’s hypothesis. On the other hand, Joseph A. Fitzmyer presented what would later be described as a ‘classic’ defence of the two-source theory.9 In addition, James M. Robinson asserted that the success of redaction-critical studies of Matthew’s and Luke’s Gospels on the assumption of Markan priority helped to corroborate the priority of Mark. His accompanying comment—that a weakness in Farmer’s resuscitation of Griesbach’s hypothesis was the absence of a convincing redaction-critical investigation of Mark’s Gospel assuming Markan dependence on the Gospels of Matthew and Luke—proved to be a significant challenge to anyone willing to abandon Streeter’s ‘fundamental solution’ to join forces with Farmer.

Even if the result of Dungan’s, Fitzmyer’s and Robinson’s papers was something of a stalemate on the question of the tenability of the two-source theory, an important outcome of the Pittsburgh conference was a general loosening of commitment to any particular solution to the synoptic problem. While defending the status quo ante, Fitzmyer nevertheless surmised that ‘the history of Synoptic research reveals that the problem is practically insoluble’. Furthermore, in ‘The Interpretation of the Gospels Today: Some Questions about Aims and Warrants’, Albert C. Outler urged gospel critics to be less polemical and to tolerate greater diversity of opinion on source-critical questions. Since 1970, there has been greater diversity of opinion among scholars, even if the implications of the origins and interrelations of the gospels discourage detached and irenic discussion between critics of different persuasions.

In 1971, Dungan presented another paper at the Colloquium Biblicum Lovaniense. Leaving aside the issue of who might have wanted to produce a gospel like Mark’s from the Gospels of Matthew and Luke, Dungan focused on identifying historical precedents for the type of compositional procedure Mark employed on the Griesbach hypothesis and suggested an appropriate socio-cultural context within which a posterior Mark would fit. In short, he aimed to discern ‘more of the historical warrant supporting the Griesbach hypothesis’ literary contentions’.14 After showing that it was standard polemical practice in antiquity—including the period of the early church—to ridicule inconsistencies, absurdities and morally repulsive behaviour by dominant figures (divine or human) in the myths and sacred writings of opponents, he argued that the content and structure of Mark’s Gospel are explicable on the grounds that its author shared the editorial proclivities of Tatian and Marcion, whose gospels reflect ‘reactionary tendencies’. In Dungan’s view,

These [reactionary tendencies] show up in two different ways: (a) the desire to harmonize the tradition, either by reducing a manifold (and growing) gospel tradition into a unity (Tatian), or by bringing Gospel and Apostle into a doctrinal uniformity (Marcion). (b) They achieved this logical simplicity through an essentially editorial kind of act, namely, by excising corruptions and redundancies from already existing and widely accepted (= authoritative) stories of the life of Jesus. Neither added anything really new. Thus Tatian and Marcion seem to me to represent two reactionary figures, or perhaps a part of a reactionary trend, in the gospel producing activity of the early Church.

Within a short time, Dungan had demonstrated that it was possible to defend Mark’s editorial procedure on Griesbach’s source theory by proposing a markedly different, yet viable, account of socio-cultural forces behind the composition of Mark’s Gospel. Among those who heard his Leuven paper was J. Bernard Orchard, who subsequently played an important role in renovating Griesbach’s hypothesis.

In 1975, a review article by R.H. Fuller, E.P. Sanders and T.R.W. Long-staff revealed the momentum achieved by Farmer’s revival of Griesbach’s hypothesis. After surveying reviews of Farmer’s book and identifying those whom Farmer had influenced, Fuller emphasized the use of reliable ‘direction indicators’, what Farmer had called ‘canons of criticism’, to distinguish early from later traditions. ‘Henceforth’, Fuller advised, ‘when we work on the synoptic tradition, all the available direction indicators must be applied afresh to every pericope’.18 He also recommended that critics not ignore the probable ‘continued circulation of oral tradition after it had assumed written form in the gospels’. Accounting for the persistent influence of oral tradition renders one’s source theory less tidy and manageable, but also more historically credible.

Sanders explored various reasons why Farmer’s book had not received the attention it deserved, particularly in Germany. Nevertheless, he noted how in the previous decade many critics had become reticent about expressing whole-hearted confidence in the two-source theory. ‘Ten years ago’, he reflected, ‘the situation would have been precisely the opposite. In the English-speaking world, as in the German, the two-source hypothesis would have been regarded as the most firmly established result of New Testament literary criticism. Thus Prof. Farmer has produced a kind of revolution’.

Longstaff, who in 1973 had completed his doctoral dissertation, ‘Evidence of Conflation in Mark? A Study in the Synoptic Problem’, advised that his research revealed new insights into how conflators used their sources. After investigating several Markan pericopes for evidence of conflation, he concluded that ‘the evidence considered in the dissertation clearly indicates the possibility and the plausibility of the proposal that Mark has conflated material taken from Matthew and Luke when writing his gospel’ He also noted that scholars influenced by Farmer’s work had organized research seminars under the auspices of both the Society of Biblical Literature and the Society for New Testament Studies.

In 1976, at the Griesbach Bicentenary Colloquium convened in Münster to commemorate Griesbach’s contributions to New Testament criticism, Farmer surveyed ‘Modern Developments of Griesbach’s Hypothesis’. In this paper, later published in two parts, he noted three important developments that helped to explain the revival of Griesbach’s hypothesis during the period between 1964 and 1976: (1) Longstaff’ s 1973 dissertation documenting evidence of conflation in six Markan pericopes, evidence consistent with Mark’s dependence on the Gospels of Matthew and Luke; (2) Dungan’s research on the socio-cultural context of much early Christian literature, in which he identified the tendency—so evident in the writings of Tatian and Marcion, and possibly in the Gospel of Mark-to produce literature free from inconsistencies and contradictions; and (3) Orchard’s recently published defence of Luke’s dependence on the Gospel of Matthew, the first such attempt to explain Luke’s method of composition on neo-Griesbachian terms.24 Alongside developments within the Society of Biblical Literature and the Society for New Testament Studies, which he regarded as having helped to test the credibility of his hypothesis, Farmer considered the research of Longstaff, Dungan and Orchard to be particularly significant in the re-emergence of Griesbach’s hypothesis.

A major topic of discussion at Münster concerned gospel synopses. Heinrich Greeven, then in the process of revising Huck’s Synopsis, provided a history of ‘The Gospel Synopsis from 1776 to the Present Day’,28 and Xavier Léon-Dufour spoke on ‘The Gospel Synopsis of the Future’. Orchard, who proposed the conference in Griesbach’s honour and played a key role in its organization, argued that it was impossible to create a neutral synopsis, that is, one whose text and arrangement are not influenced to some extent by the synopsis-maker’s source theory. Greeven and Frans Neirynck contested this, but Orchard and Dungan have argued persuasively that there is no such thing as a neutral synopsis.

Reflecting on the Münster conference, Longstaff observed that although the two-document hypothesis remained the regnant theory, Griesbach’s hypothesis was ‘generally acknowledged to provide a strongly viable alternative to the two-document hypothesis’. He noted not only the emerging ‘climate of pluralism’ in the field of gospel criticism but also the view of most conference participants that this was a positive development.31 Another outcome of the Griesbach Bicentenary Colloquium was that the published proceedings contained Orchard’s English translation of Griesbach’s late-eighteenth-century Latin dissertation, Commentatio qua Marci Evangelium totum e Matthaei et Lucae commentariis decerptum esse monstratur, which made Griesbach’s source-critical work more widely accessible.

In 1977, at the Colloquy on the Relationships Among the Gospels in San Antonio, Texas, Roland Mushat Frye further substantiated the plausibility of Mark’s dependence on Matthew and Luke by noting similarities between the Gospel of Mark and other literature known to be secondary and derivative—similarities in linguistic style, comparative length and conflation. Although the relevance of some of his analogies was questioned, Frye seems to have shown that certain arguments levelled against the theory of Markan dependence on the Gospels of Matthew and Luke—for example, that a shorter, more vivid gospel could not be dependent on longer gospels and that Mark’s alleged conflation of two sources is unparalleled, if not impossible—are specious. Discussion at the conference revealed that participants were most impressed by Frye’s comments on conflation, particularly in light of the independent corroboration his study provided for Longstaff s research (published that same year) documenting evidence of conflation in Mark. Indeed, in response to Frye, Farmer remarked, ‘In my opinion, it is what Frye has written about “conflation” that constitutes his most enduring contribution to the solution of the Synoptic Problem’.34 Although Joseph B. Tyson warned against viewing Frye’s analogies as demonstrating that Mark’s Gospel is in fact a conflation of the Gospels of Matthew and Luke, he nevertheless opined that Frye’s and Longstaff’ s studies did show that Mark possibly conflated material from the other two synoptics.

Up until this time the burden of restoring Griesbach’s hypothesis had been shouldered by English-speaking scholars, even though some continental European critics were interested in the debate. In 1977, however, Hans-Herbert Stoldt published his History and Criticism of the Marcan Hypothesis, a book that both corroborated Farmer’s explanation of why Griesbach’s hypothesis had been overthrown by nineteenth-century proponents of the two-source theory and maintained Mark’s dependence on Matthew and Luke. His work was naturally hailed by neo-Griesbachians, but others were not so positive in their evaluation of its importance.

In 1979, scholars gathered in Cambridge to evaluate the plausibility of Griesbach’s hypothesis as an alternative research paradigm. Introducing the papers gathered together after the conference and published as New Synoptic Studies, Farmer explained the purpose of this conference:

By the end of the Trinity Colloquy at San Antonio it had become clear to many that the view that the Gospels were written in the sequence Matthew, Luke, Mark was an altogether respectable hypothesis, albeit not proven. What some defenders of the status quo seemed to want was a moratorium on further criticism of Markan priority and some solid work showing the results of using an alternate research paradigm. That was the basic rationale for the Cambridge Conference.

The Cambridge Gospel Conference marked a number of significant developments in the rehabilitation of Griesbach’s hypothesis. Among these was a change of name from the ‘Griesbach hypothesis’ to the ‘two-gospel hypothesis’, thereby indicating that ‘the basic Gospel tradition of the church can be traced back to two actual Gospels, Matthew and Luke, rather than two hypothetical documents, namely Ur-Marcus and “Q”.

The conference also provided a forum for recognized neo-Griesbachians like Farmer, Orchard and Dungan to explore issues of continuing importance for their hypothesis. Responding to George Kennedy’s observation at the Trinity conference that the patristic evidence is often dealt with in cursory or sceptical fashion, Farmer re-evaluated patristic statements on the chronological sequence of the gospels.41 Commenting after the conference, he wrote, ‘Nothing that is happening in Gospel studies today is more important than the present reexamination of the patristic evidence called for by Professor George Kennedy’. Since the Cambridge conference, the patristic evidence has continued to be a fruitful topic for research.

For his part, Orchard addressed an issue vigorously debated at the Münster conference, namely, whether a synopsis can be neutral. In an audiovisual presentation, he reiterated his view that there is no such thing as a neutral synopsis, illustrating his position from his forthcoming synopsis.45 Meanwhile, Dungan continued his quest for a convincing socio-cultural context for a posterior Mark. In ‘The Purpose and Provenance of the Gospel of Mark according to the Two-Gospel (Owen-Griesbach) Hypothesis’, first presented at the Cambridge conference in 1979 and subsequently discussed in 1980 at the Colloquy on New Testament Studies in Fort Worth, Texas, he identified his purpose as follows: ‘to provide a plausible historical explanation for the very existence of the Gospel of Mark’.47 Particularly noteworthy is that his research was undertaken to provide an explanation for the phenomenon of order. According to Dungan,

From a Two-Gospel point of view, it is this objective, non-reversible evidence based on the order of pericopes that constitutes the basis of the theory: wherever Matthew and Luke had the same episode in the same order, that Mark invariably took over into his account (with a few exceptions); in between he took from one or the other in succession (yet blending their details). This statement of the literary facts is the primary datum for which I must find a historical explanation or setting.

Taking the phenomenon of order as a point of departure, then, he addressed two principal questions:

What could have been the purpose in producing such a combined account as we are now hypothesizing Mark to have been? And secondarily, since the purpose for writing Mark can hardly be understood apart from the religious situation of the actual Christians who desired to bring this Gospel into being, we are necessarily required to ask also concerning the provenance of Mark.

After noting broad similarities between 1 Peter and Mark, Dungan speculated about the apostle Peter’s mediating role between ‘Paulinists’ and the ‘Torah Party’, particularly after the appearance of Matthew’s Gospel (favoured by Torah-bound Christians) and Luke’s Gospel (preferred by Paulinists). After Peter’s death, Christians in Rome asked Mark, Peter’s former associate (1 Pet. 5:13), to record Peter’s preaching. Using Peter’s speeches as a framework, Mark drew up an account based on the Gospels of Matthew and Luke that ‘transcended the partisan struggle between Mosaic Torah-rigorists and anti-Jewish, anti-Torah libertarians’. For Dungan, the purpose of Mark’s Gospel was to mediate between traditions contained in the Gospels of Matthew and Luke that were in some sense representative of dominant factions within the primitive church, while its provenance was Rome during the relatively stable period under Flavian rule following Nero’s death and the brief period of civil war.

This is not the place to critique Dungan’s proposal concerning the purpose and provenance of Mark. However, his thesis that Mark’s Gospel is a ‘kerygmatic suture binding together the Judaizing and Paulinizing divergency in the early church’ deserves some comment. According to Peter W. Agnew,

Dungan’s scenario for the two-gospel ordering of the gospels is not a contrived, arbitrary, or manifestly fanciful rewriting of the past to serve the needs of present scholarship. It is, rather, a legitimate use of the historical imagination in building upon the careful reexamination of the objective literary evidence, accomplished by those scholars who in recent years have mounted stiff challenges to the two-document ordering of the synoptic gospels.

Dungan’s use of historical imagination is legitimate, and his historical scenario is neither contrived, arbitrary nor fanciful, but his proposal nevertheless has weaknesses. For example, while the Gentile mission undoubtedly caused disagreements within the early church (see Gal. 2; Acts 10:1–11:18, 15:1–41), to designate these as tensions between two principal factions, as Dungan does, is probably too imprecise. He also overstated the differences between Matthew’s and Luke’s Gospels when he wrote, ‘Clearly, the appearance of these two writings, so profoundly at loggerheads with each other, posed a deadly threat to the tenuous unity of the Christian Brotherhood’.54 Statements of this kind reflect Dungan’s hypothetical scenario more than the gospels themselves. It is also doubtful whether Mt. 5:17–19 was directed against Paul, as he suggested. Even if his characterization of the tension between dominant factions within the early church were accurate, it is doubtful whether editorial glosses like Mk 7:19 would pacify Christians for whom conformity to Torah was important. Finally, the alleged similarities between Mark’s Gospel and 1 Peter do not advance his thesis because, as John H. Elliott observed, ‘The tension of which 1 Peter explicitly speaks concerns not an internal rupture between Jewish and Gentile branches of the church, as Dungan suggests, but rather a conflict between the elect and holy people of God (2:4–10) and an oppressive society of non-believing “Gentiles” (2:12; 4:3)’.

Another scholar who prepared a paper for the Cambridge conference was O. Lamar Cope, author of an important study on the Gospel of Matthew. Cope discussed four passages often cited as incontrovertible evidence of Matthew’s deliberate alterations to, hence dependence on, Mark’s Gospel. In each case he illustrated how evidence alleged to reveal Matthew’s revision of Mark’s text is not only suspect, but better explained by Mark’s ‘awkward’ use of Matthew.58 Particularly important was his ability to buttress his case by appealing to external, independent sources. He documented interpretive parallels between Matthew’s Gospel and the Dead Sea Scrolls, parallel thought patterns between Matthew and the Mishnah (specifically Pirke Aboth 6:3), and historical evidence from Josephus that is more compatible with details in Matthew than in Mark and Luke. In addition, the conclusion to which his evidence points, Matthew’s priority, is more probable from a historical cross-cultural perspective. As Cope argued,

While it is not a priori the case that the ‘more Jewish’ is the more original, it is very difficult to see how the subtly structured Jewish thought patterns and exegesis could be built from the Markan accounts. It is at least more likely that the shift from a Christian-Jewish milieu as in the case of Matthew and his tradition, to the Greco-Roman milieu of Mark and Luke led to unintentional, but recognizable, alterations and omissions from the originally Jewish-oriented material.

In three sessions of the Cambridge conference, papers on Matthew by Jack Dean Kingsbury, on Luke by Harold Riley, and on Mark by T.R.W. Longstaff reconsidered the theology of the respective synoptists from the two-gospel perspective. Perhaps the most important of these was Longstaff s ‘Crisis and Christology: The Theology of Mark’, in which he argued that the consensus regarding Mark’s theological concerns—to correct misunderstanding about Jesus’ messiahship and Christian discipleship—is in certain respects more compatible with the two-gospel hypothesis than the two-document hypothesis. He also suggested that on the two-gospel hypothesis, Mark can be viewed as having been written during a period of intense persecution, not to replace Matthew and Luke, but to indicate how their treatments of Christology and discipleship should be understood. Here Longstaff was responding to a long-standing objection to Markan dependence on Matthew and Luke, namely, given the Gospels of Matthew and Luke, why was Mark necessary? To my mind, his explanation of Mark’s purpose on the two-gospel hypothesis is more credible than the view that Mark’s Gospel was intended to mediate between contradictory, or at least conflicting, traditions within and/or between the Gospels of Matthew and Luke.

Finally, the Cambridge conference highlighted new research on linguistic and redactional features of the gospel writers. For example, Franklyn J.G. Collison (now deceased) presented a paper on ‘Linguistic Usages in the Gospel of Luke’, which challenged Henry Joel Cadbury’s influential work on The Style and Literary Method of Luke. Collison was one of three scholars who undertook doctoral research under Farmer’s supervision to identify linguistic and stylistic characteristics of each gospel.63

The publication of selected papers from the Cambridge conference enabled Farmer to include other essays that strengthened the consensus-questioning perspective of the volume as a whole. For example, to supplement his own reappraisal of the patristic evidence, Farmer included a study by the Patristics scholar Giuseppe Giovanni Gamba, written in response to discussion at Cambridge, and a re-examination of Augustine’s views on gospel relations by David Peabody.65 Among other essays that added weight to New Synoptic Studies, those in part 2 are especially noteworthy for their cumulative support for Mark’s posteriority and/or Matthew’s priority.

At Cambridge, participants called for a conference in which the major source theories could be defended in a holistic way by providing as much evidence as could be presented in their support. The venue and date were subsequently determined—Jerusalem, 1984. Between the Cambridge and Jerusalem conferences, however, three smaller conferences helped to prepare for Jerusalem.

In 1980, Southwestern Baptist Theological Seminary in Fort Worth, Texas, hosted a colloquy on the theme, ‘A Time for Reappraisal and Fresh Approaches’. This colloquy consisted of seminars on the synoptic problem, the genre of the gospels and Pauline chronology. While the seminar on the synoptic problem was small compared to previous conferences at Pittsburgh, Münster, San Antonio and Cambridge, one session proved to be particularly important from the two-gospel viewpoint. In discussing the ‘History and Development of Mark’s Gospel (From Mark to Secret Mark and “Canonical” Mark)’, Helmut Koester argued that five different recensions of Mark’s Gospel are discernible from internal and external evidence. As Peabody pointed out in response, however, the evidence Koester adduced not only indicates the posteriority of canonical Mark, but can also be explained on the two-gospel hypothesis without recourse to earlier recensions of Mark’s Gospel.68

E.P. Sanders’s keynote address on ‘New Testament Studies Today’ contained a number of observations on synoptic source criticism. ‘Source criticism’, he remarked, … is in well-known disarray. The classical two-document hypothesis, though defended by such a learned advocate as Neirynck, has been found wanting by most scholars who have studied the problem afresh since Streeter’s synthesis’.70 He also clarified his own position on the synoptic problem:

The one thing that my own studies have convinced me of is that there was no Q. I am not convinced that there were no sources, just that there was no Q as traditionally defined. The principal arguments for it are not true and the defenses of it are circular, arbitrary, and consequently unconvincing.

Among other noteworthy comments on different aspects of the synoptic problem, Sanders opined, ‘I think that Farmer (supported by Dungan and Longstaff) has the best of it with regard to the argument from order’.

In 1982, the Tübinger Evangelische Stift hosted a symposium on ‘Das Evangelium und die Evangelien’. Although most papers presented at this conference expressed little interest in the synoptic problem and dealt more with the transmission of traditions by and about Jesus, the relation between the gospel of Jesus and Paul’s gospel, the genre of the gospels, and historical and redaction-critical perspectives on the gospels, E. Earle Ellis’s paper, ‘Gospels Criticism: A Perspective on the State of the Art’, did note the impact of neo-Griesbachian scholarship. While cognizant of significant difficulties for the two-source hypothesis, however, he did not endorse the two-gospel hypothesis. Rather, for Ellis,

In the present state of affairs source criticism appears either to have come full circle or to have reached something of an impasse. In any case it is difficult to disagree with M.-É. Boismard’s judgment that both the two-source theory and the Griesbach theory are too simple to account for all the literary facts in the Gospels, whether or not one can accept his resolution of the problem.

Having said this, Ellis went on to suggest that new form-critical perspectives might illuminate the question of gospel sources. However, his judgment that both the two-source and two-gospel hypotheses are too simple to account for all the literary data expresses a form-critical perspective that is less concerned with the relation between the gospels as gospels than with the relation between so-called ‘traditions’ within the gospels. Seen from this angle, no source theory is ever likely to explain all the literary data in a satisfactory way.

In 1982 and 1983, two further conferences were held at Ampleforth Abbey in England. In the preface to Synoptic Studies: The Ampleforth Conferences of 1982 and 1983, Christopher M. Tuckett advised that some of the issues addressed were recommended by the planning committee for the Jerusalem conference, for example, the patristic evidence on the authorship and chronological sequence of the gospels and the question of potential analogies to the gospels in contemporary non-Christian literature. He also noted, ‘It is clear that the problem of the relationships between the gospels has become a matter of genuine scholarly debate, and the two-document hypothesis is no longer the firm bedrock of gospel study which it was once thought to be’.76 The two conferences also provided opportunities to explore the relation between the Gospels of Matthew and Luke and to evaluate arguments based on the phenomenon of order. Dungan’s and Tuckett’s discussions proved to be particularly instructive, each shedding new light on crucial aspects of the debate over which source theory best explains the phenomenon of order.

Shortly before the Jerusalem conference, Tuckett published The Revival of the Griesbach Hypothesis: An Analysis and Appraisal. Although unconvinced that Mark’s dependence on Matthew and Luke, or Luke’s dependence on Matthew, was as credible as Matthew’s and Luke’s independent use of Mark and Q, his study indicated that the Griesbach, or two-gospel, hypothesis could not be ignored and was worthy of detailed scholarly consideration.

Early in 1984, 20 years after the publication of The Synoptic Problem, 28 scholars gathered in Jerusalem for two weeks to participate in the most comprehensive investigation of the synoptic problem within a conference setting. Three source theories were scrutinized—the two-source hypothesis, the two-gospel hypothesis and the multi-stage hypothesis. Each theory was defended by a team that presented an overview of the hypothesis, concentrating on essentials, a discussion illustrating the theory with reference to specific gospel texts and a critique of the other two hypotheses. Neirynck headed the two-source team, which also included Tuckett, J. Keith Elliott, Helmut Merkel and Frans Van Segbroeck. Farmer led the two-gospel team, which also comprised Dungan, Peabody, Allan McNicol and Philip Shuler. Marie-Émile Boismard led the multiple-stage team, which also included Rainer Riesner and Philippe Rolland.

To assist the three research teams, the planning committee had proposed that each team address a guideline list of questions relating to the synoptic problem. This was intended to ensure that all three theories were evenly matched. Dungan, who edited the proceedings of the Jerusalem conference, explained:

it was hoped that this conference would make a unique contribution to the on-going debate by treating the Synoptic Problem as holistically as possible. In the opinion of the Planning Committee, this was something that had never been carefully and systematically attempted before. Therefore, the summary list of questions was recommended to each of the Research Teams as a guideline to follow when they wrote their Position Papers, as an ideal picture of the way in which all major issues pertaining to the central question of the relationships among the Gospels fit together. By adopting the same ‘paradigm’ of closely related questions for all three Position Papers, it was believed that the relative strengths and weaknesses of the three hypotheses might reveal themselves more clearly. In this sense, the Committee expected that the Position Papers which emerged from this rigorous exercise would present to the scholarly world, for the first time, a complete, more or less similar ‘portrait’ of each of the three major hypotheses.

Given that no previous attempt had been made to explore the interrelations of the gospels in such a comprehensive manner, the guideline questions admirably document the issues involved in attempting to resolve the synoptic problem. A first set of questions related to presuppositions that are rarely articulated in arguments for particular source theories: (1) What text of the gospels is used? (2) What synopsis is used and why? (3) Should Aramaic, Hebrew or Greek be given primary importance when considering the words ascribed to Jesus in the gospels? (4) Does the theory assume a primarily oral or written transmission of the gospel tradition(s)? (5) What role was played by individuals or the Christian community in forming and preserving traditions about Jesus? (6) How were the gospels written and what are the closest analogies in Graeco-Roman or Jewish literature? (7) What was the nature and extent of the Holy Spirit’s influence on the gospel writers?

A second set of questions related to data within the gospels: (1) What are the basic facts regarding the relations between all four gospels? (2) How is ‘secondary’ material distinguished from ‘original’ material? In other words, how is redactional material distinguished from source material, and what criteria are used for such judgments? For example, what role do the criteria of ‘semitisms’ and longer versus shorter accounts of the same material play in judging which version of a particular incident is earliest? (3) How is the pattern of agreement and disagreement in the order of pericopes explained? (4) How does one’s theory explain the similarities and differences in the order of words and phrases within pericopes? (5) How does one explain doublets, the so-called Q-material, the minor agreements of Matthew and Luke against Mark and the problems associated with the Old Testament quotations in the gospels? (6) Are there any non-reversible stylistic arguments? Is there evidence that would disprove one’s theory?

A third set of questions related to evidence external to the texts of the gospels: (1) Is evidence from Tannaitic Judaism significant? (2) Is evidence from Hellenistic culture significant? (3) What role does the following evidence from the patristic period play: direct statements about the order in which the gospels were written; quotations from the gospels; manuscript evidence? (4) What role is played by non-canonical gospels?

A final question related to the value of a source theory for theology or preaching and was posed in a manner reminiscent of Karl Barth’s threefold doctrine of the Word of God. How does the theory help to understand the incarnation of the Word of God in its three key ‘moments’: Jesus as the incarnate Word of God; the gospels as incarnations of the Word of God; and the situation of the theologian or preacher as an incarnator of the Word of God?82

Since the publication of the papers from the Jerusalem Symposium, the task of evaluating the two-source, two-gospel and multi-stage theories has been placed on a better footing. At the final session of the symposium, participants drew up two lists. The first comprised four points on which there was unanimous agreement: (1) the existence of direct literary dependence between the synoptic gospels; (2) the existence and use of earlier traditions in the synoptic gospels; (3) that a literary, historical and theological explanation of the evangelists’ compositional activity, giving a coherent and reasonable picture of the whole of each gospel, is the most important method of argumentation in defence of a source hypothesis; and (4) that the Gospel of John must be studied alongside the synoptic gospels, rather than treated independently.

The second, longer list detailed areas of disagreement, thereby indicating topics for future research: (1) the phenomenon of pericope order; (2) the pattern of agreement and disagreement in words and phrases within pencopes; (3) whether it is possible to identify redactional features of a specific gospel independently of a source theory; (4) the significance of doublets; (5) the minor agreements; (6) the influence or otherwise of the gospel genre(s) on the compositional activity of the gospel writers; (7) the ‘Jesus tradition’ outside the gospels with reference to the synoptic problem, and indeed to all four gospels; (8) the principles of synopsis construction; (9) the source-critical significance of Old Testament quotations and allusions in the gospels; (10) the process of transmission of tradition; (11) potential approaches to the synoptic problem in non-European contexts; (12) the socio-historical setting of each Gospel; (13) the theological implications of each source theory; (14) stating the phenomena that comprise the synoptic problem in a more satisfactory way; and (15) computer-aided statistical analysis of the synoptic phenomena.

In the published proceedings of the symposium, the two-gospel hypothesis was elucidated and defended by Farmer, McNicol, Dungan and Peabody. From the two-source perspective, Tuckett responded to the papers by Farmer and McNicol,85 while Boismard responded from the viewpoint of the multi-stage theory. As part of his response to the two-source hypothesis, Dungan included a ‘Chart Showing Mark’s Use of Matthew and Luke’,87 which was later revised for publication as ‘Narrative Outline of the Markan Composition According to the Two Gospel Hypothesis’.

This is not the place to evaluate all the papers by the two-gospel team. Nevertheless, some comments on their importance both for source criticism generally and for this study are apposite. First, the papers by Farmer, McNicol, Dungan and Peabody provide the most comprehensive statement and defence of the two-gospel hypothesis now available. Building upon earlier research, they present a coherent literary and historical case for the priority of Matthew, Luke’s dependence on Matthew, and Mark’s dependence on both Matthew and Luke.

Second, at various points two-gospel advocates appeal to the phenomenon of order as significant evidence of Mark’s dependence on Matthew and Luke and/or attempt to explain the phenomenon of order on compositional grounds in accordance with their source theory. For example, commenting on literary evidence for the two-gospel hypothesis, Farmer asserted:

Specifically, one can see clearly that it is the evidence from Matthew which explains much of the content and almost all of the order of Mark when it differs from the content and order of Luke, and so it is with Luke when Mark disagrees with Matthew. The fact that one can thus explain the order and content of Mark, and at the same time do this without needing to appeal to the use of a hypothetical source like ‘Q’, or even better, without appealing to hypothetical sources like ‘Q’ and ‘Ur-Markus’ or ‘Deutero-Markus’, is a consideration which clearly constitutes a prime reason for regarding Mark as third.

Farmer also addressed the question of how the phenomenon of order is explained compositionally on two-gospel terms. In his view, Matthew planned the narrative framework of his Gospel in accordance with the fulfilment-of-prophecy motif from Isa. 9:1–2 (see Mt. 4:15–16). In Matthew, Jesus discloses the light of God’s salvation first in Galilee, then ‘across the Jordan’ (Mt. 19:1). Following this ‘double fulfilment’ of prophecy, Jesus goes to Jerusalem. Into this narrative outline Matthew incorporated a series of discourses. In turn, Luke adapted Matthew’s framework of a ministry in Galilee followed by a passion account in Jerusalem by inserting a lengthy central section of sayings material using a journey motif. However, Luke’s dependence on Matthew is revealed by the fact that each of Matthew’s discourses is not only represented in Luke but, with one exception, also occurs in the same relative order.

Finally, Mark adopted the procedure of generally, but not inflexibly, following the order of Matthew and Luke when they agreed in sequence. When they disagreed in order, he usually followed the order of one or the other. For Farmer,

It is a distinct merit of the Two-Gospel Hypothesis that it can attribute the unique Synoptic phenomenon of order of episodes … to authorial intent and does not need to resort to the less satisfactory appeal to a literary accident due to random chance. (Random chance is how, on the Two-Document Hypothesis, one must explain the fact that whenever Matthew departs from the order of episodes in Mark, Luke supports Mark’s order and vice versa.)

These comments reveal the perceived importance of the phenomenon of order from the two-gospel perspective.

Other advocates of the two-gospel hypothesis also commented on the importance of the phenomenon of order. McNicol remarked that his analysis of the synoptic versions of the eschatological discourse was intended to complement what he described as the ‘two foundational pillars’ of the two-gospel hypothesis, namely, ‘the external testimony of the earliest Church Fathers that Matthew was the first of the canonical gospels to be written and the internal evidence of order of pericopes which suggests that Mark is a conflation of Matthew and Luke’. In Dungan’s response to Neirynck’s position paper, after first criticizing Neirynck’s methodological starting point, he remarked, ‘We agree with Neirynck that the argument from order of pericopes is the fundamental starting point’.92

In 1986, C.S. Mann published his Anchor Bible Commentary on the Gospel of Mark. Whereas Fitzmyer’s Anchor Bible Commentary on the Gospel of Luke presupposed, even defended, the two-source theory,94 Mann’s commentary presupposed Mark’s dependence on the Gospels of Matthew and Luke. Although the text of the commentary itself does not demonstrate Mark’s dependence on the other two gospels, Mann’s introduction contains an important discussion of ‘Synoptic Relationships and the Supposed Priority of Mark’, in which he provided reasons for accepting Mark’s posteriority to both Matthew and Luke. Two years after Mann’s commentary was published, Farmer noted:

It is too soon to tell what effect this commentary will have on the course of synoptic studies. Commentaries themselves offer limited scope for advancing scholarly discussions, tending rather to presuppose those discussions. Nonetheless, commentaries, if they have been carefully composed (and Mann’s work, nine years in the doing, has been so composed), can popularize scholarly work, and in some instances actually advance scholarly discussion.

In 1987, Peabody’s doctoral dissertation on Mark’s redactional features was published as the first volume in the New Gospel Studies monograph series. A student of Farmer, Peabody used a method for identifying redactional features of Mark without assuming a source theory. As such, his work is an important resource in synoptic source criticism.98

An important milestone in the development of the two-gospel hypothesis was reached when one of its advocates was commissioned to prepare an entry on the theory for the prestigious Anchor Bible Dictionary (ABD). In addition to articles on the ‘Synoptic Problem’ by Tuckett and the ‘Two-Source Hypothesis’ by Boismard, the ABD includes an article on the ‘Two-Gospel Hypothesis’ by Dungan. In his overview of the two-gospel hypothesis, Dungan comments on five interrelated aspects: (1) methodological presuppositions, (2) internal literary evidence, (3) external historical evidence, (4) patristic evidence and (5) theological consequences or ramifications of the theory. Regarding (1), he asserts:

The proper starting point, the basic phenomenon to be explained by any source hypothesis, is the whole concatenation of agreement and disagreement among all levels of all three Synoptic Gospels. Precisely this whole web or complex pattern is the basic phenomenon to be explained; it is the Synoptic Problem.

On the basis of this observation he again criticized the approach adopted by many advocates of Markan priority—and defended by Neirynck—to compare only two gospels together at any one time, specifically, Mark with Matthew and Mark with Luke.

Dungan also disputes any approach that claims to resolve the synoptic problem solely on the basis of oral sources or tradition. Here he concurs with most gospel critics, who consider that certain phenomena such as extensive agreement in wording and narrative sequence are best explained by at least some measure of literary dependence. Contrary to proponents of Markan priority, however, Dungan insists that greater weight should be given to patristic statements on relations between the gospels and their chronological sequence. Under a separate heading he reiterates the views of Farmer, Orchard and Peabody that important patristic authorities such as Clement of Alexandria and Augustine of Hippo provide external confirmation of the two-gospel hypothesis.

Finally, echoing a point of general agreement resulting from the Jerusalem symposium, Dungan remarks, ‘A successful source hypothesis will provide a complete, intelligible, redactional analysis of the whole of each of the Gospels’. This is an important point, but one that to date has favoured the two-source theory because of the redaction-critical work in many monographs and commentaries that assume Markan priority and Q. Hence the importance of redaction-critical work presupposing the dependence of Mark on Matthew and Luke, such as the commentaries on Mark’s Gospel by Mann and Riley, the two-gospel team’s ‘Narrative Outline of the Markan Composition’ and McNicol’s book on Jesus’ Directions for the Future.103 Dungan also emphasizes that any hypothesis is simply that, not a proven fact, because source criticism is, in part, a historical discipline and because any theory must remain open to new evidence. In view of the history of the debate on the synoptic problem, this is an important reminder.

Dungan’s remarks on literary evidence within the gospels are particularly important. Rejecting appeals to such standard, yet inconclusive, criteria as style, form and theological development, he asserts, ‘All source hypotheses must rest upon two types of objective internal literary evidence: macrostructural phenomena and microstructural phenomena’. Arguing that small-scale phenomena such as words and phrases were subject to significant alteration in the textual tradition, he urges that ‘one must begin by examining the unaltered major structural features of the Gospels’, particularly the ‘relative order of pericopes’.105

Having identified the phenomenon of order as the proper point of departure for a source-critical investigation of the gospels, Dungan summarizes ‘the literary facts with respect to the phenomenon of the relative orders of pericopes in the Synoptic Gospels’. He then asserts:

This statement of the literary facts can most readily be explained by the hypothesis that Mark combined Matthew and Luke: where their pericopes coincided in order, Mark utilized that account. In between, he selected from one or the other. All source hypotheses postulate a process of conflation to some degree. The process envisioned for Mark on the 2GH resembles Arrian’s when he wrote the Anabasis of Alexander, and it is confirmed in other biographical literature where it is known a third author used two earlier authors.

Before turning to ‘microstructural evidence’, Dungan refers his reader to the two-gospel team’s ‘Narrative Outline of the Markan Composition’. Since compositional arguments from order attend not only to ‘macro-structural evidence’ but also to linguistic details within relevant pericopes, Dungan’s comments on such data are of interest. After criticizing the dubious method of assuming the two-source hypothesis in standard works by J.C. Hawkins, C.H. Turner, H.J. Cadbury and others, he summarizes the results of Peabody’s and Longstaff’ s studies on redactional characteristics and evidence of conflation in Mark’s Gospel. To his mind these studies not only isolate compositional features of Mark’s Gospel but also support the two-gospel hypothesis:

The results of this new microstructural research coincide precisely with the macrostructural evidence discovered in research on the relative order of pericopes. Whenever the microstructural evidence put forth by a particular hypothesis corroborates the macrostructural evidence upon which the same hypothesis is founded, it gains in plausibility.

My primary concern in this chapter is to test the plausibility of Mark’s compositional procedure on the two-gospel hypothesis by focusing both on the phenomenon of order and on linguistic features within relevant pericopes.

To conclude this overview of the recent history of the two-gospel hypothesis, it is worth noting that J. Enoch Powell defended the primacy of Matthew and Mark’s dependence on Matthew and Luke in his book The Evolution of the Gospel. Although he never refers to Griesbach or any recent advocate of his hypothesis, this is the source theory Powell upholds: ‘Mark evidently regarded Matthew as his authoritative source which he normally followed. Confronted, however, with too severe a difficulty in Matthew, he held himself entitled to see if Luke could help’.109

Mark 1:21–22, Mark’s Initial Transition from Matthew to Luke

As demonstrated in Chapter 2, much depends on one’s arrangement of parallel pericopes, including compositional considerations relating to pericope order. A decisive discovery for the two-gospel team was the ‘major revelation’ concerning the appropriate Markan context for situating Matthew’s Sermon on the Mount: ‘With the key of the Two Gospel Hypothesis, we believe we can see where [Matthew’s Sermon] should properly be according to Mark’s own thinking, namely, after [Mk] 1:20′. Unfortunately, this revelation was not given to the best-known British proponents of Markan posteriority, Bernard Orchard and Harold Riley, each of whom followed Aland by positioning Matthew’s Sermon on the Mount after Mk 3:19. In Chapter 2, I give reasons for placing Matthew’s Sermon after Mk 1:20—in agreement with both Neirynck and the two-gospel team—although positioning the Sermon after Mk 1:39 (and Lk. 4:44) is also defensible. The following discussion presupposes my arrangement of pericopes.

If Mark compiled his account largely from the Gospels of Matthew and Luke, he clearly favoured Matthew’s sequence up to the Sermon on the Mount. Where Luke’s order differs from Matthew’s prior to this point, Mark agrees with Matthew. For example, he agrees with Matthew against Luke in the relative order of the arrest of the Baptist and the call of the first disciples; he also agrees with Matthew against Luke in delaying his account of Jesus’ rejection at Nazareth.

The two-gospel team does not comment on Mark’s agreement with Matthew against Luke on the relative sequence of John’s imprisonment. Riley makes some noteworthy observations about the significance of the Baptist for Matthew’s literary structure, but his source-critical conclusion is unsustainable. He describes Luke’s reference to John’s imprisonment as one of his ‘anticipated dismissals’, then comments: ‘To Luke the significant thing was that it was after his Baptism that Jesus began to preach in Galilee. Luke then rearranged his material on a thematic basis to deal with acceptance and rejection, first at Nazareth, then in Galilee, and finally at Jerusalem’. In Riley’s view, however, John’s imprisonment is structurally significant for Matthew’s narrative. By connecting the notice in Mt. 4:12 of John’s arrest and Jesus’ subsequent withdrawal into Galilee with the notice in Mt. 4:17 that from that time onwards Jesus began to preach, he maintains that ‘this is the first of the two dates on which Matthew hinges Jesus’ career’. The second such ‘hinge’, in his view, is the reference to John’s death, after which Matthew records that Jesus ‘withdrew from there’ (Mt. 14:13; cf. Mt. 4:12). Depending on one’s pericope divisions, ten or eleven pericopes interpose between Mt. 14:13 and the second occurrence of the seemingly important phrase Ἀπὸ τότε ἤρξατο … at Mt. 16:21, yet Riley considers that a structural connection exists between Mt. 14:13 and Mt. 16:21. ‘To Matthew, John’s imprisonment and John’s death were turning points in Jesus’ career’.

In Riley’s view, having chosen to follow Matthew’s order rather than Luke’s in Mk 1:1–15, Mark recorded John’s arrest but either failed to see the importance this event had for Matthew or chose not to attach the same significance to it:

At neither point does Mark see the significance of Matthew’s words. Later, he records John’s death without connecting it in any way with what followed. Here at the beginning, when Matthew speaks of Jesus’ hearing of John’s arrest, Mark simply writes Now after John was arrested. Mark has kept the timing without seeing the connection.

Riley’s view of Matthew’s structure is indebted to Jack Dean Kingsbury, for whom the key to the structure of Matthew’s Gospel is the twofold use of the characteristic phrase Ἀπὸ τότε ἤρξατο ὁ Ἰησοῦς … at Mt. 4:17 and Mt. 16:21. He is able to posit a connection between the death of John in Mt. 14:1–12 and the structural marker at 16:21 only by appealing to the theory of a Proto-Matthew, to which Mt. 14:28–31, 15:29–16:12 and 16:17–19 were allegedly added later.

W.D. Davies and D.C. Allison also note a link between John and Jesus in Matthew’s Gospel:

In both 4:12 and 14:13 a turn of events in the life of John makes for a turn of events in the life of Jesus. When John is handed over, Jesus opens his public ministry in Galilee (4:12); and when John is finally put to death, it is the beginning of the end of Jesus’ public ministry (14:13). Thus John and Jesus are linked in salvation-history.

However, they do not regard Mt. 4:17 and 16:21 as structurally significant, so do not advocate any connection between Mt. 14:13 and 16:21. Nor do their comments support Riley’s proposal of a close connection between the death of John and a second major phase in Jesus’ public ministry. While suggestive at points, therefore, his conjectures are too tenuous to sustain his judgment that Mark missed the structural significance of Matthew’s connection between John the Baptist and Jesus. However, one feature of his proposal, the repeated notice of Jesus withdrawing, may have source-critical significance.

Following his transitional summary at Mk 1:14–15, Mark has reached a point where the Gospels of Matthew and Luke diverge widely. On one hand, Matthew records the call of the first disciples (4:18–22), followed by his introduction to the Sermon on the Mount (4:23–5:2) and the Sermon itself. On the other hand, Luke relates Jesus’ rejection in Nazareth (4:16–30), followed by a series of incidents in Capernaum (4:31–44), followed in turn by his unique version of the call of the first disciples (5:1–11). The two-gospel hypothesis envisages Mark reproducing the substance of Mt. 4:18–22 using much the same vocabulary, then abandoning Matthew’s narrative to follow the sequence of events in Lk. 4:31–44. In doing so, he omitted both Matthew’s Sermon on the Mount and Luke’s programmatic rejection story.

Before considering why Mark might have deserted Matthew for Luke at this point, it may be instructive to note what has been said about the question of priority between Mt. 4:18–22 and Mk 1:16–20. Davies and Allison consider Mt. 4:22 to be ‘clearly secondary’ to Mk 1:20. ‘If Matthew were original’, they aver, ‘we would be at a loss to explain Mark’s destruction of the parallelism between Mt. 4:20 and 22 or the addition of the obscure reference to hired servants’. From the two-gospel perspective, the comments of both Mann and the two-gospel team are disappointingly general and lack source-critical force. For Riley, however, the rationale for Mt. 4:18–22 is to be found in the previous pericope, which interprets Jesus’ departure from Nazareth to ‘Capernaum-by-the-sea’ as a fulfilment of the prophecy of Isa. 9:1–2 that people in darkness in ‘Galilee of the Gentiles’ have seen a great light. Since Mt. 4:13 describes Jesus’ move to Καφαρναοὺμ τὴν παραθαλασσίαν as a fulfilment of prophecy, it is fitting that his account of Jesus’ public ministry in Mt. 4:18 should begin παρὰ τὴν θάλασσαν τῆς Γαλιλαίας. ‘The placing of the story is therefore integral to Matthew’s structure’, according to Riley. ‘Its place in Mark is due simply to the fact that he is copying Matthew’.

Riley does not consider Eduard Schweizer’s view that Mk 1:14–20 establishes a structural pattern that is twice repeated in the first half of Mark’s Gospel (cf. Mk 3:7–19 and 6:6b–13). On this view, Mk 1:16–20 plays an important structural role in the first half of Mark’s Gospel and cannot be said to appear at this point for no other reason than that Mark was following Matthew. However, Riley’s discussion of the ‘logic’ behind Matthew’s placement of the call story is plausible. Certainly Matthew’s account is explicable without recourse to Mark’s text. Given that the placement of both Matthew’s and Mark’s accounts of the call of the first disciples can be adequately explained within the overall structure of their respective narratives, it is foolhardy to maintain that one’s reason for recounting this incident at the outset of Jesus’ public ministry is more appropriate than the other’s, and more foolhardy still to argue on that basis that one of these accounts is prior.

As for the view of Davies and Allison that Mt. 4:22 is ‘clearly secondary’ to Mk 1:20, their criteria for making such a judgment, namely, Mark’s destruction of the parallelism between Mt. 4:20 and 4:22 and his ‘addition of the obscure reference to hired servants’, are tenuous. Mark’s stylistic ‘roughness’ could account for his failure to reproduce the parallelism between Mt. 4:20 and 4:22, and his penchant for circumstantial detail could explain his reference to the hired hands. In addition, with regard to parallelism, what one writer considered parallelism might well be considered redundancy by another. Mark was clearly not averse to redundancy, but he may well have chosen to vary his expression from time to time. More importantly, with regard to the hired servants, it could be argued that Mark softened the reference to sons abandoning their father by noting that Zebedee had hired help. In the ancient world, leaving one’s kin and social network was unusual.127 In Discipleship and Family Ties in Mark and Matthew, Stephen C. Barton argues on the assumption of Markan priority that the reference in Mt. 4:21 to Zebedee being with James and John in the boat, together with the ‘omission’ in Mt. 4:22 of Mark’s notice about the hired hands, makes the brothers’ abandonment of their father ‘all the more radical’. But does Matthew radicalize or Mark soften? While the call story in Lk. 5:1–11 clearly seems to be secondary, it is difficult to determine the issue of priority between Mt. 4:18–22 and Mk 1:16–20.

Turning to Mk 1:21–22, Griesbach’s original explanation for Mark’s transition from Mt. 4:22 to Lk. 4:31 was that Mark wished to avoid Matthew’s Sermon on the Mount because ‘as he meant to write a short book, it seemed to him too verbose, and, besides, it comprises many things which specially pertained only to those persons who heard Christ speaking on the mountain’. This suggests that Mark regarded much of Matthew’s Sermon as relevant only to a Jewish audience or that by Mark’s time many issues addressed in the Sermon were no longer pertinent, not only because of their Jewishness but because they were not relevant to the specific needs of his audience.

After noting that Mk 1:16–20 occurs in the same order as Mt. 4:18–22, Farmer offered the following reason for Mark’s decision to change sources:

Mark seems to have thought of Matthew’s Sermon on the Mount and Luke’s sermon on the plain as marking two clearly distinguishable and quite comparable literary transitions in his sources. He began by following Matthew up to his Sermon on the Mount, and thereafter proceeded to follow the order of Luke up to his sermon on the plain. In this way Mark deviated from his sources as little as possible, following their common order whenever possible, adhering first to the order of Matthew up to a distinguishable point of literary transition and thereafter the order of Luke up to the corresponding point in Luke’s narrative.

Speaking of Mark’s problem of combining two sources with different orders, Riley wrote:

There could be more than one solution possible, but the one for which Mark opts is to bring Matthew and Luke together at outstanding points, and as the first of these he chooses the Sermon on the Mount, which occupies a considerable place in both Gospels. As therefore Matthew has no sooner spoken of the call of the four disciples than at 4:23 he begins to lead into the Sermon, while in Luke the Sermon is preceded by a number of episodes, Mark at this point turns to Luke and begins to follow his order.

The most plausible aspect of Farmer’s and Riley’s comments is that on the two-gospel hypothesis, Mark diverges first from Matthew’s order and later from Luke’s order shortly before two similar discourses, one of which could be regarded as an abbreviation of the other. Assuming Markan posteriority, Mark could have reverted to Matthew’s order much earlier than he did. For example, Lk. 4:44 provided Mark with an appropriate point of transition, both because it is a summary not unlike that of Mt. 4:23 and because it precedes an episode similar in some respects to his call of the first disciples in Mk 1:16–20. That Mark chose, ex hypothesi, to delay his return to Matthew’s order until he reached a sermon similar to that which prompted him to make his first transition might be significant, but this cannot be argued with certainty. It is difficult to discriminate between this compositional procedure and that implied by the Markan hypothesis, according to which Matthew and Luke selected different but appropriate Markan contexts for inserting their respective sermons. However, if Mark aimed to compose an account of the good news about Jesus Messiah from the Gospels of Matthew and Luke, to proceed in the way that Farmer and Riley hypothesize was reasonable.

Concerning Mark’s alleged transition at Mk 1:21, the two-gospel team comments, ‘This story [Mk 1:16–20] brings Mk to the introitus to Mt’s Sermon on the Mount 4:23ff. which Mk desires not to repeat in his narrative, preferring Lk’s dramatic healings as a way of continuing the thread of Jesus’ early activities’. Although more precise than the remarks of Farmer and Riley, this simply makes explicit what is implied by the arrangement of pericopes. If Mark compiled his Gospel from the Gospels of Matthew and Luke, he obviously preferred Luke’s Capernaum sequence (4:31–44) to Matthew’s Sermon. Lacking is a discussion of the pros and cons of such an implied compositional procedure. Are there features in Mark’s text, for example, that support Mark’s posteriority at this point or suggest that Matthew’s Sermon and Luke’s rejection account were inappropriate at this juncture in Mark’s narrative? We are not told, yet such questions must be grappled with if the two-gospel hypothesis is to gain wider acceptance.135

In support of Markan dependence, Luke’s Nazareth episode would not have suited Mark’s evident purpose of illustrating Jesus’ authority. On the other hand, Matthew’s Sermon was one—but only one—way of underscoring Jesus’ authority at this early stage in Mark’s narrative (see Mt. 7:28–29). However, an author who has decided that Jesus’ authority can be illustrated as well with deeds as with words might well choose to relate the Capernaum sequence in Lk. 4:31–44. In addition, Mt. 4:23, the first use of a summary statement used also at Mt. 9:35 and again partially at Mt. 11:1, might have suggested to Mark that this was a good point at which to give a more precise description of Jesus teaching in synagogues, proclaiming and enacting the good news of God’s reign and healing sundry diseases among the people. The Capernaum sequence in Lk. 4:31–44 not only illustrates Jesus’ teaching, preaching and healing, but also emphasizes Jesus’ authority in word (Lk. 4:32) and deed (Lk. 4:36, 39, 40–41). As Mann notes, ‘this collection [Mk 1:21–39] appears to have been made to demonstrate, as typical of the Galilean ministry, the authority of Jesus’ teaching and his challenge to the dominion of evil (cf. Matt 4:13; 7:28, 29; Luke 4:31–37)’.

Mann also contends that the discrepancy between the implied ‘they’ of Mk 1:21a and the implied ‘he’ of Mk 1:21b reveals ‘either a confusion or a conflation of traditions’. If Mk 1:21 marks a transition from Mt. 4:18–22 to Lk. 4:31–32, this discrepancy could be explained on the grounds that Mark first altered the implied ‘he’ (κατῆλθεν) of Lk. 4:31a to the implied ‘they’ (εἰσπορεύονται) of Mk 1:21a so as to refer to Jesus and his newly called disciples (Mk 1:16–20) going to Capernaum, and subsequently, in conformity with Lk. 4:31, reverted to the implied ‘he’ (ἐδίδασκεν) of Mk 1:21b when referring to Jesus teaching in the synagogue. Admittedly, this may simply be the result of Mark’s informal style, but it is somewhat surprising that Mark did not explicitly identify Jesus by name as the subject of Mk 1:21b. On the other hand, Mt. 4:23 follows on from Mt. 4:22, in which James and John are the subject of the sentence, without explicitly identifying Jesus as the subject of περιῆγεν. It is difficult, therefore, to attach source-critical significance to Mann’s observations.

Coupled with the question of why Mark changed sources at Mk 1:21, the two-gospel hypothesis must also confront the question of Mark’s major omissions, notably his omission of most of Matthew’s Sermon on the Mount. Streeter’s objection to Jameson’s view of Mark’s dependence on Matthew was that only a lunatic would excise the Sermon on the Mount and numerous parables only to add ‘unimportant details’. This difficulty is exacerbated if Mark had both Matthew and Luke before him, as Dungan acknowledged in 1970:

It seems to have been impossible for Streeter to imagine why anyone in the early church who had Matthew (much less Matthew and Luke as the Griesbach hypothesis must contend) would even want to create out of them such a truncated, sayings less abridgment as Mark represents. What purpose would such a mutilation serve that could not be surpassingly better served by either Matthew or Luke (much less both together) in a far more glorious way? This crucial question must be answered before Mark’s lateness can ever seem very plausible, no matter how neat and simple the Griesbach solution may be on purely literary grounds.

Various reasons have been offered for Mark’s omission of so much teaching material in Matthew and Luke. First and most importantly, however, the problem of omission is not unique to the two-gospel hypothesis. As Farmer noted in response to F.H. Woods’s argument against Griesbach, ‘the argument from omissions is inconclusive at best, and there is no hypothesis which does not face this problem in some form’. On the Markan hypothesis, Matthew and Luke each omitted significant amounts of Mark’s narrative,145 although Matthew did so less than Luke. Source- and redaction-critical analyses of Matthew and Luke that assume Markan priority must account for such omissions. One is Matthew’s omission of Mk 1:23–28, which is explained with varying degrees of plausibility. In Luke’s case particularly, reasons given for his omission of large sections of Mark’s Gospel are often unconvincing, especially the explanation often resorted to that Luke’s copy of Mark was defective and did not contain Mk 6:45–8:21.

Responding to Fitzmyer’s objections to Griesbach’s hypothesis, a number of which concerned Mark’s alleged omissions, Farmer appealed to Peter’s speeches in Acts. ‘There is no reference in the speeches of Peter in Acts to Jesus as a teacher’, he noted. ‘Everything tends to be reduced to its strongest kerygmatic form, with a preponderant interest in presenting a highly dramatic portrayal of Jesus’ approval by God through “mighty works and wonders”.’147 He also urged that Mark can be said to have omitted material from Matthew and Luke that was peculiarly Jewish and/or Palestinian, although this was not an inflexible policy, indeed, could not be in recounting the story of a Palestinian Jew. Finally, he noted that on his theory Mark did not intend to replace the Gospels of Matthew and Luke, so that his omissions, including Jesus’ teaching, would not deprive his intended audience of important material.149

In Matthew, Luke & Mark, Orchard contended that Mark omitted episodes in Matthew and Luke that he could not conflate, for example, the nativity and resurrection narratives and the genealogies. Less convincing is his view that it was impossible to conflate Matthew’s Sermon on the Mount and Luke’s sermon on a level place, primarily because of differences between the respective Beatitudes that begin each sermon. Given that Luke’s sermon could be understood as a ready-made abridgment of Matthew’s Sermon, either the content of these respective sermons or the type of material they contain probably explains their omission better.

In ‘Crisis and Christology: The Theology of Mark’, Longstaff reiterated Fanner’s point that if Mark wrote not to replace Matthew and Luke but to complement them in some way, either by summary or interpretive emphasis, his omission of Matthew’s Sermon on the Mount and other didactic material would not signify their loss. Indeed, he goes so far as to say that ‘since in these cases the omission of materials would not imply their loss, there is little need for special attention to the reasons for omission’. This is something of a tour de force, but his main point is valid. In addition, it is worth recalling that within the predominantly oral environment of the first century CE, to write something down was not the only—nor was it necessarily regarded as the best—way of preserving traditional material.

In part 1 of The Order of the Synoptics, Riley confronted the issue of omission by asking, ‘Why, if Mark wrote with Matthew and Luke before him and knew of so much teaching material, did he not include more of it in his book?’ After acknowledging that Mark was clearly concerned to indicate the importance of teaching in Jesus’ public ministry, he discussed Mark’s use of parabolic material and, like Farmer, came to the conclusion that Mark’s primary interest was to present the kerygma about Jesus rather than the didache of Jesus. As for Mark’s omission of the Sermon on the Mount, as well as Luke’s similar sermon, Riley noted:

both Matthew and Luke speak of the Sermon as addressed to disciples (Mt. 5:1; Lk. 6:20), and therefore represent it as a guide to Christian conduct, the one in Jewish, the other in Gentile environments; it is part of the διδαχή that follows acceptance of the κήρυγμα. It looks as if, as in the case of the parables, it is just this character of διδαχή that has meant that it was not within the range of Mark’s purpose …

Working through the other discourses in Matthew, Riley pointed out that Mark usually has parallels to material in their introductions or opening sections, but has no parallel to the more general teaching in these discourses. In his view,

There is then a consistent pattern seen in Mark’s practice whenever he comes to the discourses in Matthew. He has produced his book, as its initial words testify, to proclaim ‘the gospel of Jesus Christ’ (1:1), the good news of which Jesus Christ is the subject, but not to draw out all its implications for his disciples.

Of course, it is possible to regard Mark’s parallels to the opening sections of Matthew’s discourses as ready-made contexts for Matthew to incorporate teaching material that he wanted to include because of its value for Christian disciples, which illustrates the inconclusiveness of arguments based on alleged omission.

On the two-gospel hypothesis, the type of material often omitted by Mark suggests a broad reason for his doing so. If Mark derived most of his material from Matthew and Luke, he clearly omitted much of their didactic material. But even if Mark’s was the earliest Gospel, he must have omitted a significant amount of Jesus’ teaching. That Jesus taught and was regarded as an authoritative teacher was important to Mark.158 He displays this differently from Matthew, however, by having Jesus addressed often as ‘Teacher’ and by frequently alluding to Jesus’ teaching rather than by incorporating the substance of that teaching. If, as recent Markan scholarship suggests, the Gospel of Mark is a corrective or polemical document, intended to ‘correct’ deficient conceptions of the nature of Jesus’ messiahship and of discipleship, then it is not Jesus’ teaching that could most usefully be redacted, but the broad story of Jesus, especially those incidents that serve to explain not only the identity of Jesus, but also misconceptions about him, as well as those that allow Mark to depict the incomprehension of Jesus’ disciples. Longstaff’ s understanding of Mark, which reflects recent Markan scholarship, not only accounts for Mark’s purpose on the two-gospel hypothesis but also provides a useful perspective for explaining his omissions.

On the two-gospel hypothesis, then, Mark’s primary reason for making the transition from Mt. 4:22 to Lk. 4:31, and later from Lk. 6:11 back to Mt. 12:15–21, was that he saw no need to record the blocks of Jesus’ teaching in Matthew 5–7 and Lk. 6:20–49. Beginning at Mk 1:21, he followed Luke’s pericope order until he approached a similar sermon, which in Luke’s account occurs on a level place after Jesus’ descent from a mountain (Lk. 6:12–19). Since his second major transition—from Luke back to Matthew—occurred shortly before he reached Luke’s sermon on a level place, we now turn to Mk 3:7–12 and parallels.

Mark 3:7–12, Mark’s Second Major Transition

Two-gospel advocates maintain that after abandoning Matthew’s order at Mk 1:21, Mark followed Luke’s order up until Lk. 6:11, with one easily explained exception; he omitted Lk. 5:1–11 because he had already incorporated the functionally equivalent pericope, Mt. 4:18–22, immediately prior to changing sources. Otherwise, each pericope in Lk. 4:31–6:11 is paralleled in Mark’s account in the same sequence. Despite this close attachment to Luke’s order, Mark’s account continued to be influenced by Matthew’s version of these incidents, which suggests that Mark either kept an eye on Matthew’s parallels or was familiar enough with Matthew’s Gospel to incorporate Matthaean reminiscences into his own retelling of this early phase of Jesus’ ministry.

On the two-gospel hypothesis, Mark’s familiarity with Matthew’s Gospel or his perusal of Matthew’s parallel texts even when following Luke’s order of incidents explains why Mark is sometimes closer to Matthew than to Luke within Mk 1:21–3:6. An example is Mk 1:22, one verse after Mark’s initial transition from Matthew to Luke. Bernhard Weiss, an influential nineteenth-century critic of Griesbach, regarded this verse as inconsistent with Griesbach’s view that Mark changed sources because he wished to omit Matthew’s Sermon on the Mount. According to Weiss, Mk 1:22 indicates that on Griesbach’s theory, Mark must have abandoned Matthew only after passing over the Sermon, which for Weiss suggested that this explanation for Mark’s initial transition from one major source to another lacked cogency. However, the close parallelism between Mk 1:22 and Mt. 7:28b–29 is not difficult to explain. Even without the similar and partially exact statement at Lk. 4:32, which echoes Lk. 4:22, it would not have been unreasonable for Mark to borrow Matthew’s description of people’s response to Jesus’ teaching. Indeed, given the striking content of this statement, one need not envisage Mark turning to Matthew to copy these words, although it would not have been difficult to do so. Given the suitability of this statement at this point in Mark, as evidenced by Lk. 4:32, and its conspicuousness in Matthew’s narrative—at the conclusion of one of the most obvious ‘landmarks’ in Matthew’s Gospel—the close verbal parallels between Mk 1:22 and Mt. 7:28b–29 pose no difficulty for the two-gospel hypothesis. Indeed, given the pattern of Mark’s use of διδάσκειν and διδαχή, it is possible to argue that Mk 1:22 reflects Mark’s knowledge of Matthew’s Sermon on the Mount.

In following Luke’s narrative sequence, Mark bypassed a significant block of Matthew’s Gospel (Mt. 4:23–12:14), some of which occurs in the same relative order, for example, Mt. 8:14–17, 9:1–17 and 12:1–14. Towards the end of this block, Mark followed the concurrent orders of Matthew and Luke (Mt. 12:1–14; Mk 2:23–3:6; Lk. 6:1–11). Within the last of these pericopes, Mk 3:1–6, there is an interesting pattern of alternating agreement between Mark and the other two synoptics, which Longstaff has subjected to painstaking analysis. He noted: (1) Mk 3:1–2 is ‘closely parallel’ to Mt. 12:9–10 but ‘significantly different from’ Lk. 6:6–8a; (2) Mk 3:3–5a is ‘closely parallel’ to Lk. 6:8b–10a but ‘almost completely different from’ Mt. 12:11–12; (3) Mk 3:5b is unique Markan material; (4) Mk 3:5c is ‘closely parallel’ to both Mt. 12:13 and Lk. 6:10b; and (5) Mk 3:6 is ‘closely parallel’ to Mt. 12:14 but different from Lk. 6:11. This analysis is correct, except with respect to Mk 3:1–2. While these verses are closer to Matthew’s wording than to Luke’s, it is an exaggeration to say that here ‘Matthew and Mark are much more closely parallel than are Mark and Luke’. Mark 3:1–2 shares sixteen words with both Matthew and Luke, of which nine are identical in form in all three pericopes; a further four are identical in form with Matthew but not with Luke. Mark shares three additional words only with Matthew, all of which are identical in form but of which two are the conjunction καὶ, and three only with Luke. In short, 19 words in Mk 3:1–2 are paralleled in Mt. 12:9–10, of which 16 are identical in form, while 19 words are paralleled in Lk. 6:6–7, of which 11 are identical in form. A more accurate summary of the data is that Mk 3:1–2 closely parallels both Mt. 12:9–10 and Lk. 6:6–7, but with regard to word form is somewhat closer to Mt. 12:9–10.

With one exception—the presence of περιβλεψάμενος in Lk. 6:10, which is characteristic of Mark-Longstaff was able to make sense of Mk 3:1–6 as a conflation of Mt. 12:9–14 and Lk. 6:6–11. He also regarded the two-gospel hypothesis as better able to explain the alternating pattern of agreement within this pericope.

Following Mk 3:6, Mark is confronted by Luke’s account of the call of the Twelve. Luke 6:12 contains the phrase εἰς τὸ ὄρος, the same phrase used in Matthew’s introduction to his Sermon on the Mount (Mt. 5:1). Luke uses a mountain setting for his version of the call of the Twelve, after which Jesus descends to a level place where he encounters a great crowd of disciples and many people from various regions. On the other hand, Mt. 12:15 notes that Jesus healed many who followed him. Luke 6:17–19 has similarities with Mt. 4:25 and may have reminded Mark of this summary statement immediately before the Sermon on the Mount. However one explains the phenomena, Mk 3:7–12 contains verbal parallels not only with Lk. 6:17–19 and Mt. 12:15–16, but also with Mt. 4:24–25.

While a significant proportion of Mk 3:7–12 has literary points of contact with Mt. 4:24–25 and with Lk. 6:17–19, both its beginning (Mk 3:7a) and ending (Mk 3:12) are paralleled only in Mt. 12:15–16. If Mark constructed this pericope from the Gospels of Matthew and Luke, he may be said to have used Mt. 12:15–16 as a skeleton, to which he added details from Mt. 4:24–25 and Lk. 6:17–19. On the two-gospel hypothesis the following require explanation: (1) Mark’s alternating agreement from Luke to Matthew at Mk 3:6–7, from Matthew back to Luke at Mk 3:13, and again from Luke to Matthew at Mk 3:20—all in the space of three pericopes; (2) Mark’s ‘double parallelism’ with Mt. 4:24–25 and Mt. 12:15–16; and (3) Mark’s expansion of the material in the parallel pericopes, particularly the unique reference to the sea in Mk 3:7 and Jesus’ instruction to have a boat ready for him in Mk 3:9. Mark’s alternating agreement between Matthew and Luke is not difficult to explain, nor is his ‘double parallelism’ with Mt. 4:24–25 and Mt. 12:15–16, but his expansion of the material is less easily accounted for. On the other hand, of particular significance from the two-gospel perspective is the presence in Mk 3:7–12 of words and possibly a phrase that are characteristic of Matthew and Luke, thereby suggestive of Mark’s dependence on Matthew and Luke.

Farmer’s original explanation of Mark’s compositional procedure at this point does not adequately account for Mark’s text:

[Mark’s] inversion of the call of the Twelve and the healing of the multitudes is explained by his conflation of Luke and Matthew (Lk. 6:1–19// Mt. 12:1–21). Matthew did not have the call of the Twelve at this point in his Gospel. Therefore, Mark simply conflated what was in Matthew with the Lucan parallels, omitting the call of the Twelve until he finished his expanded version of the healing of the multitudes (Mk 3:7–12// Lk. 6:17–19 // Mt. 12:15–21), after which he added the call of the Twelve that he had just omitted from Luke, and very largely in the wording of Luke (Mk 3:13–19//Lk. 6:12–16).

Farmer failed to explain the verbal parallels between Mk 3:7–12 and Mt. 4:24–25, which was remedied to some extent in the two-gospel team’s ‘Narrative Outline of the Markan Composition’:

At this point, Mk moves across to Mt. for the brief healing summary in Mt. 12:15–16, the same summary found a few verses later in Lk. 6:17ff. However, for his version he has provided a typical Mkan location: ‘by the sea’ ‘in the boat’ (cp. Mk 4:1) ‘on account of the crowds’ (cp. Mk 2:4). He has combined the list of cities and regions in Lk. 6:17 with others taken from Mt. 4:25 to produce an even more powerful impression of the extent of Jesus’ fame. But Mk omits (with Luke) the continuation of Mt’s summary (Mt. 12:17–21), a lengthy O.T. quote.

On the two-gospel hypothesis, it is legitimate to say that Mark combined details from Mt. 4:25 and Lk. 6:17, but it is inaccurate to describe Mt. 12:15–16 as the ‘same summary’ as Lk. 6:17–19. Nevertheless, there are similarities, notably the reference to crowds (Mt. 12:15b; Lk. 6:17, 19) and to Jesus healing (Mt. 12:15c; Lk. 6:18–19), which might suggest the possibility of conflation. It is also conceivable that Lk. 6:17 reminded Mark of the similar statement in Mt. 4:25, especially since Lk. 6:12, which Mark would have read first, could have reminded him of Mt. 5:1. In other words, Lk. 6:12 and 6:17 are reminiscent of the same Matthaean pericope, indeed, one that occurs at a conspicuous juncture in Matthew’s Gospel. On the Markan hypothesis, Matthew’s use of Mk 3:7–12 in Mt. 4:25 is an ‘anticipation’, whereas on the two-gospel hypothesis, Mark’s use of Mt. 4:25 at Mk 3:7–12 is a reminiscence of a passage bypassed at a crucial juncture in his own narrative but now incorporated because of similar material in his other source. On the face of it, the latter seems more plausible. But is such a redactional procedure supported by other evidence?

The presence of the verb ἀναχωρέω in Mk 3:7 may be evidence of Mark’s dependence on Matthew. While this verb occurs ten times in Matthew, once in Jn 6:15 (in a sense similar to its usual meaning for Matthew) and twice in Acts (23:19 and 26:31), it occurs nowhere in Luke and only here in Mark. J.C. Hawkins listed ἀναχωρέω as one of the most distinctive and important words characteristic of Matthew’s Gospel. Despite reservations about identifying individual words as characteristic of a single writer, Tevis suggested that each occurrence of this verb in Matthew’s Gospel is from a single hand.169 Using Farmer’s redactional criterion that the presence of a word or phrase characteristic of another gospel writer in a pericope that reveals evidence of a literary relation indicates dependence, the presence of ἀνεχώρησεν in Mk 3:7 suggests Mark’s dependence on Mt. 12:15, especially since the two occurrences are identical in form. Furthermore, Matthew explains his choice of the verb. Jesus withdraws or retreats from there (ἐκεῖθεν) because he learns of or intuits (γνοὺς) the plot of the Pharisees to destroy him (Mt. 12:14). Mark, on the other hand, has Jesus withdraw or retreat without indicating why he used this uncharacteristic term. In this connection, it is worth noting that the presence of συμβούλιον at Mk 3:6 partially parallels the characteristically Matthaean phrase, συμβούλιον + λαμβάνω.

In ‘The Verb ΑΝΑΧΩΡΕΩ in Matthew’s Gospel’, Deirdre Good suggests that the note of hostility towards Jesus and his subsequent withdrawal in Mk 3:6–7 may have been one of the sources for the seven occurrences of a threefold pattern of hostility-withdrawal-prophetic fulfilment in the Gospel of Matthew. Her analysis of Matthew’s ‘withdrawal’ motif within this threefold pattern is illuminating, as is her discussion of Matthew’s derivation of the pattern. For example, she draws attention to passages in the lxx and in 1 and 2 Maccabees containing the elements of hostility followed by withdrawal. Most noteworthy, given Matthew’s interest in Moses typology, is the lxx translation of Exod. 2:15, where in response to Pharaoh’s hostile intentions Moses ‘withdrew (ἀνεχώρησεν) from Pharaoh and dwelt in the land of Midian’. Good also suggests that ‘the theme of the withdrawal of Wisdom in apocalyptic sources is the more immediate background for the hostility/withdrawal/fulfilment of prophecy pattern in Matthew, especially when the subject of the pattern is Jesus’.175 However, her demonstration that the hostility-withdrawal-fulfilment of prophecy pattern occurs throughout the Gospel of Matthew—beginning at Mt. 2:12–15 (where ἀναχωρέω occurs three times in as many verses, as if to alert attention to a recurrent motif) and ending at Mt. 27:5–10—and her discussion of the wealth of resources from which Matthew might have derived this pattern reduce the likelihood that Mk 3:6–7 was one of them. Since ἀναχωρέω and the threefold pattern of hostility-withdrawal-prophetic fulfilment with which it is usually associated are so characteristic of Matthew’s Gospel, the presence of ἀναχωρέω at Mk 3:7 and the partial replication of Matthew’s threefold pattern in Mk 3:6–7 suggest instead Mark’s dependence on Matthew.

Mark 3:7 records not only that Jesus withdrew, but that he withdrew πρὸς τὴν θάλασσαν μετὰ τῶν μαθητῶν. The reference to the disciples is explicable on the grounds that they are referred to at Lk. 6:13 and again at Lk. 6:17, and also by the supposition that Mark planned to record the ‘making’ of the Twelve immediately following Mk 3:7–12. The reference to the sea as the place to which Jesus retreated is less easily explained, but if ἀνεχώρησεν in Mk 3:7 indicates Mark’s dependence on Mt. 12:15, πρὸς τήν θάλασσαν may simply be Mark’s interpretation of the place to which Jesus withdrew. However, the ‘sea’ is uniquely important for Mark, especially in the first half of his Gospel. The sea-crossings of Mk 4:35–8:21 are recognized to be significant, but for Mark the sea has a more general significance for Jesus and his disciples. His first two references to the sea, Mk 1:16–20 and 2:13–14, occur in connection with the call of Jesus’ earliest disciples. Only Mark has Jesus ‘go out again beside the sea’ before calling Levi (cf. Mt. 9:9–13; Lk. 5:27–32). At Mk 3:7 Jesus retreats with his disciples to the sea before going up a mountain to appoint the Twelve. Something similar is reported at Mk 6:30–34, where Jesus takes his disciples ‘on retreat’, so to speak, and they go by boat to an isolated place.

If the sea is the locus of disciple-calling and the setting both for ‘retreat’ with the disciples and for ‘revelatory’ events witnessed only by the disciples (Mk 4:35–41; 6:45–52; 8:13–21), it is also the setting for more general teaching and healing. Mark 2:13, almost wholly unique to Mark, notes that Jesus taught a crowd beside the sea. In agreement with Mt. 13:1–2, but not Lk. 8:4, Mk 4:1 designates the seaside as the setting for Jesus’ parabolic teaching. (While he agrees with Matthew, Mark emphasizes the seaside setting by referring to the sea three times in one verse!) In Mk 5:1–20 Jesus exorcizes Legion out of a demon-possessed man and permits them to enter a herd of pigs, after which the pigs stampede into the sea. Immediately after this incident, having returned to the other side (of the sea), Jesus attracts a large crowd, at which point Mark adds this unique observation: καὶ ἠν παρὰ τὴν θάλασσαν. Although Mk 6:53–56 does not stipulate the seaside as the setting for Jesus’ healing ministry at Gennesaret, the pericope begins by noting that Jesus was recognized as soon as he left the boat. In Mk 7:31–37 the Sea of Galilee is also the setting for the healing of the deaf and dumb man. Clearly, the sea is an important setting for much that takes place in the first half of Mark’s Gospel. Given Mark’s obvious predilection for the sea as a setting for ‘engagement’ between Jesus and his disciples and as a setting for Jesus’ teaching and healing ministry, whether one regards Mark as prior or posterior, the presence of πρὸς τὴν θάλασσαν at Mk 3:7 should not be surprising. It is an aspect of Mark’s peculiar contributions to the synoptic picture of Jesus’ ministry and probably has no source-critical significance.

If one can discount the source-critical significance of πρὸς τὴν θάλασσαν at Mk 3:7, what of the reference to the boat in Mk 3:9? Peabody connects this verse with Mk 4:1 because both share the motif of a boat being readied or used to protect Jesus from the crowd. According to Petersen, ‘The fact is that in 3:9 the boat is introduced in anticipation of a contingency—the crush of the crowd—which only becomes a reality in 4:1, and that from 4:1 through 8:26 the boat is in service of the distinctive topography of sea transit’. Like Mk 4:1, Mt. 13:1 has Jesus get into a boat because of the presence of large crowds. Luke 8:4, Luke’s parallel to Mt. 13:1 and Mk 4:1, makes no reference to the seaside setting of Jesus’ parabolic teaching and therefore does not refer to a boat. However, earlier in his narrative Luke employs the motif of Jesus getting into a boat to teach an overbearing crowd (Lk. 5:1–3). Having passed over Lk. 5:1–3 because the pericope as a whole is primarily concerned with the call of the first disciples, which Mark had recounted, Mark may have wished to include a similar statement when appropriate because it accentuates Jesus’ popularity, which Mark was concerned to emphasize. (Before Mk 3:7–12, see Mk 1:28, 33, 36–37, 45; 2:1b–2, 13b, 15b; verses in italics are unique to Mark.) However, if this was Mark’s reason for referring to a boat, perhaps a more likely place to have done so was Mk 2:13, first because this is Mark’s first reference to Jesus going to the seaside after passing over Lk. 5:1–11, and second, because the pericope beginning with this verse records the call of a disciple, thereby presenting Mark with the opportunity to duplicate the pattern of Lk. 5:1–11—Jesus teaches from a boat, after which he gathers disciples. (The parallel between Simon Peter, who in Lk. 5:8 confesses to being a ‘sinful man’, and Levi, who keeps company with ‘sinners’, would strengthen the pattern.) On the other hand, knowing that Mt. 13:1 has Jesus teaching from a boat may have influenced Mark to delay introducing this motif until the last opportunity to do so before paralleling Mt. 13:1, that is, Mk 3:7–12, although on this theory Mark must contrive the seaside setting to do so (cf. Mk 2:13!). In any case, on the two-gospel hypothesis, Mark’s reminiscence of Lk. 5:1–3 is a possible explanation for the reference to the boat in Mk 3:9. In this connection, it is noteworthy that while both Mt. 13:1 and Mk 4:1 refer to Jesus teaching from a boat because of a crowd, neither explicitly states that Jesus boarded a boat to avoid being overwhelmed by the crowd. On the other hand, despite using different verbs, both Lk. 5:1 and Mk 3:9 indicate that Jesus either boarded a boat to avoid being crushed (ἐπίκει—μαι, Lk. 5:1) by the crowd or ordered one to be ready in case he was crushed (θλίβω, Mk 3:9) by the crowd.

A final, albeit not weighty, consideration in favour of regarding Mk 3:9 as a reminiscence of Lk. 5:1–3 is that on the two-gospel hypothesis, Mk 3:11 also contains a reminiscence of Lk. 4:41a. On the Markan hypothesis, Luke must have anticipated this closely parallel phrase from Mk 3:11.

Further evidence suggestive of Mark’s posteriority is his use of the noun πλῆθος in Mk 3:7 and 3:8, in both cases with the adjective πολὺ. These are the only two occurrences of the noun in Mark’s Gospel. As in Mark, the noun appears twice in John, but it occurs eight times in Luke’s Gospel and sixteen times in Acts. Hawkins listed πλῆθος as a word characteristic of Luke, and Collison thought it a probable linguistic usage of Luke. Commenting on Mk 3:7–12, Vincent Taylor remarked, ‘Mark’s style can be prolix, but it is difficult to think … that he is responsible for πολὺ πλῆθος followed by πλῆθος πολὺ. Elsewhere he never uses πλῆθος … but always ὄχλος, and it may be that πλῆθος πολὺ in [verse] 8 is an assimilation to Lk. 6:17’. Whereas the presence of ἀναχωρέω in Mk 3:7 suggests Mark’s dependence on Matthew, the presence of πλῆθος in Mk 3:7 and 8 suggests Mark’s dependence on Luke. That this linguistic evidence occurs at a point where Mark is not only envisaged to be conflating details from Matthew and Luke but also making a transition from one source to the other, strengthens the two-gospel case.

Considered together with the above linguistic data, the phrase πέραν Ἰορδάνου in Mk 3:8 may also indicate Mark’s dependence on Matthew. This phrase occurs three times in Matthew (4:15, 25; 19:1), twice in Mark (3:8; 10:1) and three times in John (1:28; 3:26; 10:40). Hawkins did not list the phrase as characteristic of Matthew, but Tevis wrote:

This phrase consists of three elements which occur in Matt 4:15, 4:25, and 19:1. Except for synoptic parallels, these elements do not occur together elsewhere in the synoptic gospels. In Matt 4:15 this phrase is part of a quotation from Isaiah and appears in the text of the lxx. However, it seems likely that in Matt 4:25 and 19:1 this phrase comes from a single writer, and possibly from a writer who wanted to show that the prophecy of Isaiah had been fulfilled. Because this phrase is fairly well distributed throughout the Gospel and occurs in parts of the Gospel that function redactionally (Matt 4:23–25 and 19:1, 2 provide brief summaries of Jesus’ teaching and healing ministry), there is a strong possibility that this phrase comes from a redactor.

He also considered that the phrase ‘may provide some indication of the direction of literary dependence’.

Although Tevis explained that he used the phrase ‘fairly well distributed’ to mean that a word or phrase occurs in two pericopes that are not in the same chapter, to say that the phrase πέραν τοῦ Ἰορδάνου is ‘fairly well distributed throughout the Gospel’ is misleading because it occurs only at two junctures in the narrative—at the beginning and end of Jesus’ Galilaean ministry. Nevertheless, as Tevis observed, it occurs within pericopes that are redactional. In Patterns of Persuasion in the Gospels, Burton L. Mack and Vernon K. Robbins note that ‘when Matthew … decided to present Jesus’ Galilean ministry as a fulfilment of a prophecy of Isaiah, he composed a unit which is integrated through repetition of key words (Matt 4:12–17)’. The key words are ‘Galilee’ (4:12, 15), ‘sea’ (4:13, 15) and ‘Zebulun and Naphthali’ (4:13, 15), a repetitive pattern reinforced by the further repetition of ‘sitting in’ and ‘light’ in Mt. 4:16. What Mack and Robbins fail to point out, however, is that motifs from Mt. 4:12–16 recur in Mt. 4:17–25. Both the ‘sea’ of 4:13, 15 and ‘Galilee’ in 4:12, 15 reappear conjoined in Mt. 4:16 as the ‘Sea of Galilee’. Matthew 4:23 states that Jesus went throughout all Galilee, and Mt. 4:25 repeats both Galilee and the phrase πέραν τοῦ Ἰορδάνου.

According to Farmer, Mt. 4:12–25 reveals Matthew’s decision to use Isa. 9:1–2, which speaks of the light of salvation dawning on those sitting in darkness in Galilee, as ‘the structural backbone or organizing principle governing the theological and geographical framework of Jesus’ prepassion ministry’. There can be little doubt that Mt. 4:12–17 is programmatic in a way similar to Lk. 4:16–30. As Donald A. Hagner notes, commenting on Matthew’s use of Isa. 9:1–2 (8:23–9:1 in lxx), ‘This quotation serves as a rubric for the entire Galilean ministry of Jesus’. Given Matthew’s desire to document the story of Jesus as a fulfilment of Old Testament prophecies,191 it comes as no surprise that Matthew interprets Jesus’ decision to settle in Capernaum in terms of fulfilment. Farmer’s point, however, is that Matthew’s theological interpretation of Jesus’ Galilaean ministry underpins his geographical framework: ‘the major geographical movements outlined in the redactional framework of this Gospel (as given in 4:18, 23; 13:1; 15:29; and 19:1) are motivated primarily by theological rather than historical considerations’. He also maintained that before Matthew could be satisfied that this prophecy was completely fulfilled, Jesus had to journey πέραν τοῦ Ἰορδάνου, which he does at Mt. 19:1. Given the subtlety of Farmer’s argument, it is worth reproducing his own words. Noting that Mt. 4:12–16 interprets ὁδὸν θαλάσσης to refer to the Sea of Galilee rather than the Mediterranean Sea, as in Isa. 9:1–2, he then wrote:

The evangelist clearly intends this identification with the Sea of Galilee. Furthermore, after completing the ministry in Galilee proper and before this prophecy could be completely fulfilled, people sitting in darkness ‘across the Jordan’, that is, in the land east of the Jordan River (Perea), had to see the light of salvation. Therefore, in this Gospel it was necessary for Jesus to go into the region across the Jordan (19:1) and for people there also to be healed in great numbers, before he proceeded to Jerusalem for the climax of his work. That the geographical framework of this section of Matthew’s Gospel (that is, 19:1–20:16) is motivated primarily if not solely by the literary-theological considerations suggested by this analysis, is confirmed by the absence of any reference to location in the tradition it uses.

On the basis of this understanding of the significance of the phrase πέραν τοῦ Ἰορδάνου in Matthew’s Gospel, Farmer argued that its presence in Mark is indicative of Mark’s dependence on Matthew rather than vice versa. Again, his argumentation is intricate:

The Gospel of Mark presupposes this same geographical outline without giving the fulfillment passage from Isaiah, which provides the basis. To reverse the relationship, making Matthew dependent upon Mark, requires one to believe in a redactional procedure without precedent. According to this view, Matthew has perceived that Mark’s geographical outline originated in a unique interpretation of Isa. 9:1–2. Matthew has decided to supply the passage and to elaborate upon it in 4:12–14 by offering, through an equally unique exegesis, a theological motivation for Jesus’ leaving Nazareth and dwelling in Capernaum.

Farmer’s interpretation is suggestive, but two weaknesses reduce the probity of his argument for Mark’s dependence on Matthew based on this literary evidence. First, his interpretation of the phrase πέραν τοῦ Ἰορδάνου is by no means certain. He maintained that Matthew thought it necessary for Jesus to travel to ‘the land east of the Jordan River (Perea)’. It is debatable, however, whether this is meant at Mt. 19:1, which states that Jesus left Galilee and went into the region of Judaea (εἰς τὰ ὅρια τῆς Ἰουδαίας πέραν τοῦ Ἰορδάνου). The phrase πέραν τοῦ Ἰορδάνου is adjectival here rather than nominal, as in Mt. 4:25 and Mk 10:1, and may therefore reflect a provenance for the composition of Matthew’s Gospel east of the Jordan, as H. Dixon Slingerland and Gerd Theissen contend. If so, the presence of the phrase πέραν τοῦ·̓ ὀρδάνου at Mt. 4:15 and 19:1 is not indicative of a theological-structural principle such as Farmer proposed.

In addition, even if Farmer’s understanding of the significance of πέραν τοῦ Ἰορδάνου at Mt. 4:15 is correct, he ignores the presence of the phrase πέραν τοῦ Ἰορδάνου in Mt. 4:25. On his interpretation, it is as likely that Matthew regarded the Isaiah passage as being fully fulfilled by having crowds from ‘beyond the Jordan’ follow Jesus as by having Jesus travel ‘beyond the Jordan’. It is by no means certain that for Matthew, Jesus had to go ‘beyond the Jordan’ before those ‘beyond the Jordan’ saw the light of salvation, because those ‘beyond the Jordan’ see the light by following Jesus while he is in Galilee. Again, therefore, the presence of πέραν τοῦ Ἰορδάνου at Mt. 19:1 may not have the structural significance that Farmer attributed to it, which in turn reduces the plausibility of his suggestion that Mark presupposes Matthew’s geographical outline.

In short, Farmer expected an ambiguous phrase to bear too much source-critical weight. Nevertheless, the presence of πέραν τοῦ Ἰορδάνου at Mk 3:8 is suggestive of Mark’s dependence on Mt. 4:25, where the phrase is meaningful because of the earlier reference in Mt. 4:15, especially when considered alongside other literary data in this pericope indicative of Markan posteriority.

Mark 6:1–13, Mark’s Alternating Agreement with Matthew and Luke

This passage precedes a short segment of triple-tradition material that occurs in the same relative order (Mt. 14:1–21; Mk 6:14–44; Lk. 9:7–17), after which Matthew and Mark agree in sequence for a long stretch until just before another block of context-parallel triple-tradition material (Mt. 16:13ff.; Mk 8:27ff.; Lk. 9:18ff.). Leading up to Mk 6:1–13, Mark’s parallels with Matthew and Luke create an interesting pattern of agreement and disagreement in order: Mk 4:1–20 agrees in order with both Matthew and Luke; Mk 4:21–25 agrees in order only with Luke; Mk 4:26–29 is unique; Mk 4:30–34 agrees in relative order with Mt. 13:31–35, relative because Mark has no parallel to the parable of the leaven in Mt. 13:33; Mk 4:35–5:43 agrees in order with Lk. 8:22–56; immediately thereafter, Mk 6:1–6 agrees in relative order with Mt. 13:53–58, whereas Mk 6:7–13 agrees in order with Lk. 9:1–6. On the two-gospel hypothesis, Mk 6:1–13 concludes a section in which Mark frequently alternated between his major sources.

What does this pattern of agreement and disagreement in order imply about Mark’s compositional procedure in this section of his Gospel? Beginning with Mk 4:1, he was evidently not concerned to relate every parable available to him (cf. Mk 4:33), but he did incorporate a mix of parables from both Matthew and Luke. In addition to the parable of the sower and its interpretation shared between Matthew and Luke, he borrowed the parables of the lamp and ‘measure for measure’ from Luke, the parable of the mustard seed from Matthew and the parable of the seed’s mysterious growth from his own special material. He then bypassed the material in Mt. 13:36–52 as well as Luke’s account of Jesus’ instruction about his true family (Lk. 8:19–21), the latter because he had already recounted this incident while following Matthew’s order at Mk 3:31–35. Having chosen not to include the remainder of Matthew’s parable collection, he could have continued with Matthew’s account of Jesus’ rejection in Nazareth (Mt. 13:53–58). However, before relating this event he incorporated three episodes from Lk. 8:22–56, interrupted his agreement with Luke’s order momentarily by switching sources for the rejection story, then returned to Luke for the mission of the Twelve. What explains this compositional procedure?

I have argued that Griesbach’s original explanation of Mark’s composition at this point is inadequate and suggested other reasons for his choice and organization of material. But how do neo-Griesbachians explain Mark’s alternation between sources here, particularly the two abrupt transitions from Luke to Matthew at Mk 6:1 and from Matthew back to Luke at Mk 6:7? In The Synoptic Problem, Farmer’s discussion focused on the positive correlation between agreement in order and agreement in wording. He pointed out that in terms of vocabulary, Mk 4:35–5:43 is significantly closer to Luke’s contextually parallel pericopes than to Matthew’s parallels, which occur in a different context. In stark contrast, however, Mk 6:1–6 is much closer in wording to Mt. 13:53–58 than to Lk. 4:16–30, which occurs in a vastly different context. Farmer noted that Mark agrees with Luke that the rejection incident occurred on the Sabbath, but he emphasized Mark’s agreement with the relative order of Matthew at this point and the close verbal parallels between Mk 6:1–6 and Mt. 13:53–58. For Farmer,

This is strikingly in keeping with a pattern of alternation in agreement of wording, where Mark agrees closely now with one of his sources and then suddenly just as closely with the other; an alternation which corresponds positively with an alternation in agreement in order occurring at exactly the same place where the agreement in wording shifts.

As for Mk 6:7–13, Farmer noted that after interrupting his allegiance to Luke’s order at Mk 6:1, Mark now resumes following Luke where he had just abandoned Luke’s sequence. In his view,

The influence on Mark of Matthew’s wording in this passage may be seen especially in verses 7, 8, 10 and 11. But on the whole, Mark’s text is conformed to the shape of Luke’s shorter version of the incident, including, for example, among many omissions, that of the admonition of Jesus to his disciples not to go into the way of the Gentiles, nor any city of the Samaritans, but to the lost sheep of the house of Israel (Mt. 10:5–6).

Surprisingly, Farmer failed to mention the various minor agreements between Matthew and Luke against Mark, three of which are significant.

Farmer’s observations are certainly consistent with his theory of Mark’s dependence on Matthew and Luke. However, they are also consistent with Markan priority. As I sought to demonstrate in Chapter 4, even at the formal level the phenomenon of correlation does not strongly favour the two-gospel over the Markan hypothesis.

The two-gospel team’s discussion of the two pericopes that comprise Mk 6:1–13 is imprecise and conjectural. Indeed, the team does little more than to make explicit what is implied by its hypothesis. Concerning Mk 6:1–6a, the two-gospel team writes:

At this point in Lk’s outline, Mk comes to the story of the Sending Out of the Twelve. Mk/Peter sees this event as the beginning of the most glorious period in Jesus’ Galilee ministry. In his ironic, contrapuntal style, however, Mk first interposes Mt’s account of the contempt and unbelief still prevailing, despite everything, in Jesus’ home town. Interrupting his use of Lk, Mk moves over to Mt. to the end of the Parable discourse (13:52) and relates the sad story of the hardness of heart in Jesus’ hometown.

This is the second occasion in which members of the two-gospel team appeal to Mark’s association with Peter to explain details in Mark’s text. The first is their comment that in Mk 4:5–41 ‘Mk retells the story with the vivid eyewitness verisimilitude of Peter’s memory’. Nowhere do they explain their grounds for appealing to this alleged connection between Mark and Peter, although one surmises that it stems from a sense of responsibility to take seriously the early external evidence that Mark recorded Peter’s preaching faithfully.206 Nor do they discuss what features of Mark’s text can plausibly be attributed to his knowledge of tradition associated with Peter, although in practice this subsidiary hypothesis serves to ‘explain’ those features of Mark’s Gospel that cannot have derived from the texts of Matthew or Luke. While Mark may have been associated with Peter, the two-gospel hypothesis is not strengthened by having to resort to such an association to explain Mark’s text. Apart from the inevitable question why Mark would rely on two written sources if he was familiar with first-hand reminiscences of one of Jesus’ disciples, even if one grants the credibility of the patristic evidence concerning the Mark-Peter connection we cannot know what to attribute to this association. We do have the Gospels of Matthew and Luke, however, and if Mark was dependent on them one must focus on those features of his text that can plausibly be attributed to his redaction of Matthaean and Lukan materials. The two-gospel team fails to do this when commenting on Mk 6:1–6a.

Its remarks on Mk 6:6b–13 are little better:

Mk moves back to Lk. for this (and the next) story, working in details from Mt. and adding features of his own (e.g., vs. 8). It is a striking example of a Mkan ‘zigzag’. Noteworthy is Mk’s version’s exclusive mention of the instructions to the Twelve to cast out demons, omitting any proclamation of the Kingdom of God mentioned in both Mt. and Lk. Mk/Peter may have considered it a potentially misleading concept to his Roman patrons.

It is difficult to know what to make of the comment about Mark’s ‘exclusive mention of the instructions to the Twelve to cast out demons’. Both Mt. 10:1 (cf. 10:7) and Lk. 9:1 say that Jesus gave the Twelve authority over demons, although Mark agrees with Matthew against Luke by referring to ‘unclean spirits’ rather than ‘demons’. The point must be that within the context of Jesus’ instructions to cast out demons, Mark alone omits any reference to the proclamation of God’s kingdom. Even so, one is not told why this is noteworthy, apart from the speculation about it being potentially misleading to Roman patrons. More plausible suggestions are required if Mark’s alleged redaction of Matthew’s and Luke’s materials is to gain credibility.

Because the two-gospel team’s source-critical comments are so unrewarding, one hopes for more from Mann and Riley. However, Mann’s comments on Mk 6:1–6a merely echo Farmer’s remarks on this pericope:

Mark here takes up the Matthean order, even following the Matthean wording closely (cf. Matt 13:53–58), whereas the Lucan version of the rejection comes much earlier in the Lucan narrative (Luke 4:16–24). But Mark is indebted to Luke for associating the incident with the Sabbath. As an example of the Markan conflation method, it is interesting to observe the agreement of Mark with Matthew in wording here, as compared with Mark’s dependence on Lucan wording in 4:35–5:43 (cf. Luke 8:22–56). In that last instance Mark had to deal with Matthean parallels which were in an entirely different order.

Mann goes beyond Farmer, however, in his notes on Mk 6:1–6a. He points out that Mark uses ἐκε͂ιθεν as a ‘connecting link with the previous narrative’ only three times, as opposed to eleven times in Matthew, thereby indicating Mark’s dependence on Mt. 13:53 at Mk 6:1. Matthew actually uses the adverb ἐκε͂ιθεν twelve times, ten of which clearly serve as a ‘connecting link with the previous narrative’. Whether it is used in the same way at Mt. 19:15 is less clear, although Mann obviously thinks so. On the other hand, ἐκε͂ιθεν occurs six times in Mark’s Gospel, and on four occasions, not three as Mann says, it connects with what goes before in the narrative. Hawkins listed ‘ἐκε͂ιθεν in narrative’ as characteristic of Matthew, but Tevis did not. As a result, perhaps Mann attaches too much importance to ἐκε͂ιθεν at Mk 6:1.

Given the relatively few times that Mark used ἐκε͂ιθεν, it is noteworthy that of the six Markan occurrences, it appears three times in Mk 6:1–13. As noted, it occurs at Mk 6:1 in parallel with Mt. 13:53, where Matthew uses it alongside his distinctive formula καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησους … It recurs at Mk 6:10, where it parallels the first of Luke’s four uses of the term, and again at Mk 6:11, where it is unparalleled in Matthew or Luke and marks an important agreement between Matthew and Luke against Mark. On the Markan hypothesis, Matthew and Luke independently edited ἐκπορευόμενοι ἐκε͂ιθεν at Mk 6:11 to read ἐξερχόμενοι ἔξω … (Mt.) / ἀπὸ (Lk.) τῆς πόλεως ἐκείνης … This is not particularly damaging to the Markan hypothesis, but when considered alongside other agreements between Matthew and Luke against Mark in this pericope, it deepens one’s suspicion about the alleged independence of Matthew and Luke.

Returning to Mk 6:1–6a, Mann comments on two features of Mark’s text that have been regarded as evidence that Matthew and Luke represent more developed versions of the tradition because they edited out aspects of Mark’s account that might be considered ’embarrassing’. The two allegedly ‘offensive’ features are the designation of Jesus as ὁ υἱὸς τῆς Μαρίας in Mk 6:3 and the narrative aside in Mk 6:5 that Jesus was unable to perform any work of power.

Concerning ὁ υἱὸς τῆς Μαρίας, Mann acknowledges that ‘to describe a man as the “son of” his mother, even when the father was deceased, is often a usage of insult’. However, he goes on to say, ‘But to suggest that Matthew’s version represents an amendment of Mark in the interests of reverence … is to ignore the evangelist’s use of insults on other occasions (cf. Matt 11:19, 12:24)’. On the other hand, Riley considers that Mark altered Matthew’s ὁ τοῦ τέκτονος υἱός to ὁ τέκτων because to have followed Matthew’s wording in the absence of any account of the virginal conception of Jesus in his own account might be misleading. In his view, ‘Mark could avoid difficulty by writing the carpenter, and for Matthew’s “Is not his mother called Mary?” could find it natural to put the son of Mary in apposition’. These comments are reminiscent of Farmer’s suggestion that Mark’s use of ‘Son of Mary’ may have been influenced by a developing doctrine of the virgin birth.217 Against any such suggestion, Mann declares, ‘We would be unjustified in seeing in the Markan phrase any implication of knowledge of a tradition of virgin birth’. It should be noted that Farmer’s suggestion was precisely to illustrate his primary point that arguments based on the theology or Christology of a given pericope are inconclusive. This is surely correct, even if Sherman Johnson is right to say that ‘it is proper to take theological development, or at least theological variety, into account in studying the Synoptic Problem’. That Johnson qualified his use of the phrase ‘theological development’ in the way he did is a concession to its slipperiness as a source-critical criterion.

As for the comment in Mk 6:5 that Jesus was unable to do any work of power, Mann’s remarks contain no argument in defence of Mark’s dependence on Matthew. Indeed, his observation that ‘Mark’s version of the astonishment of Jesus reads very much like a tradition handed on by someone intimately associated with him’ reflects his view that in addition to the Gospels of Matthew and Luke, Mark also incorporated reminiscences of Peter.221 As argued above, however, the plausibility of Mark’s dependence on Matthew and/or Luke is diminished, not strengthened, by having to appeal to Mark’s reliance on Petrine tradition.

Riley disputes the idea, commonly advanced by Markan priorists, that for christological reasons Mt. 13:58 mitigates the statement in Mk 6:5 that Jesus was unable to do any work of power. As he observes, ‘Both Matthew and Mark give the people’s unbelief as the reason for the few healings. Matthew says this quite simply; Mark gets off on the wrong foot by saying no mighty works, and then has to correct this by adding the exception of a few sick people“. In fact, because of the double negative, οὑκ ἐδύνατο ἐκε͂ι ποιῆσαι οὑδεμίαν δύναμιν, the sense of contradiction in Mk 6:5 is even greater than Riley indicates. Mark first stresses that Jesus was unable to do one single work of power, but then softens his own statement by adding the exceptive clause. One could argue that Mark initially misread Matthew, an easy thing to do when reading scriptio continua, and subsequently toned down what he had just written by adding that Jesus did manage to heal a few people. Few would be convinced by such an argument, however, and rightly so, although the equally insipid argument for Markan priority seems to have been widely accepted.224

Turning to Mk 6:6b–13, both Riley and Mann comment on the striking agreement between Matthew and Luke against Mark in recording Jesus’ prohibition of a staff. According to Riley,

Matthew and Luke agree in saying that the disciples should have no staff, and Matthew also says ‘no sandals’ (cf. Lk. 10:4); since Mark was not to include their prohibition, there was no necessity for him to mention them, just as Luke has no mention here of sandals. But the reference to them in his sources prompts him to do so: nothing for their journey except a staff (6:8) and sandals (6:9).

He also recalls Vincent Taylor’s judgment that ‘the more rigid prohibitions of Mt. and Lk. are doubtless more original’. As for Mann,

The mitigation of the severe restrictions in Matthew and Luke is note-worthy. The staff is explicitly prohibited in Matthew and Luke and is permitted in Mark. Similarly they might wear sandals is a departure from the texts Mark had before him. Possibly the mitigations have to do with the perils of the situation for which this gospel was written.

On any theory, these differences between Mark and the other two synoptics are difficult to explain. As Bo Reicke remarked in connection with Mk 6:6b–13 and parallels, ‘Attempts to explain this confusing picture of similarities and varieties by means of literary sources lead to precarious results’. Similarly, for Davies and Allison, ‘No synoptic theory can readily explain the similarities and differences exhibited by Mt. 10:9f. par. In the end one may do well to allow some role for the fluidity of oral tradition and the frailty of human memory’. Markan priorists often resort to the influence of Q to explain Matthew’s and Luke’s disagreement with Mark here,230 but, as Michael Goulder points out, Luke fails to mention the staff at Lk. 10:4, which ex hypothesi represents the Q-version of the mission story. ‘If then [the staff] was forbidden in Q … why has Luke omitted it in his version of the Q discourse? If it was not in Q, but has been assimilated to the things forbidden in Mt. 10:10 … what is to be done with the MA [minor agreement]?’

As for advocates of Markan posteriority, none of their suggestions about Mark’s alleged use of Matthew and Luke is particularly strong, especially Mark’s contradiction of Matthew’s prohibition of both staff and sandals. Mann’s suggestion that Mark’s mitigation of Matthew’s prohibitions may reflect the situation for which his Gospel was written may well be correct, but it is unverifiable and therefore unlikely to convince critics of Mark’s dependence on Matthew and Luke. It is also worth noting that here Mark does not act in a way consistent with the neo-Griesbachian claim that his Gospel was intended to mediate in some sense between the Gospels of Matthew and Luke. For example, according to Farmer,

Mark is basically a self-consistent version of Matthew and Luke. The evangelist was aided in his purpose by following Luke where Luke followed Matthew, added nothing from Matthew that conflicted with Luke, and nothing from Luke that conflicted with Matthew, thus producing a gospel free from contradictions with either.

Here, however, Mark contradicts both Matthew and Luke!

Finally, in ‘Repeated Language in Matthew: Clues to the Order and Composition of Luke and Mark’, Peabody asserts that the vocabulary of Mk 6:7 indicates dependence on Mt. 9:35:

The use of καὶ περιῆγεν … διδάσκων at Mark 6:7/Matt 9:35 (cf. Matt 4:23) is also evidence of the secondary nature of Mark to Matthew … These words are clearly traceable to a formula only found in the Gospel of Matthew … The words καὶ περιῆγεν in fact the root verb, περιάγω itself, is found nowhere in the Markan Gospel outside of Mark 6:7/Matt 9:35/Matt 4:23. The presence of the long redactional formula in the Gospel of Matthew at Matt 4:23 and 9:35 is due to some preference on the part of the redactor(s) of Matthew, probably the author, while its presence in the Gospel of Mark may readily be accounted for by copying. It constitutes evidence at the compositional level of Mark’s literary dependence upon Matthew.

Having stated the case for Markan dependence on Matthew at Mk 6:7, Peabody then compares what must be assumed about Matthew’s compositional procedure on the Markan hypothesis. If Matthew’s use of περιάγω must be explained by dependence on the text of Mark, Matthew must have anticipated its appearance in Mk 6:7 at Mt. 4:23 and constructed his redactional phrase καὶ περιῆγεν … διδάσκων around this sole occurrence of the term. This is Neirynck’s explanation for the presence of καὶ περιῆγεν … διδάσκων at Mt. 4:23, an explanation that, while conceivable, is inconsistent with compositional conventions of the first century ce.

However, it is debatable whether source-critical significance can be attached to this phrase. In Peabody’s favour, despite the fact that the phrase καὶ περιῆγεν … διδάσκων occurs only twice in Matthew’s Gospel, Mt. 4:23 and Mt. 9:35 are redactional summaries that can probably be attributed to the final redactor(s) or author. Therefore, he is right to say that the presence of this redactional formula is due to ‘some preference on the part of the redactor(s) of Matthew, probably the author, while its presence in the Gospel of Mark may readily be accounted for by copying’. On the other hand, advocates of Markan priority can reply that when they envisage Matthew constructing redactional summaries from Markan phrases in different contexts, this is not the result of a mechanical process of conflating words and phrases from disparate sections of Mark’s Gospel, but the natural result of close familiarity with Mark. Furthermore, it is conceivable that Matthew repeated words or phrases that occurred only once in his sources. As Davies and Allison ask in connection with another phrase in Mt. 9:35, ἐν ταῖς συναγωγαῖς αὑτῶν: ‘Why can we not imagine that the First Evangelist repeated words or phrases found only once in his sources? Why must characteristically Matthean vocabulary be unparalleled in the other gospels?’ Although Peabody’s argument is not conclusive, it should be kept in mind alongside other Matthaean redactional features that occur in Mark’s Gospel.

In conclusion, although Mk 6:1–13 does not contain overwhelming evidence of Markan dependence on Matthew, neither does it contain overwhelming evidence of Matthew’s or Luke’s dependence on Mark. What Davies and Allison say about Mt. 10:5–25 and parallels is also applicable to Mk 6:1–13 as a whole. Although they argue that Mt. 10:5–25 is a conflation of materials from Mark and Q, they also admit:

Mt. 10:5–25 is one of the many reasons the synoptic problem is in fact a problem. Mk 6:8–11 as well as Lk. 9:2–5 and 10:3–16 could, without too much difficulty, be explained as abbreviations of Matthew’s text; for a desire to abbreviate, to eliminate narrow Jewish concerns, and to omit irrelevant instructions could have led to what we now have in the Second and Third Gospels. Furthermore, Luke, followed by Mark (as on the Griesbach hypothesis), might have found most of 10:17–25 to be more appropriate for an eschatological discourse (thus elucidating why the parallels to Mt. 10:17–25 appearmostly in Mk 13 and Lk. 21). In our estimation, Mt. 10 and its synoptic counterparts do not count against Griesbach.

Beyond this, however, Mk 6:1–13 and parallels contain significant evidence against the independence of Matthew and Luke. In addition to the agreements against Mark in Mt. 10:10 // Lk. 9:3 (forbidding a staff) and Mt. 10:14 // Lk. 9:5 (ἐξερχόμενοι ἔξω … (Mt.) / ἀπὸ (Lk.) τῆς πόλεως ἐκείνης …), there is the further agreement at Mt. 10:1 // Lk. 9:1 against Mk 6:7. Following Goulder, Gundry has argued that this is evidence of Lukan dependence on Matthew:

Lk. 9:1 agrees with Mt. 10:1 against Mk 6:7 in adding a reference to healing: καὶ νόσους θεραπεύειν. As pointed out by Goulder, the healing of diseases rather than people distinguishes Mt. 4:23; 9:35 against their parallels Mk 1:39; 6:6b; Lk. 4:44; 8:1 and against all other thirty-seven occurrences of θεραπεύω in the Gospels and Acts, including eighteen occurrences in Luke—Acts—except for Lk. 9:1, which therefore agrees in this respect not only with Matthew against Mark but also with Matthew against all other Lucan usage.

Evidence of Luke’s familiarity with Matthew—whether in addition to Mark and Q (Gundry) or instead of Q (Goulder)—is not irrelevant for the two-gospel hypothesis. As E.W. Lummis noted,

When once it is conceded, even hypothetically, that Lk. was acquainted with Mt. as well as with Mk, the assumption that any notice which is common to all three gospels was derived, by Lk. and Mt., from Mk no longer holds good. It may be that Lk. took the passage from Mt., and that Mk is secondary to Mt., or to Lk., or to both Mt. and Lk. (Here the arguments of Griesbach, and those of the Tübingen critics, may well be reconsidered.)

In short, while Mk 6:1–13 and its synoptic parallels may not contain overwhelming evidence of Markan dependence on Matthew and Luke, the two-gospel hypothesis accounts for this pericope as well as—if not better than—the traditional two-source theory.


In ‘Modern Developments of Griesbach’s Hypothesis’, Farmer wrote, ‘The great merit of the Griesbach hypothesis was and remains that it offers a credible explanation for the phenomenon of order—lacking in the alternative accounts’. He also affirmed:

The force of the argument from order for the Griesbach hypothesis cannot be fully appreciated when it is stated abstractly. But when it is considered in relation to a careful examination of the text of the gospels, as for example in Griesbach’s original Demonstration of 1790, and with reference to the internal structure of given pericopes, as in Longstaff’s dissertation, it becomes a persuasive argument …

In response to these claims Tuckett countered:

Farmer’s argument for the validity of the GH based on the ordering of material in Mark needs a great deal more demonstration if it is to be made convincing … The GH does offer an explanation of the facts at one level, in that it can point to the way that Mark must have followed his two sources alternately. But this observation needs to be supplemented with a coherent set of reasons for Mark’s having proceeded in the way he is alleged to have done.

Since Tuckett made these remarks, various advocates of Mark’s dependence on Matthew and Luke have tried to provide the kind of detailed demonstration for which he called. To date, however, attempts to explain Mark’s compositional procedure on the two-gospel hypothesis have been too general to be convincing. In part, this may be because Mark’s compositional procedure on the two-gospel hypothesis is in one sense less complicated than either Matthew’s or Luke’s compositional procedure on the Markan hypothesis. Assuming Markan priority, one must explain Matthew’s and Luke’s disagreements in order as transpositions or rearrangements of the order of a principal source; assuming Markan posteriority, however, Mark’s disagreement with the order of one of his two main sources can be explained simply as the result of his agreement with the order of his other principal source. Nevertheless, Tuckett was right to demand more detailed redaction-critical demonstrations of Mark’s compositional procedure at each point where he is alleged to have ceased following the order of one of his sources to follow the order of the other.

Regarding Mark’s alleged transition from Matthew to Luke at Mk 1:21, there is no major impediment to seeing Mark act as the two-gospel hypothesis envisages. Mark ceased following the order of Matthew after Mt. 4:18–22 and immediately before Matthew’s introduction to his Sermon on the Mount so as to incorporate Luke’s Capernaum sequence in Lk. 4:31–44, which served his purpose of illustrating the authority of Jesus in word and deed. Matthew 4:23–9:35 also demonstrates the authority of Jesus in word and deed, but takes much longer to do so. At this point, the two-gospel hypothesis does not explain the texts of the gospels better than the Markan hypothesis; however, Mark’s dependence on Matthew and Luke explains the respective texts as well as Matthew’s and Luke’s dependence on Mark.

In Mk 3:7–12, however, Mark’s dependence on Matthew and Luke provides a better explanation of the synoptic parallels than vice versa because this summary pericope contains words and a phrase that are characteristic of Matthew or Luke but appear in Mark only at this point. While it is possible to explain Matthew’s and Luke’s parallels as derivative from Mk 3:7–12, the reverse seems more plausible.

In addition, on the two-gospel hypothesis, Mark’s first two major transitions at Mk 1:21–22 and 3:7–12 can be understood to have been motivated by the same redactional concern, namely, to omit the lengthy blocks of teaching in Mt. 5:3–7:27 and Lk. 6:20–49, which are similar enough to have been construed as two versions of the same sermon.

Mark 6:1–13 does not contain words or phrases strongly suggestive of Markan dependence on Matthew and/or Luke, but there is no literary data that cannot be satisfactorily explained by the two-gospel hypothesis. In addition, there is significant evidence in Matthew’s and Luke’s parallels to Mk 6:1–13 suggestive of Luke’s dependence on Matthew.


Most of the following conclusions reiterate those reached in the preceding chapters. However, my final conclusion seeks to identify what can legitimately be deduced from the evaluation of compositional arguments from order in Part III.

1.    The phenomenon of order is an ambiguous literary datum because the pattern of agreement and disagreement in the order and arrangement of pericopes depends on how one subdivides the gospel materials into pericopes and what criteria one uses to determine parallelism between pericopes. Thus, in any analysis of the phenomenon of order, it is vitally important to recognize that there is no ‘right’ or objective arrangement of parallel pericopes. Consequently, critics who use arguments based on the phenomenon of order should advise what arrangement of parallel pericopes they presuppose in their arguments and what criteria they have used to determine parallelism between pericopes.

2.    No matter how ambiguous the task of subdividing the gospel narratives into pericopes is, pericope divisions should be made on the basis of what can be known about literary and rhetorical forms in antiquity. Research on the chreia as a literary form in antiquity is important in this respect. Likewise, recent work on narrative poetics is helpful for segmenting the gospels into smaller literary units. However, it is doubtful whether full agreement can ever be reached on this issue because the gospel writers varied so much in how they subdivided or combined smaller units of narrative.

3.    With regard to the issue of parallelism between pericopes, while verbal correspondence is an important criterion, narrative function is also important. For this reason, pericopes or partial pericopes such as Lk. 3:19–20, 4:16–30 and 5:1–11 should be regarded as ‘functional parallels’ and treated as disagreements in order.

4.    When arranging pericopes in parallel, a crucial judgment that must be made concerns the placement of Matthew’s Sermon on the Mount relative to Mark’s and Luke’s pericope orders. One can appeal to various criteria for making this decision, but none settles the question decisively. As a result, since arguments can be made for different schemes of parallelization, it is important to acknowledge that different arrangements of parallel pericopes display substantially different patterns of agreement and disagreement in narrative order, including differences in the number of pericopes adjudged to disagree in relative order. In addition, different arrangements of parallel pericopes display different pericopes as disagreements in order. The synoptic pattern of agreement and disagreement in order is not static; rather, it changes according to which configuration of parallel pericopes one adopts. This has significant implications for any argument from order.

5.    Assuming direct literary relations between the synoptic gospels, Butler was correct to claim that the absence of agreement in order between Matthew and Luke against Mark is not evidence of Markan priority. However, Tuckett is also correct that the absence of agreement in order between Matthew and Luke against Mark is compatible with the theory of Markan priority, even if not conclusive evidence for it. My research indicates that at the formal level, it is impossible to discriminate between competing source theories on the basis of the phenomenon of order. This applies not only to the absence of agreement in order between Matthew and Luke against Mark, but also to the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke (the so-called ‘Markan zigzag’). Advocates of the two-gospel hypothesis have argued that the pattern of alternating agreement in order between Matthew-Mark and Mark-Luke is inexplicable on the theory of Matthew’s and Luke’s independent use of Mark, but this is unsustainable. In the same way that advocates of Markan priority have ceased to appeal to the absence of agreement in order between Matthew and Luke against Mark as evidence for their source theory, advocates of Markan posteriority should discontinue appealing to the pattern of alternating agreement in order as evidence for their source theory. All formal arguments based on the phenomenon of order are inconclusive.

6.    Various difficulties hinder one’s attempt to verify the alleged positive correlation between agreement in order and agreement in wording in Matthew and Mark on one hand and in Mark and Luke on the other, including the preliminary difficulty of determining which pericopes are ‘out of order’. Even after making allowances for these difficulties, both my examination of several Markan pericopes and their parallels in the Gospels of Matthew and Luke and my statistical tests of the ‘phenomenon of correlation’ were inconclusive. The ‘phenomenon of correlation’ does not seem to favour the two-gospel hypothesis over the Markan hypothesis.

7.    Although formal arguments from order are inconclusive and therefore do not help to discriminate between competing source theories, this does not imply that the phenomenon of order is insignificant for resolving the synoptic problem. However, it does imply that the phenomenon of order can only be significant if one uses another mode of argumentation. In Part III, I attempted to evaluate, on compositional or redactional grounds, which is more credible—the independent rearrangement of Mark’s order by Matthew and Luke (the Markan hypothesis) or Mark’s alternating dependence on the Gospels of Matthew and Luke (the two-gospel hypothesis). Although this might seem a relatively straightforward task, it is complicated by the fact that these two theories differ in their assumptions about the compositional procedures of the secondary evangelists, which allegedly account for the disagreements in order between the synoptic gospels. In short, the kind of reasons given on the Markan hypothesis to explain Matthew’s and Luke’s transpositions may well be different from those given on the two-gospel hypothesis to explain Mark’s transitions between the Gospels of Matthew and Luke.

8.    My analysis of Lk. 3:19–20, 4:16–30 and 5:1–11 indicates that Luke’s Gospel is secondary. In each of these pericopes, Luke’s account contains evidence that he rearranged a traditional order of events that is at least reflected in the Gospels of Matthew and Mark. Beyond this, however, it is difficult to argue on the basis of these three pericopes whether Luke is dependent on Mark alone, on Matthew alone or on both Matthew and Mark, although in the case of Lk. 4:16–30 Luke’s dependence on Matthew (whether alone or in conjunction with Mark) seems more plausible than Luke’s dependence solely on Mark.

9.    My discussion of Mt. 4:23–5:2, 8:1–4, 8:18–34 and 9:18–26 indicates that in general terms, Matthew’s disagreements in order within the section Mt. 4:23–9:35 can be explained fairly coherently as transpositions of Markan pericopes to suit his narrative plan. However, when one examines the relevant pericopes and their parallels in detail, it is more difficult to demonstrate that Matthew’s disagreements in order are rearrangements of Mark’s narrative sequence. Influential discussions defending Matthew’s alleged transpositions of Markan materials as part of his compositional plan often attribute anachronistic redactional procedures to Matthew or rely on suspect source-critical criteria. Furthermore, details within these pericopes and their respective parallels are usually ambiguous from a source-critical viewpoint. In each case, although it is possible to interpret Matthew’s pericope as a redacted version of its Markan parallel, neither the different context nor verbal and material differences between them require this interpretation.

10.    My analysis of Mk 1:21–22, 3:7–12 and 6:1–13 indicates that at the compositional level, none of these pericopes is better explained by the Markan hypothesis than by the two-gospel hypothesis. Indeed, in the case of Mk 3:7–12, internal evidence favours Markan dependence on both Matthew and Luke. In addition, significant verbal correspondences between Matthew’s and Luke’s parallels to Mk 6:1–13 militate against the alleged independence of Matthew and Luke.

11.    From a compositional perspective, does the phenomenon of order favour the Markan hypothesis or the two-gospel hypothesis? My source-critical analysis of ten pericopes (and parallels) in the synoptic tradition indicates that the evidence is ambiguous. While order differences between the synoptic gospels can often be explained quite well as Matthew’s or Luke’s transpositions of Markan pericopes, in no case must one conclude that this is the only possible explanation. An impartial conclusion is difficult to reach, partly because most of the literature on the synoptic gospels during the past 150 years has either argued for or tacitly assumed Markan priority without paying attention to alternative perspectives in an even-handed way. It is not the case, as some advocates of Markan priority have maintained, that once one examines the phenomenon of order in detail, one cannot help but regard disagreements in order as rearrangements by Matthew and Luke of Mark’s original sequence of incidents. Rather, at the compositional level, both the Markan hypothesis and the two-gospel hypothesis are able to offer satisfactory explanations for the phenomenon of order. Until more comprehensive arguments from order are forthcoming—in which at least two potential explanations for disagreements in pericope order are assessed in an even-handed way—critics must go beyond appeals to the phenomenon of order to discriminate between competing source theories.

Appendix 1

The Segmentation of the Synoptic Gospels

The Gospel According to Matthew