The Synoptic Problem 2022: Proceedings of the Loyola University Conference 9789042950344, 904295034X, 9789042950351

Table of contents :
Cover
Title
Copyright
TABLE OF CONTENTS
INTRODUCTION
THE SYNOPTIC PROBLEM – A PUZZLE WITH A FUNDAMENTALLY LITERARY SOLUTION
WHAT IF? HISTORICAL AND LITERARY HYPOTHESES ON MATTHEAN POSTERIORITY
TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS: COHERENCE AND DISTINCTNESS?
THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022: PROSPECTS IN RETROSPECT
MEMORY AND ORDER IN
PHILO, AND LUKE
CAN MEMORY RESCUE THE FARRER HYPOTHESIS? THE
AND FH LUKE’S MATTHEW-UTILIZATION
THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM
THE LITERARY CHARACTER OF THE GOSPEL TRADITION
DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT
MARCION’S GOSPEL DID COME FIRST! A REJOINDER TO R.V. HUGGINS
THE WIDOW AT NAIN, THE SINFUL WOMAN WHO LOVED MUCH, AND THE EXTENT OF Q
THE (MINOR) MINOR AGREEMENTS: A STATISTICAL ANALYSIS OF MATTHEW’S AND LUKE’S PARAPHRASTIC TENDENCIES
LIST OF CONTRIBUTORS
INDEX OF AUTHORS
INDEX OF REFERENCES

Citation preview

Biblical Tools and Studies 

T

he Synoptic Problem 2022 PROCEEDINGS OF THE LOYOLA UNIVERSITY CONFERENCE

Edited by

Olegs Andrejevs Simon J. Joseph Edmondo Lupieri Joseph Verheyden

PEETERS

THE SYNOPTIC PROBLEM 2022

BIBLICAL TOOLS AND STUDIES Edited by J. VERHEYDEN, KU Leuven

Associate Editors G. BAZZANA, Harvard Divinity School – A. BERLEJUNG, Leipzig K.J. DELL, Cambridge – J. FREY, Zürich – C.M. TUCKETT, Oxford

Biblical Tools and Studies – Volume 44

THE SYNOPTIC PROBLEM 2022 PROCEEDINGS OF THE LOYOLA UNIVERSITY CONFERENCE

EDITED BY

OLEGS ANDREJEVS SIMON J. JOSEPH EDMONDO LUPIERI JOSEPH VERHEYDEN

PEETERS LEUVEN – PARIS – BRISTOL, CT 2023

Cover: Τῆς καινῆς Διαθήκης ἅπαντα. Εὐαγγέλιον Novum Iesu Christi D.N. Testamentum ex bibliotheca regia. Lutetiae: ex officina Roberti Stephani, 1550. in-folio. KU Leuven, Maurits Sabbe Library, P225.042/F° Mt 5,3-12

No part of this book may be reproduced in any form, by print, photoprint, microfilm or any other means without written permission from the publisher. A catalogue record for this book is available from the Library of Congress

ISBN 978-90-429-5034-4 eISBN 978-90-429-5035-1 D/2023/0601/4 © 2023, Peeters, Bondgenotenlaan 153, B-3000 Leuven (Belgium)

TABLE OF CONTENTS

Olegs ANDREJEVS – Simon J. JOSEPH – Edmondo LUPIERI – Joseph VERHEYDEN Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

PROLEGOMENA Paul FOSTER The Synoptic Problem – A Puzzle with a Fundamentally Literary Solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15

AFTER MARKAN PRIORITY: GENERAL ESSAYS Simon J. JOSEPH The End of Q Studies: Contemporary Reflections . . . . . . . . . 49 Ronald V. HUGGINS Reminiscences and Analysis: How Closing the Door on the Farrer Hypothesis Opened the Door to Proposing Matthean Posteriority . . 73 Edmondo LUPIERI What if? Historical and Literary Hypotheses on Matthean Posteriority 101 Hildegard SCHERER Testing Biblical Traditions in the Synoptic Materials: Coherence and Distinctness? . . . . . . . . . . . . . . . . . . . . . . . . . 115 Olegs ANDREJEVS The Two-Document Hypothesis and its Main Alternatives in 2022: Prospects in Retrospect . . . . . . . . . . . . . . . . . . . . . . 131 ANCIENT COMPOSITIONAL PRACTICES, MEDIA, AND THE SYNOPTIC PROBLEM Robert A. DERRENBACKER “Unfinished” Mark “Replaced” by Matthew and Luke? Some Recent Studies and their Implications for the Synoptic Problem . . . . . . 179

VI

TABLE OF CONTENTS

Eric EVE Memory and Order in Thomas, Philo, and Luke . . . . . . . . . Alan KIRK Can Memory Rescue the Farrer Hypothesis? The Temple Scroll and FH Luke’s Matthew-Utilization . . . . . . . . . . . . . . . . . Chakrita M. SAULINA Competitive Traditions: Luke’s and Matthew’s (Con)textualization of the Beelzebul Controversy . . . . . . . . . . . . . . . . . . . Alan GARROW “Frame and Fill” and Matthew’s Use of Luke . . . . . . . . . . . Robert K. MACEWEN The Centurion’s Servant and the Synoptic Problem . . . . . . . . John C. POIRIER Luke and the Wax Tablet Revisited: An Assessment of Supposed Difficulties . . . . . . . . . . . . . . . . . . . . . . . . . . . .

195

219

233 277 297

335

NEW AVENUES Matthias KLINGHARDT The Literary Character of the Gospel Tradition . . . . . . . . . . Ronald V. HUGGINS Did Marcion’s Gospel Come First? A Response to M. Klinghardt . Matthias KLINGHARDT Marcion’s Gospel Did Come First! A Rejoinder to R.V. Huggins . David B. SLOAN The Widow at Nain, the Sinful Woman Who Loved Much, and the Extent of Q . . . . . . . . . . . . . . . . . . . . . . . . . . Jeffrey M. TRIPP The (Minor) Minor Agreements: A Statistical Analysis of Matthew’s and Luke’s Paraphrastic Tendencies . . . . . . . . . . . . . . . .

359 387 397

401

417

LIST OF CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . 447 INDEX OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . 451 INDEX OF REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . 459

INTRODUCTION

The last decade produced three separate collections dedicated to the synoptic problem: the 2011 proceedings of the Oxford conference1; the 2016 conversation between representatives of four different hypotheses2; and the 2018 proceedings of the Roskilde conference3. Leading up to this volume’s release it was therefore not surprising to hear the following question in conversations with colleagues in the field: why was another synoptic problem volume necessary so soon? We see the answer as at least threefold. First of all, the Two-Gospel Hypothesis (2GH) has once again, at least for the time being, receded in academic representation4. This means that for the first time since the advent of critical New Testament studies the momentum in the field belongs to a triumvirate of Markan-priority-based theories. Accordingly, the discussion in this volume could largely presuppose the priority of Mark and focus on other issues (the sole exception arrives courtesy of Matthias Klinghardt, who envisions Marcion’s gospel as predating all canonical gospels)5. Second, the 1. P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011. 2. S.E. PORTER – B.R. DYER (eds.), The Synoptic Problem: Four Views, Grand Rapids, MI, Baker, 2016. 3. M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018. 4. The last two decades have produced a number of works substantially problematizing the 2GH. They include M. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002; R.A. DERRENBACKER, Jr., Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005; A. DAMM, Ancient Rhetoric and the Synoptic Problem (BETL, 252), Leuven – Paris – Walpole, MA, Peeters, 2013; D.G. REID, Miracle Tradition, Rhetoric, and the Synoptic Problem (BiTS, 25), Leuven – Paris – Bristol, CT, Peeters, 2016. Many others could be listed, including commentaries, further monographs, collected volumes, and individual essays and articles. All of these recent publications capitalize on the research published in the 1970s, 1980s, and 1990s, presenting a formidable body of work for potential future 2GH revivalists to engage with. The 2GH was still prominently featured in FOSTER et al. (eds.), New Studies in the Synoptic Problem (n. 1) and PORTER – DYER (eds.), The Synoptic Problem: Four Views (n. 2). 5. Cf. M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021 (and his contributions in this volume).

2

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

opposition to the Two-Document Hypothesis (2DH)6 has now mainly coalesced around two Mark-without-Q alternatives: the Farrer Hypothesis (FH)7 and its mirror image, the Matthean Posteriority Hypothesis (MPH). The latter theory had not received much attention in synoptic problem conferences and proceedings until the present publication. It therefore makes something of a belated entry into the conversation in this volume. Third, in recognition of a sustained and growing interest over the last two decades, this volume dedicates an entire section to the role played in the synoptic problem discussion by ancient compositional practices and media realities (with a number of other authors weighing in on these matters in their contributions as well). The present collection has humble origins. The original idea, formulated in the summer of 2019, was to hold a conference on the synoptic problem at Loyola University Chicago in the spring of 2020. This was envisioned as a relatively small in-person colloquium to feature the Department of Theology’s New Testament program and for the benefit of its graduate students. Accordingly, the original list of invitees – most of whom are represented in this volume – was limited to North American scholars who were willing to travel to Chicago for a live event. This group, now the kernel of the present collection, included Olegs Andrejevs (Loyola University Chicago), Simon J. Joseph (University of California, Los Angeles), Alan Kirk (James Madison University, Virginia), Edmondo Lupieri (Loyola University Chicago), John C. Poirier (independent scholar), David B. Sloan (Ashland Theological Seminary, Ohio), and Jeffrey M. Tripp (Rock Valley College, Illinois, a Loyola University Chicago alumnus). A few others were originally signed on to participate but had to leave the project for various reasons. The Covid-19 pandemic broke out in early 2020, grinding normal life as the world knew it to a halt. All conference plans were placed on an indefinite hold. Instruction shifted online, the campus buildings were closed to all but the personnel with special permission, and travel funds were frozen. It became clear already in the spring of 2020 to those of us in the US who were following the scientific opinion of medical professionals that the situation was unlikely to improve until mass vaccination became available. This meant

It can be noted, of course, that Klinghardt still identifies Mark’s gospel as the earliest canonical gospel. 6. The term Two-Source hypothesis (or Two-Source theory) remains in use as well. 7. Referred to in earlier research (and occasionally today) as the Farrer-Goulder hypothesis (FGH). However, GOODACRE, Case (n. 4), p. 14, observes that the “-Goulder” (G) piece somewhat unfairly ties the overall theory to a particular articulation of it.

INTRODUCTION

3

that the fall of 2020 would bring little change. In fact, classes remained online and in-person events remained impossible through the spring of 2021. Accordingly, the decision was made to turn this project into a collected volume. The original participants listed above all graciously agreed to write essays that could be submitted for publication. With logistics no longer being a part of the equation, it became possible – and, in fact, logical – to reach out to colleagues from overseas. Over the summer of 2020, the current roster of contributors gradually came together, with participants from five countries and three continents. The second wave of invitees included Robert A. Derrenbacker, Jr. (Trinity College, University of Melbourne), Eric Eve (Emeritus Fellow of Harris Manchester College, University of Oxford), Paul Foster (University of Edinburgh), Alan Garrow (Sheffield Centre for Interdisciplinary Biblical Studies), Ronald V. Huggins (independent scholar), Matthias Klinghardt (Technische Universität Dresden), Robert K. MacEwen (Tyndale Theological Seminary, the Netherlands), Chakrita M. Saulina (Seattle Pacific University), and Hildegard Scherer (Universität Duisburg-Essen). A number of other scholars were invited but were unable to commit at the time. Given the number of contributors, it was desirable to identify a recurring theme that would be prominently – even if not universally – featured in the volume. Accordingly, it was suggested to the expanded roster of participants that among the project’s principal interests would be the role played by ancient compositional practices and media in the study of the synoptic problem. This was a natural direction for inquiry, given the volume of publications on these subjects in the 2010s8 and the fact that the most recent

8. Listed chronologically: 2011 C.M. TUCKETT, The Current State of the Synoptic Problem, in FOSTER et al. (eds.), New Studies in the Synoptic Problem (n. 1), 9-50; R.A. DERRENBACKER, Jr., The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, ibid., 435-458; F.G. DOWNING, Writers’ Use or Abuse of Written Sources, ibid., 523-548; A. KIRK, Memory, Scribal Media, and the Synoptic Problem, ibid., 459-482; ID., Orality, Writing, and Phantom Sources: Appeals to Ancient Media in Some Recent Challenges to the Two Document Hypothesis, in NTS 58 (2011) 1-22. 2012 J.C. POIRIER, The Roll, the Codex, the Wax Tablet and the Synoptic Problem, in JSNT 35 (2012) 3-30. 2013 E. EVE, Behind the Gospels: Understanding the Oral Tradition, London, SPCK, 2013; R.A. DERRENBACKER, Jr., Texts, Tables and Tablets: A Response to John C. Poirier, in JSNT 35 (2013) 380-387; F.G. DOWNING, Waxing Careless: Poirier, Derrenbacker and Downing,

4

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

synoptic problem conference whose proceedings were published largely did not focus on these questions9. Of course, not all of the original contributors – who had already invested much work into their papers for an in-person colloquium that never materialized – could produce all-new essays. Furthermore, a few contributors from the second wave of invitees had interesting

in JSNT 35 (2013) 388-393; F. WATSON, Gospel Writing: A Canonical Perspective, Grand Rapids, MI, Eerdmans, 2013. 2015 E. EVE, Memory, Orality and the Synoptic Problem, in EC 6 (2015) 311-333. 2016 J.W. BARKER, Ancient Compositional Practices and the Gospels: A Reassessment, in JBL 135 (2016) 109-121; E. EVE, Writing the Gospels: Composition and Memory, London, SPCK, 2016; A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226; A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016. 2017 R.A. DERRENBACKER, Jr., Matthew as Scribal Tradent: An Assessment of Alan Kirk’s Q in Matthew, in JSHJ 15 (2017) 213-23; F.G. DOWNING, Plausibility, Probability, and Synoptic Hypotheses, in ETL 93 (2017) 313-337; M. GOODACRE, Q, Memory and Matthew: A Response to Alan Kirk, in JSHJ 15 (2017) 224-233; A. KIRK, The Synoptic Problem, Ancient Media, and the Historical Jesus, in JSHJ 15 (2017) 234-259; C.J. MONAGHAN, Why Have We Reduced the Oral Tradition to Silence? The Oral Tradition’s Role in the Formation of the Minor Agreements, in ABR 65 (2017) 1-16. 2018 J.S. KLOPPENBORG, Oral and Literate Contexts for the Sayings Gospel Q, in C. HEIL – G. HARB – D.A. SMITH (eds.), Built on Rock or Sand? Q Studies: Retrospects, Introspects and Prospects (BiTS, 34), Leuven – Paris – Bristol, CT, Peeters, 2018, 49-72; D.A. SMITH, From Parable to Logion: Oral and Scribal Factors in the Composition of Q, ibid., 73-97; R.A. DERRENBACKER, Jr., Orality and Scribality in Q: A Response to the Papers of John S. Kloppenborg, Giovanni B. Bazzana, and Daniel A. Smith, ibid., 121-127; C.M. TUCKETT, Watson, Q, and ‘L/M’, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 3), 115-138; F. WATSON, Seven Theses on the Synoptic Problem, in Disagreement with Christopher Tuckett, ibid., 139-147. 2019 K.-H. CHANG, Questioning the Feasibility of the Major Synoptic Hypotheses: Scribal Memory as the Key to the Oral–Written Interface, in JSNT 41 (2019) 407-432; J.S. KLOPPENBORG, Macro-Conflation, Micro-Conflation, Harmonization and the Compositional Practices of the Synoptic Writers, in ETL 95 (2019) 629-643. 2020 A. GARROW, Plausibility, Probability, and Synoptic Hypotheses: A Response to F. Gerald Downing, in ETL 96 (2020) 131-137. 2021 E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021. 9. MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 3).

INTRODUCTION

5

proposals in other areas of inquiry. This meant that the volume’s scope could not be too narrow. Still, the eventual collection offered here features seven essays addressing the synoptic problem in the context of ancient compositional practices and media (Derrenbacker, Eve, Garrow, Kirk, MacEwen, Saulina, Poirier), with a number of other contributors touching on the pertinent questions in their essays even as their main focus lies elsewhere. This theme, then, either lies in the foreground or forms a part of a larger conversation in the majority of the essays included in this volume. In the opening paper of the 41st Colloquium Biblicum Lovaniense (1992), Frans Neirynck programmatically listed the “basic problems in the study of the gospels” as they presented themselves at the time: “Matthean redaction of Mark, minor agreements and Lukan redaction, the sayings source Q, and the Gospel of Mark”10. In 2022, two of these problems can be expanded: no longer only “Matthean redaction of Mark” and “minor agreements and Lukan redaction”11, but also “Lukan redaction of Mark” and “minor agreements and Matthean redaction”. In hindsight, it now seems clear that the emergence of the MPH will be cited as one of, if not the defining feature of the 2010s as far as the synoptic studies are concerned12. This volume features not only contributors advocating for the MPH, but also substantive interaction with this hypothesis by proponents of other theories. With the highest number of pro- and MPH-leaning contributions published by various authors in one place to date (Garrow, Huggins, Lupieri, MacEwen, Saulina, Tripp), this volume will likely be remembered as the moment the MPH crossed into the mainstream of synoptic studies. Thereby its theorists have made good on the momentum the MPH somewhat unexpectedly generated in the 2010s. We can now turn to the individual contributions13. The collection is divided into three main sections, prefaced by an opening essay. *  *  * 10. F. NEIRYNCK, Literary Criticism, Old and New, in Colloquium Biblicum Lovaniense: Opening of the 41st Session. Huldezitting Frans Neirynck 18 augustus 1992, pp. 5-16, here p. 13. The passage cited here is missing from the enlarged version of Neirynck’s farewell speech that was published the next year in the proceedings of the Conference (BETL, 110, pp. 11-38) and is reprinted in ID., Literary Criticism, Old and New, in ID., Evangelica III: 1992-2000. Collected Essays (BETL, 150), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001, 65-92. 11. ID., Literary Criticism (n. 10), p. 13. 12. See the history-of-scholarship survey (through 2015) in R.K. MACEWEN, Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS, 501), London – New York, Bloomsbury T&T Clark, 2015, pp. 6-26. 13. Some of the summaries offered below incorporate the abstracts originally submitted by the authors.

6

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

The essay by P. FOSTER was selected for a separate section of its own because of its focus on the foundation shared by all the contributions in this volume. In the paper, Foster explores the form in which the evangelists received their traditions. He contends that the formation and interrelations of the synoptic gospels should be understood principally as a literary phenomenon. Foster engages critically with some of the presuppositions behind the oral paradigm and the discipline of performance criticism in the context of synoptic gospel studies. This is followed by a discussion of the triple and double synoptic traditions, the possible origin of the Markan materials, and the Sondergut traditions. *  *  * The next section of this volume is titled “After Markan Priority: General Essays”. The enclosed essays analyze specific Markan-priority-based hypotheses (Andrejevs, Huggins, Lupieri), phenomena (Scherer), and shifts in the field (Huggins, Joseph). The title of S.J. JOSEPH’s essay is a double entendre, an appeal to double meaning in reflecting on the current state of Q Studies. It is not a call to end Q Studies, but rather an attempt to review the field’s past accomplishments and present conditions – which some have interpreted as signifying its imminent and terminal demise. In this essay, Joseph questions prevailing assumptions regarding Q studies’ ideological and theological significance by exploring a select number of its intersections with contemporary critical reflection on ethnicity, race, gender, and class. The next essay is the first of six in this volume supporting the MPH. The reader is taken to the University of Toronto / Toronto School of Theology in the late 1980s/early 1990s, where R.V. HUGGINS’s efforts to see the FH as a viable alternative to the 2DH were repeatedly undermined by Luke’s greater primitivity in a series of double-tradition passages. Accordingly, his essay here gives an account of how the failure of the FH (in the author’s eyes) helped pave the way toward his ultimately endorsing the MPH14. The essay focuses in particular on H.O. Guenther’s Q Seminar as a key setting for the process, but also describes the larger context of synoptic and Jesus studies at the time relating to it.

14. Cf. R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1992) 1-22.

INTRODUCTION

7

E. LUPIERI offers the collection’s second MPH essay. The author begins by crediting Pierson Parker’s now classic 1981 JBL article as the original inspiration for pursuing this synoptic hypothesis four decades ago15. Proceeding from a conviction that Matthew and Luke must have known each other, Lupieri then focuses on a number of considerations, including (but not limited to) the Minor Agreements of Matthew and Luke against Mark, the temptation narratives, and the infancy narratives of Matthew and Luke. Noting that doubts in the 2DH have increased, H. SCHERER proposes to systematically test the DT material for its coherence and distinctiveness over against Markan and other synoptic traditions. The work on this task was started in Scherer’s 2016 Habilitationsschrift, which scrutinized synoptic social categories16. In this essay Scherer turns to another promising parameter, analyzing and comparing references to scripture that are clearly marked by citation formulae or by onomastic reference. In his paper, O. ANDREJEVS responds to the 2DH’s recent critics from the classically argued position of this hypothesis, focusing especially on publications that appeared between 2010-2022. After addressing a number of general claims, Andrejevs assesses various categories of data whose validity has recently been challenged by the 2DH’s critics, paying special attention to source utilization on Markan priority. Against an undifferentiated approach that relies simply on the existence and number of the Minor Agreements, he calls for their renewed study – in light of Matthean and Lukan redactional tendencies and the phenomenon he calls “unsustained Minor Agreements”. Also discussed are the phenomena of alternating primitivity, FH/MPH evangelists’ rejection of congenial materials, and the triple/ double tradition doublets. *  *  * The volume’s next and largest section is titled “Ancient Compositional Practices, Media, and the Synoptic Problem” (Derrenbacker, Eve, Kirk, Saulina, Garrow, MacEwen, Poirier).

15. P. PARKER, A Second Look at the Gospel before Mark, in JBL 100 (1981) 389-413. Parker (see esp. p. 410) proposes a complex picture in which proto-Matthew and proto-Luke are utilized by Mark and the canonical Matthew/Luke. The MPH element of this hypothesis consists of proto-Luke’s influence on the canonical Matthew. 16. H. SCHERER, Königsvolk und Kinder: Der Entwurf der sozialen Welt im Material der Traditio duplex (BBB, 180), Göttingen, Bonn University Press – V&R Unipress, 2016.

8

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

The work that has heightened the field’s awareness of the factors listed in this section’s heading is R.A. DERRENBACKER’s Ancient Compositional Practices and the Synoptic Problem17. In this contribution, the same author engages the assumption of Mark’s Gospel as an ‘unfinished’ (or at least ‘lessfinished’) book. After surveying the recent studies by J. Andrew Doole, Matthew D.C. Larsen, and Chris Keith18, Derrenbacker proposes that Mark may have nevertheless survived at least in part due to its inherent ‘bookishness’. Another ‘incomplete’ and ‘unfinished’ potential source of Matthew and Luke, namely Q, may not have fared as well, lacking some of Mark’s literary characteristics. Implications of the ‘unfinished’ character of Matthew’s and Luke’s sources are examined for each of the Markan-priority-based solutions to the synoptic problem. In his essay, E. EVE explores the possible role of memory in FH Luke’s ordering of material taken from Matthew by examining two partial parallels. Eve begins by setting the discussion in the context of the proposals for Luke’s ordering of Matthean material offered in his 2021 monograph19. He then discusses the reordering of Synoptic material in the Gospel of Thomas on the assumption of Thomas’s (memory-based) dependence on the gospels, while also noting what difference the assumption of Thomasine independence might make. He subsequently builds on existing work on the structure of Philo’s allegorical treatises to argue how Philo’s use of memory cues might resemble those proposed for FH Luke. The threads of the discussion are drawn together in the conclusion, arguing that while neither example provides an exact parallel, together they do support the ability of FH Luke to order his Matthean (and L) material as he would need to have done. A. KIRK and Eve have recently been engaged in an exciting exchange20. Accordingly, Kirk’s reaction to Eve’s 2021 monograph makes for a logical segue at this point. In his essay, Kirk tests the claim that FH Luke’s rather unsystematic use of Matthew can be explained with help of memory-based source utilization. He explores the source utilization techniques employed in the Temple Scroll (11Q19) and finds that that document’s compositional technique issues from media conditions and from the topical and visualspatial (concentric) schematic of 11Q19. This provides yet another analogy 17. Supra, n. 4. 18. J.A. DOOLE, What Was Mark for Matthew? (WUNT, II/344), Tübingen, Mohr Siebeck, 2013; M.D.C. LARSEN, Gospels before the Book, Oxford, Oxford University Press, 2018; C. KEITH, The Gospel as Manuscript: An Early History of the Jesus Tradition as Material Artifact, Oxford, Oxford University Press, 2020. 19. EVE, Relating (n. 8). 20. KIRK, Q in Matthew (n. 8); EVE, Relating (n. 8).

INTRODUCTION

9

to Matthew’s use of Mark and (conventionally reconstructed) Q, supplementing Kirk’s analogies already offered in his 2016 book. In contrast to 11Q19, FH Luke’s use of Matthew is found to lack a compositional schematic that his (memory-based) movements around Matthew might be mapped to. C.M. SAULINA’s essay opens a block of three contributions endorsing the MPH. She directs attention to distinct phenomena in the gospels, e.g., common order and verbal agreements, that, in contrast to common features in classical texts, position the evangelists within the author-tradent spectrum. Saulina builds on Chris Keith’s concept of ‘competitive textualization’ within a ‘textual consciousness’, recognizing evidence of a strong relationship between the evangelists and their (written) sources21. The footprints of these sources, she suggests, can be retraced by analyzing gospel intertextuality. By focusing on the intertextuality between the Beelzebul Controversy passage and its surrounding contexts, Saulina seeks to establish the order of influence among the three synoptists. The different levels of examination adopted in the paper lead her to conclude that the MPH offers a better explanation of the phenomena than the FH or the 2DH. In his essay, A. GARROW argues that ancient authors who rewrote a narrative using two or more sources are consistent in their adoption of a ‘frame and fill’ technique. Here, one source is chosen as the base, or ‘frame’, and elements from additional sources are then used to supplement, or ‘fill’, that frame. According to Garrow, examples of this approach may be found in Plutarch, Livy, Diodorus, Josephus, Qumran, Tatian, and, in related activity, in Ammonius of Alexandria and the Eusebian Canons. This consistent pattern provides a broad backdrop against which to assess the credibility of the compositional techniques required by the three Markan Priority solutions to the synoptic problem. The implications are found to be least favorable to the FH and most favorable to hypotheses in which Matthew directly used Luke. R.K. MACEWEN’s paper uses the Healing of the Centurion’s Servant (Matt 7,28a; 8,5-13//Luke 7,1-10; 13,28-29) as a test case for exploring the synoptic problem. This pericope has a number of features that make it valuable for studying the interrelations among the Gospels. It is a rare example of a narrative pericope in the double tradition. It is a pericope in which (a) material that is separated in Luke is found together in Matthew; (b) in which there is a high degree of verbatim agreement between Matthew and Luke, including in the speech of a character who is not Jesus; and (c) in 21. C. KEITH, The Competitive Textualization of the Jesus Tradition in John 20:30-31 and 21:24-25, in CBQ 78 (2016) 321-337.

10

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

which Matthew, in comparison to Luke, is both abbreviated in terms of narrative and expanded by the inclusion of theologically rich material, phenomena that are also frequently observable in Matthew’s parallels with Mark. MacEwen’s essay explores and evaluates how three competing solutions to the synoptic problem account for this set of phenomena, with emphasis on redaction criticism and ancient scribal practices. MacEwen concludes that the MPH provides the best explanation for the phenomena of this pericope. Closing out this section, J.C. POIRIER responds to the critics of his 2012 JSNT article22. There, Poirier argued that M. Goulder’s suggestion that Luke had moved backward through a scroll of Matthew was not only technically feasible and circumstantially sensible, but also that one can see a similar proceeding in Clement of Alexandria’s use of Philo. Poirier had further proposed in 2012 that there is an altogether better way to account for FH Luke’s handling of Matthew, namely, to allow him access to a collection of wax tablets. The present essay responds to three 2DH proponents who engaged Poirier’s 2012 arguments. Poirier seeks to show that the criticisms lodged by R.A. Derrenbacker, F. Gerald Downing, and A. Kirk do not succeed23, and that an improved understanding of ancient compositional practices provides strong support for the FH. *  *  * The volume’s closing section is titled “New Avenues” (Klinghardt [with a response by Huggins and a rejoinder by Klinghardt], Sloan, Tripp). This group of papers focuses on innovative (or at least less explored) approaches to the synoptic problem. In the present volume’s sole non-Markan priority-based contribution, M. KLINGHARDT reprises his thesis of Marcion’s gospel as the oldest of all the gospels and the source for the canonical Gospels24. After a brief outline of his theory’s foundational assumptions, Klinghardt discusses the literary character of the Gospel tradition and its methodological implications for the synoptic problem. This is followed by R.V. HUGGINS’s assessment of Klinghardt’s hypothesis and by KLINGHARDT’s rejoinder to Huggins.

22. POIRIER, Roll (n. 8). 23. DERRENBACKER, Response (n. 8); DOWNING, Waxing Careless (n. 8); KIRK, Q in Matthew (n. 8). 24. KLINGHARDT, Gospel (n. 5).

INTRODUCTION

11

In his essay, D. SLOAN addresses the problem the standard 2DH has explaining Luke’s apparent tendency toward conflation within Q material despite the same evangelist’s avoidance of conflations within the Markan material. The author proposes that this problem disappears if the extent of Q is reconsidered. Rather than suggesting an alternation between two sources in the non-Markan blocks (Luke 3,1–4,30; 6,20–8,3; 9,51–18,14; 19,1-27; 22,14-38), Sloan suggests that Luke is working with one source throughout these blocks and that the so-called “L material” in these blocks is really Q material that Matthew has omitted. As a test case, he considers two “L” passages in Luke 7,11-17 and 7,36-50. Sloan argues that 2DH Luke’s source here is the same source that Luke used for 7,1-10 and 7,18-35 – namely, Q – and suggests that Matthew knew this same source and had reasons for omitting these two passages. It is suggested that the extent of Q has been drastically underestimated by modern scholarship and that a new approach to identifying Q material is necessary. In the volume’s sixth and final MPH-leaning essay, J.M. TRIPP returns to the subject of the Minor Agreements or points of lexical agreement between Matthew and Luke against Mark within the triple tradition. The author attempts to break the current impasse in the discussion of this phenomenon by turning to statistical hypothesis testing, which, he contends, is expressly designed to distinguish coincidental events from significant ones. Counting the Minor Agreements through the pedagogical lens of firstcentury paraphrase exercises, he finds the number of Minor Agreements to be significantly higher than expected in independent texts. In some cases, conditional probability suggests to Tripp a particular direction of influence: from Luke to Matthew. After the volume’s acceptance in the series “Biblical Tools and Studies”, Joseph Verheyden (Katholieke Universiteit Leuven, Belgium) accepted the invitation to join the team of editors. The other editors are very grateful to him and the editorial board of BiTS for accepting this volume for publication; they also owe thanks to Prof. Verheyden for his assistance in the preparation of the publication. Together, the four editors would like to thank all the contributors from both the first and second waves for their interest in and sustained commitment to this project, and for their hard work. Dr. Robert A. Di Vito, chair of the Department of Theology at Loyola University, has supported and encouraged this project from its inception. In lieu of an in-person conference, he hosted a dinner for the participants at the 2022 Annual Meeting of the Society of Biblical Literature in San Antonio, TX. We would like to thank Dr. Vito for

12

O. ANDREJEVS – S.J. JOSEPH – E. LUPIERI – J. VERHEYDEN

helping us see this project through to its successful conclusion. Finally, our sincere gratitude goes out to Peeters Publishers for their support of this volume. Loyola University Chicago United States of America University of California, Los Angeles United States of America Loyola University Chicago United States of America Katholieke Universiteit Leuven Belgium

Olegs ANDREJEVS [email protected] Simon J. JOSEPH [email protected] Edmondo LUPIERI [email protected] Joseph VERHEYDEN [email protected]

PROLEGOMENA

THE SYNOPTIC PROBLEM – A PUZZLE WITH A FUNDAMENTALLY LITERARY SOLUTION

I. INTRODUCTION The Synoptic Problem is still not solved – at least not to everybody’s satisfaction. That is not fresh news. The majority of scholars still divide between the most widely supported solution the Two-Source Theory (and its variant the Four-Source Theory), and other solutions such as neoGriesbach theory, the Farrer or Mark-without-Q theory, and the alternative of Matthean Posteriority. Again, this is well-known. What has changed over recent decades has been attempts to jettison such literary solutions altogether. These are often represented as no more than quasi-algebraic expressions with its own variables, leading to equations such as Matt = Mark + Q. This perhaps is a more fundamental challenge to previous attempts to clarify gospel relationships through literary models, and it may even call into question the possibility of finding a comprehensive solution to the Synoptic Problem. II. ORALITY

AND

PERFORMANCE CRITICISM

The idea that Jesus-traditions came to the evangelists through oral media has had a long history in accounts of how gospel writers received the material out of which their narratives were constructed. In this vein, the work of Birger Gerhardsson is correctly identified as a watershed in advocating for the oral transmission of gospel traditions1. For Gerhardsson the model of oral transmission of traditions could be identified in rabbinic Judaism, and thus could be seen as a way to account for the transmission of gospel traditions.

1. B. GERHARDSSON, Memory and Manuscript: Oral Tradition and Written Transmission in Rabbinic Judaism and Early Christianity, Uppsala, C.W.K. Gleerup, 1961.

16

P. FOSTER

The most common criticism of Gerhardsson’s theory was the justifiable counter-argument that the early Jesus movement was not analogous to the proposed model of a scholarly rabbinic academy, and moreover it was likely the case that what rabbis were memorising orally was a set of texts. While oral models did not dominate the discussion, Gerhardsson was not a lone voice. With significant differences, others attempted to account for the evolution of gospel traditions through the medium of orality. For instance, Lohr attempted to account for much of the material in the Gospel of Matthew on the basis of orality2. Of more lasting impact was the work of Werner Kelber. He sought to give a better account of the process by which oral traditions were transformed into the literary products of written gospels. Thus, he stated “Mark’s writing manifests a transmutation more than mere transmission”3. Unlike the form-critics, these scholars were less interested in the origins of traditions that at several stages later coalesced into the canonical gospels. Instead, their theories attempted to explain the formation of the Synoptic Gospels and the immediate traditions the evangelists utilised as being best understood by postulating that these traditions were accessed through an oral medium. Typically, attention focused on the Gospel of Mark, because such ideas were more easily questioned in relation to Matthew’s and Luke’s narratives. One point that will stressed throughout this discussion is that it is not being questioned that some traditions, perhaps at several stages behind the gospels might have first been delivered and then circulated in oral form. That of course is especially the case with any material that has a good claim to having originated with Jesus of Nazareth. The central concern is to argue that the Synoptic Gospels are best understood as being primarily composed from written sources and from the evangelists’ compositional activities exhibited in their redactional work. This paper responds to those who have contended that the production and interrelationships of the Synoptic Gospels are better understood by factoring in a large-scale oral component in the transmission of traditions and the composition of these literary narratives.

2. C. LOHR, Oral Techniques in the Gospel of Matthew, in CBQ 23 (1961) 403-435. 3. W. KELBER, The Oral and the Written Gospel: The Hermeneutics of Speaking and Writing in the Synoptic Tradition, Mark, Paul, and Q, Philadelphia, PA, Fortress, 1983, p. 91.

THE SYNOPTIC PROBLEM

17

1. Assembling Oral Traditions? More recently, and in opposition to theories based primarily on written sources of tradition, ideas of traditions coming to the evangelists in oral form have been given fresh impetus. One of the defining examples of this trend from around the turn of the millennium was the SNTS Presidential address of Prof. James D.G. Dunn, entitled “Altering the Default Settings: Re-envisaging the Early Transmission of the Jesus Tradition”4. Dunn’s main contention was that “classic solutions to the Synoptic Problem, and the chief alternatives, have envisaged the relationships between Gospel traditions in almost exclusively literary terms”5. He diagnosed the cause of this outlook as being the “literary mindset (‘default setting’) of modern Western culture [which] prevents those trained in that culture from recognizing that oral cultures operate differently”6. Behind this assessment stands the claim that modern scholarship was unreflectively blinkered in the way it projected its own textual and print-based understandings onto a first century society that did not handle traditions in that manner. It is of course a non sequitur to assert that first century culture was not dependent on a print-based culture. More apposite is the rejection of it being a text-based culture. Then as now, but not in identical forms, society was a mix of text and orality. Words were spoken, and words were written. The relationship between these two media for communication of words was certainly not discrete, yet it was definitely complex. Spoken words were at times written down – sometimes verbatim, sometimes in summary form. Such spoken words could be recorded contemporaneously with the act of speaking, or recalled and written with varying degrees of accuracy of the passage of time. Alternatively, words could appear first as written documents to be read aloud at a later date in dramatic performance, at skillful public reading events, or to be read silently for individual consumption. The ancient world was a complex interface between the oral and the textual. It is a mistake to have any pre-determined default setting, be it literary or oral. The multiplicity of forms in which words could be used and the complexity of the relationship between speech and text precludes any automatic preference for one default-setting over another. Rather, each example of preserved tradition must be carefully scrutinized to 4. The lecture was presented as the Presidential address at 57th annual meeting of the Studiorum Novi Testamenti Societas at the University of Durham on Wednesday, 7th August 2002. It was later published as J.D.G. DUNN, Altering the Default Settings: Re-envisaging the Early Transmission of the Jesus Tradition, in NTS 49 (2003) 139-175. 5. Ibid., p. 139. 6. Ibid.

18

P. FOSTER

best account for its textual or oral origin, and it is unsurprising if in the quest to understand such origins that binary answers do not always prove to be the most satisfying accounts. Dunn’s Presidential address – itself an oral event, but read from a written text, which after the skilful reading of the text was itself preserved in written form in the pages of the journal New Testament Studies and is now available on several electronic platforms – was an opening salvo in an attempt to recentre orality as a better way to account for the development and transmission of the Jesus tradition. Dunn drew upon the theoretical work of Walter Ong7, and other early attempts to conceptualize the gospel traditions as primarily oral in nature8. Thus, Dunn states in relation to Jewish core traditions that “knowledge of Torah for most people would have been by hearing, aural, rather than by reading”9. That observation is of course totally true, but equally misleading. The reason for that is because the act of hearing was based upon written texts, that were not only carefully produced by skilled scribes, but which were checked for standardisation and then studied and scrutinised by those trained in the production of texts. Of course, public reading took place, but this was frequently public reading of meticulously produced texts. Dunn’s arguments are in many ways extended and given more detail by his former PhD student Terence Mournet. Writing specifically in relation to the double tradition material Mournet concludes that “several of the key arguments … have led many scholars (rightly) to conclude that there is a textual source behind the common Matthean and Lukan double tradition”10. Admittedly, Mournet believes too much of the double tradition material is attributed to the common written source (this is discussed in section IV). However, his own investigation agrees with the overarching observation that the core of the double tradition existed in written form prior to its incorporation into the gospels of Matthew and Luke. As this paper focuses on the composition, interrelationships and transmission of the Synoptic Gospels, this seems to bear out the central claim that those processes are to be understood principally in literary and textual terms. That is not to deny that several of these traditions, perhaps a couple of stages prior to their incorporation into gospel narratives, may have existed in oral form. However, the 7. W.J. ONG, Orality and Literacy: The Technologizing of the Word, London, Methuen, 1982. 8. KELBER, The Oral and the Written Gospel (n. 3). 9. DUNN, Altering the Default Settings (n. 4), p. 148. 10. T.C. MOURNET, Oral Tradition and Literary Dependency (WUNT, II/195), Tübingen, Mohr Siebeck, 2006, p. 287.

THE SYNOPTIC PROBLEM

19

work of the synoptic evangelists should not be conceived as the assemblage of units of free-floating oral traditions. 2. Performance Criticism A specific and focused instance of the notions of orality as being essential to the solution to the Synoptic Problem and the formation of the Jesus tradition draws upon insights from the field of performance criticism. Again, what is not in dispute is that those gospel traditions that can be traced back to Jesus of Nazareth almost certainly originated orally and not in written form. While that is not contentious, that is not the issue that is at stake in this discussion. Here the concern is with the formation of the written gospels. In particular, in this discussion the focus is primarily on the Synoptic Gospels. However, what is being argued could be equally applied more widely to the whole phenomenon of gospel writing, without imposing any unnecessary dichotomy between those writings that would come to be classified as canonical gospels and those that would be viewed as non-canonical11. As an academic discipline performance criticism came to the fore around the late twentieth century and the turn of the twenty-first century. This is reflected by the emergence of a body of literature that attests this approach emerging as an academic discipline in its own right12. In terms of origin, performance criticism developed as an approach to the study of drama, and, more specifically, as a facet of theatre studies. Presenting a centrist perspective, C. Edward McGee states that “[p]erformance criticism treats the written text of a play as a script to be realized in performance”13. Such perspectives have been drawn into the field of New Testament studies as a means for understanding either the origin or the function (or both) of 11. In this respect, the perspective concurs with Francis Watson’s observation that canonical boundaries impose later categories on gospel writings at a time when those divisions had neither been created or imposed. See F. WATSON, Gospel Writings: A Canonical Perspective, Grand Rapids, MI, Eerdmans, 2013. More recently Watson states in relation to the Epistula Apostolorum, “Before the concept of a fourfold gospel achieved dogmatic status, the collective decision about any gospel text remained in principle open. If EpAp can be shown to pre-date the canonical dogma, it can take its place within the diverse landscape of early Christian literature as yet undivided by the canon’s binary logic”. ID., An Apostolic Gospel: The “Epistula Apostolorum” in Literary Context (SNTS MS, 179), Cambridge, Cambridge University Press, 2021, p. 7. 12. At a general level see the collection of essays in P. AUSLANDER (ed.), Performance: Critical Concepts in Literary and Cultural Studies, London, Routledge, 2003. 13. C.E. MCGEE, Performance Criticism, in I. MAKARYK (ed.), Encyclopedia of Contemporary Literary Theory, Berlin – New York, De Gruyter, 1993, 133-139, p. 133.

20

P. FOSTER

New Testament writings. In this vein, David Rhoads makes a maximalist case for the role of performance in relation to the reception of New Testament texts. He states, A performance was an integral part of every early Christian experience of the compositions that now comprise the writings of the New Testament. The New Testament writings were either written “transcriptions” of oral narratives composed in performance or they were composed in writing (perhaps orally by dictation) for use in oral performance. These compositions were presented orally. Always there was a performer or storyteller. Always the performances were heard / experienced rather than read. Always there was a communal audience. Always there was a physical location and a socio-historical circumstance that shaped the performance and the reception. Frequently, perhaps more often than not, no written text was even present to the event. Why have we not given greater attention to the performance dimension of the ancient world and to the experience of biblical performances by ancient Christian audiences?14

Thus, one of the key assertions of this approach is that the gospels are artefacts that were composed in, or for performance. The view that a gospel was composed in performance is espoused by Antoinette Clark Wire in relation to the Gospel of Mark15. Here, Wire sensibly isolates the Gospel of Mark, which is widely recognized as being the earliest surviving witness to the gospel tradition. The ethnographic research of Milman Parry and Albert Lord leveraged insights gained from recordings of Serbo-Croatian singers in order to propose Homeric epics such as the Iliad and the Odyssey were composed orally through the performances of minstrels16. Thus, Wire sees the written “transcript” of Mark as recording an instance of performance, or being a base text on which later performances were based. Later in this discussion, the case will be cautiously made that the Gospel of Mark is better understood as being a literary composition, which drew upon some precursor traditions that are most likely to have already been in circulation in literary form and written in Greek. Here, however, what is challenged is the central claim of Rhoads and Wire that the Gospel of Mark was performed and composed through such events, rather being composed through an act of writing and then subsequently read.

14. D. RHOADS, Performance Criticism: An Emerging Methodology in Second Testament Studies – Part I, in BTB 36 (2006) 118-133, pp. 118-119. 15. A.C. WIRE, The Case for Mark Composed in Performance (Biblical Performance Criticism Series, 3), Eugene, OR, Cascade Books, 2011. 16. M. PARRY, in A. PARRY (ed.), The Making of Homeric Verse: The Collected Papers of Milman Parry, Oxford, Clarendon, 1971; A. LORD, Singer of Tales, Cambridge, MA, Harvard University Press, 22000.

THE SYNOPTIC PROBLEM

21

This claim has been identified by Hurtado as resulting from “a number of oversimplifications (and so distortions) of relevant historical matters, and that consequently, the inferences drawn about the composition and use of texts in early Christianity are the more dubious”17. One of the key insights that Hurtado draws upon was framed by Holt Parker in relation to the poetic texts of the Republican and Augustan period. Parker cautioned against the “exoticizing” of such literature as being destined or designed for performative contexts18. As Hurtado notes, Parker “showed amply that poets and other writers of the time wrote for readers, indeed individual readers more precisely, and not simply, or even particularly, for auditors of group/oral ‘performances’”19. Furthermore, the notion that variability in performance was prized above fixity of the written form is another aspect of the theory that is open to question. Rather than understanding the Gospel of Mark as a mere script, which according to Shiner evoked “applause of different types, including waving of hands, loud exclamations, and rhythmic clapping”20, the text is more accurately understood as being written in order to be read, either by a lector in a public setting or by an individual for private consumption. That first setting of a lector reading a relatively fixed text to gathered audience has been amply demonstrated as being a common feature of GraecoRoman literary culture. In this vein, Brian J. Wright states that “communal reading events were already a prevailing practice over a wide geographical range in the first century CE. In other words, communal reading events were a widespread phenomenon in the Roman Empire during the first century CE”21. This conclusion is supported by an impressive array of evidence22. Examples of individual or private reading can also be found. In the New Testament, Mark addresses the individual read with the authorial aside, “let the reader understand” (ὁ ἀναγινώσκων νοείτω, Mark 13,14). Perhaps even more compelling is the example of the Ethiopian eunuch undertaking 17. L.W. HURTADO, Oral Fixation and New Testament Studies? ‘Orality’, ‘Performance’ and Reading Texts in Early Christianity, in NTS 60 (2014) 321-340, p. 322. 18. H.N. PARKER, Books and Reading Latin Poetry, in W.A. JOHNSON – H.N. PARKER (eds.), Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford, Oxford University Press, 2009, 186-229, p. 187. 19. HURTADO, Oral Fixation and New Testament Studies (n. 17), p. 324. 20. W. SHINER, Proclaiming the Gospel: First-Century Performance of Mark, Harrisburg, PA, Trinity Press International, 2003, p. 5. 21. B.J. WRIGHT, Communal Reading in the Time of Jesus: A Window into Early Christian Reading Practices, Minneapolis, MN, Fortress, 2017, p. 207. 22. In his chapter five Wright discusses a host of first-century examples of communal reading in both Graeco-Roman and Jewish texts. Further examples, from a wider period, are documented in the appendix to the volume. Ibid., pp. 63-115, 211-231.

22

P. FOSTER

individual reading is his chariot, “he was reading the prophet Isaiah” (ἀνεγίνωσκεν τὸν προφήτην Ἠσαΐαν, Acts 8,28). Again, this is a single individual reading a text. Examples of this phenomenon from the period can be multiplied when one looks beyond the confines of biblical texts23. What this range of evidence established is that in the first two centuries of the Christian era there is evidence both for the phenomenon of public reading of texts to a gathered group of co-religionists, and also there exists evidence for the practice of individual reading of early Christian texts. Both of these practices can be seen as taking place in the wider cultural society of the period. By contrast, while some have inferred that early Christians transmitted their traditions through the medium of dramatic performance, and that texts served simply as a basis for this performative event, there is no surviving evidence that supports early Christian texts being used in this way in the same period. It is difficult to know precisely when to date the first recorded evidence of Christian biblical texts being performed in any meaningful sense. Perhaps it is not until one finds the emergence of the phenomenon of passion or mystery plays. It is perhaps the case that this tradition grew out of medieval liturgies designed for use during the major Christian festivals where multiple voices were used for the characters in the text. If this is the case then the earliest certain examples post-date the composition of the gospels by at least a millennium. By contrast, the earliest use of the gospels reveals that they were read in a fairly fixed form, and it will be argued that they were also written by authors drawing on literary predecessor texts with the expectation that these texts would be read either in a collective community, or privately by individuals24. III. TRIPLE TRADITION PASSAGES The correct way to compare and enumerate similarity and difference in the synoptic tradition is not a straightforward task. The division of an entire gospel narrative into separate units or pericopae involves some degree of subjectivity. Moreover, counting the degree of correspondence involves decisions concerning how to handle case changes, transpositions of word order, insertion of redactional material, and even minor spelling variation. 23. For instance, see the following discussions: B.M.W. KNOX, Silent Reading in Antiquity, in Greek, Roman and Byzantine Studies 9 (1968) 421-435; F.D. GILLIARD, More Silent Reading in Antiquity: Non omne verbum sonat, in JBL 112 (1993) 689-696; and A.K. GAVRILOV, Techniques of Reading in Classical Antiquity, in CQ 47 (1997) 56-73. 24. HURTADO, Oral Fixation and New Testament Studies (n. 17), p. 340.

THE SYNOPTIC PROBLEM

23

Some of these issues are more substantial than others. However, together these factors and others complexify the task of providing an accurate measure of the degree of similarity and variability between parallel synoptic passages. Rather than re-open and redefine a new set of statistical measures, the following discussion will make use of Morgenthaler’s existing analysis25. As with any attempt to describe the synoptic tradition some of the individual ways to quantify the data could be questioned. That cannot totally be avoided. On the positive side, however, utilising Morgenthaler’s data provides a set of statistical measures that has established itself as a standard within the field. For that reason, as well as its accessibility to wider scholarship, it will be used as the basis for the comparative analysis that follows. Morgenthaler divides the Markan narrative (Mark 1,1–16,8a) into 128 units26. These units vary considerably in length. Several of the shortest contain a single Markan verse27. The shortest of these contains only eight Markan words. The longer units are far more extensive. Several of the units are more than one hundred words in length, with the three longest being in excess of three hundred Markan words. In ascending order these three units are the Gerasene Demoniac (Mark 5,1-21) with 346 words; the story of Jairus’ daughter containing the intercalation of the woman with the haemorrhage (Mark 5,22-48) with 353 words; and the controversy over handwashing (Mark 7,1-23) with 359 words. Units of greater extent provide more opportunity for variation between parallel accounts due to their length. Furthermore, short single verse units are often pithy aphorisms, which by their nature are not always as susceptible to the same types of reworking as narrative accounts. Therefore, it must be acknowledged from the outset that limitations exist due to the nature of the dataset itself. This is not fatal for analysis. It simply means that close consideration must be given to multiple factors when drawing inferences based on this comparative data. Of the 128 units into which Morgenthaler divides the Markan text, he finds 118 with parallels in Matthew, and 96 with parallels in Luke. However, the Markan material omitted by Matthew and Luke differs. This means that there are 91 passages that may be counted as strict triple tradition units. That is material found in Mark that has a parallel version both in Matthew and in Luke. A selection of these passages will form the basis of the analysis that follows. 25. R. MORGENTHALER, Statistische Synopse, Zürich – Stuttgart, Gotthelf, 1971. 26. Ibid., pp. 227-229. 27. Examples of single-verse units include Mark 1,39.45; 3,22; 6,17; 7,31; 8,31; 9,42; 11,11; 12,12; 14,26.65; 15,1.33.

24

P. FOSTER

Illustrative of the difficulties in accurately quantifying the level of similarity or difference is the Markan passage dealing with the time of the parousia (Mark 13,30-32), along with its Matthean and Lukan parallels (Matt 24,3436//Luke 21,32-33). Matt 24,34-36

Mark 13,30-32

ἀμὴν λέγω ὑμῖν ὅτι οὐ μὴ ἀμὴν λέγω ὑμῖν ὅτι οὐ μὴ παρέλθῃ ἡ γενεὰ αὕτη ἕως παρέλθῃ ἡ γενεὰ αὕτη μέχρις οὗ ταῦτα πάντα ἂν πάντα ταῦτα γένηται. γένηται. ὁ οὐρανὸς καὶ ἡ γῆ παρε- ὁ οὐρανὸς καὶ ἡ γῆ παρελεύσεται, οἱ δὲ λόγοι μου λεύσονται, οἱ δὲ λόγοι μου οὐ μὴ παρελεύσονται. οὐ μὴ παρέλθωσιν. Περὶ δὲ τῆς ἡμέρας ἐκείνης Περὶ δὲ τῆς ἡμέρας ἐκείνης καὶ ὥρας οὐδεὶς οἶδεν, οὐδὲ ἢ τῆς ὥρας οὐδεὶς οἶδεν, οἱ ἄγγελοι τῶν οὐρανῶν οὐδὲ οἱ ἄγγελοι ἐν οὐρανῷ οὐδὲ ὁ υἱός, εἰ μὴ ὁ πατὴρ οὐδὲ ὁ υἱός, εἰ μὴ ὁ πατήρ. μόνος

Luke 21,32-33 ἀμὴν λέγω ὑμῖν ὅτι οὐ μὴ παρέλθῃ ἡ γενεὰ αὕτη ἕως ἂν πάντα γένηται. ὁ οὐρανὸς καὶ ἡ γῆ παρελεύσονται, οἱ δὲ λόγοι μου οὐ μὴ παρελεύσονται.

Morgenthaler presents the comparison between these three passages in the following form. He first notes that the Markan tradition contains 49 words, the Matthean tradition 50 words, and the Lukan tradition 27 words. He then notes that of Matthew’s 50 words, 40 are shared with Mark, which is an 80% agreement. In comparison, of Luke’s 27 words, 24 are shared with Mark, which implies an 88.9% agreement. However, there is a problem here. Luke has completely omitted the material in Mark 13,32, that is 22 of the 49 Markan words, or 44.9% of the words contained in Mark 13,30-3228. In many ways the saying in Mark 13,32, while being thematically related to the topic of the passage, is nonetheless not as tightly linked as the material in Mark 13,30-31. If those two verses had been taken as forming the unit of tradition for comparison, then in relation to the triple tradition parallel there would be 27 Markan words, 28 Matthean words, and 27 Lukan words. Matthew has 25 words in common with Mark, but within these shared words Mark writes ταῦτα πάντα, whereas Matthew has the inverted for πάντα ταῦτα. This is a difference, but how is one to measure that, if at all, given that word order is less significant in Greek? Similarly, the final word in the two-verse parallel is the future middle indicative παρελεύσονται in Mark 13,31 whereas in Matthew it is the aorist active subjunctive 28. MORGENTHALER, Statistische Synopse (n. 25), p. 239.

THE SYNOPTIC PROBLEM

25

παρέλθωσιν (Matt 24,35). Again, there is a judgment to be made as to the degree of significance of this change. Matthew’s form may reflect the grammatical conventions of expressing an emphatic negative future condition by employing the double negative with the aorist subjunctive, οὐ μὴ παρέλθωσιν. Mark’s construction, by contrast, is tolerable, but reflects a slightly barbarised use of Greek29. Further complicating the issue is the fact that there exist a number of significant textual variants in the Matthean text. The original hand of Codex Sinaiticus omits Matt 24,35 in its entirety. That is almost certainly a scribal slip. More significantly a number of manuscripts read παρελεύσονται as the Matthean form of the verb30. While there is an impressive range of manuscripts that read the future middle indicative παρελεύσονται form in Matt 24,35, this is probably due to the evolution of the Greek language which tended to drop the emphatic negative subjunctive in favour of using future forms of the verb. It is less likely that the Matthean text was assimilated to the Markan form, since the Matthean form of the text tended to predominate. Hence these manuscripts coincidentally appear to conform the form of the verb to the Markan parallel probably under the influence of the contemporary usage of their own day. This is probably an indication that at this much later stage of transmission the Synoptic Gospels were revised in a literary and scribal environment rather than through an oral and performative medium. While this example has revealed the complexity of measuring the level of textual agreement or difference several important observations can be made in regard to this unit of tradition contained in Mark 13,30-32 and parallels. First, the level of verbatim agreement is high across those elements of the tradition that are replicated in the Matthean and Lukan versions of the tradition. Second, the differences in wording can potentially be explained on the basis of improving the stylistic form of the text. This can be most readily explained in terms of literary reuse of the Markan text. Third, Luke’s omission of Mark 13,32 is most likely due to a theological sensibility concerning the statement that the son does not know the hour of the Parousia31.

29. For a discussion of the emphatic negation subjunctive that also discusses the case of Mark 13,31 and Matt 26,35 see D.B. WALLACE, Greek Grammar: Beyond the Basics, Grand Rapids, MI, Zondervan, 1995, p. 468. 30. The manuscripts that read παρελεύσονται include ℵ2a K W Γ Δ Θ ƒ1.13 565. 579. 700. 1241. 1424. ℓ 2211 𝔐 lat, whereas the form παρέλθωσιν is read by ℵ2b B D L 33. 892 e; Irlat vid. 31. The omission of Mark 13,32 is discussed by R. PESCH, Naherwartungen. Tradition und Redaktion in Mk 13 (Kommentare und Beiträge zum Alten und Neuen Testament), Düsseldorf, Patmos, 1968, p. 195.

26

P. FOSTER

A second example of a different kind that is helpful for discussing the relationship between the synoptic accounts is the story of the Gerasene demoniac (Mark 5,21). This is the third longest of the Markan tradition units according to Morgenthaler’s division32. This tradition is handled in markedly different ways in Matthew (Matt 8,28-34) and Luke (Luke 8,26-39). The Markan version of this story contains 346 words, the Matthean version is heavily abbreviated to 146 words, while the Luke version although abbreviated in length is not shortened to the same degree at 308 words. According to Morgenthaler of Matthew’s 146 words there are 55 words identical with the Markan version of the tradition, or 37.7%. Whereas with Luke’s 308 words there are 124 identical to Mark’s form of the tradition, or 40.3%. The degree of correspondence to Markan words is, therefore, broadly similar in Matthew and Luke despite their different levels of abbreviation of the Markan story. However, since Luke has preserved more of the base text the degree of correspondence could be judged to be much higher. Many have seen the Markan story as having a complex tradition history and reflecting a convoluted set of intertwined themes. Thus, Joel Marcus writes: “[t]here are several features in the story that suggest it is not an original unity but has grown by a process of accretion that has left some loose ends dangling”33. Due to these narratological loose threads this was a story that benefitted from editorial reworking. Moreover, the Matthean version of the story reflects several Matthean redactional modifications and concerns. Of particular note, is that here is an example of Matthew’s tendency to double characters in a story. In this case Matthew modifies the Markan story from a healing of one demoniac to that of two. This change has ramifications for measuring the degree of correspondence, since there is a concomitant change of singular verb forms to plural forms. For instance, Matthew reads καὶ ἰδοὺ ἔκραξαν λέγοντες (Matt 8,29a) in place of καὶ κράξας φωνῇ μεγάλῃ λέγει (Mark 5,7a). Also what is said immediately following this is modified to reflect the plural subjects: τί ἡμῖν καὶ σοί (Matt 8,29b) instead of τί ἐμοὶ καὶ σοί (Mark 5,7b). Therefore, the change in the number of demoniacs introduces consequential changes in the ensuing narrative. By purely counting identical words this suggests a lower degree of correspondence even in those portions of the tradition that Matthew reproduces. Matthew also appears to have a concern with the potential inaccuracy of Markan geography. The inland location of Gerasa, about thirty-three miles 32. MORGENTHALER, Statistische Synopse (n. 25), p. 228. 33. J. MARCUS, Mark 1–8: A New Translation with Introduction and Commentary (AB, 27), New Haven, CT, Yale University Press, pp. 199, 347

THE SYNOPTIC PROBLEM

27

south east of the Sea of Galilee, seems to defy the logic of the story with Jesus’ arrival by boat. Matthew alters the place name to Gadara, which is still problematic because it is six miles inland but perhaps less difficult than the location of Gerasa34. In many ways, what is important here is not an exact account of what motivated these various changes – some large, others fairly minute. Rather, what each of these examples reveals is that across the triple tradition passages the changes that do occur are best understood as the result of an editor or redactor working with a predecessor text and introducing changes at a literary level, and not through the process or oral transmission. IV. DOUBLE TRADITION PASSAGES While the term “double tradition” could refer to any passage that is found in two Synoptic Gospel accounts such as Matthew and Mark but not Luke, or Mark and Luke but not Matthew, the term has come to designate a more specific set of material. The term “double tradition” is used here in its standard technical sense to denote material common to Matthew and Luke, but which is not found in the common source Mark. Frequently, this double tradition material has formed the basis of the Q hypothesis and it is the bedrock for establishing the contents of a reconstructed Q source. Here, however, the term will be used initially to denote the material shared by Matthew and Luke which is not in common with Markan traditions. The purpose is to analyse this material in order to assess whether more sense can be made of this material on the basis of the assumption that it came to the Matthean and Lukan evangelists through an oral or a literary channel, or even via some combination of these two processes. Thus, the term “double tradition” is being used here precisely to denote material common to Matthew and Luke without a Markan parallel, and furthermore, in this context, the designation does not assume or require any source critical hypothesis.

34. Davies and Allison note that, “[i]t is, therefore, possible that the story concerned not the Gadarenes or the Gerasenes but rather the Gergasenes, that is the people of Gergesa. Gergesa was on the eastern coast of the city of Galilee, and the pre-Markan tradition could have confused it with Gerasa (a city much better known). How then explain Matthew’s reading? If the First Evangelist knew that Gerasa was too far inland for the requirements of the story, he might have substituted a place closer to the lake and have come up with Gadara”. W.D. DAVIES – D.C. ALLISON, A Critical and Exegetical Commentary on the Gospel according to Saint Matthew (ICC), Edinburgh, T&T Clark, 1988-97, vol. 2, p. 77.

28

P. FOSTER

Without over-generalizing, scholars have often seen more of this double tradition material as potentially originating with the historical Jesus, than other elements of the gospel tradition such as the triple tradition or especially Johannine material. For what it is worth, the Jesus Seminar classed Jesus material in four categories using colours to denote these categories. The traditions with the highest likelihood of originating with Jesus were assigned red as the colour. This corresponded to a saying that was understood with the rubric “Jesus undoubtedly said this or something very like it”35. There are also traditions that describe Jesus’ actions rather than his sayings. They could also be classified as red, with the rubric “I would include this item unequivocally in the database for determining who Jesus was”36. Of the fifteen traditions given a ranking of “red”, all were sayings or words of Jesus. More importantly for current purposes of these fifteen sayings eleven occur in double tradition material (described as Q). Of the other four, two are unique Lukan parables (the Samaritan, Luke 10,30-35; and the Shrewd manager, Luke 16,1-8a), one is a unique Matthean parable (the Vineyard labourers, Matt 20,1-15), and one is triple tradition material with a parallel in the Gospel of Thomas (Render to Caesar, Mark 12,17b and parallels)37. Of the seventy-six sayings given a pink rating, “Jesus probably said something like this”38, forty-two occur in double tradition material (with some having multiple attestation in other sources such as the Gospel of Thomas)39. The implications of this data need to be carefully considered. If these sayings originated with Jesus, then they almost certainly originated in occasions of oral teaching. Without entering into the debate of whether or not Jesus spoke some Greek, it strongly appears to be the case that he taught in Aramaic, as attested by a few words and phrases preserved in Aramaic in the gospel tradition40. Again, what is not in dispute is the almost self-evident claim that authentic Jesus material in the gospels originated in occasions of oral teaching. That is not the focus of this discussion. Instead, the concern is to explain the interrelationships gospels and the medium of their composition. Here double tradition material is highly illuminating.

35. R.W. FUNK – R.W. HOOVER (eds.), The Five Gospels: The Search for the Authentic Words of Jesus, New York, Macmillan, 1993, p. 36. 36. Ibid. 37. Ibid., p. 549. 38. Ibid., p. 36. 39. Ibid., pp. 549-553. 40. That is not to assume that Jesus necessarily said the words preserved in Aramaic. Rather the evangelists believed that such Aramaic words placed on the lips of Jesus gave the story an air of verisimilitude, presumably because it was widely known or believed that Jesus was primarily an Aramaic speaker.

29

THE SYNOPTIC PROBLEM

Consideration of various individual units of double tradition is illuminating. It is helpful to begin with one of the classic pieces of evidence used in this discussion, the parallel pericopae dealing with the preaching of John the Baptist. This is not, in fact, the double tradition with the highest level of verbatim correspondence. Rather, what makes this striking is the high level of correspondence over an extended portion of text, alongside the differences in the respective evangelists’ redactional introductions. Matt 3,7-10

Luke 3,7-9

Ἰδὼν δὲ πολλοὺς τῶν Φαρισαίων καὶ Σαδδουκαίων ἐρχομένους ἐπὶ τὸ βάπτισμα αὐτοῦ εἶπεν αὐτοῖς· γεννήματα ἐχιδνῶν, τίς ὑπέδειξεν ὑμῖν φυγεῖν ἀπὸ τῆς μελλούσης ὀργῆς; ποιήσατε οὖν καρπὸν ἄξιον τῆς μετανοίας καὶ μὴ δόξητε λέγειν ἐν ἑαυτοῖς· πατέρα ἔχομεν τὸν Ἀβραάμ. λέγω γὰρ ὑμῖν ὅτι δύναται ὁ θεὸς ἐκ τῶν λίθων τούτων ἐγεῖραι τέκνα τῷ Ἀβραάμ. ἤδη δὲ ἡ ἀξίνη πρὸς τὴν ῥίζαν τῶν δένδρων κεῖται· πᾶν οὖν δένδρον μὴ ποιοῦν καρπὸν καλὸν ἐκκόπτεται καὶ εἰς πῦρ βάλλεται.

Ἔλεγεν οὖν τοῖς ἐκπορευομένοις ὄχλοις βαπτισθῆναι ὑπ’ αὐτοῦ· γεννήματα ἐχιδνῶν, τίς ὑπέδειξεν ὑμῖν φυγεῖν ἀπὸ τῆς μελλούσης ὀργῆς; ποιήσατε οὖν καρποὺς ἀξίους τῆς μετανοίας καὶ μὴ ἄρξησθε λέγειν ἐν ἑαυτοῖς· πατέρα ἔχομεν τὸν Ἀβραάμ. λέγω γὰρ ὑμῖν ὅτι δύναται ὁ θεὸς ἐκ τῶν λίθων τούτων ἐγεῖραι τέκνα τῷ Ἀβραάμ. ἤδη δὲ καὶ ἡ ἀξίνη πρὸς τὴν ῥίζαν τῶν δένδρων κεῖται· πᾶν οὖν δένδρον μὴ ποιοῦν καρπὸν καλὸν ἐκκόπτεται καὶ εἰς πῦρ βάλλεται.

The following differences need to be observed. In Matthew’s version of this tradition, the subjects of the Baptist’s withering speech are his frequent pairing of opponents, the Pharisees and Sadducees (six times). It appears to be the case that Matthew has modified the earlier wording of the double tradition to have Pharisees and Sadducees castigated by John for their hypocritical behaviour41. That wording would replace what is presumably the earlier stage of the tradition with its non-specific reference to “crowds” ὄχλοις (Luke 3,7). However, in the direct speech that follows, which recounts the words of the Baptist, there are only three differences across sixty-four Lukan words, or sixty-three Matthean words. These are first Matthew used the singular form καρπὸν ἄξιον, whereas Luke has the plural expression καρποὺς ἀξίους. The second difference is the reading δόξητε in Matt 3,9, whereas Luke uses the term ἄρξησθε in Luke 3,8. Both verbs are second person plural aorist 41. Fleddermann comes to the same judgment. He states, “[s]ince Luke mentions neither group, we should trace both the Pharisees and Sadducees to Matthew”. H.T. FLEDDERMANN, Q: A Reconstruction and Commentary (BiTS, 1), Leuven – Paris – Dudley, MA, Peeters, 2005, p. 214.

30

P. FOSTER

subjunctive forms. Third, the final minor difference is that Luke uses two consecutive conjunctions, δὲ καί (Luke 3,9), whereas Matthew has only the first, δέ (Matt 3,10). Determining the originality of either of these three alternatives is not straightforward. It appears that the singular form καρπὸν ἄξιον has a stronger claim to originality on the basis of the reuse of the singular form Matt 3,10//Luke 3,9, although contrariwise one could argue that Matthew has made the pericope more consistent by introducing this change. The double conjunction δὲ καί appears to be a form favoured by Luke, so the originality of the single conjunction δέ is favoured. The choice between the two verbs, δόξητε in Matt 3,9 or ἄρξησθε in Luke 3,8, is more difficult. Harry T. Fleddermann favours the Matthean form both on linguistic grounds42, and also due to intention of the tradition. In regard to that second indicator he argues, “Matthew’s verb accurately expresses the intention of Q in this sentence, for it unmasks the religious claim to privilege that Q attacks in this passage – ‘do not presume to say’”43. In many ways it is relatively unimportant for the present discussion to determine the exact wording of the earliest form of this tradition. However, this discussion does highlight the closeness of the parallels even in small linguistic details, and furthermore that even the minor differences can best be explained on the basis of the redactional preferences of the two evangelists. This level of similarity has important implications for determining the origin of the synoptic tradition as it came to the evangelists. The high level of similarity is to be found in the Greek text of a tradition that may have originated in Aramaic. This strongly suggest that the tradition had attained a level of fixity by the stage it was transmitted in Greek. It is not plausible to suggest that Lukan and Matthean versions are the result of oral transmission of a source that was independently translated into respective Greek versions by the evangelists or their tradents. There is a clear and demonstrable literary relationship between these parallel accounts of the same tradition. Moreover, there is no evidence to support the supposition that the Greek version of the tradition ever underwent circulation in oral form. Discussing this unit of tradition, John Kloppenborg observes that This can be explained only on the supposition that Matthew and Luke used a written source. Hypotheses that have tried to trace the double tradition material exclusively to oral tradition have either ignored or minimized these agreements or have made romantic, but quite unrealistic, assumptions about the nature and faithfulness or oral tradition44.

42. Ibid., pp. 216-217. 43. Ibid., p. 217. 44. J.S. KLOPPENBORG, Excavating Q: The History and Setting of the Sayings Gospel, Minneapolis, MN, Fortress, 2000, p. 57.

31

THE SYNOPTIC PROBLEM

Here the concern is not to advocate for the Q hypothesis. Alternative literary explanations of the dependence of Luke and Matthew regardless of the direction of dependence could also explain the high level of verbatim agreement. For instance, a literary explanation is required if one adopts the hard form of the Farrer Hypothesis, as advocated by Michael Goulder45. The preaching of the Baptist would be a tradition that derived from the redactional creativity of Matthew, but of which the written form was copied and lightly modified by Luke. It is also helpful to consider double tradition material that purports to preserves the words of Jesus rather than those of the Baptist. A representative example of broadly the same length as the preaching of John the Baptist is the dominical saying concerning the return of the unclean spirit (Matt 12,43-45//Luke 11,24-26). This saying was ranked pink for its Lukan version and grey for its Matthean version by the Jesus’ seminar46. Despite this relatively low ranking, even in relation to the Matthean version of the tradition it is noted that “[t]he inversion of ordinary Judean and Christian opinion about exorcism makes it difficult to attribute this saying to anyone other than Jesus. Its perspective is compatible with Jesus’ attitude elsewhere towards defilement”47. Therefore, this tradition has a reasonable claim to originating with Jesus in some form. The parallel versions are as follows: Matt 12,43-45

Luke 11,24-26

Ὅταν δὲ τὸ ἀκάθαρτον πνεῦμα ἐξέλθῃ ἀπὸ τοῦ ἀνθρώπου, διέρχεται δι’ ἀνύδρων τόπων ζητοῦν ἀνάπαυσιν καὶ οὐχ εὑρίσκει. τότε λέγει· εἰς τὸν οἶκόν μου ἐπιστρέψω ὅθεν ἐξῆλθον· καὶ ἐλθὸν εὑρίσκει σχολάζοντα σεσαρωμένον καὶ κεκοσμημένον. τότε πορεύεται καὶ παραλαμβάνει μεθ’ ἑαυτοῦ ἑπτὰ ἕτερα πνεύματα πονηρότερα ἑαυτοῦ καὶ εἰσελθόντα κατοικεῖ ἐκεῖ· καὶ γίνεται τὰ ἔσχατα τοῦ ἀνθρώπου ἐκείνου χείρονα τῶν πρώτων. οὕτως ἔσται καὶ τῇ γενεᾷ ταύτῃ τῇ πονηρᾷ.

Ὅταν τὸ ἀκάθαρτον πνεῦμα ἐξέλθῃ ἀπὸ τοῦ ἀνθρώπου, διέρχεται δι’ ἀνύδρων τόπων ζητοῦν ἀνάπαυσιν καὶ μὴ εὑρίσκον· [τότε] λέγει· ὑποστρέψω εἰς τὸν οἶκόν μου ὅθεν ἐξῆλθον· καὶ ἐλθὸν εὑρίσκει σεσαρωμένον καὶ κεκοσμημένον. τότε πορεύεται καὶ παραλαμβάνει ἕτερα πνεύματα πονηρότερα ἑαυτοῦ ἑπτὰ καὶ εἰσελθόντα κατοικεῖ ἐκεῖ· καὶ γίνεται τὰ ἔσχατα τοῦ ἀνθρώπου ἐκείνου χείρονα τῶν πρώτων.

45. M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989. 46. FUNK – HOOVER (eds.), The Five Gospels (n. 35), p. 552. 47. Ibid., p. 189.

32

P. FOSTER

The most obvious difference is the final aphorism in the Matthean version of this tradition, οὕτως ἔσται καὶ τῇ γενεᾷ ταύτῃ τῇ πονηρᾷ (Matt 12,45b). This appears to be an obvious redactional addition designed to link this story more tightly to the foregoing pericope which uses the denouncing form of address γενεὰ πονηρὰ καὶ μοιχαλίς (Matt 12,39). The level of correspondence is self-evident. There are several minor differences. These include additions or adjustments of vocabulary (indicated in bold) as well as transpositions48. This leads Davies and Allison to observe that, “[g]iven the close verbatim agreement, one must suppose either a common written source (Q) or one evangelist’s knowledge of the other”49. Although Matthew has added a final saying to tie this tradition to preceding material in his narrative arrangement, the image of the unclean spirit aligns the perspective in the double tradition on the need of disciples to withstand any desire to return to evil. The call for constancy and single-minded commitment are consonant with the outlook of the double tradition, and perhaps with Jesus himself. However, given the extremely high alignment of the two forms in Greek of a tradition which may have originated in the Aramaic teaching of Jesus in some form, one is left with the conclusion that the stability of the Greek tradition had occurred at least at the initial translational stage of this tradition, after it was translated and transmitted as a literary unit in Greek. Therefore, both Matthew and Luke appear to have known and received this tradition in a literary form written in Greek. Numerous other examples with similarly high levels of verbal correspondence in the double tradition could be presented. However, it would be misleading to suggest that all double tradition parallels exhibit equally high levels of correspondence50. It is instructive to consider alongside this a parallel tradition with low verbatim agreement. Two of the large parables typically assigned to the double tradition might be worthy of consideration. However, there is doubt as to whether the Parable of the Talents/Pounds 48. These changes are explained in detail by FLEDDERMANN, Q: A Reconstruction and Commentary (n. 41), p. 489. 49. DAVIES – ALLISON, Matthew (n. 34), vol. 2, p. 333. 50. On Kloppenborg’s estimate there are 63 double tradition pericopae. These are divided into quintiles of percentage agreement. Of these 11 have agreement in the range 80-98%, 15 in the range 60-79%, 16 in the range 40-59%, 14 in the range 20-39%, and 8 in the range 0-19%. Presumably the figure 0% is simply used to complete the division into quintiles as a tradition with zero correspondence could not be classified as a parallel version of the same underlying tradition. Although the parallel between Matt 12,11b ἔσται ἐξ ὑμῶν ἄνθρωπος ὃς ἕξει πρόβατον ἓν καὶ ἐὰν ἐμπέσῃ τοῦτο τοῖς σάββασιν εἰς βόθυνον, οὐχὶ κρατήσει αὐτὸ καὶ ἐγερεῖ; and Luke 14,5b τίνος ὑμῶν υἱὸς ἢ βοῦς εἰς φρέαρ πεσεῖται, καὶ οὐκ εὐθέως ἀνασπάσει αὐτὸν ἐν ἡμέρᾳ τοῦ σαββάτου; reproduce much the same idea with little correspondence. KLOPPENBORG, Excavating Q (n. 44), p. 63.

33

THE SYNOPTIC PROBLEM

(Matt 25,14-30//Luke 19,12-27) or the Wedding Banquet (Matt 22,1-14// Luke 14,16-24) reflect versions of the same written tradition or whether they each reflect two separate but related parables, perhaps spoken by Jesus in different moments of his teaching ministry. For this reason, a more representative tradition drawn from widely accepted double tradition material, but having lower agreement is the promise to the disciples that they will sit on twelve thrones in the coming eschatological age (Matt 19,27-30// Luke 22,28-30). This selection also has the advantage of being of a similar length as the previous two examples in this section, which is thus beneficial for comparative purposes. This tradition does contain several differences that present genuine difficulties in accounting for the motivations for those different forms and therefore making it challenging to identify the earlier or more primitive version of this pericope. It is precisely those complicating factors that make this a useful example to consider. Here, because they are fewer in number, the similarities are presented in bold typeface. Matt 19,27-30

Luke 22,28-30

Τότε ἀποκριθεὶς ὁ Πέτρος εἶπεν αὐτῷ· ἰδοὺ ἡμεῖς ἀφήκαμεν πάντα καὶ ἠκολουθήσαμέν σοι· τί ἄρα ἔσται ἡμῖν; ὁ δὲ Ἰησοῦς εἶπεν αὐτοῖς· ἀμὴν λέγω ὑμῖν ὅτι ὑμεῖς οἱ ἀκολουθήσαντές μοι ἐν τῇ Υμεῖς δέ ἐστε οἱ διαμεμενηκότες μετ’ παλιγγενεσίᾳ, ὅταν καθίσῃ ὁ υἱὸς τοῦ ἐμοῦ ἐν τοῖς πειρασμοῖς μου· κἀγὼ διατίθεμαι ὑμῖν καθὼς διέθετό μοι ὁ ἀνθρώπου ἐπὶ θρόνου δόξης αὐτοῦ, πατήρ μου βασιλείαν, ἵνα ἔσθητε καὶ πίνητε ἐπὶ τῆς τραπέζης καθήσεσθε καὶ ὑμεῖς ἐπὶ δώδεκα μου ἐν τῇ βασιλείᾳ μου, καὶ καθήσεσθε θρόνους κρίνοντες τὰς δώδεκα φυλὰς ἐπὶ θρόνων τὰς δώδεκα φυλὰς κρίνοντες τοῦ Ἰσραήλ. τοῦ Ἰσραήλ. καὶ πᾶς ὅστις ἀφῆκεν οἰκίας ἢ ἀδελφοὺς ἢ ἀδελφὰς ἢ πατέρα ἢ μητέρα ἢ τέκνα ἢ ἀγροὺς ἕνεκεν τοῦ ὀνόματός μου, ἑκατονταπλασίονα λήμψεται καὶ ζωὴν αἰώνιον κληρονομήσει. πολλοὶ δὲ ἔσονται πρῶτοι ἔσχατοι καὶ ἔσχατοι πρῶτοι.

Even measuring the level of correspondence is difficult in this case. The material in Matt 19,27-30 forms a unit, but it is equally obvious that Matt 19,29-30 derives from Mark 10,29-31, and Luke replicates a version of this Markan tradition elsewhere (Luke 18,29b-30). Moreover, the opening of the Matthean pericopae appears to be a reworking of Mark 10,29a

34

P. FOSTER

(cf. Luke 18,29a). Therefore, the double tradition material contained in Matthew appears to consist of Matt 19,28b only, with the remainder of the Matthean tradition derived from Mark. Hence this is an example where the Matthean evangelist has done something he frequently does – namely conflated a pithy double tradition saying into a larger Markan context. By contrast, the Lukan version consists of a three-verse unit of which the middle verse (Luke 22,29) has no Matthean parallel. Fleddermann attributes heavy redactional work to Luke both in the opening to the tradition and through the insertion of the middle verse. Thus, he states “Luke had to alter the original construction because he added the sentence about Jesus conferring a kingdom (Luke 22,29)”51. The middle verse (Luke 22,29) is also excluded by Heil and others from being part of the text of a reconstructed Q52. Of course, strictly speaking, it is not double tradition material, since it is a tradition that does not occur in both Matthew and Luke. However, it is not possible to exclude the case that Luke 22,29 was in fact part of the tradition known to Luke and Matthew, and that Matthew simply did not reproduce this saying since he simply harvested the unit for the saying about judging the twelve tribes. That excision was motivated by the fact that the harvested saying produced the best thematic support for the Markan tradition into which in was intercalated. However, even when all these uncertainties are taken into account, and although reconstruction of the underlying tradition may remain allusive in this case, there remains a striking verbal correspondence in the parallel versions of the twelve tribes saying. By itself, this might not be substantial enough to establish a literary rather than an oral channel of transmission to the evangelists. When, however, this example is taken in collaboration with the earlier examples drawn from double tradition material where the high levels of verbatim agreement suggested literary dependence there may be a plausible way forward. If for much of the material shared uniquely by Matthew and Luke a strong case can be mounted for a literary relationship between the two forms of the double tradition, and if for a case like Matt 19,28b//Luke 22,28-30 a literary relationship remains a possibility, then that explanation is to be favoured over the subsidiary and isolated hypothesis that this saying came to the respective evangelists in oral form.

51. FLEDDERMANN, Q: A Reconstruction and Commentary (n. 41), p. 865. 52. P. HOFFMANN – S.H. BRANDENBURGER – U. BRAUNER – T. HIEKE, and (volume editor) C. HEIL, Q 22:28, 30. You Will Judge the Twelve Tribes of Israel (Documenta Q), Leuven, Peeters, 1999.

THE SYNOPTIC PROBLEM

35

Consequently, the double tradition material provides some of the strongest evidence for the Synoptic Gospel tradition circulating among the evangelists in literary form, regardless of which synoptic theory one supports. First, a number of the traditions have strong claim to originate with the setting of the ministry of Jesus of Nazareth. Given that figures such as John the Baptist and Jesus are highly likely to have taught in Aramaic during the Galilean phase of the ministry, the fixity of the tradition cannot be explained by the accurate memorization of the words of Jesus or John in Aramaic, and subsequently the independent but corresponding translations of those words into oral Greek sayings. The earliest Greek phase of the gospel tradition has a strong claim to being literary, and the differences in the versions attested in Matthew and Luke can be seen as arising from the respective theological interests of the evangelists, from their differing compositional techniques, and from the literary conventions of the day which cherished some adaptation of predecessor texts by those creating new compositions. The question concerning which literary synoptic theory best accounts for the data of the double tradition is not addressed fully in this discussion. Nonetheless, a few observations emerge. There can be little doubt of the prior existence of Mark before the composition of Matthew and Luke. This is especially clearly seen in the final case with the saying concerning judging the twelve tribes. Whatever was Matthew’s source for this saying, it has been conflating with a pre-existing and more extensive tradition drawn from Mark 10,29-31. In terms of whether the double tradition material is the result of Matthew drawing on Luke’s additions to Mark, Luke drawing on Matthew’s additions to Mark, or both Matthew and Luke drawing upon a common source, the combined evidence of the three examples appears to be suggestive and helpful. In the case of the parallel versions of the Baptist’s preaching in the few cases of textual variation Matthew appeared to preserve a form of the tradition that was more original than that found in Luke. The evidence in the two forms of the tradition concerning the unclean spirit was mixed. It is more likely that Matthew preserves the original finite verb οὐχ εὑρίσκει (Matt 12,43) and that Luke has replaced this with a participle form μὴ εὑρίσκον (Luke 11,24). By contrast, it appears that Matthew has modified the original form by adding the term σχολάζοντα (Matt 12,44). Thus, at different points both Matthew and Luke appear to preserve the more original form of the tradition. Lastly with the saying concerning judging the twelve tribes, regardless of the decision concerning the origin of Luke 22,29, Matthew appears to have heavily modified the tradition harvesting it for the saying concerning the twelve tribes and increase the parallelism by referring to the

36

P. FOSTER

numerically precise twelve thrones. Here, throughout this unit, Luke appears to preserve the earlier form of the tradition contained in Luke 22,28.30. One is then driven to consider the often-cited criterion of “alternating primitivity”53, often discussed by supporters of the Q hypothesis that is a key element of the Two-Source Hypothesis. It appears that the phenomenon of sometimes Matthew having the more primitive version of double tradition material, while at other times it is Luke is best explained by appeal to a common source. V. THE ORIGIN OF THE MARKAN GOSPEL MATERIAL When discussing the use of triple tradition or double tradition material by Matthew or by Luke there is a clear methodological advantage. One is able to see how these two evangelists have, in the case of triple tradition material, recast and repackaged Markan material through direct comparison with the source. The double tradition also permits a high level of insight into the way these traditions were handled irrespective of which of the dominant synoptic theories is advocated. For either Matthean posteriority, or for the Farrer Hypothesis it is possible to compare the tradition in question with the earlier gospel from which it is seen to have derived. For those who hold to the Q hypothesis there is an extra step required. Since, on that account, both Matthew and Luke represent later reworkings of a now no-longer extant source, it is necessary to reconstruct the source on the basis of close comparison of the later versions of this double tradition material and through taking into account the respective redactional tendencies of the two evangelists. Obviously, there is a degree of subjective judgment in this process. However, there are some fine examples of how this reconstructive work can be carried out54, as well as the production of a compendium of alternative decisions on elements of the various units of tradition55. 53. For a discussion of the phenomenon see C.M. TUCKETT, Q and the History of Early Christianity: Studies on Q, Edinburgh, T&T Clark, 1996, p. 10. 54. Important examples include but are not limited to J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000; FLEDDERMANN, Q: A Reconstruction and Commentary (n. 41). 55. On the continuing work of collecting a wide-ranging set of scholarly decisions concerning the pre-Matthean and pre-Lukan form of the double tradition see the series, P. HOFFMANN – J.S. KLOPPENBORG – J. VERHEYDEN – C. HEIL (general eds.), Documenta Q, Leuven, Peeters, 1996-. There are several volumes currently in various stages of production.

THE SYNOPTIC PROBLEM

37

When it comes to describing the origin of the material contained in the Gospel of Mark the situation is much more complex and theories of origins of this material are diverse and often share little in common. The reason for this is obvious. Scholars engaged in this enquiry are seeking to delve behind what is probably the earliest extant stage of the gospel tradition56. Perhaps the most common proposal is that of a pre-Markan passion narrative, which was utilised by the author of the Gospel of Mark. However, opinions on the nature of the use of pre-Markan traditions in the Markan passion narrative varies greatly. At one end of the spectrum among those who detect pre-Markan tradition in the passion narrative are scholars who advocate that there might be a number of isolated and free-floating traditions employed by Mark, but that the evangelist was responsible for creating a connected and continuous narrative57. At the other end of the spectrum, some have argued for a single unified source that the evangelist takes over wholesale with little editorial input58. In essence the level of confidence in separating pre-Markan material from Markan redactional shaping in the passion narrative depends on several factors. These include the level of certainty one has in the ability to discern differences in style and vocabulary between putative source material in the passion narrative in comparison with other Markan material, the ability to identify narrative tensions between material in the passion narrative and preceding sections of the gospel, and recognition of proposed changes of perspectives in regard to theological themes59. Ultimately, a strong commitment to any claim to be able to separate the parameters of a pre-Markan passion narrative almost certainly goes beyond the limits of the available evidence. Notwithstanding this, it does appear plausible that Mark received some of the elements of his passion account from prior tradents, and that he did not create the entire story de novo. Marion L. Soards’ assessment is a little more positive. He concludes his discussion with the following assessment,

For more details see: https://neues-testament.uni-graz.at/de/fwf-projekte-zu-q/internationalesq-projekt/documenta-q/ 56. While the double tradition may well pre-date Markan material it is only known through its representation in later literary documents. 57. E. LINNEMANN, Studien zur Passionsgeschichte (FRLANT, 102), Göttingen, Vandenhoeck & Ruprecht, 1970. 58. W. SCHMITHALS, Das Evangelium nach Markus (ÖTKNT, II/1-2), Gütersloh, Mohn; Würzburg, Echter, 1979. 59. These factors and others are also described by M.L. SOARDS, Appendix IX: The Question of a Premarcan Passion Narrative, in R.E. BROWN, The Death of the Messiah: From Gethsemane to the Grave. A Commentary on the Passion Narratives in the Four Gospels (ABRL), Garden City, NY, Doubleday, 1994, vol. 2, pp. 1492-1524.

38

P. FOSTER

We may safely conclude that Mark uses a source in writing his PN. We know that source, however, only as incorporated in Mark. The greatest challenge that lies before us is not the separation of tradition from Marcan redaction: for, as our earlier work shows, that task may finally be an impossible one. Rather, we must investigate the rich layers of tradition that come to us in the form of the Marcan PN. This conclusion does not mean that we may simply discard any notion of editorial activity. It demands, however, that a preoccupation with the data of that editorial work not be our first concern60.

Thus, Soards is confident that a significant part of the Markan passion narrative derives from a pre-gospel source. Yet, the source and the evangelist’s redactional activity are so intertwined that in most cases separation is not possible. More recently, important questions have been raised concerning the criteria that are typically employed by those who seek to isolate a pre-Markan passion narrative. Reading Mark as a biography, Bond argues in line with other ancient bioi that perceived changes in content, vocabulary, narrative chronology, and themes “derive not from the use of the source, but the form of the material”61. This, however, does not mean that Bond rejects outright the presence of pre-Markan elements in the Markan passion narrative. Rather, she seeks to give greater weight to the role of Mark as an author of a bios and to see him continuing that work throughout the entirety of the narrative. However, alongside this careful compositional work, Bond posits that Mark was familiar with a number of pre-Markan traditions, and even goes so far as to state, “[s]ome of these interpretations may even have begun to circulate with a basic narrative framing”62. Other sections of Mark’s gospel have been identified as possibly stemming from blocks of pre-Markan tradition. Among the sections that are more frequently seen are drawing on prior material are the cycle of controversy stories in Mark 2,1–3,663, and the series of parables in Mark 4,1-3464. Without analysing these respective arguments in detail, the problems that surround these theories are largely similar to those involving a pre-Markan passion narrative. While it is probably plausible to see Mark drawing on a range of earlier Jesus traditions, and even to posit that these circulated in 60. Ibid., pp. 1523-1524. 61. H.K. BOND, The First Biography of Jesus: Genre and Meaning in Mark’s Gospel, Grand Rapids, MI, Eerdmans, 2020, p. 111. 62. Ibid., p. 112. 63. See J. DEWEY, Markan Public Debate: Literary Technique, Concentric Structure, and Theology in Mark 2:1–3:6 (SBL DS, 48), Chico, CA, Scholars, 1980. 64. P.H. SELLEW, Early Collections of Jesus’ Words: The Development of Dominical Discourses, Diss. Harvard Divinity School, Cambridge, MA, 1986.

THE SYNOPTIC PROBLEM

39

Greek in written form, precise identification of source material and its delineation from Mark’s compositional and redactional activity remains speculative and unconvincing. As Brown further observes is relation to the issue of Markan sources, “the uncertainty of source reconstruction makes their analyses very debatable”65. In regard to what can be said concerning the pre-Markan gospel tradition only a few observations can be drawn. First, the foregoing discussion shows the plausibility of gospel traditions in circulation in Greek and in written form prior to the composition of Mark’s Gospel. The discussion has focussed on material that is part of the triple tradition, and therefore this is the case in relation to material that does not stem from the double tradition. The long-standing debate concerning whether Mark had knowledge of Q is not the point of this analysis. However, there are several Mark-Q overlaps texts. These have been the focus of much discussion66. It is possible to explain these as arising due to Mark’s direct knowledge of Q, or stemming from a stage of tradition history prior to the period of the independent assimilation of this material into Q and Mark respectively. If the latter is correct, rather than the hypothesis of Mark’s knowledge of Q, then this might take one a stage further back in the circulation of some Jesus material that was to coalesce as gospel tradition. This, however, is highly speculative and nothing is based upon this possibility for the sake of this discussion. Furthermore, Mark shows distance from the Aramaic phase of the Jesus tradition by only using terms from that language in his gospel for a specific purpose. Aramaic words occur in a number of contexts in the Gospel of Mark (Mark 3,17; 5,41; 7,11.34; 14,36; 15,22.34). In one way or another, each of these instances introduces a Greek translational equivalent. The most minimal attempt to supply a translation is in relation to the use of “Abba” where Mark simply juxtaposes this with its Greek equivalent, αββα ὁ πατήρ (Mark 14,36). Presumably, by the time of the composition of the Gospel of Mark, the term αββα had become part of the sociolect of the early Jesus movement (cf. Rom 8,15; Gal 4,6) and thus little explanation was required67. The purposes for the other Aramaic words vary. On the one 65. R.E. BROWN, An Introduction to the New Testament (ABRL), New York, Doubleday, 1997, p. 156. 66. For instance, see H.T. FLEDDERMANN, Mark and Q: A Study of the Overlap Texts (BETL, 122), Leuven, Leuven University Press – Peeters, 1995. 67. A further example might include the use of the term “Hosanna” in Mark 11,9f. See V. TAYLOR, The Gospel according to St. Mark: The Greek Text with Introduction, Notes and Indexes, London, Macmillan, 21966, p. 296. Typically, proper names of individuals are not discussed in relation to Mark’s use of Aramaisms.

40

P. FOSTER

hand, the place name Golgotha for which Mark supplies the translation “Place of a Skull”, τὸν Γολγοθᾶν τόπον, ὅ ἐστιν μεθερμηνευόμενον Κρανίου Τόπος (Mark 15,22), may represent a fairly mundane use of a place name although Mark no doubt wanted readers to recognize its rather ominous sounding translation in Greek. Other occurrences may be motivated by various factors, and might not reflect theories that these words were either uttered by Jesus, or that the gospel tradition ever existed in a thoroughgoing Aramaic literary stage. Perhaps these words were intended to add an air of verisimilitude to their respective stories. Another possibility, as Marcus notes in relation to the phrase ταλιθα κουμ (Mark 5,41) is that “[t]he retention of Aramaic here is partly for effect: the exotic foreign words increase the sense of mystery about the miracle that is about to occur”68. The most extended piece of Aramaic occurs at the moment of Jesus’ final cry from the cross. Here Mark supplies both the Aramaic form of words, and a fairly literal rendering into Greek, ἐβόησεν ὁ Ἰησοῦς φωνῇ μεγάλῃ· ελωι ελωι λεμα σαβαχθανι; ὅ ἐστιν μεθερμηνευόμενον· ὁ θεός μου ὁ θεός μου, εἰς τί ἐγκατέλιπές με; (Mark 15,34). Collins argues that based on the Gethsemane story that was part of Mark’s passion source, the evangelist was not content to attribute a wordless cry to Jesus. Thus, Collins states, “[t]aking his cue from the depiction of Jesus in Gethsemane, Mark has Jesus speak the language of scripture in his last words”69. Moreover, Mark requires the Semitic version of the scriptural citation for his narrative sequence. The cry in Aramaic permits Mark to portray some of the onlookers as misconstruing Jesus’ cry (ελωι = “Eloi”) as invoking Elijah (Ἠλίας) for assistance (Mark 15,35-36). Therefore, the occurrence of Aramaic expressions in the Gospel of Mark is neither an indication of the oral transmission of gospel traditions nor of a phase when pre-gospel sources circulated in Aramaic literary forms. Instead, Aramaic is used for a number of reasons. These include mundane uses of names or terminology (Mark 3,17; 7,11; 15,22), as terms that had entered into the sociolect of the early Jesus movement (Mark 11,9; 14,36), for the exoticisation of miracle stories (Mark 5,41; 7,34), or in order to place a scripture cry on the lips of Jesus that leads to a confusion in the understanding of hears (Mark 15,34). All this data is consistent with the literary formation of the gospel tradition in Greek. Moreover, the necessity for Mark to render translations of each of these Aramaic terms attests the presence of

68. MARCUS, Mark 1–8 (n. 33), p. 363. 69. A. YARBRO COLLINS, Mark (Hermeneia), Minneapolis, MN, Fortress, 2007, p. 753.

THE SYNOPTIC PROBLEM

41

Greek readers who did not have the level of bilingualism necessary to navigate between the two languages. Finally, those who argue for the transmission of the gospel tradition in performative oral contexts have to account for the aside in the Markan eschatological discourse, ὁ ἀναγινώσκων νοείτω (Mark 13,14). With this incidental comment, the author of the gospel demonstrates the expectation that the message would be received by a “reader”. It is not possible to tell if what is envisaged in community or individual reading. Regardless of this, Mark anticipates the communication and reception of the text of his gospel to occur through the medium of the reading of a written Greek text. Wright favours seeing this as a reference to communal reading70, and given what is known of literacy levels in the ancient world and the collective nature of early Christian meetings that may indeed be the better interpretation. Furthermore, drawing upon the wider context of the passage, Wright makes the following observation “Mark’s initial expectation that his gospel will be read communally, then, is not difficult to imagine, especially when Mark also recounts Jesus’ prediction that places of communal reading events (synagogues in general) will be one of the places his disciples will be beaten in the future (13:9)”71. Hence, the Gospel of Mark provides strong evidence that the author expected it to be received via acts of reading. Analysis of the text also suggests that certain traditions that it utilised came to the evangelist as written Greek material. Here, however, caution must be exercised. Beyond that assertion, the means to distinguish between source traditions and redactional material do not typically exist. In part, this is due to Mark’s authorial skill in taking earlier material and weaving it with his own theological insights and creativity to produce a flowing narrative with its own central purpose. VI. SONDERGUT TRADITIONS Typically, Sondergut traditions are understood to be those gospel traditions found in either the Gospel of Matthew or the Gospel of Luke, which were not derived from either triple or double tradition material, nor are they likely due to the editorialising of the evangelist. In other words, the term refers to traditions that may have come to the author through an unknown channel, rather than to have been created by the Matthew or Luke themselves. In 70. WRIGHT, Communal Reading in the Time of Jesus (n. 21), p. 125. 71. Ibid., p. 126.

42

P. FOSTER

terms of wider source critical theories, in the Four-Source Hypothesis this material would have formed the basis of the putative L and M sources72. Since little is known of these traditions prior to their appearance in the gospels of Matthew or Luke, it is necessary to exercise caution when stating anything concerning their origins or form. For certain scholars these singly attested traditions might in some cases be Q traditions preserved only by one evangelist. Petros Vassiliades identifies a set of criteria for assigning Sondergut material to the Q source73. However, many of these criteria appear to involve judgments on which significant disagreement could occur. For instance, the Matthean saying concerning going the second mile (Matt 5,41) does occur sandwiched between two double tradition sayings and it is thematically related to those sayings (Matt 5,39b-40.42//Luke 6,29-30). So, it is possible that Luke omitted what would have been a central saying in a block of Q teaching. On the other hand, Matthew does like to conjoin thematically related material from his sources, or even to redactionally enlarge the perspectives of source material. Given that the majority of the larger block of material in Matt 5,21-48 is uniquely Matthean, one can see the evangelist at work in this section and Matt 5,41 might also be a Matthean creation or a combination of a tradition from another source. The same problem arises when discussing certain Lukan traditions such as the lost coin (Luke 15,8-10). It is difficult to determine whether this is a Q tradition that Matthew has not replicated74, or whether this reflects Luke’s tendency to link related parables regardless of whether Luke created this tradition or it came to him through a different channel75. The point here is not that of establishing which of these opinions is correct, but rather to show that it is difficult to determine the correct account of the tradition history of these units. What is noteworthy, 72. For the classic statement of the Four-Source Hypothesis see B.H. STREETER, The Four Gospels: A Study in Origins, London, Macmillan, 1924, esp. pp. 198, 208-214, 254-268. See also P. FOSTER, The M-Source: Its History and Demise in Biblical Scholarship, in ID. – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 591-616. 73. P. VASSILIADIS, The Nature and Extent of the Q Document, in NT 20 (1978) 49-73. 74. Among others, this is the view of C. HEIL, Lukas und Q. Studien zur lukanischen Redaktion des Spruchevangeliums Q (BZNW, 111), Berlin – New York, De Gruyter, 2003, pp. 153-158. 75. By contrast, Fleddermann views this text as a Lukan creation based on vocabulary and motifs draw from the preceding parable of the lost sheep (Matt 18,12-14//Luke 15,4-7). Thus, Fleddermann states, “[a]s a pragmatic, opportunistic redactor Luke borrow motifs and vocabulary from Q and combines them with his own characteristic motifs and vocabulary when he wants to expand Q with additional material”. FLEDDERMANN, Q: A Reconstruction and Commentary (n. 41), p. 773.

THE SYNOPTIC PROBLEM

43

however, is that both of these proposals understand the tradition as originating in written form either as a Lukan composition or as part of the pre-gospel source Q. More problematic is the large block of material in Matthew 23. There appears to be much happening here in terms of tradition history. This material can be divided into three discrete, but related blocks. First, the warnings against religious leaders (Matt 23,1-12) is primarily unique Matthean material (Matt 23,2-3.5.7b-10), but not exclusively so since Matt 23,6-7a.11 have Markan parallels (Mark 12,38b-39; 9,35) and Matt 23,4 has a parallel with Luke 11,46. Despite this careful weaving of some pre-existing perspectives the material appears largely a Matthean composition, replete with Matthean perspectives. Others have suggested the material that is unique to this unit came to the evangelist as a pre-formed literary piece. Thus, Davies and Allison state, Matthew, one suspects, took up a pre-Matthean text on Christian piety, broke it in two, and deposited the first half in chapter 6, the other half in chapter 23. This would explain not only the intriguing parallels but the fact that both passages seem to be composed primarily not of sayings of Jesus but of Christian admonitions. In any case it is nearly certain that both originated in the same Jewish-Christian community, one with close ties to the synagogue76.

While one might debate whether this material was Matthean composition and thus came into existence as the evangelist’s literary product, or a pre-existing text, in both scenarios a literary model offers the most plausible explanations for the origin of this material as it is found in Matthew 23. The second part of Matthew 23 is formed from the series of seven woes in Matt 23,13-3377. There are parallels between several of the Matthean woes and similar traditions in the Gospel of Luke. However, the relationship is complex, and in this case it seems difficult to argue for the dependence of one gospel upon the other. However, there are difficulties that the parallels pose for the Two-Source Hypothesis also. Both Matthew and Luke have divergent forms of this tradition. So, one is left with the notion that distinct redactional aims were functioning for the two evangelists. Thus, 76. DAVIES – ALLISON, Matthew (n. 34), vol. 3, p. 266. 77. Seven woes are to be enumerated since Matt 23,14 is almost certainly a later addition to the Matthean text. As Bruce Metzger observes, “That ver. 14 is an interpolation derived from the parallel in Mark 12,40 or Luke 20,47 is clear (a) from its absence in the earliest and best authorities of the Alexandrian and the Western types of text, and (b) from the fact that the witnesses that include the passage have it in different places, either after ver. 13 (so the Textus Receptus) or before ver. 13”. B.M. METZGER, A Textual Commentary on the Greek New Testament, Stuttgart, New York; Deutsche Bibelgesellschaft, United Bible Societies, 2 1994, p. 50.

44

P. FOSTER

Kloppenborg reasons that “Matthew’s focus on hypocrisy and Luke’s interest in almsgiving are part of larger redactional strategies visible … in their editing of Q”78. However, several scholars believe the differences here are too great to allow the case of independent editing of a common source. Therefore, some argue that Luke draws on the Q version of this tradition, while Matthew employs a Jewish-Christian document. In respect to this latter source, Newport argues that it is largely a unified and coherent source, which “has been lightly edited by a later redactor (almost certainly Matthew himself)”79. Thus, once again while the tradition history is uncertain, nonetheless the various alternatives seem to demand a literary explanation to account for the channel by which this material came to Matthean evangelist. The final element in Matthew 23 (Matt 23,34-39) comprises two double tradition units that Matthew has brought together. These are Matt 23,34-36// Luke 11,49-51 and Matt 23,37-39//Luke 13,34-35. Here there is a reasonably high level of verbatim agreement80. So here again, it is possible to see that these traditions have come to Matthew in written form. The evangelist’s main contribution is to bring together these two blocks of double tradition material. As was stated at the outset of this section, it is wise to refrain from making dogmatic statements concerning the origin of Sondergut material in either Matthew or Luke. Despite this difficulty in identifying exact channels of transmission, it appears most plausible in the cases examined to conclude that the traditions considered reached the evangelist in written form, or that they are the product of the evangelist’s written redactional activity. VII. CONCLUSION This discussion concerns itself with a particular area of investigation. That is namely the origin and development of the synoptic tradition. Specifically, it is interested in probing the question concerning the form in which the evangelists received the traditions they employed to form their respective gospel narratives. Of course, several of these traditions had a complex pre-gospel life. Moreover, those that have a strong claim of being able to be traced back to the historical Jesus almost certainly originated in 78. KLOPPENBORG, Excavating Q (n. 44), p. 76. 79. K.G.C. NEWPORT, The Sources and Sitz im Leben of Matthew 23 (SupplJSNT, 117), Sheffield, Sheffield Academic Press, 1995, p. 182. 80. See FLEDDERMANN, Q: A Reconstruction and Commentary (n. 41), pp. 544, 701.

THE SYNOPTIC PROBLEM

45

Aramaic and in oral form. Beyond those twin assertions it is difficult to say anything further about the earliest stages of those traditions. By contrast, when one comes to examine the form in which material reached the synoptic evangelists a far greater degree of confidence is possible. For Matthew and Luke, it is possible to be confident that the majority of the traditions they used reached them in Greek and in written form. Some of the other material in their gospel accounts is due to the written redactional activity of these two evangelists. This editorial activity should be understood through a literary and compositional lens81. As has been noted, it is more challenging to describe the form in which the traditions employed by Mark reached the earliest of the synoptic evangelists. There is, however, little to support the theory that the gospel was composed orally, or in the context of performance82. There are no early examples of Christian literature being composed in this manner. By contrast, although caution is required, there are reasons to suspect that units of material came to Mark already written in Greek. It is now no longer possible to separate those units of tradition from Markan editorial or compositional work. This attests to Mark’s skill as one composing his own narrative account of the ministry and death of Jesus. What this suggests is that exotic theories of chains of oral transmission, which supposedly even undergirded the translation jump when some of these traditions may have moved from Aramaic to Greek are unnecessary and remain implausible in accounting for the formation of the Synoptic Gospel tradition83. Rather, the question of the formation of the Synoptic Gospels must first and foremost be considered in relation to literary mechanics of the transmission of traditions in Greek. On this point proponents of the Two-Document Hypothesis, the Farrer Hypothesis, and advocates of Matthean posteriority are in broad and well-founded agreement. School of Divinity, University of Edinburgh United Kingdom

PAUL FOSTER [email protected]

81. For instances, see N. PERRIN, What Is Redaction Criticism? (Guides to Biblical Scholarship, New Testament Series), Philadelphia, PA, Fortress, 1969, pp. 65-67. 82. For a recent restatement of the role of performance in the transmission of gospel traditions see K.R. IVERSON, Performing Early Christian Literature: Audience Experience and Interpretation of the Gospels, Cambridge, Cambridge University Press, 2021. 83. See again HURTADO, Oral Fixation and New Testament Studies (n. 17), pp. 322-324.

AFTER MARKAN PRIORITY: GENERAL ESSAYS

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

I. INTRODUCTION What is the “end” (τέλος) of Q (Studies)? In Rom 10,4, Paul famously claims that “Christ is the end of [the] Law” (τέλος νόμου Χριστὸς), leaving behind a legacy of exegetical ambiguity insofar as it remains unclear whether what Paul intended to say was that Christ was the goal or purpose of the Law or whether Paul thought that Christ was the terminus of the Law, that is, its abrogation and ultimate demise1. The title of the present essay appeals to this double entendre in its contemporary reflection on the current state of Q Studies2. It is not, then, a call to end Q Studies – or a kind of prophetic prediction of the (imminent) end of Q – although it does comment on such calls in current synoptic studies debates. Rather, this essay represents an attempt to review the field’s past accomplishments and present conditions by questioning prevailing assumptions regarding its 1. For the former, see G.E. HOWARD, Christ the End of the Law: The Meaning of Romans 10:4ff, in JBL 88 (1969) 331-337; W.S. CAMPBELL, Christ the End of the Law: Romans 10.4, in E.A. LIVINGSTONE (ed.), Studia Biblica 1978: III. Papers on Paul and Other New Testament Authors. Sixth International Congress on Biblical Studies (SupplJSNT, 3), Sheffield, JSOT, 1978, pp. 73-81; R. BADENAS, Christ: The End of the Law: Romans 10.4 in Pauline Perspective (SupplJSNT, 10), Sheffield, JSOT, 1985, pp. 81-115; G.N. DAVIES, Faith and Obedience in Romans: A Study in Romans 1–4 (SupplJSNT, 39), Sheffield, Sheffield Academic Press, 1990, pp. 185-189. Along with Campbell, see E.P. SANDERS, Paul, the Law, and the Jewish People, Minneapolis, MN, Fortress, 1983, pp. 39-40; R. JEWETT, The Law and the Coexistence of Jews and Gentiles in Romans, in Interpr 39 (1985) 341-356; T.R. SCHREINER, Paul’s View of the Law in Romans 10:4–5, in WestTJ 55 (1993) 113-135; J.P. HEIL, Christ, the Termination of the Law (Romans 9:30 – 10:8), in CBQ 63 (2001) 484-498. On the possibility of combining both meanings, see C.K. BARRETT, The Epistle to the Romans (BNTC, 6), Oxford, Oxford University Press, 1991, pp. 197-198; M.D. HOOKER, Christ: The ‘End’ of the Law, in D.E. AUNE – T. SELAND – J.H. ULRICHSEN (eds.), Neotestamentica et Philonica: Studies in Honor of Peder Borgen (SupplNT, 106), Leiden – Boston, MA, Brill, 2005, 126-146. 2. Rhetorical appeals to the “end” of particular themes, topics, and traditions often play on this allusive ambiguity of meaning. See, for example, W.C. SMITH, The Meaning and End of Religion, Minneapolis, MN, Fortress, 1991; S. HARRIS, The End of Faith: Religion, Terror, and the Future of Reason, New York, W.W. Norton, 2004; J.W. LOFTUS (ed.), The End of Christianity, Amherst, NY, Prometheus, 2011.

50

S.J. JOSEPH

ideological and theological significance, as well as its intersections with contemporary critical reflections on ethnicity, race, gender, and class. The past, of course, comes first. Since the founding of the International Q Project in 1983, and up until quite recently, the Two-Document Hypothesis (2DH) has enjoyed a relatively unrivalled dominant position in synoptic studies. One could call this the Golden Age of Q Studies3. There was a time when one could (almost) assume that the 2DH had become the solution to the Synoptic Problem, that its formative stratigraphical layers could be identified (as sapiential)4, that its geographical provenance in rural Galilee could be adduced5, that it provided relatively reliable access to the historical Jesus6, that generic comparisons between Q and the Gospel of Thomas provided both the justification and the evidentiary warrant for the existence of Q7, that the “difference” Q made could be a useful analytical tool in the redescription of Christian origins8, with Q being deployed to dissociate a Galilean Jesus movement from the Judean Jerusalem ekklēsia9, 3. My own Q Studies began upon entering the doctoral program in Religion/New Testament at Claremont Graduate University in 2004. James M. Robinson had recently retired, but he agreed to direct my Q Studies as well as serve on my dissertation committee in 2010. The subsequent closing of the Institute for Antiquity and Christianity, the restructuring of the School of Religion, and Robinson’s passing in 2015, represent, to my mind, the end of an era. 4. J.S. KLOPPENBORG, The Formation of Q: Trajectories in Ancient Christian Wisdom Collections (Studies in Antiquity and Christianity), Philadelphia, PA, Fortress, 1987. 5. ID., City and Wasteland: Narrative World and the Beginning of the Sayings Gospel (Q), in Semeia 52 (1991) 145-160; ID., Literary Convention, Self-Evidence, and the Social History of the Q People, in Semeia 55 (1991) 77-102; ID., Conflict and Invention: Recent Studies on Q, in ID. (ed.), Conflict and Invention: Literary, Rhetorical and Social Studies on the Sayings Gospel Q, Valley Forge, PA, Trinity Press International, 1995, 1-14; J.L. REED, The Social Map of Q, ibid., 17-36; W.E. ARNAL, Jesus and the Village Scribes: Galilean Conflicts and the Setting of Q, Minneapolis, MN, Fortress, 2001; G. BAZZANA, Kingdom of Bureaucracy: The Political Theology of Village Scribes in the Sayings Gospel Q (BETL, 274), Leuven – Paris – Bristol, CT, Peeters, 2015. 6. J.M. ROBINSON, The Jesus of the Sayings Gospel Q, in IAC Occasional Papers 28, Claremont, CA, Claremont Graduate School, 1993; ID., The Gospel of Jesus: A Historical Search for the Original Good News, New York, Harper One, 2005; A. LINDEMANN (ed.), The Sayings Source Q and the Historical Jesus (BETL, 158), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001. For a note of caution, see J.S. KLOPPENBORG, The Sayings Gospel Q and the Quest of the Historical Jesus, in HTR 89 (1996) 307-344. 7. S.J. PATTERSON, The Gospel of Thomas and Jesus (Foundations and Facets), Sonoma, CA, Polebridge, 1993; ID., The Lost Way: How Two Forgotten Gospels Are Rewriting the Story of Christian Origins, San Francisco, CA, Harper One, 2014. 8. R. CAMERON – M.P. MILLER (eds.), Redescribing Christian Origins (SBL SS, 28), Atlanta, GA, Society of Biblical Literature, 2004. 9. The Redescribing Christian Origins project appealed to Q as an alternative, “different” source and perspective on early Christian origins. See D.E. SMITH, What Do We Really Know about the Jerusalem Church? Christian Origins in Jerusalem according to Acts and Paul, in

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

51

that analogical comparisons with Q’s formative stratum and the Greek philosophical school of Cynicism could illuminate the early Jesus movement10, and that we could re-construct a “critical edition” of Q – based on the assumption that the redactional fingerprints of Matthew and Luke’s use of Mark could guide the interpretation of what they may have done with Q – alongside other text-critical productions in the field of New Testament studies. Yet ever since the publication of The Critical Edition of Q in 2000 – certainly one of the major milestones in contemporary Q Studies – a number of developments have indicated that all is not well in Q’s Kingdom. Much of this exuberant hegemony came to an end with the publication of Mark Goodacre’s Case Against Q11. An advocate of the Farrer Hypothesis (FH), which posits Markan Priority combined with Luke’s use of Matthew12, Goodacre launched an assault on Q that seemed eminently reasonable insofar as it appealed to the principle of Ockham’s Razor of not unnecessarily multiplying hypotheticals and recognized that Luke admitted to utilizing “many” sources, with Mark – and Matthew – being candidates for inclusion. Matthew and Luke’s infancy narratives, after all, are structurally similar in significant ways despite their differences. The Farrer Hypothesis – insofar as it represents a reasonable, relatively simple, and apparently economical solution to the Synoptic Problem – has proven to be a formidable opponent of the 2DH. For many, it is now far easier to doubt the existence of Q than to defend it13. CAMERON – MILLER (eds.), Redescribing Christian Origins (n. 8), 237-252; B.L. MACK, A Jewish Jesus School in Jerusalem, ibid., 253-262; L.H. MARTIN, History, Historiography, and Christian Origins: The Jerusalem Community, ibid., 263-275. 10. B.L. MACK, The Lost Gospel: The Book of Q and Christian Origins, New York, HarperSanFrancisco, 1993; L.E. VAAGE, Galilean Upstarts: Jesus’ First Followers according to Q, Valley Forge, PA, Trinity Press International, 1994; ID., Q and Cynicism: On Comparison and Social Identity, in R.A. PIPER (ed.), The Gospel Behind the Gospels: Current Studies on Q (SupplNT, 75), Leiden – New York – Köln, Brill, 1994, 199-229. 11. M.S. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002. See now also, e.g., F. WATSON, Q as Hypothesis: A Study in Methodology, in NTS 55 (2009) 397-415; ID., Gospel Writing: A Canonical Perspective, Grand Rapids, MI, Eerdmans, 2013, 117-155; E. EVE, Relating the Gospels: Imitation, Memory, and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021. 12. A. FARRER, On Dispensing with Q, in D.E. NINEHAM (ed.), Studies in the Gospels: Essays in Memory of R.H. Lightfoot, Oxford, Blackwell, 1955, 55-88. 13. J.G. CROSSLEY, Jesus and the Chaos of History: Redirecting the Life of the Historical Jesus, London, Oxford University Press, 2015, p. 45, for example, eschews any particular solution to the Synoptic Problem, preferring to leave “to one side the question of whether it [Luke 11,2-4] is a Q passage or Luke using Matthew”. Similarly (ibid., p. 70), when discussing sayings

52

S.J. JOSEPH

For many, Q can now be safely demoted to the status of an intellectual relic, a make-believe text of exegetical threads that once held together illusory dreams of access to an original, ancient past. For many, Q Studies should have no future, but should rather be relegated to the dustbins of outdated theories. And so it is that after a century of scholarly consensus on Q’s theoretical existence and historical significance as an ancient Jewish source of early Jesus tradition, the lost source is now increasingly being regarded as a lost cause. It may be tempting to think that this backlash against Q is simply the inevitable succession of a superior hypothesis (the FH) slowly eliminating its primary rival (the 2DH). To be sure, dispensing with Q does seem to be a more economical hypothesis than positing the hypothetical existence of a lost source, but transferring the Double Tradition to the Gospel of Matthew by positing new pre-Matthean sources simply re-names Q as Matthew14. At the same time, appreciating Luke’s creativity as an author does not necessarily explain why Luke changed so much of Matthew15. The case against Q can be seen both as an academic challenge to a scholarly construct as well as a public discourse disseminated via social media and the Internet. In this latter context, Q Studies (and Q scholars!) have suffered from a public relations problem. I have noted elsewhere the curious correlation between Q skepticism and the rise of biblio-blogging16. It is tempting, therefore, to speculate about how Q skepticism makes strange bedfellows in forging alliances between Synoptic Problem specialists, Internet atheists and free-thinkers, Jesus Mythicists, conservative Evangelicals, and the general public.

related to poverty and wealth, he makes “a point which stands irrespective of whether we accept some form of Q or not”. 14. R. CARRIER, On the Historicity of Jesus: Why We Might Have Reason for Doubt, Sheffield, Sheffield Phoenix Press, 2014, pp. 269-270, 470-474; ID., The Backwards and Unempirical Logic of Q Apologetics, https://www.richardcarrier.info/archives/19380 (January 30, 2022) [accessed February 11, 2022], argues that Q “should have been abandoned by the Biblical Studies field decades ago” because “we can prove conclusively that Luke used Matthew as a source”. Carrier’s argument – that because Luke’s use of Matthew is sufficiently probable and detectable there is simply no need for Q – fails to provide persuasive explanatory arguments for Luke’s use of Matthew. Rather, Luke’s use of Matthew is assumed until proven otherwise. 15. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, p. 308: the FH’s allegedly “greater economy is a sham: its claims for parsimony at one end must always be paid for by complicated accounts of Luke’s utilization at the other end”. Cf. O. ANDREJEVS, Reception of the Twelve in Matthew and Luke: Comparing the Current Synoptic Hypotheses, in ExpT 133 (2022) 233-244. 16. S.J. JOSEPH, Exit the ‘Great Man’: On James Crossley’s Jesus and the Chaos of History, in JSHJ 16 (2018) 3-22, p. 7.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

53

If it is difficult to infer any particular ideological motive informing the current climate of Q skepticism, there is no question that the FH’s programmatic denial of the existence of a non-canonical and non-orthodox source of early Jewish Jesus tradition appeals both to those who affirm the canonical authority of the Gospels and those who deny them as evidence for the existence of an historical Jesus17. The “conceptual stakes” involved in eradicating Q as an inconvenient site of “difference” can also be measured in terms of how the Farrer Hypothesis allows for Luke’s use of Matthew to represent a pro-Pauline, Gentile-friendly universalist “Christian” perspective to correct and/or replace Matthew’s more “Jewish” reading of the Jesus movement18. If Luke knew and used Matthew, then the New Testament’s suppression of its Jewish roots becomes more pronounced. Luke’s literary creativity would now be pressed into service as an apologetic reaction to Matthew, an attempt to re-write the “most Jewish Gospel” from the perspective of a Pauline Gentile. Luke’s use of Matthew would thus represent an artful subversion of the Jewish roots of the Jesus movement. To be sure, increasing skepticism regarding the Two-Document Hypothesis19, sustained critiques of the evidentiary presuppositions underlying the hypothesis20, facile assumptions conflating Q and Jesus21, and multiple competing hypotheses delineating Q’s provenance, compositional history, and sociological profile have undermined scholarly confidence in Q. Accordingly, there will probably never be an “end” to Q skepticism. Yet scholarly interest in the Synoptic Problem continues apace22. For many, Q continues 17. On at least part of the extant Didache as providing “examples” of Q-material, see A. GARROW, An Extant Instance of ‘Q’, in NTS 62 (2016) 398-417. 18. On “conceptual stakes,” see J.S. KLOPPENBORG, Conceptual Stakes in the Synoptic Problem, in M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018, 13-42. 19. BAZZANA, Kingdom of Bureaucracy (n. 5), p. 2: “scholars are becoming increasingly hesitant to make of the Sayings Gospel an object of sustained historical inquiry, since both the existence and the precise redactional contours of this writing must remain necessarily hypothetical”. Bazzana suggests that “the proponents of such skeptical views can hardly hide the fundamentally ideological motives that motivate their objections. In particular, Q’s representation of a decidedly non-kerygmatic Jesus is obviously disturbing for traditional reconstructions of the historical development of early Christianity”. 20. GOODACRE, The Case Against Q (n. 11); WATSON, Q as Hypothesis (n. 11); ID., Gospel Writing (n. 11), pp. 117-155; EVE, Relating the Gospels (n. 11). 21. See M. GOODACRE, Why Are Scholars So Keen on Q?, a short video made for the Center for the Study of Christian Origins, New College, University of Edinburgh (November 25, 2017), http://www.christianorigins.div.ed.ac.uk/2017/11/25/why-are-scholars-so-keenon-q-mark-goodacre/ [accessed January 11, 2018]. 22. J. SCHRÖTER, Memory, Theories of History, and the Reception of Jesus, in JSHJ 16 (2018) 85-107, p. 102: “What the historian cannot do, however, is abdicate the role of offering some explanation of the sources and relating them in one way or another to the past narrated by them”.

54

S.J. JOSEPH

to be a textual puzzle of endless fascination23. Q, after all, is a (re-)constructed text requiring the theoretical sophistication to recognize that one is working with a literary and theoretical construct that is discursively “real” insofar as it is a text agreed to by its critics and interpreters. Despite the bluster coming from some quarters, then, “[d]efinitive arguments” for or against any particular source-critical solution to the Synoptic Problem have not been forthcoming24. Consequently, “it is best to avoid such terms like ‘disproved’ and ‘discredited’ in relation to any synoptic hypothesis. Such terms are at best rhetorical, attempting to manufacture unearned credibility for one’s own hypothesis, and at worst simply fallacious”25. At present, Q Studies represents a contested field characterized by inconclusive – but “troubling” – results26. Q’s rhetorical power as the earliest “source” of Jesus tradition continues to render it a polarized site of “difference”27. It is tempting, therefore, to conclude that contemporary Q skepticism has less to do with the apparent and perceived inadequacies of the 2DH’s proposed solutions to synoptic puzzles and far more to do with the last three decades of critical explorations of Q’s “difference” that continue to raise concerns about its potential challenges to the traditional narrative of Christian origins28. II. THEOLOGY AND IDEOLOGY In the summer of 2019, a European conference investigated whether “Q scholars are driven by unscholarly projections or prejudices”29. A wideranging number of studies conducted critical analyses of the history of 23. See, for example, MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 18); C. HEIL – G. HARB – D.A. SMITH (eds.), Built on Rock or Sand? Q Studies: Retrospects, Introspects and Prospects (BiTS, 34), Leuven – Paris – Bristol, CT, Peeters, 2018. 24. KLOPPENBORG, Conceptual Stakes in the Synoptic Problem (n. 18), p. 13: “I am not of the view that definitive arguments for or against any of these hypotheses or variations on these are in fact possible. At best, we can propose compositional scenarios that account for most of the data, most of the time, by what we think of as the ‘most probable’ explanations”. 25. Ibid., pp. 21-22. 26. W.E. ARNAL, The Trouble with Q, in Forum: Foundations and Facets (2013) 7-79. 27. D.A. SMITH, What Difference Does Difference Make? Assessing Q’s Place in Christian Origins, in W.E. ARNAL – R.S. ASCOUGH – R.A. DERRENBACKER, Jr. – P.A. HARLAND (eds.), Scribal Practices and Social Structures Among Jesus Adherents: Essays in Honour of John S. Kloppenborg (BETL, 285), Leuven – Paris – Bristol, CT, Peeters, 2016, 183-211. 28. R.F. WALSH, Q and the ‘Big Bang’ Theory of Christian Origins, in B.S. CRAWFORD – M.P. MILLER (eds.), Redescribing the Gospel of Mark (ECL, 22), Atlanta, GA, Society of Biblical Literature, 2016, 483-533. 29. M. TIWALD (ed.), The Q Hypothesis Unveiled: Theological, Sociological, and Hermeneutical Issues behind the Sayings Source (BWANT, 225), Stuttgart, Kohlhammer, 2020.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

55

research related to the Synoptic Problem30, British skepticism31, Protestant German history32, French and Roman Catholic interests33, and the sociological profile of Q’s “itinerant radicalism”34. More recently, a similar volume based on papers presented at the Leuven Centre for the Study of the Gospels in 2017 explored the theoretical and theological interests embedded in synoptic studies35. A number of papers insinuated that “theological” interests were present in this sub-discipline, but none of the contributors discussed, disclosed, or problematized their own theological locations. Mark Goodacre, for example, has never publicly disclosed his own socio-religious location vis-à-vis Christianity, but has confidently claimed that other scholars are “so keen on Q” because it gives them “a route straight back to the historical Jesus … the chance to hear the actual words of the real Jesus”36. Similarly, Ronald Charles criticizes John Kloppenborg for not disclosing his subject location vis-à-vis theology and ideology, but fails to disclose his own religious, ideological, and theological locations as well37. To be sure, some scholars have made public statements of self-disclosure regarding their personal, ideological, and theological orientation(s) to Q Studies38. On the other hand, the categorical ambiguity of the term “theology” is all too capable of being weaponized in contemporary scholarship39. 30. J. SCHRÖTER, Key Issues Concerning the Q Hypothesis: Synoptic Problem, Verbal Reconstruction, and the Message of Jesus, ibid., 18-40. 31. C.M. TUCKETT, The Reception of Q Studies in the UK: No Room at the Inn?, ibid., 62-85. 32. L. BORMANN, Das Interesse an Markuspriorität, Logienquelle und Zweiquellentheorie im deutschen Protestantismus des 19. Jahrhunderts, ibid., 41-61. 33. M. TIWALD, The Investment of Roman Catholics in the 2DH and Q, ibid., 129-145; J. VERHEYDEN, Introducing ‘Q’ in French Catholic Scholarship at the Turn of the 19th and 20th Century: Alfred Loisy’s Évangiles synoptiques, ibid., 146-174. 34. G. THEISSEN, Itinerant Radicalism: The Origin of an Exegetical Theory, ibid., 175-190; M. FRENSCHKOWSKI, Itinerant Charismatics and Travelling Artisans – Was Jesus’ Travelling Lifestyle Induced by His Artisan Background?, ibid., 191-222. 35. J.S. KLOPPENBORG – J. VERHEYDEN (eds.), Theological and Theoretical Issues in the Synoptic Problem (LNTS, 618), London – New York, Bloomsbury T&T Clark, 2020. 36. GOODACRE, Why Are Scholars So Keen on Q? (n. 21). 37. R. CHARLES, Q as a Question from a Postcolonial Point of View, in Black Theology: An International Journal 7 (2009) 182-199, does, however, state that he is “a student from Haiti” (p. 194) and a “minority scholar” (p. 196). Nonetheless, his critique of Kloppenborg suggests that “this kind of overlook comes as a great disappointment” (p. 188). Cf. ibid., p. 194: “he [Kloppenborg] does not provide or take any clear position where he himself stands theologically and ideologically”. 38. S. PARKS, Gender in the Rhetoric of Jesus: Women in Q, Lanham, MD, Lexington – Fortress Academic, 2019, p. 6, for example, programmatically acknowledges her own “situatedness as a white ecumenical Christian Canadian, a critical biblical scholar, and a feminist”. 39. S.E. ROLLENS, review of Simon J. Joseph, The Nonviolent Messiah: Jesus, Q, and the Enochic Tradition, in RBL (June 2015), https://www.sblcentral.org/home/bookDetails/9811 [accessed March 2022], repeatedly refers to my “theological perspective”,

56

S.J. JOSEPH

It is important, therefore, to maintain a distinction between “confessional theology” and “non-confessional theology”, the latter being an indispensable critical exploration of the emic terms, categories, and language utilized by ancient authors in ancient texts40. At the same time, it is important to recall that even purportedly a-theological or non-theological orientations are not immune from pursuing ideological interests. Leif Vaage, for example, has recently expressed his disappointment in the direction Q Studies have taken as well as the field’s apparent failure to revolutionize the study of Christian origins41. The “end” of Vaage’s own Q Studies thus represents a reflective and retrospective overview of the field since the 1980s/90s. Recalling how Q “seemed to augur an interesting opportunity to test the strength of more than one foundational feature or ‘theorem’ of modern biblical scholarship regarding earliest Christianity”42, Vaage describes how “the topic of Q once had promised the possibility of … the prospect of a very different kind of early Christianity, no longer reflecting just the ‘kerygmatic’”43. Following the publication of John Kloppenborg’s Formation of Q, which identified Q’s earliest stratum as sapiential, it seemed that what was “at stake” in the discussion was “the primacy of apocalyptic as the mother of early Christian theology or some form of eschatology as the necessary horizon of early Christian discourse”. Yet since this debate failed to reach “a clear conclusion”, Q studies have subsequently “rendered such an ‘excited’ exchange essentially passé”44. Q Studies have suffered a “slow but steady eclipse” of scholarly “interest”. Vaage argues that a significant impetus in this lagging interest was the “identification of the sapiential genre of Q’s formative stratum with the social institution of ancient scribalism”. This apparent departure from (or development of) the argument in “theological interpretation”, “theological investigation”, and theological “endgame”, which she claims to be to “reveal a nonviolent historical Jesus”. 40. On this critical distinction, see D. WIEBE, The Failure of Nerve in the Academic Study of Religion, in StudRel/SciRel 13 (1984) 401-422, p. 403: “I find it helpful to distinguish ‘confessional theology’ from ‘non-confessional theology’ … Confessional theologies presume the existence of some kind of Ultimate, Transmundane Reality whereas non-confessional theologies recognize the cultural reality of ‘the gods’ … without presuming that such an account is possible only on the supposition that ‘the Ultimate’ exists”. Cf. F. WATSON, A Reply to My Critics, in C.S. HAMILTON – J. WILLITTS (eds.), Writing the Gospels: A Dialogue with Francis Watson (LNTS, 606), London – New York, Bloomsbury T&T Clark, 2019, 227-248, p. 248: “theological relevance has to do with subject matter alone, not with the interpreter’s personal orientation towards it”. 41. L.E. VAAGE, How I Stopped Being a Q-Scholar, in ARNAL et al. (eds.), Scribal Practices and Social Structures among Jesus Adherents (n. 27), 213-231. 42. Ibid., p. 220. 43. Ibid., p. 221. 44. Ibid., p. 222.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

57

The Formation of Q attempted to identify the specific agents and actors “behind” Q45, but this quest for Q’s Galilean village scribes, although appealing to “scribal” and documentary material culture where it could be found (Egypt) proceeded by analogy46, led Q Studies astray and into “a much more restricted historiographical terrain”47. That is, the critical study of Q became the study of a scribal product that had “nowhere else to go but back into the heads of those responsible for it”, a function of their “frustrated imagination”, a social and textual experiment that resulted in failure. For Vaage, the field’s failure to properly reckon with the analogically comparative similarities between Q and Cynicism signifies why he “stopped being a Q-scholar”48. Vaage argues that the Q/Jesus/Cynic comparison(s) struck a nerve that was “not historiographical, but theological”49. Positing a sustained “resistance by modern biblical scholars to taking seriously their own observations of a certain similarity”, Vaage concedes that he “stopped” working on Q because “the possibility of discerning – via Q, via Cynicism – a larger, more heterogeneous, still expansive, decidedly ex-centric, ideally less flatly hegemonic historical reality” has “not proven to be the order of the day”50. Since Q Studies have allegedly failed to provide a truly different picture of Christian origins, this sub-field in New Testament studies has become “thoroughly uninteresting”, “increasingly conventional” in character, and “basically interrupt[s] nothing”51. 45. Vaage (ibid., p. 223) notes how “There is no evidence in The Formation of Q that the book’s basic argument … depends in any way on one kind of historical subject having been the agent of the observed discursive patterns”. Furthermore, “the question of discourse essentially plays no role in the argument of that book”. It was only “afterwards starting with the 1991 essay, ‘Literary Convention’”, that Kloppenborg “began to address of question of who it was who most likely wrote these words”. 46. ARNAL, Jesus and the Village Scribes (n. 5); BAZZANA, Kingdom of Bureaucracy (n. 5). 47. VAAGE, How I Stopped Being a Q-Scholar (n. 41), p. 224: “I still fail to perceive the self-evidence” of Q being composed by “sapiential instructors, or scribes”. This “equation” between Q’s genre and its “putatively scribal point of departure” has persuaded “other scholars for reasons” which he “either fail[s] to fathom or to follow”. 48. Vaage (ibid., p. 226) notes that he has not “come to think otherwise”, but he has “ceased belabouring this possibility” of a Cynic-like Q. He is simply not interested in “the current pattern of scholarly conversation ‘around the text’ of Q”. 49. Cf. J.S. KLOPPENBORG VERBIN, Excavating Q: The History and Setting of the Sayings Gospel, Minneapolis, MN, Fortress, 2000, p. 422. Kloppenborg noted that the Cynic hypothesis had “yet to be made effectively” (ibid., p. 431). Cf. VAAGE, How I Stopped Being a Q-Scholar (n. 41), p. 227: “the case for Galilean village scribes as the primary producers of Q as a historical thesis also and equally has yet to be made beyond the imputation of its self-evidence, or the putative homology between the document’s ancient nature qua writing and its erstwhile writers”. 50. Ibid., p. 228. 51. Ibid., p. 229.

58

S.J. JOSEPH

This elegiac lament over a field that never flowered – a road never taken, a tree that never bloomed – identifies a critical moment when Q failed to live up to its full potential to destabilize dominant ideological discourses in the study of Christian origins. For Vaage, this critical juncture occurred when Q’s scribal-geographical provenance in Galilee localized its emergence as reflective of a particular social context. For other Q specialists, however, identifying the specific authors “behind” Q and their interests, social location, geographical context, and “scribal” status was a significant advance in Q Studies52. At present, then, the field continues to be polarized between those who maintain an allegiance to the Galilean “village scribe” hypothesis, those who either eschew locating the text/author in a specific social, cultural, and geographical context, and those who seek “alternative” cultural contexts within which to locate or “map” Q. Although the Galilean provenance of Q remains a speculative hypothesis, it maintains a powerful hold on the conceptual imagination of those working in this field53. Similarly, attempts to stratify the composition of Q remain a lasting legacy of Q Studies, especially since the publication of The Formation of Q. Indeed, two recent monographs devoted to re-evaluating the formative stratum of Q indicate that persuasive arguments identifying both sapiential and apocalyptic elements in this stratum continue to be made, illustrating that the interpretation of Q remains a live issue. In Apocalypticism in the Synoptic Sayings Source54, Olegs Andrejevs reorients our attention toward the thoroughly apocalyptic thought-world and formative impulses in the composition of Q. In The Formative Stratum of the Sayings Gospel Q, Llewellyn Howes re-examines the “formative stratum” of Q, argues for the inclusion of a number of traditionally excluded passages, re-evaluates its “message” as focused on “actual daily survival”, and its compositional integrity as a “unified document or tradition”55. Both monographs, however, locate Q firmly – and rightly – within a first-century Jewish milieu. 52. ARNAL, Jesus and the Village Scribes (n. 5); S.E. ROLLENS, Framing Social Criticism in the Jesus Movement: The Ideological Project in the Sayings Gospel Q (WUNT, II/374), Tübingen, Mohr Siebeck, 2014; BAZZANA, Kingdom of Bureaucracy (n. 5); C. GOUSOPOULOS, Sub-Elite Readers and the Transmission of Christian Literary Texts, in ETL 97 (2021) 37-59. 53. PARKS, Gender in the Rhetoric of Jesus (n. 38), p. 27, begins her recent study of Q by calling it “an early Galilean sayings collection”. Elsewhere, she recognizes it as an “assumption” (p. 32), and relegates my “alternative view, namely, that a Galilean setting does not at all prove a Galilean provenance”, to an endnote (p. 44, n. 27). Cf. S.J. JOSEPH, Jesus, Q, and the Dead Sea Scrolls: A Judaic Approach to Q (WUNT, II/333), Tübingen, Mohr Siebeck, 2012, pp. 33-93. 54. O. ANDREJEVS, Apocalypticism in the Synoptic Sayings Source: A Reassessment of Q’s Stratigraphy (WUNT, II/499), Tübingen, Mohr Siebeck, 2019. 55. L. HOWES, The Formative Stratum of the Sayings Gospel Q: Reconsidering Its Extent, Message, and Unity (WUNT, II/545), Tübingen, Mohr Siebeck, 2021. Howes affirms

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

59

III. THE JEWISHNESS OF Q Q represents an ethnically Judean text and tradition56. As with the historical Jesus’ Jewishness, few Q scholars doubt Q’s Jewish/Judean origins57. Yet despite Q’s Jewish context being a broad consensus58, the fact that Q’s Jewish profile is a relatively recent discovery – a discursive novum that needs to be emphasized – shows that Q Studies have been over-determined by “Christian” assumptions, whether historical, sociological, theological, strucKloppenborg’s insight that an apocalyptic redactional layer was superimposed onto a “formative” sapiential stratum of instructional wisdom. In fact, this monograph is “an attempt to reinforce and buttress” his stratigraphic model (p. 5). Howes also follows Kloppenborg’s “criteria” to distinguish between the two layers, namely, “characteristic forms”, “motifs”, and “implied audience”. Howes is aware of how his own “situatedness” and “positionality on the African continent” has contributed “in no small way” to his reading of these texts in terms of their focus on “poverty” (p. 3), reading “from below”, and concludes that Q was “in the first place directed at poor people and struggling peasants in ancient Judaism” (p. 3). The “poor” of Q are the poor, the economically struggling and economically destitute, of first-century Judea and Galilee. Howes argues that “the ultimate result that the formative stratum is a sapiential stratum and that the main redaction harbours apocalyptic eschatology and prophecy justifies the subsequent use of these findings as criteria when considering the placement of individual traditions in Q” (p. 6). Howes insists that the “formative stratum” does not “develop futurist eschatology or apocalypticism at all” (p. 6). 56. KLOPPENBORG VERBIN, Excavating Q (n. 49), p. 256; W.E. ARNAL, The Q Document, in M. JACKSON-MCCABE (ed.), Jewish Christianity Reconsidered: Rethinking Ancient Groups and Texts, Minneapolis, MN, Fortress, 2007, 119-154, p. 129; M. CROMHOUT, Jesus and Identity: Reconstructing Judean Ethnicity in Q, Eugene, OR, Cascade, 2007, p. 260; C.M. TUCKETT, Q and the ‘Church’: The Role of the Christian Community within Judaism according to Q, in ID., From the Sayings to the Gospels (WUNT, 328), Tübingen, Mohr Siebeck, 2014, 219-231. Cf. M. TIWALD (ed.), Q in Context I: The Separation between the Just and the Unjust in Early Judaism and in the Sayings Source (BBB, 172), Göttingen, Bonn University Press – V&R Unipress, 2015. ARNAL, Jesus and the Village Scribes (n. 5), p. 202, suggests that Q re-presents the local “deracination” of Galilean scribes, resulting in its author’s attempt to increase their rhetorical powers of persuasion by appealing to the common socio-cultural capital of Judean ethno-religious norms and customs. 57. A notable exception is Harry T. Fleddermann who has argued that Q is a “Gentile” Gospel written by and for Gentile followers of Jesus. See H.T. FLEDDERMANN, Q: A Reconstruction and Commentary (BiTS, 1), Leuven – Paris – Dudley, MA, Peeters, 2005. 58. CROMHOUT, Jesus and Identity (n. 56); J.M. ROBINSON, LOGION SOPHON: On the Gattung of Q, in ID. – H. KOESTER (eds.), Trajectories through Early Christianity, Philadelphia, PA, Fortress, 1971, 71-113; H.E. TÖDT, The Son of Man in the Synoptic Tradition, Philadelphia, PA, Westminster, 1965; O.H. STECK, Israel und das gewaltsame Geschick der Propheten: Untersuchungen zur Überlieferung des deuteronomistichen Geschichtsbildes im Alten Testament, Spätjudentum und Urchristentum (WMANT, 23), Neukirchen-Vluyn, Neukirchener, 1967; D. LÜHRMANN, Die Redaktion der Logienquelle (WMANT, 33), NeukirchenerVluyn, Neukirchener, 1969; S. SCHULZ, Q: Die Spruchquelle der Evangelisten, Zürich, TVZ, 1972; P. HOFFMANN, Studien zur Theologie der Logienquelle (NTAbh, NF 8), Münster, Aschendorff, 31982, 332-333; ID., QR und der Menschensohn: Eine vorläufige Skizze, in F. VAN SEGBROECK – C.M. TUCKETT – G. VAN BELLE – J. VERHEYDEN (eds.), The Four Gospels 1992: FS Frans Neirynck (BETL, 100), Leuven, Leuven University Press – Peeters, 1992, 421-456; KLOPPENBORG, Literary Convention (n. 5); ID., Excavating Q (n. 49), p. 256.

60

S.J. JOSEPH

tural, and/or systemic to the field and guild. There is no denying the fact that Biblical Studies tends to be characterized by predominantly Christian interests. Christian identity has long been implicated in racial, political, economic, and colonial power structures, whether or not they are openly acknowledged59. These structures are now embedded in Western culture, providing both residual privilege and socio-cultural capital as well as the (white, male, Christian) option of not having to self-reflect on one’s ethnicity, race60, class, gender, and/or “whiteness”61. We find, then, a correlative parallel between Q Studies and Jesus Research insofar as both discursive fields re-cognize the nominal Jewish matrices of their respective objects of study but are noticeably reluctant to make much more of this nominal recognition. Q, like the historical Jesus, may be Jewish, but “not that Jewish”62. Q does not provide us with compelling or definitive evidence of a Gentile mission63. Rather, its ethno-religious horizon remains Judean despite its subsequent adoption into the Gentile-oriented Gospels of Matthew and Luke. Indeed, Q represents a classic example of Jewish sectarian identity formation64. The author redefines who belongs to Israel65, criticizes other Jews66, rejects the villages that have rejected Jesus’ message67, condemns

59. D.G. HORRELL, Paul, Inclusion and Whiteness: Particularizing Interpretation, in JSNT 40 (2017) 123-147, p. 124: “I have spent the last 25 years as a New Testament scholar without ever reflecting on how my racial or ethnic identity shapes what I do, what I see, what I ask and do not ask”. Cf. ID., Ethnicity and Inclusion: Religion, Race, and Whiteness in Constructions of Jewish and Christian Identities, Grand Rapids, MI, Eerdmans, 2020. 60. R. DIANGELO, White Fragility, in International Journal of Critical Pedagogy 3 (2011) 54-70; EAD., White Fragility: Why It’s So Hard for White People to Talk About Racism, Boston, MA, Beacon, 2018, referring to the implicit racial biases that “white” people harbor, often unwittingly, in perpetuating systemic socio-economic injustice. 61. R. NEWTON, Racial Profiling: Reading Whiteness & the Bible in Postmodernity, in Religion Compass 14 (2020); E. TUPAMAHU, The Stubborn Invisibility of Whiteness in Biblical Scholarship, in Political Theology (November 12, 2020), https://politicaltheology.com/ the-stubborn-invisibility-of-whiteness-in-biblical-scholarship/[accessed December 28, 2020]. 62. Cf. J.G. CROSSLEY, A ‘Very Jewish’ Jesus: Perpetuating the Myth of Superiority, in JSHJ 11 (2013) 109-129. 63. On the question of a “Gentile mission”, see J.S. KLOPPENBORG, ‘A Parting of the Ways’ in Q?, in TIWALD (ed.), Q in Context I (n. 56), 123-143. On the rhetorical construction of the “Gentile” (Christian) vis-à-vis the “Jew”, see now T.L. DONALDSON, Gentile Christian Identity from Cornelius to Constantine: The Nations, the Parting of the Ways, and Roman Imperial Ideology, Grand Rapids, MI, Eerdmans, 2020. 64. S.J. JOSEPH, A Social Identity Approach to the Rhetoric of Apocalyptic Violence in the Sayings Gospel Q, in HR 57 (2017) 28-49. 65. Q 3,8; 13,29.28.30; 14,11.16-23. 66. Q 11,42.39b.43-44.46b.52.47-48. 67. Q 10,10-12.13-15.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

61

“this generation”68, and pronounces judgment on Israel69. The author stereotypes opponents, labels Pharisees “empty tombs”, and draws symbolic boundaries of social closure around their in-group. The “last will be first and the first last”. Some will enter the Kingdom and others will be excluded. Q specialists have long identified the rejection of Israel as a dominant compositional motif70. The Jews of “this generation” have rejected the message of John and Jesus71. The Jewish Galilean towns of Chorazin, Bethsaida, and Capernaum have rejected Jesus72. The city of Jerusalem and its leaders have rejected Jesus. John and Jesus are the “children of Wisdom” rejected by “this generation”, a generation condemned for killing and persecuting the prophets73. Even the disciples – sent out like “sheep among wolves” – are to anticipate rejection and persecution for their efforts74. According to the author of Q, the rejection of Jesus cannot be forgiven: a judgment is coming, but it will be a judgment without warning75. The rejection and failure of Jesus’ mission to Israel has provoked divine judgment76. The one who is to come will separate the “wheat” from the “chaff” and burn the chaff in an everlasting fire (Q 3,16b-17)77. The Temple is now “desolate”, “forsaken” (Q 13,34-35). Q’s polemic is directed toward the non-responsive part of Israel78. Q’s conflict with “this generation” is a Jewish 68. Q 7,31; 11,16.29-30.49-51. 69. Q 13,34-35. 70. J.S. KLOPPENBORG, The Formation of Q Revisited: A Response to Richard Horsley, in D.J. LULL (ed.), SBL 1989 Seminar Papers, Atlanta, GA, Scholars, 1989, p. 207, regards “this generation” as “co-extensive with Israel” because it invokes and accompanies criticism of Jerusalem, woes against Pharisees and Galilean towns and threats of exclusion. This, however, does not make the Q people non-Jewish or any less Jewish than the Qumran community. The severe criticism of Israel is not to be seen as a “rejection” of Israel. LÜHRMANN, Redaktion (n. 58), pp. 24-48, 30, 93; SCHULZ, Q (n. 58), p. 340; H. SCHÜRMANN, Die Redekomposition wider “dieses Geschlecht” und seine Führung in der Redenquelle (vgl. Mt 23,1-39 par Lk 11,37-54): Bestand – Akoluthie - Kompositionsformen, in SNTU 11 (1986) 33-81, p. 73; KLOPPENBORG, Formation (n. 4), pp. 148, 167, see “This generation” as a term for all Israel: the Q people have given up on Israel and may be embracing a Gentile mission. 71. MACK, The Lost Gospel (n. 10) p. 136, argues that the Q people adopted this shift in tone because the movement lost and/or could not secure the loyalty of its members. This caused “social stress”. 72. Q 7,1-10; 10,13-15. ARNAL, Jesus and the Village Scribes (n. 5), p. 162. 73. Q 7,31-35; 11,49-51. 74. Q 10,3.10-12. 75. Q 12,39-40.42-46; 13,29.28. 76. MACK, The Lost Gospel (n. 10), p. 45. 77. Q 3,16b-17. 78. C.M. TUCKETT, Q and the History of Early Christianity: Studies on Q, Edinburgh, T&T Clark, 1996, pp. 199, 201. Contra STECK, Geschick (n. 58), pp. 286-288, who sees these sayings as aimed at the “Erweckung” of Israel. R.A. HORSLEY, Social Conflict in the Synoptic Sayings Source Q, in KLOPPENBORG (ed.), Conflict and Invention (n. 5), 37-52, points

62

S.J. JOSEPH

conflict within Judaism79. While Jesus and his followers probably did meet sustained resistance from Jews, rhetorical appeals to persecution do not necessarily represent actual experiences of persecution80. Rejection and/or persecution could have been perceived as real whether or not actual rejection or persecution ever took place. These appeals to rejection and persecution served purposes other than mere reporting of the historical facts, including justifying the righteous invocation of divine judgment on the oppressors. Rhetorical pronouncements of apocalyptic violence allowed authors to criticize, condemn, and attack others without resorting to physical violence. Rhetorical appeals to suffering and persecution also served the purpose of associating Q’s contemporary conflicts with the ancestral theme, motif, and tradition of “the rejected prophet(s)”, now imagining Jesus and his followers as subject to unjust suffering and an eschatological reversal of divine vindication. Judaic approaches to Q recognize these rhetorical appeals to suffering, persecution, and rejection as a Jewish discourse of sectarian rivalry utilizing ancestral scriptural traditions to buttress and augment contemporary claims. This does not mean that the author of Q invented this tradition out of whole cloth81. In a cultural context characterized by sectarian rivalries that valued the experience of persecution as proof of fidelity to divine truth, out that Q 22,28-30, Q 13,28-29, and Q 13,34-35 do not support a reading of Q vs. “all Israel”. This reading of Israel’s “unfaithfulness” appears to be “a residue of a Christian theological agenda”. Horsley (ibid., p. 39) has argued that a Christian theological reading (of the Q group vs. all of Israel) has been supported by redactional strata in Q because the “apocalyptic” layer is seen as addressing “all Israel”. In ID., Wisdom Justified by All Her Children: Examining Allegedly Disparate Traditions in Q, in SBL 1994 Seminar Papers, Atlanta, GA, Scholars, 1994, 733-751, p. 741, Horsley claims this is “typical of scholarship on Q, but not in the text”. 79. T.S. FERDA, Jesus and the Galilean Crisis: Interpretation, Reception, and History, diss. University of Pittsburgh, 2016, p. 373 n. 1142, notes that “this generation” refers to nonresponsive Jews, but disagrees with my alleged assessment that it is just “scribes, Pharisees, and Temple elite”, which he calls “a strange universalizing of its own kind”. In JOSEPH, Jesus, Q, and the Dead Sea Scrolls (n. 53), p. 92, I make no such “strange universalizing” claim. On the contrary, I state that it is “not ‘Israel’ as a whole” that is being criticized: this is “a Jewish conflict within Judaism” (emphasis in original). 80. S.E. ROLLENS, Persecution in the Social Setting of Q, in M. TIWALD (ed.), Q in Context II: Social Setting and Archaeological Background (BBB, 173), Göttingen, Bonn University Press – V&R Unipress, 2015, 149-164, p. 149, points out that “we are largely dealing with textual reflections” of “historical circumstances”. Accordingly, if “the collection of ideas related to suffering and rejection seems to be a common core among many early Christian texts … how do we know that any of these actually reflect historical circumstances?” (p. 152). 81. Rollens (ibid., p. 163), concludes that “due [to its appeal to Deuteronomistic theology] … Q itself is not even concerned with the socio-historical dimensions of its experience of persecution. That is to say, this theological framework is itself the reason for rejection … the legacy of the past prophets is the only explanation necessary”. Rollens thus performs a curious kind of interpretative reversal: it is not the experience of rejection that leads to its rhetorical

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

63

it is not surprising that Q appeals to this tradition to justify its claims about Jesus vis-à-vis his Jewish contemporaries. It is not until Q was incorporated into the Gospels of Matthew and Luke, however, that this polemical rhetoric, originally directed toward Jews, would be read as a conflict between Jews and Christians. If the Jewishness of Q has become a near-universal consensus, that does not mean that the kind of Jewishness represented by Q enjoys anything even remotely close to a consensus. To be sure, as Martin Hengel points out, “the usual distinction between Palestinian and Hellenistic Judaism needs to be corrected” as it “tends to give a mistaken account of the new situation of Judaism in the Hellenistic period”82. There were “Greek-speaking Jewish Christians from 30/31 up to the time of Paul – indeed he himself was one of them”. The “Aramaic-speaking primitive community in Jerusalem and the Greek-speaking Jewish-Christian community in Jerusalem, Caesarea, Damascus, Antioch, and Rome do not so much come one after another in time as stand side by side, so that it would also be possible in theory that traditions from the Greek-speaking ‘primitive community’ were also taken over by the Aramaic-speaking community”83. Ethnically Jewish Greekspeaking followers of Jesus were indeed in an excellent position to remember, record, and distribute collections of Jesus’ sayings. The fact, however, that they could have done so based either in the Diaspora, the Galilee, in Judea, or in Jerusalem – which had an outreach extending throughout the Diaspora – has been relatively neglected in Q Studies. It is difficult not to conclude that the primary reason this latter hypothesis has been neglected is a direct result of the weight given the hypothesis of a Galilean (“village scribal”) provenance of Q, an hypothesis which seems increasingly insecure. Consider, for example, the two most prominent leaders of the Jesus movement in the mid-first century CE: Peter and James. Simon Peter allegedly began his adult life as a Galilean fisherman in Capernaum (ca. 30 CE), but apparently did not remain in that occupation, as his presence and likely relocation to Jerusalem by the mid-30s CE suggests. It is also pertinent to note that Paul locates Peter in Antioch shortly after the Jerusalem council (ca. 50 CE). Similarly, James, “the brother of the Lord”, apparently began his adult life as a rural Nazarene, but his central leadership role in the post-Easter expression and performance; rather, it is the “theological” presupposition of rejection that informs, directs, and constructs a rhetorical appeal to rejection. 82. M. HENGEL, Judaism and Hellenism: Studies in Their Encounter in Palestine During the Early Hellenistic Period, Philadelphia, PA, Fortress, 1974, p. 104. 83. ID., Christology and New Testament Chronology, in ID., Between Jesus and Paul: Studies in the Earliest History of Christianity, Eugene, OR, Wipf and Stock, 2003, p. 37.

64

S.J. JOSEPH

movement suggests that he too relocated to Jerusalem as early as 33 CE and remained there until his death in 62 CE. Here, then, we have the two most prominent Jewish leaders of the Jesus movement transformed from being rural Galilean Jews into relatively cosmopolitan leaders of a Diaspora-wide mixed Jewish/Gentile movement centered in Jerusalem within the first two decades of the movement. Why, then, should we treat “Galilee” as a culturally and geographically isolated and hermetically sealed territory within which “Galilean village scribes” represent the only potentially creative cultural producers and “influencers” of the Jesus tradition in Q? We need a more complex and robust model of cultural interactionism that re-locates Q within a thoroughly sectarian Jewish milieu in conversation and collaboration with what we can actually know about the Jesus movement ca. 50 CE. The Galilee has long been romanticized as the idyllic location of Jesus’ early ministry84, a geographical guarantor preserving an isolated pure stratum of untouched Jesus tradition85, and/or a cultural and ethno-racial cipher used to promote a non-Jewish or Aryan Jesus86. Contemporary Q scholars may not be prone to such extravagances, but insofar as Q Studies maintain Q’s theological “difference” from the Christian kerygma in terms of its geographical distance from Judea and Jerusalem, the “conceptual stakes” implicated in identifying Q as “the product of a rural Jesus movement … [that] does not range much outside of Jewish Palestine” continue to be complicated by “centering” it in Galilee87. The unfortunate result of this lacuna in Q Studies is that few scholars have recognized that since Q arguably represents the chronologically earliest appearance of a (titular) Son of Man Christology in the synoptic tradition, it should be critically compared to its Palestinian Jewish contemporary and complement, the Enochic Book of the Parables. The current consensus is that the Parables represents a first-century pre-Christian Jewish text describing the imminent arrival of a messianic Son of Man. Despite an emergent consensus that the Parables should be dated to the turn of the common 84. E. RENAN, Vie de Jésus, Paris, Michael Levy, 1871 (1863). 85. A. VON HARNACK, The Sayings of Jesus: The Second Source of St. Matthew and St. Luke (New Testament Studies, 2), London, Williams & Norgate, 1908, p. 171. 86. W. BAUER, Jesus der Galiläer, in G. STRECKER (ed.), Aufsätze und kleine Schriften, Tübingen, Mohr, 1967, 91-108; W. GRUNDMANN, Jesus der Galiläer und das Judentum, Leipzig, Wigland, 1940; P. HEAD, The Nazi Quest for an Aryan Jesus, in JSHJ 2 (2004) 55-89; S. HESCHEL, The Aryan Jesus: Christian Theologians and the Bible in Nazi Germany, Princeton, NJ, Princeton University Press, 2008. 87. KLOPPENBORG, Conceptual Stakes in the Synoptic Problem (n. 18), p. 38 (emphasis in original), adding that Q’s world is “centred in the Galilee, on the towns of Capernaum, Bethsaida, Khorazin and Nazara” (p. 40).

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

65

era88, and represents a pre-Christian Jewish text that can be dated earlier than the ministry of Jesus and influenced some of the writings in the New Testament89, there is still a noticeable tendency to segregate the Parables from Q Studies and avoid the question of whether it influenced the composition of Q and/or the Synoptic Gospels90. Similarly, the fact that Q contains a number of parallels to Essenic texts and traditions has been neglected insofar as the latter tend to be seen as “Judean” rather than “Galilean”. The Messianic Apocalypse (4Q521), for example, contains a number of similarities between the activities of Jesus as described in Q and the messianic expectations of the Essenes. The text is generally regarded as “non-sectarian” because it does not contain some of the hallmark features of “sectarian” identity-markers, but it was found at Qumran and represents a copy of a text otherwise unknown in ancient Jewish literature. It is reasonable, therefore, to regard it as illustrative of beliefs compatible with Essenic messianism. The text describes the blessings that will occur when “his messiah” arrives: liberating the captives, giving sight to the blind, straightening the bent … For he will heal the wounded, revive the dead, and proclaim good news to the poor.

Like 4Q521, Q describes Jesus citing a number of the same scriptural passages to confirm his identity as “the one who is to come” the blind see, and the lame walk, the lepers are cleansed, and the deaf hear, and the dead are raised, and the poor have good news preached to them.

88. G. BOCCACCINI, The Enoch Seminar at Camaldoli: Re-entering the Parables of Enoch in the Study of Second Temple Judaism and Christian Origins, in ID. (ed.), Enoch and the Messiah Son of Man: Revisiting the Book of Parables, Grand Rapids, MI, Eerdmans, 2007, p. 16. The proposed dates actually range from 20 BCE to 100 CE. For the latter, see M.A. KNIBB, The Date of the Parables of Enoch: A Critical Review, in NTS 25 (1979) 345-359. A possible, albeit ambiguous, allusion to the Parthians (1 En 56,5-7) suggests a date after their invasion into Roman territory in 40 BCE. A passing reference to hot springs that “serve the kings” also seems to refer to Herod the Great at Callirhoe (1 En 67,4-13; Josephus, A.J. 17.6.5.171-173; B.J. 1.33.5.657-658), which suggests a date around the turn of the era or the early to mid-first century CE. The Parables do not refer to or suggest knowledge of the Revolt or the destruction of the Temple, so it would seem that a date before 70 CE is likely. 89. J.H. CHARLESWORTH – D.L. BOCK (eds.), Parables of Enoch – A Paradigm Shift (JCT, 26), London – New York, T&T Clark, 2013. See also BOCCACCINI (ed.), Enoch and the Messiah Son of Man (n. 89). 90. S.J. JOSEPH, The Promise of Providence and the Problem of the Parables: Revisiting Prayer in the Sayings Gospel Q, in D.A. SMITH – C. HEIL (eds.), Prayer in the Sayings Gospel Q (WUNT, 425), Tübingen, Mohr Siebeck, 2019, 57-87. For a notable exception, see ANDREJEVS, Apocalypticism in the Synoptic Sayings Source (n. 54).

66

S.J. JOSEPH

The parallels suggest relationship91. Attempts have been made to explain the similarities by positing a “common tradition” or by positing indirect contact. The question, however, is not whether there is an exact correspondence. We do not need “pure or perfect parallels” to posit relationship. We “just need historical evidence”92. The evidence suggests that the authors of the early Jesus tradition were in conversation with Essene-like Jews about the signs of the messianic age and whether or not Jesus fulfilled them93. There has been some resistance to this assessment. Remarkably, some scholars even prefer resorting to caricature rather than concede any relationship between the Jesus tradition and Essenic literature94. Here it is important to recognize the distinction between the wider multi-regional rural and urban Essenic movement and “Qumran”, where Q would clearly be out of place95. 91. O. FREIBERGER, Considering Comparison: A Method for Religious Studies, New York, Oxford University Press, 2019; J. CARTER, Comparison in the History of Religions: Reflections and Critiques, in MTSR 16 (2004) 3-11. Cf. J.Z. SMITH, Adde Parvum Parvo Magnus Acervus Erit, in HR 11 (1971) 67-90; ID., Imagining Religion: From Babylon to Jonestown, Chicago, IL, University of Chicago Press, 1982, pp. 19-35; ID., Drudgery Divine: On the Comparison of Early Christianities and the Religions of Late Antiquity, Chicago, IL, University of Chicago Press, 1990. 92. D.B. MARTIN, Response to Downing and Fredriksen, in JSNT 37 (2015) 334-345, p. 342: “The demand by some NT scholars that parallels, in order to do any ‘historiographical work’ at all, must be perfect parallels is just bad historiography. We do not need pure or perfect parallels. We just need historical evidence”. 93. L. FRAYSSE, Le mystère des esséniens (4): Jésus était-il essénien?, in Réforme (June 4, 2020), https://www.reforme.net/religion/histoire/2020/06/04/le-mystere-des-esseniens-4jesus-etait-il-essenien/ [accessed June 2020], quotes Michael Langlois as saying “Mais selon toute vraisemblance, le noyau dur des premiers chrétiens, des proches de Jésus, était composé en majorité d’esséniens”. 94. C.K. ROTHSCHILD, review of Simon J. Joseph, Jesus, Q, and the Dead Sea Scrolls: A Judaic Approach to Q, in DSD 22 (2015) 129-131, claims that my study argues that the Baptist was “an Essene raised at Qumran” and that my goal is to “escort readers” into the world of my “imagination” where the Baptist “comes of age at Khirbet Qumran”, “mastering the texts of Cave 4”, and “flouting apprenticeship to his father, a Zadokite priest” before he decides “to peddle a version of atonement”. Jesus, Q, and the Dead Sea Scrolls, however, does not claim that John was an “Essene” or ever lived at Qumran, let alone “mastered” the “texts of Cave 4”. Moreover, Rothschild’s suggestion that the monograph is “a well-constructed house … for those willing to entertain such theses”, but “moot” for those who “reject the existence of Q” and the “Essene authorship of the sectarian scrolls”, is an astonishing statement coming from someone who wrote a monograph on Q and continues to “assume the existence of Q”. See C.K. ROTHSCHILD, Baptist Traditions and Q (WUNT, 190), Tübingen, Mohr Siebeck, 2005; EAD., Refusing to Acknowledge the Immerser (Q 7.31-35), in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 18), 165-183, p. 165, n. 1 (emphasis added). 95. PARKS, Gender in the Rhetoric of Jesus (n. 38), p. 32, describes my Jesus, Q, and the Dead Sea Scrolls as one of a number of “Alternative theories of provenance [that] place Q in as disparate settings as Qumran and Jerusalem”. I do not, however, “place Q” in or at Qumran. Rather, I argue that Q can be compared to and was influenced by the kind of Essenic/Enochic Judaism represented in the texts recovered from Qumran.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

67

Of course, one hopes for continued engagement with all the available evidence96. Despite the fact, however, that comparative studies of the Jesus movement and the Essenes have been a regular feature in biblical scholarship since the Enlightenment97, with the discovery of the Dead Sea Scrolls – associated with the Essenes – providing evidentiary support for such studies98, Q Studies have not integrated this history of research and so continue to represent an inadequate accounting of the cultural range of early Jewish literature99. *  *  * What is the “end”, then, of Q Studies? In recent years, the field of Biblical Studies has increasingly recognized that ethnicity and race represent ideological sites of power and privilege. For some, the Synoptic Problem itself is potentially implicated in the problematic conceptualization of “whiteness”100. 96. C. WASSÉN, review of Joseph, Jesus, the Essenes, and Christian Origins, in DSD 27 (2020) 323-325, p. 324, notes my objections to the “common tradition” model as an explanatory theory; the similarities are “a noteworthy point”. Cf. M. TIWALD, Die Logienquelle: Text, Kontext, Theologie, Stuttgart, Kohlhammer, 2016, p. 89: “Einen etwas anderen Zugang wählt S. J. Joseph, der darauf abzielt, dass Q ‘the work of skilled scribes located in or adjacent to an urban setting’ sei. Aufgrund einer Tendenz der Qumrantexte, ‘[that] redefines who belongs to ‘Israel,’ möchte der Einflüsse der Essener auf die Logienquelle geltend machen und optiert daher für eine Entstehung der Logienquelle in Judäa. Allerdings: Die Frage nach dem ‘wahren Israel’ war im Frühjudentum weit verbreitet, direkte Abhängigkeiten zwischen Qumran und Jesusbewegung sind dafür nicht vonnöten und auch ansonsten nicht zu belegen”. Tiwald suggests that my proposal for Essene influence(s) on Q is derived primarily from the Qumranic “tendency” to represent itself as “true Israel”, and that such notions were widespread in Early Judaism. 97. See K.F. BAHRDT, Ausführung des Plans und Zwecks Jesu. In Briefen an Wahrheit suchende Leser, Berlin, August Mylius, 1784-1792, 11 vols.; K.H. VENTURINI, Natürliche Geschichte des großen Propheten von Nazareth, Copenhagen, 1800-1802, 4 vols.; S. WAGNER, Die Essener in der wissenschaftlichen Diskussion: Vom Ausgang des 18. bis zum Beginn des 20. Jahrhunderts: Eine wissenschaftsgeschichtliche Studie, Berlin, Alfred Töpelmann, 1960. For recent attempts to identify John the Baptist as a (former) member of the Essenes, see J. MARCUS, John the Baptist in History and Theology (Studies on Personalities of the New Testament), Columbia, SC, University of South Carolina, 2018, pp. 27-45; J. BERGSMA, Jesus and the Dead Sea Scrolls: Revealing the Jewish Roots of Christianity, New York, Image, 2019, pp. 31-67. 98. S.J. JOSEPH, Jesus, the Essenes, and Christian Origins: New Light on Ancient Texts and Communities, Waco, TX, Baylor University Press, 2018. 99. É. PUECH, review of Joseph, Jesus, the Essenes, and Christian Origins, in RB 127 (2020) 451-453, p. 453, agrees that many Essenes, although perhaps not Qumran-Essenes, joined the Jesus movement, given their absence in the New Testament and in rabbinic literature, yet notable influence on the former: “Il est probable que plusieurs, qumraniens exclus, sont entrés dans la nouvelle Voie, comme le laissent entendre l’absence de leur mention mais leur influence dans le NT, ainsi que leur absence dans le judaïsme rabbinique”. 100. TUPAMAHU, The Stubborn Invisibility of Whiteness in Biblical Scholarship (n. 61), suggests that “whiteness” is the assumed normative social location in Biblical Studies and that

68

S.J. JOSEPH

Indeed, white Christian privilege has now come under increased scrutiny insofar as it plays a systemic and structural yet often “invisible” role in providing actual as well as residual and lingering ethno-racial and class privilege to its member-constituents whether or not an active, confessional “Christian” identity is affirmed101.

the Synoptic Problem is an example of “the persistent invisibility of whiteness in New Testament studies”. As a “white ‘problem’”, the literary relationship(s) between the Gospels of Mark, Matthew, and Luke is “a retrojection of whiteness into biblical texts”, “a uniquely white modern European phenomenon”, because it reflects “modern white social imagination” and “white logic” and can be likened to the modern conception of “copyright”, which controls and limits literary plagiarism. Tupamahu further associates the concepts of copyright, property, and “whiteness”, implicating the Synoptic Problem as complicit in a covert racial discourse operating within “a white individual property imagination”. The “problem”, then, is one of unauthorized (“white” male) property and copyright violation and the Synoptic Problem “is an effort to protect the property rights of gospel authors”. Whatever the value of this alleged resonance might be – and there are good reasons to question Tupamahu’s reduction of the “problem” to a puzzling literary relationship predominantly of interest to a European scholarly tradition – there is little doubt that the Synoptic Problem has been “almost exclusively a white men’s obsession”. See, however, the following critical essays on the Synoptic Problem by non-white authors: S.O. ABOGUNRIN, The Synoptic Gospel Debate: A Re-Examination from an African Point of View, in D.L. DUNGAN (ed.), The Interrelations of the Gospels. A Symposium Led by M.-É. Boismard – W.R. Farmer – F. Neirynck, Jerusalem 1984 (BETL, 95), Leuven, Leuven University Press – Peeters, 1990, 381-407 (in support of the 2GH); K.-H. CHANG, Questioning the Feasibility of the Major Synoptic Hypotheses: Scribal Memory as the Key to the Oral–Written Interface, in JSNT 41 (2019) 407-432 (in support of the 2DH). See also Chakrita Saulina’s contribution in this volume (in support of the MPH). Cf. G. CAREY, Looking for White in the Synoptic Problem, in Political Theology (December 10, 2020), https://politicaltheology.com/looking-for-white-in-the-synoptic-problem/[accessed December 28, 2020]: “This White author has personally experienced the Synoptic Problem as my own problem. A high school convert who did not grow up in church, I did what I was told and began reading the Bible all the way through … Being a product of American culture, I perceived diverse accounts of a single story as a potential problem … It undermined my assumption – is this a White assumption or not? – that the Gospels provided straightforward access to Jesus’s words and deed. This was a crucial early step in my journey toward critical theology”. The same author concludes: “But is it a White problem? I think that’s complicated”. 101. K.Y. JOSHI, White Christian Privilege: The Illusion of Religious Equality in America, New York, New York University Press, 2020. On “whiteness”, see D.R. ROEDIGER, The Wages of Whiteness: Race and the Making of the American Working Class, London, Verso, 1991; R. DYER, White, London – New York, Routledge, 1997; R. FRANKENBERG, Displacing Whiteness: Essays in Social and Cultural Criticism, Durham, NC, Duke University Press, 1997; R. DELGADO – J. STEFANCIC (eds.), Critical White Studies: Looking behind the Mirror, Philadelphia, PA, Temple University Press, 1997; T.K. NAKAYAMA – J.N. MARTIN (eds.), Whiteness: The Communication of Social Identity, Thousand Oaks, CA, Sage, 1999, pp. vii-xiv. In biblical studies, see G. YANCY (ed.), Christology and Whiteness: What Would Jesus Do?, London, Routledge, 2012; D. KIMBER BUELL, Anachronistic Whiteness and the Ethics of Interpretation, in K.M. HOCKEY – D.G. HORRELL (eds.), Ethnicity, Race, Religion: Identities and Ideologies in Early Jewish and Christian Texts and in Modern Biblical Interpretation, London – New York, Bloomsbury T&T Clark, 2018, 149-167.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

69

It has recently been observed that the sub-field of synoptic studies has been dominated by “white men”102. Sarah Rollens suggests that interest in Q, in particular, is now lagging because it is perceived to be a maledominated field of study. Rollens notes the paucity of female Q scholars – indeed, of new Q scholars in general – in an institutional analysis of Q studies as “a sub-field that cannot let go of its old-fashioned models of scholarship”, “hesitates to entertain innovative forms of scholarship”, and cannot “assess how its ‘old boys club’” requires correction, with the result being that “Q scholarship is slowly dying out due to decreasing numbers of interested scholars”103. She suggests that Q studies are seen as “outdated and unexciting” due to “a narrowly conceived stereotype of what a typical Q scholar is” or does – and not because Q Studies is seen as a controversial field of speculative hypotheses about Q’s “difference”, composition, stratification, Jewishness, relationship to the Gospel of Thomas, and/or Galilean (“village scribal”) provenance. It seems, however, that focusing on race, whiteness, and/or male domination in Q studies runs the risk of failing to articulate the most significant discursive factors affecting contemporary Q Studies104. Male domination is simply not a central factor affecting lagging contemporary interest in Q. This is not to deny that gender – like religion, ideology, race, ethnicity, and class – does play a role in the guild, but the complexity of interrelated factors mitigate and counter-balance against any one factor playing a determinative role105. A paucity of women in Q Studies cannot be reduced to simplistic equations of gender bias when so many additional factors must be taken into consideration, including class privilege, access to, and adept use of social networks and institutional power structures in a field of intersecting and competing social and institutional power-dynamics. In a world where humanities and religion departments are increasingly under

102. CAREY, Looking for White in the Synoptic Problem (n. 101); TUPAMAHU, The Stubborn Invisibility of Whiteness in Biblical Scholarship (n. 61). 103. S.E. ROLLENS, Where Are All the Women in Q Studies? Gender Demographics and the Study of Q, in TIWALD (ed.), The Q Hypothesis Unveiled (n. 29), 224-255. 104. Rollens (ibid., p. 227) analyzes my own work as apparently illustrative of a structural and systemic gender imbalance in Q Studies, noting how “For example, Simon Joseph … has single-handedly written eight peer-reviewed articles on Q”. Yet these articles were, of course, subject to programmatically ethnicity-, gender-, and identity-blind peer reviews. 105. On (rhetorical) gender (equality) in Q, see PARKS, Gender in the Rhetoric of Jesus (n. 38). ROLLENS, Where Are All the Women in Q Studies? (n. 103), p. 224, n. 5, recognizes that “Other expressions of diversity” are important, including “ethnic background, racial identification, sexual orientation, geographical origin, [and] socio-economic status”, but “here, we only focus on gender dynamics”.

70

S.J. JOSEPH

pressure to explain their utility and justify their very existence, the hypothetical existence of Q – beleaguered by facile Internet polls, conservative Evangelical skepticism, suspicion about its deviant “difference”106, and infighting within the sub-discipline itself – does not seem to be a priority for many scholars, whether male or female. IV. CONCLUSION Returning to the “end” of Q Studies, then, one can make the following observations. There is now a noticeable reticence – at least in some circles – to affirm the existence of Q. In and of itself, this could be taken as a sign of healthy debate in a complex field of interdisciplinary specialization. On the other hand, one also notes a reluctance to engage in Q Studies at all, as well as a number of vocal calls for an “end” to Q Studies. At any rate, it is clearly the case that a number of (ex-Q) scholars have indeed put an “end” to their own Q Studies. Nonetheless, it is also clear that rumors of the demise of Q, not unlike those ascribed to the traditional criteria of authenticity, have been much exaggerated107. Obviously, the hypothetical existence of Q continues to be the major battleground, with synoptic studies representing minor and major skirmishes on many fronts. To my mind, however, an “end” of Q Studies – with the word “end” signifying here the biblical sense of the goal or purpose of this academic pursuit – is to further clarify the particular Jewish cultural milieu within which Q emerged. There is also a pressing need to re-evaluate and re-assess the Galilean “village scribal” hypothesis insofar as it insulates Q from Judea and the Jerusalem community, where Greek literary scribal culture thrived in cultural proximity to Essenic (and Enochic) social networks. And last, but not least, it is time to affirm Q’s sapiential

106. Note the recurring emphasis in the work of the 2DH’s critics on the alleged nineteenth century biases of the theory’s founders and early advocates. See, for example, in V. WITTKOWSKY, Luke Uses/Rewrites Matthew: A Survey of the Nineteenth-Century Research, in M. MÜLLER – J.T. NIELSEN (eds.), Luke’s Literary Creativity (LNTS, 550), London – New York, Bloomsbury T&T Clark, 2016, 3-25; W. KAHL, Inclusive and Exclusive Agreements – Towards a Neutral Comparison of the Synoptic Gospels, or: Minor Agreements as Misleading Category, in ibid., 44-78, p. 65; M. MÜLLER, Were the Gospel Authors Really ‘Simple Christians without Literary Gift’ (Albert Schweitzer)? Arguments for the Quest for Sources behind the Gospels, in ID. – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 18), 79-96. 107. C. KEITH – A. LE DONNE (eds.), Jesus, Criteria, and the Demise of Authenticity, NewYork – London, T&T Clark, 2012.

THE END OF Q STUDIES: CONTEMPORARY REFLECTIONS

71

and apocalyptic profile in conversation with historical Jesus studies, in all its (their) social, ethnic, economic, and political contexts. This “end” of Q Studies is promising, but not yet visible on the horizon. Its future is unwritten. It seems best, then, to leave it as is, a work in progress. University of California, Los Angeles United States of America

Simon J. JOSEPH [email protected]

REMINISCENCES AND ANALYSIS: HOW CLOSING THE DOOR ON THE FARRER HYPOTHESIS OPENED THE DOOR TO PROPOSING MATTHEAN POSTERIORITY I. INTRODUCTION I owe a debt of gratitude to Austin Farrer for his attempt to dispense with Q by arguing that Luke used Matthew (the Farrer Hypothesis = FH). Had I not been attracted by the FH’s promise of a way of making Q unnecessary, I should very likely have never come around to proposing the Matthean Posteriority Hypothesis (MPH)1. Had I not personally wrestled on a passage-by-passage basis with the many stubborn difficulties faced by those who want to argue that Luke used Matthew, I should have never fully appreciated how most of those difficulties vanish the moment one turns the sequence around and assumes that Matthew used Luke rather than Luke, Matthew. Q is the name assigned to a hypothetical source used by Matthew and Luke in addition to Mark to explain material they share that did not come from Mark. It is part of a solution to the Synoptic Problem known as the Two-Document Hypothesis (2DH). The presumed existence of Q rests on the conviction that Matthew and Luke could not have known one another’s Gospel. As Christopher M. Tuckett writes: “At one level, the Q hypothesis is simply a negative theory, denying the possibility that one evangelist made direct use of the work of the other”2. The reason many think this is the case is the numerous significant and sometimes seemingly irreconcilable differences existing between the two Gospels3. Yet at the same time Matthew and

1. R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1992) 1-22. 2. C.M. TUCKETT, The Existence of Q, in R.A. PIPER (ed.), The Gospel behind the Gospels: Current Studies on Q (SupplNT, 75), Leiden – New York – Köln, Brill, 1995, 19-47, p. 21. 3. Their very different genealogies being a conspicuous example, especially between King David and Joseph (Matt 1//Luke 3). Matthew’s traces his version through David’s son Solomon (Matt 1,6), but Luke traces his through David’s more obscure son Nathan (Luke 3,31). Further, Joseph’s father is named Jacob in Matthew (1,16) but Heli in Luke (3,23).

74

R.V. HUGGINS

Luke share a significant body of other material (the “double tradition”) that they did not get from Mark, material that often agrees verbatim in Greek4. But what if this were not so? What if the problems posed by the significant differences between Matthew and Luke were not insurmountable after all, and Q could ultimately be shown to be an unnecessary postulate? This was the argument put forward by Farrer in 1955 in his article “On Dispensing with Q”5. Q was unnecessary, he said, because Luke knew and used Matthew. Farrer’s view was further elaborated by his student Michael D. Goulder6, and then by Mark Goodacre, who wrote his dissertation on Goulder and whose ongoing defense of the FH is largely responsible for its current popularity7. My own flirtation with the FH dates back to the late 1980s when I was pursuing a doctorate with a major emphasis in Synoptics at the University of Toronto/Toronto School of Theology, which was strongly oriented at the time toward the dominant Two-Document Hypothesis (2DH)8. Although I had done my master’s thesis on the Gospel of Matthew9, I had no intention of studying Synoptics at Toronto. Indeed, shortly before going there while sitting one day in the quiet, airy reading room of Western Washington State University’s Wilson Library, paging through Rudolf Bultmann’s History of the Synoptic Tradition, I had firmly determined not to study Synoptics. Upon my arrival in Toronto, however, Richard N. Longenecker brushed aside this determination during our first meeting and informed me that I “had to” do Synoptics because “that’s what we do here”10. And so it was that I found myself in the first semester of my doctoral program in Heinz O. Guenther’s Q Seminar (Winter Session 1987-88).

4. See, e.g., Matt 7,7-8 = Luke 11,9-10, where 24 identical Greek words appear in the identical order. 5. A.M. FARRER, On Dispensing with Q, in D.E. NINEHAM (ed.), Studies in the Gospels: Essays in Memory of R.H. Lightfoot, Oxford, Blackwell, 1955, 55-88. 6. See M.D. GOULDER, Five Stones and a Sling: Memoirs of a Biblical Scholar, Sheffield, Sheffield Phoenix Press, 2009, and Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989. 7. Goodacre’s dissertation was published as M.S. GOODACRE, Goulder and the Gospels: An Examination of a New Paradigm (SupplJSNT, 133), Sheffield, Sheffield Academic Press, 1996. See also esp. ID., The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002. 8. A number of prominent Q scholars, including Wendy J. Cotter, John S. Kloppenborg, Robert A. Derrenbacker, Alan Kirk, and Daniel A. Smith, studied at Toronto around that time. 9. R.V. HUGGINS, Matthew’s Apocalyptic Timetable: A Reassessment of the Genre of Matt 1:1-17, M.C.S. Thesis, Regent College, Vancouver, BC, 1987. 10. As I write the sad news reached me of Professors Longenecker’s 7 June 2021 death.

REMINISCENCES AND ANALYSIS

75

As it turned out Longenecker had done me an enormous favor. The late 1980s and early ’90s were exciting times for both Historical Jesus and synoptic studies, and Toronto was about as good a place as any to be, especially in the case of the latter. But they were also perplexing times. In 1988 Marcus J. Borg, a New Testament Scholar whose star was just then rising, announced that a “major renaissance is occurring in North American Jesus studies”11. That same year N.T. Wright, whose towering celebrity as a Jesus scholar still lay in the future, first introduced to a broad audience his idea that a “Third Quest for the Historical Jesus” was underway12. The old “New Quest” announced by James M. Robinson in the late 1950s, it was said, had pretty much run its course, but now a vigorous new “Third Quest” was emerging. Wright’s idea was tantalizing, but the boundaries between the New Quest and Third Quest were never very clear. Three years previous, in March 1985, the Jesus Seminar had gotten underway with its “ringmaster-entrepreneur” Robert W. Funk grandly promising that its work, “if carefully and thoughtfully wrought, [would] spell liberty for … millions”13. John P. Meier, who had not yet published the first volume of his magisterial Marginal Jew: Rethinking the Historical Jesus series, would count the Jesus Seminar as Third Questers14, but Wright regarded them as essentially New Quest revivalists15. With typical bravado, Funk embraced the term “renewed questers” for himself and the Jesus Seminar, while branding pretty much everyone else as “pretend questers”16, including Third

11. M.J. BORG, A Renaissance in Jesus Studies, in TTod 45 (1988) 280-292, p. 280. 12. In S. NEILL – T. WRIGHT, The Interpretation of the New Testament 1861-1986, Oxford, Oxford University Press, 21988, pp. 379-403. Wright had earlier discussed this in his Towards a Third Quest? Jesus Then and Now, in ARC 10 (1982) 20-27. 13. R.W. FUNK, The Issue of Jesus, in Forum 1 (1985) 7-12, p. 8. The description “ringmaster-entrepreneur” was given to Funk in L.T. JOHNSON, The Real Jesus: The Misguided Quest for the Historical Jesus and the Truth of the Traditional Gospels, San Francisco, CA, HarperSanFrancisco, 1996, p. 6. 14. J.P. MEIER, The Present State of the ‘Third Quest’ For the Historical Jesus: Loss and Gain, in Bib 80 (1999) 459-487, p. 459. 15. N.T. WRIGHT, The Challenge of Jesus: Rediscovering Who Jesus Was and Is, Downers Grove, IL, InterVarsity Press, 1999, pp. 29-30. Early on Wright had become aware of Borg’s first book (M.J. BORG, Conflict, Holiness and Politics in the Teachings of Jesus, Lewiston, NY, Edwin Mellen Press, 1984) and featured Borg in his discussion of Third Quest authors (NEILL – WRIGHT, Interpretation [n. 12], pp. 387-391). This later created something of a tension when Wright categorized Borg’s associate, fellow Jesus Seminar member and co-author J.D. Crossan as essentially a revived New Quester. Wright would later describe Borg as “straddling the Jesus Seminar on the one hand and the post-Schweitzer ‘Third Quest’ on the other” (Jesus and the Victory of God, Minneapolis, MN, Fortress, 1996, p. 75). 16. R.W. FUNK, Honest to Jesus: Jesus for a New Millenium, San Francisco, CA, HarperSanFrancisco, 1996, p. 64.

76

R.V. HUGGINS

Questers like Meier, Raymond E. Brown, and presumably also Wright. These Funk accused of engaging in a mere “apologetic ploy”17. In the meantime, in the fall of 1983, Robinson initiated a research project at the Institute for Antiquity and Christianity for the purpose of reconstructing Q in conjunction with a Society of Biblical Literature Study of Q Consultation, which met for the first time at the SBL annual meeting in Dallas in that same year18. This led in 1985 to the SBL Q Seminar which met from 1985-1989 before becoming the International Q Project (IQP) which began its work of reconstructing Q in 1989 and completed it in 199619. This became the basis, with additional tweaking by the editors, for the publication in 2000 of The Critical Edition of Q (CritEd)20. The Jesus Seminar and the IQP both featured the 2DH as foundational to their work. And, in the early days, involvement in the two significantly overlapped. John S. Kloppenborg, for example, who had done his MA and Ph.D. at Toronto, and was at the time teaching at the University of Windsor, Ontario (before returning to Toronto in 1988), was both the co-chair of the International Q Project (with Robinson) and a charter member of the Jesus Seminar21. Guenther, Kloppenborg’s Doktorvater, whose perspective on Q I was now learning, was also involved in both the Jesus Seminar and the IQP, as was Michael G. Steinhauser, one of the directors of my doctoral program. Leif E. Vaage, another well-known Jesus Seminar/IQP crossover, would not come to Toronto until 199122. Thus, so far as I experienced it

17. Ibid., p. 65. 18. J.M. ROBINSON, The Sayings of Jesus: Q. The Vosburgh Lectures, October 26-27, 1983, in C. HEIL – J. VERHEYDEN (eds.), The Sayings Gospel Q: Collected Essays by James M. Robinson (BETL, 189), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, 177192, p. 192. For more details on the founding of the IQP, see F. NEIRYNCK, The International Q Project, in ETL 69 (1993) 221-225. 19. J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000, p. lxviii. See E.K. BROADHEAD, The (Mis)definition of Q, in JSNT 68 (1997) 3-12. 20. For a detailed account of the differences between the original IQP text and the published CritEd, see F. NEIRYNCK, The Reconstruction of Q and IQP / CritEd Parallels, in A. LINDEMANN (ed.), The Sayings Source Q and the Historical Jesus (BETL, 158), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001, 53-147. 21. J.M. ROBINSON, The International Q Project Work Session, 17 November 1989, in JBL 109 (1990) 499-501, p. 499, and https://www.westarinstitute.org/membership/westarfellows/fellows-directory/. 22. For a list of members of the Jesus Seminar in 1988, see R.W. FUNK – B.B. SCOTT – J.R. BUTTS, The Parables of Jesus: Red Letter Edition (Jesus Seminar Series, 1), Sonoma, CA, Polebridge, 1988, pp. 95-97.

REMINISCENCES AND ANALYSIS

77

at least, Toronto very much landed on the side of the revived New Quest rather than on that of the Third Quest as conceived of by Wright. The exception was Joseph Plevnik, S.J. who was neither Jesus Seminar nor IQP and who, so far as I recall, never actively promoted any particular solution to the Synoptic Problem, but made, in my case at least, an even stronger case for the 2DH by simply interrogating his students regarding what they were seeing in the text. But Guenther, who was by all accounts a most remarkable teacher, came first. Guenther’s influence was due in no small part to his fastidious care for his students23. He made the case for the 2DH very forcefully, but wed it closely to an extremely negative assessment of the Jesus tradition, facilitated by notions of how the evolution of Q within the developing Q community must have taken place. If he accepted any saying in the Gospels as originating with Jesus, I never knew it. Generally very mild in temper, gentle even, the most passionate I ever saw Guenther was when Jeffrey Weima, now of Calvin College, suggested that a particular synoptic saying of Jesus, might actually go back to the historical person Jesus. Guenther actually began pounding the table repeatedly to punctuate each word as he emphatically said: “I/will/have/no/historical/Jesus/in my/seminar!”24

At the time I could clearly see the elegance and explanatory force of the 2DH, but was left wondering whether Guenther’s singularly negative assessment of the Jesus tradition necessarily followed from it. It was this, along with the fact that I was hearing a very forceful albeit one-sided defense of the 2DH, that persuaded me I needed to explore alternative solutions, of which the FH seemed most promising.

23. Kloppenborg’s description of the reason he chose Guenther to direct his dissertation perfectly describes my own experience with him (Professor Heinz O. Guenther [1926-2003], in StudRel/SciRel 32 [2003] 182-184, p. 183): “For a graduate seminar paper I had written, he returned almost a dozen pages of careful, close and critical engagement with my argument and use of evidence, taking issue with some conclusions, suggesting ways to bolster other conclusions, and pointing me to other secondary literature. This was a model of graduate supervision and it was an index of the seriousness with which Guenther took his responsibility to aid young scholars to make that transition from student to independent researcher”. My article Matthean Posteriority (n. 1) was my paper for Guenther’s Q Seminar. I submitted it to a journal the same day and in the same form as I turned it in to Guenther. I still have the manuscript containing several pages of his comments. 24. Whenever I have recalled the story over the years to Weima, he has always stressed what a positive experience overall it was working with Guenther.

78

R.V. HUGGINS

My first encounter with the FH was through Farrer’s article. I do not recall whether I was aware of the fact that Goulder had also defended the position in his 1974 book Midrash and Lection in Matthew25. And although I was favorably inclined toward Farrer from other things I had read of his, I was less than impressed by his attempt at dispensing with Q. Yet I thought perhaps the view might still hold promise if worked out on a passage-bypassage basis as a sort of separate lens to be used alongside the 2DH. One thing the FH had in its favor over, for example, William R. Farmer’s revived Griesbach Hypothesis, was that it supported Markan Priority, of which I was already firmly convinced. From the first, however, I took for granted that if the FH were to be accepted as a viable alternative to the 2DH it had to account for the double tradition at least as adequately, and preferably more adequately, than the 2DH. So, from the beginning I considered irresponsible those solutions (like Goulder’s) that tried to account for the double tradition by asserting that Matthew and/or Luke simply invented whatever they did not get from Mark26. Such an approach evaded the question of other source dependencies, which, in the vein of E.P. Sanders and Margaret Davies, I could not accept as legitimate27. If Luke and Matthew were both so ready to invent the 25. M.D. GOULDER, Midrash and Lection in Matthew (The Speaker’s Lectures in Biblical Studies, 1969-71), London, SPCK, 1974, pp. 452-471. Earlier defenders of the view included E.W. LUMMIS, How Luke was Written, Cambridge, Cambridge University Press, 1915, p. vii; J.H. ROPES, The Synoptic Gospels, Cambridge, MA, Harvard University Press, 1934, p. 67; A.W. ARGYLE, Evidence for the View That St. Luke Used St. Matthew’s Gospel, in JBL 83 (1964) 390-396; J. DRURY, Tradition and Design in Luke’s Gospel: A Study in Early Christian Historiography, London, Darton, Longman & Todd; Atlanta, GA, John Knox Press, 1976, pp. 25, 36, 120-173; and E.P. SANDERS – M. DAVIES, Studying the Synoptic Gospels, London, SCM; Philadelphia, PA, Trinity Press International, 1989, p. 117. Sometimes H.G. JAMESON, The Origin of the Synoptic Gospels, Oxford, Blackwell, 1922, is wrongly included in this group because Jameson believes Luke used Matthew. But he also defends Matthean Priority (p. 6), which puts him closer to the position of B.C. BUTLER, St. Luke’s Debt to St. Matthew, in HTR 32 (1939) 237-308, and The Originality of St. Matthew: A Critique of the TwoDocument Hypothesis, Cambridge, Cambridge University Press, 1951, p. 170. GOODACRE, Goulder and the Gospels (n. 7), p. 118, n. 75, faults me for not including mention of Goulder in my original article. 26. I had also already come to feel that appeals to midrash as a justification for extensive invention on the part of the synoptic authors were too vague to be meaningful. In his Midrash and Lection (n. 25), Goulder used the term to describe Matthew’s process of inventing most of the material he did not get from Mark. But then, after P. Alexander criticized his use of the term as not consistent with the usage of the Rabbis, Goulder avoided using it in connection with his claim that Luke invented most of what he did not get from Matthew and Mark (cf. Luke [n. 6], p. 128). My own view of the limited usefulness of appeals to midrash was not directly influenced by Goulder, but started in connection with R.H. GUNDRY, Matthew, Grand Rapids, MI, Eerdmans, 21994. 27. Studying the Synoptic Gospels (n. 25), p. 117: “Thus far Goulder has not persuaded us that one can give up sources for the sayings material. With this rather substantial modification, however, we accept Goulder’s theory: Mathew used Mark and Luke used them both”.

REMINISCENCES AND ANALYSIS

79

constituent pieces of what went into making up their gospels, then why did they trouble themselves with weaving Mark into the midst of their inventions? We see this when we consider the extent of the invention that would be required from the perspective of the FH. First, Matthew would have needed to invent some 465 versus to add on to Mark (M = 230 and Q = 235), and then on top of that, Luke would have had to make up a further 400 verses (L= 400) to add on to Mark and Mathew28. That adds up to 865 verses of invented material all together, more than are found in the entirety of Mark’s Gospel (= 660). In addition to this, however, there was still the issue of the apparent direction of synoptic dependencies, which, as I was beginning to suspect, seemed less complicated when one assumed an independent Q that Matthew and Luke made of what they would, rather than a Luke relying on Matthew for double-tradition material. In this article, then, I will present the various passages and issues that, taken together, led me to conclude that the 2DH did a much better job of making sense of the double tradition than did the FH. Afterward I will briefly discuss how I initially came to the Matthean Posteriority Hypothesis (MPH) and explain how its treatment of the same issues and passages did not suffer the disadvantages of the FH when compared to the 2DH. The boundaries of the present article therefore are necessarily temporal, I only deal with the items that, to the best of my recollection, contributed to my rejecting the FH, avoiding as much as possible my more recent reflections on them, which have become more refined but, on the whole, have not essentially changed. I am already aware that this will not be possible to accomplish perfectly because I cannot recall perfectly what I knew and when. The only absolutely firm guide is the 1992 article where I initially proposed the MPH, which was identical to the essay I turned in for Guenther’s Q Seminar. In any case, the temporal limits are significant because, as my title implies, it was really the process of trying to make the FH work and having it fail, or seem to fail, that set me on the road to looking at things the other way round and proposing the MPH. The first step, however, was coming to a clear conviction that the 2DH was far simpler and more plausible than the FH.

28. The numbers for M and L are from E. BROADHEAD, The Extent of the Sayings Tradition (Q), in LINDEMANN (ed.), The Sayings Source Q (n. 20), 719-728, p. 723, and those for Q are from J.S. KLOPPENBORG, Q: The Earliest Gospel: An Introduction to the Original Stories and Sayings of Jesus, Louisville, KY – London, Westminster John Knox, 2008, p. 51.

80

R.V. HUGGINS

II. BASELINES To review the matter in its simplest form, the 2DH explains the doubletradition material by saying that Matthew and Luke independently used Q. The FH tries to dispense with Q by redefining the double tradition as material that Luke included from Matthew. The 2DH allows that in any given double-tradition passage Matthew or Luke might be following their shared source (Q) more closely. This was an argument for the existence of Q that was already articulated long before Farrer proposed his solution. B.H. Streeter, for example, wrote in the 1920s: “Sometimes it is Matthew, sometimes it is Luke, who gives a saying in what is clearly the more original form”29. The phenomenon is commonly referred to as “alternating primitivity”. Alternating primitivity does not come into play, however, in the case of the FH, which assumes that where Matthew and Luke differ in the double tradition it is because Luke has changed what he got from Matthew. Apart from special-case situations where some other factor intrudes, Matthew’s versions of double-tradition passages are always “more primitive” than Luke’s. As Goulder once wrote: “at all points the Lucan form is secondary, carrying over Matthean expressions and theology, and adapting them with Lucan expressions and theology”30. To be sure it can legitimately be objected that Goulder stated the matter too categorically. Might we not say rather that the FH implies that Matthew is usually more primitive than Luke in double-tradition passages rather than always? This is what I had in mind when I spoke of other factors intruding. Suppose, for example, that Luke had available to him versions of sayings or parables, prayers or liturgical formulae that had become standard within his own Christian community before he ever gained access to the Gospel of Matthew. Would that not imply that Luke might have preferred to use already familiar forms, even if “more primitive” than the ones presented to him in Matthew? As we shall see, Farrer made such an argument in connection with the Lord’s Prayer. Such questions, in my view, are perfectly legitimate, so long as we keep in mind that making such arguments sometime (as here) amounts to reidentifying or renaming underlying hypothetical sources rather than dispensing with them31. 29. B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, p. 183. 30. GOULDER, Midrash and Lection (n. 25), p. 452. 31. At least so long as the stated goal is to lessen appeals to hypothetical documents as such. One could argue that it doesn’t matter how many additional hypothetical sources we

REMINISCENCES AND ANALYSIS

81

Not all defenders of the FH consider this issue of “alternating primitivity” worthy of consideration. Farrer himself, for example, breezily dismissed the entire issue by insisting that there was “a deceptive simplicity about the proposed method of argument which evaporates as soon as we try to apply it. There is scarcely an instance in which we can determine priority of form without invoking questionable assumptions”32. By the time I first read this I was already convinced that Farrer was plainly incorrect. Alternating primitivity is just one aspect of the larger question of primitivity, and one cannot enter into discussing the Synoptic Problem at all without raising questions of relative primitivity of gospel parallels. This goes for defenders of any of the solutions. A case in point is the FH’s and 2DH’s joint acceptance of Markan Priority. For the 2DH the issue of relative primitivity is an essential building block for establishing Markan Priority. So, if the issue really has no relevance, is entirely dubious, “evaporates as soon as we try to apply it”, then on what basis does the FH ground its acceptance of Markan Priority, or for that matter of Lukan Posteriority? The very act of talking about how one Gospel took over something from another and modified it amounts to a claim about relative primitivity. But, with regard to Farrer, I already knew that a bad argument for a thesis was not the same thing as a bad thesis, and so I remained intrigued by the FH per se, as something separate from its defenders, as something that, starting from where I was, I might be able to think through to a place where I could view it as credible on the grounds of discovering better arguments. The first hurdle for me was the fact that even granting the 2DH’s idea of alternating primitivity, it was not as though it viewed Luke and Matthew as preserving the more primitive form of double-tradition passages equally often. Luke was generally conservative in the way he used his sources and therefore was viewed as retaining, for example, the original order of Q more must appeal to as a way of making the particular hypothetical document Q as generally conceived go away. But that is ultimately a rather pointless pursuit that really amounts to trying to make Q seem to go away by saying it had more parts than previously thought. The double tradition is still there, only now it is attributed to multiple hypothetical sources rather than just one. 32. FARRER, Dispensing with Q (n. 5), p. 64. See more recently, along these lines, F. Watson, who insists that “Each of the competing hypotheses can produce plausible examples of apparent ‘primitivity’ or ‘secondariness’, as required” (Gospel Writing: A Canonical Perspective, Grand Rapids, MI – Cambridge, Eerdmans, 2013, p. 162). But it was not only a matter of the different hypotheses being able to generate examples of apparent primitivity or secondariness as needed, but of whether the examples were equally plausible to those put forward in defense of other hypotheses. Had it been as simple a matter as Watson implies, I should have had no problem finding my way around the issue of primitivity, since at the time I was positively inclined toward accepting the FH.

82

R.V. HUGGINS

often than Matthew33. But I was quickly becoming convinced that Luke was also more conservative than Matthew in retaining the form of doubletradition passages as well, the opposite of what ought to be the case if the FH were correct. 1. “Our Father in the Heavens” (Matt 6,9) // “Father” (Luke 11,2) During the course of my time in Guenther’s Q Seminar, the Jesus Seminar, at their meeting in Atlanta on 13-16 October 1988, voted on the Lord’s Prayer and decided that only the word “Father” went back to Jesus. The rest of the first line, “Our … in the heavens”, was from Matthew not Jesus34, and the rest of the prayer was probably not from Jesus35. The vote naturally created a stir in the public press foreshadowing what was to come in terms of the Seminar’s reputation for attention seeking36. Such a claim concerning the first line could not help but come across as sensational to the average reader, especially when added to the further suggestion that the rest of the prayer didn’t come from Jesus. Who were these people who had the audacity to imagine they could vote the Lord’s Prayer off the lips of Jesus? But contextualized within the larger discussion of synoptic relationships as seen through the lens of the 2DH the decision regarding the first line was far less controversial than it seemed. Luke’s version of the first line of the Lord’s Prayer has only “Father”. Since this is a double-tradition passage it is quite 33. See already, H.J. CADBURY, The Style and Literary Method of Luke (HTS, 6), Cambridge, MA, Harvard University Press, 1920, p. 73. See further below. 34. H. TAUSSIG, The Lord’s Prayer, in Forum 4 (1988) 25-41, p. 30; R.J. MILLER, The Lord’s Prayer and Other Items from the Sermon on the Mount, in Forum 5 (1989) 177-186, p. 185; R.W. FUNK, The Lord’s Prayer: Does it Go Back to Jesus?, in The Fourth R 2 (1989) 1, 4-6, p. 1. When R.W. FUNK – R.W. HOOVER (eds.), The Five Gospels: The Search for the Authentic Words of Jesus, Sonoma, CA, Polebridge, 1993, came out, someone apparently bungled it and printed “Our Father” in red and “in the heavens” in black (p. 148). For a full account, see my Jesus, Abba, and the Seminar, in Midwestern Journal of Theology 8 (2009) 41-56. There were other differences between the early announcements and the final result as well. 35. MILLER, Lord’s Prayer (n. 34), p. 185; A.L. GOLDMAN, Religion Notes: Lord’s Prayer Isn’t His, Panel of Scholars Says, in The New York Times, October 23, 1988, 46. For any given saying each member would vote with a bead of one of four different colors, the significance of each being summarized as follows: “[1] Red: That’s Jesus! … [2] Pink: Sure sounds like Jesus. … [3] Gray: Well, maybe. … [4] Black: There’s been some mistake. No way!” Cf. R.W. FUNK, The Emerging Jesus, in The Fourth R 2 (1989) 1,11-15, p. 1. Essentially the same list (minus “No way!” in the last item) appears in FUNK – HOOVER, The Five Gospels (n. 34), p. 37. 36. GOLDMAN, Lord’s Prayer Isn’t His (n. 35), p. 46; M. BORG, What Did Jesus Really Say?, in Bible Review 5/5 (1989) 18-25, p. 22. On the larger media circus surrounding the Jesus Seminar see JOHNSON, The Real Jesus (n. 13), pp. 1-27.

REMINISCENCES AND ANALYSIS

83

easy to imagine that Luke preserved the original form of the prayer in Q while Matthew embellished it in ways typical of him (see below). But from the perspective of the FH the situation was the reverse. Matthew’s “Our Father in the heavens” had been pared down in Luke to include only the word “Father.” Farrer had hesitated to suggest that Luke had taken it upon himself to, as it were, cut the fat out of the Lord’s Prayer, on the grounds that its “words … must be sacred to a Christian”37. So he appeals instead to a hypothetical shorter local version that Luke ostensibly preferred to the form he encountered in Matthew. And he admits that Luke’s local version might have been more primitive than the form we find in Matthew38. In this way Farrer shifted the burden of the evidence from one hypothetical document to another. Goulder pursued a different tack. He insisted that “as the Matthean version of the Prayer shows strong traces of the Matthean style, so the Lucan version, in so far as it differs from the Matthean, shows strong traces of Lucan style”39. Curiously, the first example he offered of this was πάτερ from the first line of the Lord’s Prayer, which, he said, “is a characteristically Lucan address to God”. However that may be in a limited sense, it is certain that πάτερ is far more characteristically Matthean than Lukan. The Gospel of Luke with its 19,482 words calls God Πάτερ 16 times40, but Matthew with its 18,293 words – more than a thousand less than Luke – does so 43 times41. Matthew’s Sermon on the Mount refers to God as Πάτερ 17 times, i.e., more often than Luke does in his entire Gospel. Luke’s Sermon on the Plain refers to God as Πάτερ only once (Luke 6,36). Furthermore, there is a difference between a characteristic word and a characteristic phrase. The whole first line of the Lord’s Prayer in Matt 6,9 is characteristically Matthean, even though the word πάτερ is not in any sense uniquely Matthean. The appending of the words “in the heavens” (ἐν [τοῖς] οὐρανοῖς) to “Father” (Πάτερ) in the line is frequently repeated with minor variation elsewhere in Matthew42. This occurs almost exclusively in his 37. FARRER, Dispensing with Q (n. 5), pp. 64-65. Goulder rejects the “variant-version hypothesis” in The Composition of the Lord’s Prayer, in JTS 14 (1963) 32-45, pp. 32-33. 38. FARRER, Dispensing with Q (n. 5), p. 65. 39. GOULDER, Lord’s Prayer (n. 37), p. 33. 40. Number of words in Luke from A.W. PITTS, History, Biography, and the Genre of Luke-Acts: An Exploration of Literary Divergence in Greek Narrative Discourse (BIS, 177), Leiden – Boston, MA, Brill, 2019, p. 106. 41. Number of words in Matthew from R.H. STEIN, Mark (BECNT), Grand Rapids, MI, Baker Academic, 2008, p. 63, n. 1. 42. The appending of the words “in the heavens” (ἐν [τοῖς] οὐρανοῖς) to “Father” (Πάτερ) in the opening line of Matthew’s Lord’s Prayer (Matt 6,9) is frequently echoed with minor

84

R.V. HUGGINS

unique material (×7) or in double-tradition material (×5). In the latter cases it stands in the place of something else in Luke: “Most High” (Ὑψίστος Luke 6,35 = Matt 5,45)43; “Father from heaven” (sing.) (ὁ πατὴρ ὁ ἐξ οὐρανοῦ Luke 11,13 = Matt 7,11); “the angels of God” (τῶν ἀγγέλων τοῦ θεοῦ [×2] Luke 12,8 = Matt 10,32-33), and “God” (ὁ θεός Luke 12,24 = Matt 6,26). There is only one triple-tradition passage with Matthew linking “Father” with “in the heavens”. In that passage Luke 8,21 agrees with Mark 3,35 in having “God”, which Matthew replaces with “my Father in the heavens”. Did Luke root out or modify all examples of “Father” + “in the heavens”, or did Matthew introduce them? Matthew provides a further avenue to pursue this question in another favorite phrase of his: “The Kingdom of the heavens” (ἡ βασιλεία τῶν οὐρανῶν [×32]). None of the synoptic gospels are averse to using the word “heaven” (οὐρανός), but Matthew (×78) uses it more than twice as often as Luke (35×) and more than four times as often as Mark (×19). Mark and Luke generally prefer “heaven” in the singular, although there are a few exceptions44. Matthew’s use of “Kingdom of the heavens” runs right through his unique material45, and his double-46 and triple-tradition material47. This makes it possible to compare Matthew’s use of “Kingdom of the heavens” more generally with Luke’s and/or Mark’s. Luke uses “Kingdom of God” in double-tradition passages, where, again, Matthew usually has “Kingdom of the heavens”48. variation elsewhere in Matthew: 5,16.45 (Mt); 6,1 (Mt); 6,26 (Mt/Lk); 7,11 (Mt/Lk); 7,21 (Mt); 10,32.33 (Mt/Lk); 12,50 (Mt/Mk/Lk); 16,17 (Mt); 18,10.14 (Mt). 43. Luke refers to God as ὕψιστος (Most High) or attaches it as an adjective to God in 1,32.35.76 (Lk); 6,35 (Mt/Lk); 8,28 (Mt/Mk/Lk); Acts 7,48; 16,17. Matthew has πατρὸς ὑμῶν τοῦ ἐν οὐρανοῖς at Matt 5,45//Luke 6,35. In the single triple-tradition passage Jesus is called “the Son of the Most High God” in Mark 5,7 and Luke 8,28 and “son of God” in Matt 8,29. 44. See, e.g., Luke’s use of the plural in 10,20; 12,33; 18,22. The last mentioned agrees with Matt 19,21 in having ἐν οὐρανοῖς over against Mark’s ἐν οὐρανῷ (Mark 10,21). 45. Matt 5,10.19bis.20; 7,21; 13,24.44.45.47.52; 16,19; 18,23; 19,12; 20,1; 25,1. 46. Matt 5,3//Luke 6,20; Matt 8,11//Luke 13,29; Matt 11,11//Luke 7,28; Matt 11,12// Luke 16,16; Matt 13,33//Luke 13,20; Matt 22,2//Luke 14,15; Matt 23,13//Luke 11,52. 47. Matt 3,2//Mark 1,4//Luke 3,3; Matt 4,17//Mark 1,15//Luke 4,15(?); Matt 10,7// Mark 6,7-8(?)//Luke 9,2; Matt 13,11//Mark 4,11//Luke 8,10; Matt 13,31//Mark 4,30// Luke 13,18; Matt 18,1//Mark 9,34-35//Luke 9,46; Matt 18,3//Mark 9,36-37//Luke 9,48; Matt 18,4//Mark 9,36-37//Luke 9,48; Matt 19,14//Mark 10,14//Luke 18,16; Matt 19,23// Mark 10,23//Luke 18,24. 48. Matt 5,3//Luke 6,20; Matt 8,11//Luke 13,29; Matt 11,11//Luke 7,28; Matt 11,12// Luke 16,16; Matt 13,33//Luke 13,20; Matt 22,2//Luke 14,15. The reason I say “usually” is because Matthew himself unexpectedly echoes Luke’s (Matt 12,28 = Luke 11,20) and Mark’s and Luke’s (Matt 19,24 = Mark 10,25//Luke 18,25) “Kingdom of God”. He also has the phrase once in a triple-tradition passage where the other two do not (Matt 21,43 = Mark 12,11//

REMINISCENCES AND ANALYSIS

85

When Matthew encounters the phrase “Kingdom of God” in Mark, he almost always replaced it, usually with “Kingdom of the heavens”49. In contrast, Luke almost always retains Mark’s “Kingdom of God” as over against Matthew’s “Kingdom of the heavens” or whatever else Matthew has instead. Luke also inserts “Kingdom of God” on a few occasions into triple-tradition passages where Matthew and Mark do not include it50. Like Πάτερ + ἐν [τοῖς] οὐρανοῖς the phrase “Kingdom of the heavens” is distinctly Matthean. Luke and Mark do not use it. From the perspective of the FH, we would have to say both phrases are fastidiously avoided by Luke. But why would Luke wish to replace “Our Father in the heavens” with “Father”? As we have seen he has no aversion to Πάτερ or even to ἐν [τοῖς] οὐρανοῖς, which he has three times (Luke 10,20; 12,33; 18,22). Whether one accepts Farrer’s suggestion that Luke preferred his local, more primitive, version of the Lord’s Prayer, or Goulder’s claim that Luke had removed “Our … in the heavens” as part of his general program of trimming out unnecessary material, the fact remains that the parts that were removed turned out to be just the sort of parts Matthew would have been inclined to add. And this for me at the time represented part of a bigger issue in which the FH seemed to imply that Luke’s aversion was not merely to a lack of brevity or anything of that sort, but to distinctively Matthean redactional features. Why did Luke, when taking over material from Matthew, labor at stripping away specifically Matthean details51?

Luke 20,17), and once in his own unique material (Matt 21,31). Finally, Mark uses the phrase without parallel in his own unique material at 4,26 and 12,34, the former introducing a Kingdom parable not found in the other two Gospels. 49. Matt 4,17//Mark 1,15; Matt 13,11//Mark 4,11; Matt 13,31//Mark 4,30; Matt 19,14// Mark 10,15; Matt 19,23//Mark 10,23. Matthew also replaces it with “life” (ζωή Matt 18,9// Mark 9,47), “my Father’s Kingdom” (Matt 26,29//Mark 14,25), or simply removes it (Matt 27,57//Mark 15,43). 50. Luke 9,11//Matt 14,14//Mark 6,34; Luke 18,30//Matt 19,29//Mark 10,29; Luke 21,31//Matt 24,33//Mark 13,29. 51. This double tradition phenomenon is comparable to what Streeter had complained about relating to the triple tradition when he wrote (The Four Gospels [n. 29], p. 183): “If then Luke derived this material from Matthew, he must have gone through both Matthew and Mark so as to discriminate with meticulous precision between Marcan and non-Marcan material; he must then have proceeded with the utmost care to tear every little piece of the non-Marcan material … from the context of Mark in which it appeared in Matthew – in spite of the fact that contexts in Matthew are always exceedingly appropriate – in order to re-insert it into a different context of Mark having no special appropriateness”. In order for the FH to work, the claim that Luke’s placement “has no special appropriateness” would need to be countered. At the time I saw Streeter’s statement as more directly relevant to Luke’s treatment of double-tradition material than triple-tradition material.

86

R.V. HUGGINS

2. “Woe to you Scribes Pharisees Hypocrites” (Matthew 23//Luke 11) One does not have to argue that Matthew was the inventor of material he didn’t get from Mark to notice that he never hesitates to expand, or contract, or elaborate on the Jesus tradition at will throughout his Gospel, and that not excluding the sayings of Jesus. Indeed, one can say without exaggeration that such is a distinctive feature of his compositional procedure: elaboration and the imposition of symmetries at both the macro and micro levels. On the macro level this is apparent in such things as the conspicuous threefold structure that comes into play in his Gospel: a three-part genealogy at the beginning, a Trinitarian baptismal formula at the end, and, as some have argued, the entire Gospel divided into three by the phrase ἀπὸ τότε ἤρξατο ὁ Ἰησοῦς … (“From that time Jesus began…”) at Matt 4,17 and 16,21 (buttressed, in T.B. Slater’s opinion, by a chiasm)52. Matthew had also gathered together the teachings of Jesus into five major blocks, each concluding with a nearly identical formula that begins καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησοῦς… (“And it happened when Jesus had finished…”) (Matt 7,28; 11,1; 13,53; 19,1; 26,1)53. The FH needed a Luke that was a very active dismantler of Matthew’s symmetries and conflator of that material with other unaccounted for hypothetical non-Markan sources. The passage that especially crystalized this problem for me was the series of woes presented in Matthew 23, each with the same stereotyped introduction repeated six times without variation: Οὐαὶ ὑμῖν, γραμματεῖς καὶ Φαρισαῖοι ὑποκριταί… (“Woe to you Scribes, Pharisees, hypocrites…”)54. The section represents part of Matthew’s expansion of Jesus’s attack on the Scribes in Mark 12,37b-40 in which he adds a series of woes from another source. Since Mark’s Jesus is condemning the behavior of the Scribes at that point, it is a perfectly obvious place for Matthew to have included his additional series of woes. 52. E. KRENTZ, The Extent of Matthew’s Prologue: Toward the Structure of the First Gospel, in JBL 83 (1964) 409-414 (= Der Umfang des Matthäus-Prologs. Ein Beitrag zum Aufbau des ersten Evangeliums, in J. LANGE [ed.], Das Matthäus-Evangelium [Wege der Forschung, 525], Darmstadt, Wissenschaftliche Buchgesellschaft, 1980, 316-325); J.D. KINGSBURY, Matthew: Structure, Christology, Kingdom, Philadelphia, PA, Fortress, 1975, pp. 7-25; F. NEIRYNCK, ΑΠΟ ΤΟΤΕ ΗΡΞΑΤΟ and the Structure of Matthew, in ETL 64 (1988) 21-59 (= ID., Evangelica II, ed. F. VAN SEGBROECK [BETL, 99], Leuven, Leuven University Press – Peeters, 1991, 141-182); T.B. SLATER, Notes on Matthew’s Structure, in JBL 99 (1980) 436. 53. B.W. BACON, Studies in Matthew, New York, Henry Holt, 1930, pp. 81-82. But see also, already, F. SCHLEIERMACHER, Ueber die Zeugnisse des Papias von unsern beiden ersten Evangelien, in Theologische Studien und Kritiken 5 (1832) 746-749, and Sir J.C. HAWKINS, Probabilities as to the So-Called Double Tradition of St. Matthew and St. Luke, in W. SANDAY (ed.), Studies in the Synoptic Problem by Members of the University of Oxford, Oxford, Clarendon, 1911, 95-138, p. 121. 54. Apart from the δέ placed between Οὐαί and ὑμῖν in Matt 23,13.

REMINISCENCES AND ANALYSIS

87

Luke’s parallels to Matthew’s woes appear rather in a conflict story from his greater interpolation. An unnamed Pharisee invites Jesus to come and dine with him and his friends, but is astonished when Jesus does not wash his hands before eating (Luke 11,33-34). Jesus responds with a series of woes (Luke 11,37-52), three against the Pharisees, and three against the Lawyers (νομικοί). Assuming the FH, Luke strips them all of their Matthean stereotyped introductions which he replaces with irregular ones. (1) Luke 11,39: Νῦν ὑμεῖς οἱ Φαρισαῖοι, “Now you Pharisees” (= Matt 23,25). (2) Luke 11,42: Ἀλλὰ οὐαὶ ὑμῖν τοῖς Φαρισαίοις, “But woe to you Pharisees” (= Matt 23,23). (3) Luke 11,44 οὐαὶ ὑμῖν “Woe to you” (= Matt 23,27). (4) Luke 11,47 οὐαὶ ὑμῖν, “Woe to you” (= Matt 23,29). (5) Luke 11,52 οὐαὶ ὑμῖν τοῖς νομικοῖς “Woe to you Lawyers” (= Matt 23,13). Luke also supplies the addition οὐαὶ ὑμῖν τοῖς Φαρισαίοις, “Woe to you Pharisees” at 11,43, which he tacks on to a saying taken from Matt 23,6. From the perspective of the 2DH it seemed to me likely that the whole of Luke 11,37-52 appeared in essentially the same form in Q as in Luke55, making allowances, of course, for the addition of Lukan stylistic touches, such as ἐν … τῷ + infinitive (11,37)56, εἶπεν δέ57 and πρός + accusative for speaking to (11,39)58, and καὶ αὐτοί (11,46)59. Then Matthew would have come along in typical fashion and cherry-picked the woes, fitted them up with stereotyped introductions, and inserted them into an obvious place in the Markan outline. If we accept the perspective of the FH, we know Luke’s source of the woes: Matthew 23. We also know that the woes did not originate as part and parcel of the larger conflict story where they appeared in Luke. And we know that Luke changed the form of Matthew’s woes. But none of this reduces our reliance on hypothetical sources. Indeed, it further complicates the issue. From the perspective of the 2DH, the woes come to both Matthew and Luke from Q. But from the perspective of the FH, we begin by having no idea where Matthew got the woes that he added to Mark which Luke 55. (Minimal) Q 11,?39a?.42.39b.41.43-44 (I preserve the CritEd’s sigla). The 2DH can say little about the original boundaries of Q. 56. J.C. HAWKINS, Horae Synopticae: Contributions to the Study of the Synoptic Problem, Oxford, Clarendon, 21909, p. 32; N. TURNER, Style (A Grammar of New Testament Greek, by J.H. Moulton, vol. IV), Edinburgh, T&T Clark, 1976, p. 47; J. JEREMIAS, Die Sprache des Lukasevangeliums: Redaktion und Tradition im Nicht-Markusstoff des dritten Evangeliums (KEK, Sonderband), Göttingen, Vandenhoeck & Ruprecht, 1980, p. 205; J.A. FITZMYER, The Gospel according to Luke I–IX (AB, 28), Garden City, NY, Doubleday, 1981, p. 110. 57. HAWKINS, Horae Synopticae (n. 56), p. 31; JEREMIAS, Sprache (n. 56), p. 33; CADBURY, Style (n. 33), p. 169; FITZMYER, Luke I–IX (n. 56), p. 110. 58. HAWKINS, Horae Synopticae (n. 56), p. 36; TURNER, Style (n. 56), p. 54; FITZMYER, Luke I–IX (n. 56), pp. 111, 116. 59. HAWKINS, Horae Synopticae (n. 56), p. 33; FITZMYER, Luke I–IX (n. 56), p. 110.

88

R.V. HUGGINS

latter took over. We also have no idea where Luke got the conflict-at-dinner story into which he fitted the woes allegedly taken over from Matthew. Nor is it easy to comprehend why Luke felt it necessary to move the woes from their location in the Markan outline where Matthew had placed them in the first place (Mark 12,37b-40 = Luke 20,45-47). If it is asked why Luke “primitivized” Matthew’s series of woes by stripping away their stereotyped introductions, it might be argued that he did so to make them fit more easily into the conversational format in which they appear in Luke’s story. But that is only part of what we mean by primitivizing. Another part involves Luke allegedly taking something plainly stated in Matthew and recasting it into something whose meaning is obscure, as for example, in Matt 23,13 = Luke 11,52: Matt 23,13: “But woe to you, scribes and Pharisees, hypocrites! For you lock people out of the kingdom of the heavens (τὴν βασιλείαν τῶν οὐρανῶν). For you do not go in yourselves, and when others are going in, you stop them”.

Luke 11,52: “Woe to you lawyers! For you have taken away the key of knowledge (τὴν κλεῖδα τῆς γνώσεως); you did not enter yourselves, and you hindered those who were entering”.

Luke does not say what people are being kept out of, but Matthew is explicit: “the Kingdom of the heavens”. Luke also does not explain what the “Keys of Knowledge” are. Neither the phrase not the words κλείς or γνῶσις are characteristic of Luke. The former word appears only here and the latter only once elsewhere in the prayer of Zechariah in Luke 1,77, where it refers to the knowledge of salvation (γνῶσιν σωτηρίας). Matthew uses the word κλείς only one time as well, in the famous passage where Jesus promises Peter “the keys to the Kingdom of the heavens” (Matt 16,19 τὰς κλεῖδας τῆς βασιλείας τῶν οὐρανῶν). But he never uses γνῶσις. Matthew’s woes are aimed at the Scribes and Pharisees (γραμματεῖς καὶ Φαρισαῖοι). If the FH were correct, we would need to explain why Luke replaced γραμματεῖς with νομικοί. If νομικοί was original to Q we should have no problem explaining why Matthew might replace it with γραμματεῖς. In the first place it is the γραμματεῖς in particular that Mark singles out for criticism in Mark 12,38, which would make it entirely natural for Matthew to prefer that term over Q’s νομικός. In addition, Matthew uses the term νομικός only once (Matt 22,35)60, in reference to the man whom Jesus describes in Mark as “not far from the Kingdom” (Mark 12,34). Both 60. Matthew’s single use may have been slipped in later through the influence of Luke 10,25.

REMINISCENCES AND ANALYSIS

89

Matthew and Luke often make reference to the Scribes as being hostile to Jesus (Matt ×23 // Luke ×19), but apart from this passage Luke refers to νομικοί only three other times (Luke 7,30; 10,25; 14,3). So again, why would Luke replace the γραμματεῖς he found in Matthew with νομικοί? It seemed to me then that these points ought not be gingerly explained away, but faced square on for what they really are: signs of apparent primitivity. Not only did the 2DH require less material accounted for by appeals to invention or to additional hypothetical documents, but assuming that Luke did in fact remain closer to the Q original than Matthew seemed to me to make more sense. It really did look like it was Matthew that introduced the stereotyped introductions in the course of adapting the woes for inclusion in his expansion of Mark 12,37b-40. It was also Matthew that introduced γραμματεῖς to replace νομικοί, and he again who replaced Luke’s confusing reference to the taking away of “the keys of knowledge” with a clear reference, framed in his own distinctive terminology, to locking people out of “the Kingdom of the heavens”. Finally, so far as Luke was concerned, there seemed to be enough going on in his version of the passage to suggest that Luke was not the inventor of the story of the supper dispute but that it came rather from his source, which in the case of the 2DH would be Q. 3. Matthew Clarifies the Identity of the Son of Man (Matt 5,11; 10,32; 16,21) Another case of Matthew’s clarifying things that struck me at the time was his specific identification of the Son of Man as Jesus himself in two double-tradition passages and one triple-tradition passage. One of Luke’s beatitudes blesses those who are excluded and called evil “on account of the Son of Man” (Luke 6,22). Matthew’s parallel version has instead “on my account”, with ἐμοῦ referring to Jesus (Matt 5,11). In Luke 12,8 Jesus says that “everyone who acknowledges me before others, the Son of Man also will acknowledge before the angels of God. In Matthew, Jesus says instead “…I also will acknowledge before my Father in the heavens” (Matt 10,32). The first of Mark’s three passion predictions begins with “the Son of Man must undergo great suffering…” (Mark 8,31). Luke 9,22 follows Mark closely here, but again Matthew has instead: “Jesus began to show his disciples that he [Jesus] must go to Jerusalem and undergo great suffering… (Matt 16,21)”61. 61. In one instance the situation is reversed. In the question leading to Peter’s confession of Jesus’s messiahship at Caesarea Philippi, Luke follows Mark in having “Who do people/

90

R.V. HUGGINS

4. The Sign of Jonah (Luke 11,29//Matt 12,38-42) One of the earliest difficulties I ran up against with regard to the FH was also one of the most straightforward: the Sign of Jonah: “This generation is an evil generation; it asks for a sign, but no sign shall be given to it except the sign of Jonah” (Luke 11,29). The explanation of the sign given by Matthew is that “just as Jonah was three days and three nights in the belly of the sea monster, so for three days and three nights the Son of Man will be in the heart of the earth” (Matt 12,40). In agreement with many defenders of the 2DH before and since, Guenther regarded Luke’s version to be closer to Q’s than Matthew’s, and the link between the sign of Jonah and Jesus’s death and resurrection as Matthew’s interpretive embellishment62. But Guenther went further than some in insisting that Matthew’s explanation was not only secondary, it was also wrong. For Guenther, the Q version had intended no reference at all either to Jonah’s extraordinary conveyance to Nineveh in the belly of a fish, or to Jesus’s death, burial and resurrection. Rather the sign of Jonah subsisted in Jonah’s call for repentance and/or announcement of impending judgment. As Jonah had preached judgment against Nineveh, so too had Jesus against “this generation”. The Ninevites had repented, “this generation” had not. But viewed from the perspective of the FH, it was Guenther that could not be right. Matthew is so clear in what he says (cf. Matt 12,40) that Luke could simply not have misunderstood the point. In short, if one accepts the FH, the rationale of some defenders of the 2DH for driving a wedge between the two evangelists’ versions in terms of their meaning and between the Jesus of the New Testament Kerygma and the Q Jesus conveniently vanishes. I certainly saw the attractiveness of the FH in offering this solution, but ultimately had to side with the 2DH in viewing Matthew as having added the explanation of 12,40. It was not that I saw the necessity the crowds say that I am?” (Mark 8,27//Luke 9,18), but in Matthew Jesus asks instead: “Who do people say that the Son of Man is?” (Matt 16,13). Despite the fact that in this latter case the move is from me in Mark/Luke to Son of Man in Matthew, it is still Matthew who introduces the change. For an instance in which Luke has “Son of Man” in a triple-tradition passage where Matthew and Mark do not, see Luke 22,48//Matt 26,50//Mark14,45. 62. A. HARNACK, The Sayings of Jesus: The Second Source of St. Matthew and St. Luke, London, William & Norgate; New York, G. P. Putnam’s Sons, 1908, p. 23; R. BULTMANN, The History of the Synoptic Tradition, trans. John Marsh, New York, Harper & Row, 21968, pp. 112-113; W.G. KÜMMEL, Promise and Fulfillment: The Eschatological Message of Jesus, Naperville, IL, Alec R. Allenson, 1957, p. 68; R.A. EDWARDS, The Sign of Jonah in the Theology of the Evangelists and Q (Studies in Biblical Theology, II/18), London, SCM, 1971, pp. 96-99; J.M. ROBINSON, The Son of Man in the Sayings Gospel Q, in HEIL – VERHEYDEN (eds.), The Sayings Gospel Q (n. 18), 405-425, p. 413.

REMINISCENCES AND ANALYSIS

91

of the wedge, which relied on a particular view of who “this generation” referred to and when they were active, only that Luke preserved the Q saying in its more primitive form and that Matthew had either provided the explanation of the sign himself or derived it from some other tradition accessible to him. My problem was that of Harnack who, it seemed to me, was right when he said of Matthew’s explanation that it “would never have been omitted by Luke if he had read it in his source”63. Why would Luke, encountering in his source a clear prediction by Jesus himself of his coming death, burial, and resurrection, elect to suppress it64? 5. Salt of the Earth//Light of the World (Matt 5,13-16) In his Sermon on the Mount, Matthew coordinates two sayings, one about salt and the other about light, using the parallel introductory phrases: “You are the salt of the earth …You are the light of the world” (Ὑμεῖς ἐστε τὸ ἅλας τῆς γῆς … Ὑμεῖς ἐστε τὸ φῶς τοῦ κόσμου). Mark has his own version of both sayings in two different places in his Gospel (Mark 9,49-50; 4,21), as does Luke in two different places in his greater interpolation (Luke 14,34-35; 11,33). Luke also has a second version of the light saying paralleling its location in Mark (Luke 8,16 = Mark 4,21). Matthew’s and Luke’s versions of both sayings share details not included in Mark’s, pointing, according to the 2DH, to the existence of a Q version. So, for example, Matthew and Luke share the nearly identical line: ἐὰν δὲ [Luke: καὶ]65 τὸ ἅλας μωρανθῇ, ἐν τίνι ἁλισθήσεται/ἀρτυθήσεται (“but if salt has become insipid, how shall it be salted/seasoned?”). Mark for his part has ἐὰν δὲ τὸ ἅλας ἄναλον γένηται, ἐν τίνι αὐτὸ ἀρτύσετε (“but if salt becomes unsalty, 63. HARNACK, Sayings (n. 62), p. 23. Goulder satisfied himself by suggesting that “the fact that Jesus was really only one day and two nights in the tomb leads him [Luke] to suppress the resurrection reference”, and that “Matthew allegorizes … but then Luke deallegorizes” (Midrash and Lection [n. 25], p. 335). And Farrer, who did not seem to register the issue as relevant at all, says only: “St. Luke simplifies the sign of Jonah by omitting the distracting allegory on the whale’s belly and the Easter sepulchre (Matt. 12.40)” (Dispensing with Q [n. 5], p. 73). 64. An obvious counter example is where Mark reports at the conclusion of the transfiguration how Jesus ordered Peter, James, and John “to tell no one about what they had seen, until after the Son of Man had risen from the dead” (Mark 9,9-10). Matthew repeats this prediction of the resurrection (Matt 17,9), but Luke excludes it from where it would be expected to appear in Luke 9,36. But the situation in this case is complicated by Luke’s inclusion of a reference to Moses’ and Elijah’s discussing Jesus’s “departure, which he was about to accomplish at Jerusalem” (Luke 9,31), which neither Matthew nor Mark have. 65. Luke’s addition of καί after δέ is typical of him. E.g., HAWKINS, Horae Synopticae (n. 56), p. 30; JEREMIAS, Sprache (n. 56), pp. 78-79; FITZMYER, Luke I–IX (n. 56), p. 110.

92

R.V. HUGGINS

how shall you season it?”). Matthew and Luke also include reference to unsalty salt being considered useless and thrown away, a detail not found in Mark. Luke agrees with Mark in having “Salt is good” in place of Matthew’s “You are the Salt of the Earth”. A similar situation obtains with the light saying. Matthew’s and Luke’s versions (Matt 5,14-15//Luke 11,33; 8,16) contain a detail not shared by Mark, i.e., the reason a lamp is placed on a stand66. And again, neither Luke nor Mark have Matthew’s opening coordinating phrase “You are the light of the world” or Matthew’s following statement “A city built on a hill cannot be hid”. Looking at this from the perspective of the 2DH, what seemed to be happening was fairly straightforward. The form of the sayings in Q and Mark were relatively similar in content except that the Q version was slightly longer in both cases. They were not originally connected in Q or Mark, but were drawn together by Matthew as part of his speech building process and provided with the nicely symmetrical parallel introductory statements: “You are the salt of the earth … You are the light of the world…”. Matthew could have been conflating the two sayings with another source that contained the two nicely balanced phrases, or he could have provided them himself. In either case the move represented a typical example of what Vincent Taylor called “Matthew’s love of orderly arrangement”67. From the perspective of the FH, everything is looked at the other way around, with Luke deriving material from Matthew, breaking it apart, removing the two nicely balanced introductions, and placing its two sayings in two different places in his greater interpolation. Matthew’s love of orderly arrangement, being everywhere conspicuous in his Gospel, remains a constant regardless of whether one comes at it from the perspective of the 2DH, the FH, or any other proposed solution to the Synoptic Problem, so the idea of his drawing the two sayings and neatly coordinating them fits right into that. Consequently, what initially drew my attention to this passage was that, from the perspective of the FH, it represented yet another case of Luke’s obliterating Matthean symmetries in the process of taking over materials from him. In other words, it was a parallel to what we have already seen Luke allegedly doing with the woes of Matthew 23. The most obvious question for me in hopefully considering the FH was what compelled Luke to remove the parallel introductory 66. The same detail may have been included in Luke’s version of the saying paralleling Mark (Luke 8,16//Mark 4,21), but it is missing from both Vaticanus and 𝔓75. 67. V. TAYLOR, The Gospels: A Short Introduction, London, Epworth, 61948, p. 76.

REMINISCENCES AND ANALYSIS

93

phrases in order to insert the two sayings in the reverse order in two unrelated places in his greater interpolation (Matt 5,13-15 = Luke 14,34-35 + 11,33). Why could not Luke have contented himself with retaining Matthew’s nicely constructed double saying in his own Sermon on the Plain, where these sayings would have appeared between Luke 6,26-2768? In order to consider the FH viable, I had to provide some sort of compelling reason to suppose that Luke did what he would have needed to do with Matthew’s material in order to make it appear as it does in his Gospel. Whatever may be said about the salt and light sayings considered in themselves, for me they represented a single example of something bigger, of Luke’s dismantling of Matthew’s Sermon on the Mount and scattering pieces, as it were, to the wind, so that they came to rest in 6 different chapters of his Gospel (Luke 6, 11–14 and 16) and with the bits scrambled around in terms of order69. It is good to remember in this connection that both the Sermon on the Mount and Mission Discourse are twice as long in Matthew as in Luke with the deconstructed material mostly scattered in Luke’s greater interpolation70. Luke’s overall project of dismantling Matthew’s material and resettling its pieces in other places in his Gospel, therefore, struck me as unwieldy and unnaturally labor intensive, not to say generally pointless, due to the fact that, as Streeter said, “contexts in Matthew are always exceedingly appropriate”71. 6. Dr. Jekyll and Mr. Hyde Also significant for me in this connection was the fact that when, according to the FH, Luke shifts over from following Mark to following Matthew, he suddenly undergoes a strange Dr. Jekyll and Mr. Hyde transformation 68. Answering questions like this involves more than simply showing that the particular piece of tradition makes sense in the setting which each evangelist has given it. Indeed, that such should be the case is only to be expected. And if there was a Q it would be expected for that document as well. It is not enough, for example, to assert for the passage under discussion that Luke removed Matthew’s stereotyped “You are the Salt of the earth” because Luke’s Jesus was not speaking to the disciples, or that Luke removed Matthew’s “You are the light of the world” in order to join the rest of the saying more easily with the lamp of the body saying taken from later in Matthew’s Sermon on the Mount (Luke 11,34-36//Matt 6,22-23). This would amount to a mere description attached to an assertion of a direction of dependence, and such statements can be as easily framed in defense of any synoptic solution one favors. 69. See the comparative chart in my Matthean Posteriority (n. 1), p. 11. 70. B.H. STREETER, On the Original Order of Q, in SANDAY (ed.), Studies in the Synoptic Problem (n. 53), 141-164, p. 143. 71. ID., The Four Gospels (n. 29), p. 183.

94

R.V. HUGGINS

in terms of his redactional habits. As H.J. Cadbury noted just over a century ago, “The starting point for any study of Luke’s method of using sources is a comparison of Luke and Mark”72, because of the “closeness with which Luke follows Mark, as compared with the freedom of paraphrase and embellishments in other ancient writers”73. Joachim Jeremias similarly noted that Luke “has kept almost pedantically to the sequence of the Markan pericopes”74 and called him “ein Feind von Umstellungen” (“an enemy of rearrangement”)75. Hence, the idea that Luke was more likely to have preserved the original order of Q than Matthew. As Streeter had said, “we should a priori expect that where Matthew and Luke differ the original order of Q is to be presumed to be that of Luke unless in a particular case a reason to the contrary can be assigned”76. When we approach the matter in this way from the perspective of the 2DH, we see a uniform compositional pattern emerge in which “sections of Luke derived from Mark and those of other origins are arranged in continuous blocks and not interspersed as in the Gospel of Matthew”77, and in which, as T.W. Manson observed, “Matthew conflates his sources … Luke selects from them”78. The recognition that the Luke who used Mark was the same author with the same redactional preferences as the one using his other sources made it possible to make the case, despite Matthew’s more general disregard for the order of his sources, that the double-tradition material was derived from a shared source of common arrangement79. This becomes clear, for example, even in the Sermon on the Mount (Plain). When one lays out Luke’s parallel passages according to Matthew’s order, there is less agreement in the sequence of passages than when one lays out Matthew’s parallels according to Luke’s order80. Thus, when Taylor writes that “Luke 72. CADBURY, Style (n. 33), p. 73. 73. Ibid. See Cadbury’s tally of changes in the order of sections and within sections (pp. 76-78). 74. J. JEREMIAS, New Testament Theology. Part One: The Proclamation of Jesus (The New Testament Library), trans. John Bowden, London, SCM; New York, Scribner’s, 1971, p. 40. 75. ID., Die Abendmahlsworte Jesu, Göttingen, Vandenhoeck & Ruprecht, 31960, p. 92. ET: The Eucharistic Words of Jesus (The New Testament Library), London, SCM; New York, Scribner’s, 1966, p. 98. See also ID., Neutestamentliche Theologie. Erster Teil: Die Verkündigung Jesu, Gütersloh, Mohn, 1971, p. 48 (not in the English edition). 76. STREETER, On the Original Order of Q (n. 70), p. 147. 77. CADBURY, Style (n. 33), p. 76. 78. T.W. MANSON, The Sayings of Jesus as Recorded in the Gospels according to St. Matthew and St. Luke Arranged with Introduction and Commentary, London, SCM, 1949, p. 15. 79. See esp. V. TAYLOR, The Order of Q, in ID., New Testament Essays, Grand Rapids, MI, Eerdmans, 1972, 90-94; ID., The Original Order of Q, ibid., 95-118, p. 97. 80. In Matthew 5, for example, when parallels are set out following Luke’s order, Matthew’s sequence is continuous with the exception of the intrusion of Matt 7,10 after Matt 5,41. In contrast, when Matthew’s order is followed, there appears to be a number of intrusions into

REMINISCENCES AND ANALYSIS

95

has preserved the order of Q and had followed it with great fidelity”, he is offering his statement not as a mere presupposition, but as something that seems to be demonstrable81. Little has changed on this score since I first wrestled with this issue some thirty years ago, because the texts of Matthew and Luke remain what they are. Harry T. Fleddermann, for example, has stressed more recently that “Luke’s redactional procedure on the macrolevel makes it possible for us to reconstruct the original order of Q” 82, noting that “If Luke had followed a procedure anything like Matthew’s, the original order of Q would probably always lie beyond our grasp”. It is in fact true generally that defenders of the 2DH assume that Luke preserves the original order of Q to a much greater extent than Matthew. Thus, when The Critical Edition of Q was published in 2000, it included a list of sayings not in Lukan sequence that offered only eleven places where Matthew’s order was thought to be more primitive than Luke’s83. It was for this reason that the convention was adopted of referring to Q passages according to where they appear in Luke’s Gospel rather than Matthew’s, so that, for example, the first line of the Lord’s Prayer (Matt 6,9// Luke 11,2) is referred to as Q 11,2 rather than Q 6,984. From this I came to clearly see that when one allowed that both Luke and Matthew were generally consistent in the way they treated Mark and their other sources the evidence flowed smoothly in the direction of support for the 2DH. But when one tried to argue that Luke was relying on Matthew instead of Q the Jekyll and Hyde transformation of Luke as redactor came along as part of the package. Was there any way to explain why, when shifting over from following Mark to following Matthew, the conservative redactor Luke suddenly and inexplicably transformed himself into a veritable tradition-scattering, freewheeling, editorial party animal? I certainly could not discover any. Rather I had to admit that, as Streeter said long ago, “A theory which would make an author capable of such a proceeding would only be tenable if, on other grounds, we had reason to believe he was a crank”85.

Luke’s outline from four different chapters of his greater interpolation in the order 14, 11, 16, 12, 16. 81. TAYLOR, The Original Order of Q (n. 79), p. 117. 82. Ibid. 83. CritEd (n. 19), p. lxxxix. 84. The convention was first introduced in J.M. ROBINSON, The Sermon on the Mount/ Plain: Work Sheets for the Reconstruction of Q, in SBL 1983 Seminar Papers (1983) 451-454, p. 451. The manuscript was distributed in preparation for the Annual Meeting of the Society of Biblical Literature which met in Dallas, TX, in 1983. 85. STREETER, The Four Gospels (n. 29), p. 183.

96

R.V. HUGGINS

III. BUT WHAT IF…? In the examples discussed above my question was which synoptic solution was more convincing, the FH or the 2DH. And in each case described the evidence firmly moved me in the direction of admitting that Luke’s versions of double-tradition material were very often the more primitive ones. This undercut a foundational assumption of the FH and persuaded me to conclude that the 2DH provided a better solution to the Synoptic Problem. But the other side of that coin was the fact that Matthew was very frequently, indeed usually, secondary to Luke in the double tradition. I do not recall the exact moment when I asked myself how things might appear if I were to look at the data the other way around, that is to say, from the perspective of trying to dispense with Q by proposing that Matthew used Luke. But I do remember a passage in Streeter that firmly pushed me in that direction: “The obvious suggestion that Luke knew Matthew’s Gospel (or vice versa) and derived from it some of his materials breaks down for two reasons”86. The second reason was that “Sometimes it is Matthew, sometimes it is Luke, who gives a saying in what was clearly the more original form”87. That was an argument that was frequently repeated by defenders of the 2DH, but it did not seem particularly strong from the perspective of what I was seeing in the double tradition itself. I suspected at the time that this was “due in large part to their [defenders of the 2DH] beginning from the assumption that the lost document Q stands behind both Matthew and Luke. From this point of departure follows logically the expectation that at some points Matthew might prove to be closer to Q than Luke”88. In other words, they were finding what they had already decided must be there. But it was Streeter’s first reason that struck me. In order to refute the idea that “Luke knew Matthew’s Gospel (or vice versa)”89, Streeter proceeded to show why Luke could not have known Matthew, and then left it at that, apparently assuming that if you have successfully established that Luke did not know Matthew, you have at the same time also shown that Matthew did not know Luke. The obviousness of Streeter’s error alerted me to an important fact: If Luke used Matthew, he was basically breaking

86. Ibid. 87. Ibid. 88. HUGGINS, Matthean Posteriority (n. 1), p. 3, n. 4. 89. STREETER, The Four Gospels (n. 29), p. 183.

REMINISCENCES AND ANALYSIS

97

down Matthew’s material and scattering it about. But if Matthew was using Luke, he would have been doing the exact opposite, i.e., drawing pieces together from Luke in order to create the carefully crafted pieces that we see making up his Gospel. Once this had occurred to me it very quickly became clear that the seemingly impenetrable wall that I had been running up against in all of my efforts to make the FH work now simply vanished away when the passages discussed above were considered from the perspective of Matthew’s using Luke. The reason was that there was little difference in viewing the evidence from the perspective of the 2DH or from that of the MPH apart from a somewhat different idea of the source upon which Matthew and Luke were drawing for their double-tradition material. I still remember the strange sense of the elation of discovery I experienced as I saw the synoptic evidence transform itself before my eyes when viewed through the lens of Matthean Posteriority. It was akin to an evidential log jam breaking up or a curtain opening leaving me to wonder how it came to be that so many had tried to dispense with Q by having Luke use Matthew, but almost nobody had tried the much more obvious and less problematic road of having Matthew use Luke90. Beyond providing new insight into the FH, I now began to see things I had previously missed when looking at the synoptic evidence through the lens of the 2DH as well. From the perspective of the MPH, Luke is always or almost always more primitive than Matthew. Whether this is ultimately correct or not, it reveals an often-unacknowledged problem inherent to identifying minimal Q. Given that Matthew is much more actively selective in choosing, editing, and distributing material from the non-Markan source he shares with Luke, minimal Q as currently pursued will always have a Matthean bias. To put this another way, Matthew has the final word on the shape of minimal Q. Hence it is wrong-headed to imagine that one can in any sense discover the original extent of Q by simply tallying up places where Matthew and Luke agree in the double tradition and then perhaps 90. As I looked for others, at the time I only found two. I learned that this was the opinion of B. Bauer from A. SCHWEITZER, The Quest of the Historical Jesus: A Critical Study of its Progress from Reimarus to Wrede, London, Adam & Charles Black, 1911, p. 141. Then there was also H.P. WEST, A Primitive Version of Luke in the Composition of Matthew, in NTS 14 (1967) 75-95. West argued for the dependence of Matthew not on Luke, but on Primitive Luke (not to be confused with Proto-Luke). Primitive Luke was a document that lacked the infancy stories and served as a source for canonical Luke, Matthew, and Marcion’s Luke. Despite this difference West had clearly gone before me in thinking through some of the key issues I was then coming to terms with. I would not learn until quite a bit later that C.G. Wilke had also defended the view.

98

R.V. HUGGINS

adding, as Kloppenborg does for example, “a few pieces of special material justified by rigorous argument”, amounting to “a modest expansion of Q from 235 to 266 verses – only 31 additional verses”91. We can see this with particular clarity when we consider the alternating L/Q/L/Q pattern that runs through Luke’s greater and lesser interpolations, where the assumption that we are seeing two distinct sources represents an anachronistic illusion since in those sections Q is defined by what Matthew chose to use and L by what he did not choose to use92. No attempt to reconstruct Q therefore will be successful where this issue of Matthean bias is overlooked. A related point of clarity that emerged quite unexpectedly when I considered the Synoptic Problem from the perspective of the MPH was that the face of Luke transformed itself into something that looked quite different than it had before. Previously when considering Luke’s sources, I would tend to think in terms of three: Mark, L, and Q. Looking through the lens of the MPH, however, the distinction between L and Q vanished leaving only two sources, Mark and a second source which Luke inserts in something like five blocks into the Markan outline93. This was an entirely positive clarification since, as we said, the distinction between L and Q is largely imposed on Luke from without on the basis of Matthean redactional preferences. This clarification is as crucial for the 2DH as it is for the MPH, because it shifts attention away from first thinking of Luke’s non-Markan source from the perspective of minimal Q, which is essentially a hybrid mix of overlapping passages in Matthew and Luke, and toward engaging first the phenomena presented in the text of Luke’s non-Markan source itself. Having thereby recovered the proper starting point it becomes possible to look at the greater interpolation, for example, first as a single block of tradition, after which a closer inquiry can be made based on internal considerations as to whether or not it is possible to discern separate sources there. This does not negate the value of producing a minimal Q by comparing Matthew and Luke, but the MPH makes clear that examining Luke’s sources on their own terms prior to introducing Matthew into the mix is a necessary preliminary step that, due in no small part to the absence of consideration of the MPH, has until recently been largely overlooked. The 91. KLOPPENBORG, The Earliest Gospel (n. 28), p. 51. 92. This confusion has also led to topsy-turvy judgments such as Robinson’s: “Since the Gospel of Matthew was rooted in the Q community, it often continued Q’s language more faithfully than did the gentile Christian Gospel of Luke”. J.M. ROBINSON, The Gospel of Jesus: In Search of the Original Good News, San Francisco, CA, HarperSanFrancisco, 2005, pp. 23-24. 93. HUGGINS, Matthean Posteriority (n. 1), pp. 5-6.

REMINISCENCES AND ANALYSIS

99

MPH, then, has also made way for the reevaluation of other issues relevant to the consideration of the 2DH, but these are beyond the scope of the present article. Once I had thought through the issues of the MPH adequately to present it in the form of a preliminary proposal, I wrote it up and turned it in as my essay for Guenther’s Q seminar and on the same day submitted it to a journal94. Consistent with his usual practice, Guenther returned it with several pages of carefully handwritten notes attached (see footnote 23) featuring the idea that an essay that attempted to dispense with Q was “beyond the pale” for a Q seminar. As I reached the end of his notes, I was relieved to find that he did not ask me to submit a second, more Q-friendly essay, nor did he penalize the one I’d turned in with a bad grade. Sometime later when Guenther was getting ready to retire, we ran into each other one day in the Emmanuel College Library. On that occasion he took me aside and advised me in a fatherly way that life was too short to pursue one’s own solution to the Synoptic Problem. It was better to embrace one of the existing ones and expend my labor there. Was Guenther right? One might well argue that life is too short to pursue any solution to the Synoptic Problem. Vita brevis ars longa! So it has always seemed to me that if one is going to tackle the Synoptic Problem at all, one might as well pursue the solution one thinks most plausible, whether it be one’s own or another’s. And besides, I have always firmly believed that the MPH, in addition to providing a solution to the Synoptic Problem that is more plausible than the FH, also has the potential of polishing the lens, as it were, through which defenders of the 2DH look at the synoptic evidence. It is a win-win all around. Independent scholar United States of America

Ronald V. HUGGINS [email protected]

94. First to New Testament Studies, who passed on it on the grounds that it contributed nothing new, and then to Novum Testamentum who published it. It took me some time to finish after the seminar itself had ended.

WHAT IF? HISTORICAL AND LITERARY HYPOTHESES ON MATTHEAN POSTERIORITY

I. INTRODUCTION Like most colleagues of my generation, I matured in an academic context where the Two-Source Hypothesis was usually accepted, and the existence of Q as a written source that was independently accessed by Matthew and Luke was the most logical explanation for their parallel passages not deriving from Mark. But back in the 1980s, while working on the synoptic material on John the Baptist, I became convinced that it was not possible to imagine a reciprocal independence of Matthew and Luke, in general and in particular regarding the Baptist’s traditions1. This is not to deny the existence of different pre-synoptic sources besides Mark, among which we could still hypothesize one and call it Q. The point here is to stress the impossibility of reconstructing Q or other pre-synoptic sources on the basis of certain agreements between Matthew and Luke, since the two should in some ways have known each other. While I gladly recognize my indebtedness to the work of Pierson Parker, which inspired me to “see” Matthew as dependent upon Luke2, my contribution here will be simply to discuss some of the literary and historical reasons that have persuaded me of such dependence, even if the general picture of the authorial formation of the Gospels is probably more complex and difficult to explain than traditionally admitted. A first caveat comes from the textual situation itself. Our critical editions are a modern reconstruction of how the text might have appeared around the time of Constantine, and we can be sure that no manuscript ever existed, 1. In those years I expressed my ideas with two books in Italian: E. LUPIERI, Giovanni Battista nelle tradizioni sinottiche (Studi biblici, 82), Brescia, Paideia, 1988; and the first part of ID., Giovanni Battista fra storia e leggenda (Biblioteca di cultura religiosa, 53), Brescia, Paideia, 1988, pp. 1-192. 2. See esp. P. PARKER, A Second Look at The Gospel before Mark, in JBL 100 (1981) 389-413.

102

E. LUPIERI

the text of which corresponded precisely to any of them. This does not impede, however, our attempts to make sense of what we have. Simply put, though, complete certitude regarding some textual details is probably beyond our reach. Further, two more aspects must be taken into consideration. On the one hand, it is becoming more and more clear that the formation and redaction history of the Gospels is complex, with the concrete possibility that a particular text may have undergone multiple editions3. And on the other hand, we might be obliged to take into consideration the existence of scribal interventions, usually harmonizing similar passages from one Gospel to those of another4. Having said this, we can take as the textual basis for our discussions any acceptable critical edition of the New Testament5, the text of which we can tentatively consider the final redaction of each Gospel, the closest possible entity to an “authorial creation”6. As the next step, we can and should ask ourselves what can be said regarding the pre-synoptic traditions, the existence of which seems impossible to deny7. What is wrong with Q, then? 3. Not only the formation of our texts must have been a complex process, but also their final redaction may have been modified over time, as – at a macroscopic textual level – the multiple “endings” of Mark and the double conclusion of John seem to show. Also, after P. WALTERS, The Assumed Authorial Unity of Luke and Acts: A Reassessment of the Evidence (SNTS MS, 145), Cambridge, Cambridge University Press, 2009, it is more and more difficult to accept the traditional idea that one individual person wrote both books, even if Acts does present itself as the second volume written by the same author of Luke. While doubting authorial unity, I will consider Luke and Acts to be representative of the same group of Christ followers, strongly influenced by Pauline teaching and supporting the same line of thinking, in spite of a series of internal corrections of narrative details and some differences in focus. I would consider an internal correction, as an example, to be the specifications for the time and place of the Ascension in Acts 1, while a difference in focus would be the fact that nowhere in Acts does the infancy of Jesus have any weight in the history of salvation (but the centrality of the Temple and the function of John the Baptist are maintained). Finally, if we suppose that a person or a group of people continued to exercise authorial functions on Luke, we cannot exclude the possibility of some late interventions in its text which may or may not reflect an awareness of Matthew. 4. At least for Mark, the model is often Matthew, as shown, for example, by Mark 11,23[26]. The status of the manuscripts shows that changes took place until deep into the 2nd and 3rd centuries – and were very successful. For not-so-naïve scribal interventions, see C.G. PARDEE, Scribal Harmonization in the Synoptic Gospels (NTTSD, 60), Leiden – Boston, MA, Brill, 2019. 5. I am used to the Nestle-Aland 26th edition (Novum Testamentum Graece, etc., Stuttgart 1979ff.) and will continue using it. 6. For our goals, it is arbitrary whether we think of an individual author or of a scribal enterprise made up of a group of learned followers of Jesus; what matters is the ideal unity and internal coherence (as accepted in the 1st century CE) of each text. 7. If not for other quite obvious reasons, then because Luke 1,1-4 explicitly mentions their existence. I prefer to talk about pre-synoptic (and occasionally of para-synoptic) traditions, more than sources, developed before and during the formation of the final redaction of our

WHAT IF?

103

I have two orders of problems: a) the structure, nature, and extension of Q and b) its reconstruction. II. CARL LACHMANN In 1835, Carl Lachmann wrote a beautiful article, in his elegant Latin8, considered by some to be the earliest methodological foundation for the existence of Q, which was then hypothesized in the following years. This assessment of Lachmann’s work is not exactly true, however9. Against Griesbach, Lachmann defends Markan priority, and, regarding the sources, he develops a very articulate case. First, he distinguishes between auctores or narratores of the pre-synoptic traditions (which therefore he thought were at least partially oral) and scriptores (our evangelists, as “redactors” of the written texts we have). Then he hypothesizes the existence of six traditional narrative sections (partes) which cover most of the Gospels’ narrative (apart from the Passion narrative, with which he does not deal directly) and which he calls corpora, corpuscula, libri, libelli (and, therefore, he thinks of a plurality of written sources). For the Infancy narratives, he thinks of different oral traditions, known only to few10, while for the Temptation narrative he doubts the existence of only one specific corpusculum, given the diversities among the three Gospels11. He always allows the possibility that the

written texts. I will not usually discuss this very slippery aspect of the general question, but we must accept that Gospel narratives might have been recited (and not only read aloud) among the various groups of Christ followers. The impact of these performances on the written traditions has yet to be sufficiently analyzed and understood. See recent discussions in N.A. ELDER, The Media Matrix of Early Jewish and Christian Narrative (LNTS, 612), London – New York, Bloomsbury T&T Clark, 2019, and esp. his forthcoming book Gospel Media: Reading, Writing, and Circulating Jesus Traditions, Grand Rapids, MI, Eerdmans, 2023. 8. C. LACHMANN, De ordine narrationum in evangeliis synopticis, in Theologische Studien und Kritiken: Beiträge zur Theologie und Religionswissenschaft 8 (1835) 570-590. I strongly suggest going back and reading Lachmann: though we may disagree, his argumentative intelligence is still interestingly valid after almost two centuries. 9. Actually, one reason that I did not look beyond the Q hypothesis sooner was the tradition that Lachmann had supported it: if the father of modern textual criticism had accepted it, who was I to have doubts? The virulent reaction against that idea by H.H. STOLDT, History and Criticism of the Marcan Hypothesis, trans. D.L. Niewyk, Macon, GA, Mercer University Press, 1980 (first German ed. 1977), opened my eyes and brought me to read the original article by Lachmann. 10. LACHMANN, De ordine (n. 8), p. 584: “Nam quae Matthaeus et Lucas de Iesu pueritia referunt, ea satis adparet paucis narratoribus propria fuisse nec vulgo cognita”. 11. Ibid.: “Tentatio diabolis utrum primo historiarum corpusculo semper et iam a primis narrandi initiis infuerit, dubitari potest; nam et Lucas hic alium quam Matthaeus auctorem habuit, et Marcus sui prorsus dissimilis est”.

104

E. LUPIERI

evangelists made use of diversi auctores or their own iudicium when they add or omit passages, even if there is a vulgaris traditio that appears here and there, particularly in the sixth section. This seems to me the closest thing to Q, but it is so nebulous as to usually be spoken of in the negative and, in any case, far from being a written document like the corpora12, it sounds like a “common opinion” among the believers. Maybe others will feel relieved, as I did, in knowing that Lachmann had not hypothesized the existence of Q. But what is wrong, then, with Q? Given Markan priority, why shouldn’t Matthew and Luke have independently used a second written source to compose their Gospels? III. THE TEARS OF PETER

AND

OTHER DETAILS ON CRUMBS AND LEAVEN

If we accept that Matthew and Luke had only Mark in front of them in numerous passages, especially for the Passion narrative, I remain in need of someone who can convince me that two different persons – or two groups of persons – intending to re-write a detail like the Markan version of Peter’s desperate crying, καὶ ἐπιβαλὼν ἔκλαιεν (Mark 14,72), both arrived independently at καὶ ἐξελθὼν ἔξω ἔκλαυσεν πικρῶς (Matt 26,75//Luke 22,62). Please notice the homoioarkton, the instances of assonance (esp. gutturals and liquids), and the rhythm of the sentence, all of which seem to reproduce Peter’s desperate sobbing. These were people who knew their Greek and were able to construct a little literary gem. Either they found it in an earlier source13, or one knew (of) the other and reproduced it. While I cannot exclude Matthew (who also likes assonance) as the creative mind behind the text, this still looks and sounds like Luke to me. In this case, the resulting texts are identical in Matthew and in Luke, but there are also cases which seem to convincingly show Matthew conflating Mark and Luke.

12. Ibid., p. 588: “Scriptores hic, ut solent, vel diversos auctores secuti vel suo quisque usus iudicio, nunc aliquid addunt, nunc omittunt. … Non facile aliquis dicat quo loco vulgaris traditio evangelica rettulerit … Neque negaverim Mattheaum non nulla in amplificatas suas et auctas orationes relata ex proprio loco summovisse. Sed parabola illa [Matt 25,14-30 = Luke 19,11-27] … adeo discrepantibus exemplis scripta est ut eam non ex illa vulgari traditione sed ab utroque evangelista a diversis auctoribus repetitam esse existimem”. Please notice that Lachmann also thinks that Matthew is the one who modifies the preexisting tradition the most. 13. But what source, since Q should not contain the Passion narrative? Was it written or recited and by whom?

WHAT IF?

105

There are two particular examples that I like, among many possible ones. The first one deals with (bread)crumbs falling from the table. In Mark 7,28 we read: Κύριε, καὶ τὰ κυνάρια ὑποκάτω τῆς τραπέζης ἐσθίουσιν ἀπὸ τῶν ψιχίων τῶν παιδίων. It is part of the dialogue in the scene with the “Greek Syrophoenician” woman, avoided by Luke, probably, because Jesus’s harsh rebuke went squarely against Luke’s openings and sympathies for the GrecoRoman world. Luke, though, appears to reuse some of it in the (exclusively Lukan) parable of Lazarus and the rich man, where we read: καὶ ἐπιθυμῶν χορτασθῆναι ἀπὸ τῶν πιπτόντων ἀπὸ τῆς τραπέζης τοῦ πλουσίου· ἀλλὰ καὶ οἱ κύνες ἐρχόμενοι ἐπέλειχον τὰ ἕλκη αὐτοῦ (Luke 16,21). In the passage parallel to Mark, Matthew, for reasons opposite to those of Luke, while appreciating the Markan narrative, must not have liked that such a miracle was performed on a “Greek woman” and transforms her into a “Canaanite” one (Matt 15,22 versus Mark 7,26). According to him, her words are: Ναί, κύριε, καὶ γὰρ τὰ κυνάρια ἐσθίει ἀπὸ τῶν ψιχίων τῶν πιπτόντων ἀπὸ τῆς τραπέζης τῶν κυρίων αὐτῶν (Matt 15,27). I recognize that the detail of the crumbs “falling from the tables” could be an independent serendipitous addition by Matthew to the Markan context and developed independently by Luke in a different, only Lukan, context, but the coincidence of the scenes (including the mention of dogs) seems strong to me and sufficient to justify a literary contact between the two narratives. A second example that I find quite convincing has to do with the symbolic use of “leaven”. In Mark 8,15 we read: Ὁρᾶτε, βλέπετε ἀπὸ τῆς ζύμης τῶν Φαρισαίων καὶ τῆς ζύμης Ἡρῴδου14. Luke 12,1 eliminates Herod and explains the symbolic meaning of the phrase, besides ameliorating its Greek: Προσέχετε ἑαυτοῖς ἀπὸ τῆς ζύμης, ἥτις ἐστὶν ὑπόκρισις, τῶν Φαρισαίων. Moreover, Luke eliminates the whole Markan context with the discussion about the bread between Jesus and his disciples who do not understand what he is talking about. On his part, Matthew also believes that the Markan context should be ameliorated: he keeps the substance of the discussion between Jesus and the disciples, but has Jesus explaining twice who the adversaries are. Ὁρᾶτε καὶ προσέχετε ἀπὸ τῆς ζύμης τῶν Φαρισαίων καὶ Σαδδουκαίων (Matt 16,6) and προσέχετε δὲ ἀπὸ τῆς ζύμης τῶν Φαρισαίων καὶ Σαδδουκαίων (Matt 16,11). At this point the disciples finally understand: Tότε συνῆκαν ὅτι οὐκ εἶπεν προσέχειν ἀπὸ τῆς 14. Some manuscripts, and even 𝔓45, have “the Herodians” instead of “Herod”, but no Markan manuscript lacks the second kind of leaven. It is not impossible to hypothesize that this second kind of leaven was added during the formation of the Markan text, but I see no reason for it, given the declining political power of the Herods and the Herodians.

106

E. LUPIERI

ζύμης τῶν ἄρτων ἀλλὰ ἀπὸ τῆς διδαχῆς τῶν Φαρισαίων καὶ Σαδδουκαίων (Matt 16,12). The verb προσέχω (never used by Mark) is equally split in its use between Matthew and Luke, while the connection between Pharisees and Sadducees is typically and exclusively Matthean (Matt 3,7; 16,1.6.11.12bis), but the impression one gets is that Matthew begins with the Markan ὁρᾶτε and then continues with Luke’s προσέχετε, while agreeing with Luke in avoiding any reference to Herod in this context. It is well known that the alternative hypothesis, that all these not-soMinor Agreements derive from Q, is usually rejected because the extent and nature of Q would be entirely different from those supposed and accepted for its reconstruction (since it should include large narrative sections and the Passion). IV. SOME TEMPTATIONS All the passages we have discussed, and the ones we will be discussing, have already been the object of long scholarly analyses which have led to antithetical conclusions regarding the history of the formation of the synoptic gospels. Another “classic” object of many discussions (including Lachmann’s, as we saw above) is the synoptic narrative of the temptations, situated by all three Synoptics after John’s baptism and before the so-called public life of Jesus. A complete analysis of the three passages would be impossible here. I will only discuss some aspects, to explain a) one general aspect of the findings we are faced with in this kind of research and b) why I think what I now believe is possible15. The passages under discussion are Mark 1,12-13 (+14-15), Luke 4,1-13 (+14-15), and Matt 4,1-11 (+12-17). The sections in parentheses represent the continuations of the narratives which contain elements useful to be compared. As a general rule, we should accept that discrepancies in the texts do not mean independence and reciprocal ignorance or use of different sources. Diversities usually can and should be explained as corrections and changes; criticism too does prove reciprocal knowledge and dependence. We must admit authorial intentions and not just naïve redactional activities. Multiplying sources to transform ancient “authors” into redactors can be a cryptoapologetic stand: It is better to have Evangelists who are not so brilliant but

15. Which is different from what I believed in 1988. See my Giovanni Battista nelle tradizioni sinottiche (n. 1), pp. 21-23, n. 2.

WHAT IF?

107

are always faithful to their sources, than to admit – God forbid! – that they invented something or willingly modified their traditions to prove their theological, Christological, or ecclesiological ideas16. Not surprisingly, what emerges immediately from a comparison of the three synoptic versions of the Temptation narrative is the absolute coherence, verbal and on the level of ideas, of each text with the lexicon and ideas of the Gospel to which it belongs. Beginning with καὶ εὐθύς (Mark 1,12), the whole Markan passage is in full harmony with the wording and the style of the rest of the Gospel. Here I would like to draw attention to the Markan use of the verb ἐκβάλλω, equally present in other passages of all three Synoptics, usually to describe the driving away of a devil through an exorcism17. Here it is used to stress the full power of the Spirit upon Jesus, who is “thrown” by it “into the desert”. After that, no verb of movement appears in the Markan Temptation. Both Luke and Matthew avoid the verb here. As a result, while they both stress the presence of the Spirit, Jesus seems to be more compos sui or, at least, to be “led” or “guided”, but not thrown. He ὑπέστρεψεν … καὶ ἤγετο in Luke 4,1 and ἀνήχθη in Matt 4,1. After that, the Devil takes over for a while. In Luke, the Devil appears to be already present for the first temptation (no verb of movement for him), then he “lifts [Jesus] up” for the second temptation (ἀναγαγών in Luke 4,5), and “leads” and “makes [him] stand” for the third one (ἤγαγεν … καὶ ἔστησεν in Luke 4,9). In Matthew, the Devil “approaches” Jesus for the first temptation (προσελθών in Matt 4,3) and then “takes” him into the two following temptations, with a verb much loved by Matthew, especially in the Infancy narrative (παραλαμβάνει … καὶ ἔστησεν in Matt 4,5 and παραλαμβάνει in Matt 4,8). By comparing the use of verbs, we can also hypothesize that ἀνήχθη of Matt 4,1 was inspired by ἀναγαγών of Luke 4,5, but we should all agree that, whatever the explanation, παραλαμβάνει … καὶ ἔστησεν of Matt 4,5 shows a Matthean redactional intervention on ἤγαγεν … καὶ ἔστησεν of Luke 4,9, which represents the traditional text. The conclusion of the scene again shows Matthew conflating Mark and Luke. Mark 1,13 ends with: καὶ οἱ ἄγγελοι διηκόνουν αὐτῷ. Luke 4,13 drops the angels and tells only that ὁ διάβολος ἀπέστη ἀπ’ αὐτοῦ ἄχρι καιροῦ18. Matt 4,11 first follows Luke and then goes back to Mark, bringing 16. We should always remember how, back in 1983, even a conservative scholar like Robert H. Gundry was obliged to resign from the Evangelical Theological Society because he dared say that Matthew, in his Infancy narrative, had created probably non-historical narratives, like the Massacre of the Infants in Matt 2,16-18. 17. But scarcely used by John and never in Paul, if not in a quotation in Gal 4,30. 18. This is probably an internal reference to the active presence of Satan in the Passion narrative: Luke 22,3 and 22,31.

108

E. LUPIERI

the “angels” back on the scene: τότε ἀφίησιν αὐτὸν ὁ διάβολος, καὶ ἰδοὺ ἄγγελοι προσῆλθον19 καὶ διηκόνουν αὐτῷ. The very order of the three temptations reflects the ideological and literary preferences of the authors. There is a crescendo that culminates in Matthew on an “extremely high mountain” (Matt 4,8) and in Luke in the Temple of Jerusalem (Luke 4,9). This should not come as a surprise, if we think of the Matthean apocalyptical love of the “mountain” as a locus of revelations, as shown in Matt 5,1 or 28,16 (a love not shared by Luke) and, on the other hand, the centrality of the Temple as a locus of revelation in the Lukan tradition as a whole, from the Infancy narrative to Acts of the Apostles (at least until Acts 22,17-21). But there is more to be said here. The detail of the mountain is probably also connected to the fact that, while both Mark 1,13 and Luke 4,2 say that Jesus remained “in the desert for forty days”, and Luke adds that “he did not eat anything in those days” (ibid.), Matt 4,2 further expands the narrative by explaining that Jesus, being in the desert, “fasted for forty days and forty nights”. In this way Matthew creates another explicit analogy with Moses, who fasted on Mount Sinai for the exact same amount of time (Exod 34,28 and Deut 9,9)20. On his part, Luke shows no knowledge of any mountain, but likes to bring Jesus into the Temple. From the Lukan Infancy Narrative we learn that Jesus used to go on a pilgrimage to Jerusalem and enter the Temple annually, until at the age of twelve he remained there alone for three days (Luke 2,41-50). From a Lukan perspective, nothing impedes the thinking that Jesus continued to go to Jerusalem and to the Temple, as is also implied by John. Here, at 4,13, he is left alone again, this time by the Devil, who goes away. Nothing of this sort would have been accepted by Matthew. For him, Jesus is persecuted since his birth by the Jewish religious and political authorities who impede God’s project of salvation for his people, rejecting the Messiah sent to them. Jesus is born in Judaea, but must take refuge first in a pagan land, Egypt, the fatherland of theriolatry, the most heinous form of idolatry, and then in the semi-pagan Galilee (Matt 4,15); he finally enters Judaea for his baptism by John and, for the first time in his life, goes to Jerusalem. I find not a little irony (against the priestly leadership in Jerusalem) that in Matthew 19. The verb is of common use, but it is a redactional addition by Matthew, as earlier in Matt 4,3. 20. The analogy of Jesus with Moses is notoriously stressed many times by Matthew, beginning with the persecution of the child (which includes the Massacre of the Infants) and his salvation in Egypt in the Infancy narrative. The Great Commission, though, with the “commandments” given by Jesus and to be “kept” by “all the peoples/Gentiles” (Matt 28,19-20), will show beyond doubt the substantial superiority of Jesus.

WHAT IF?

109

it is the Devil who brings Jesus in and out of the Temple, as if the Devil already somehow is in charge of the “holy city” and its Temple21. What happens later is also Christologically meaningful. According to Mark 1,14, Jesus, “after the handing over of John, went … to Galilee”. The menace against a champion of God obliges Jesus to immediately abandon Judea and take refuge in Galilee22. According to the Synoptics, he would subsequently enter Judaea and Jerusalem only for the final part of his earthly mission. And there, in the various Passion narratives, we find a large discussion regarding who “hands over” Jesus to whom, while our traditions contain no trace of John the Baptist being handed over to Antipas. It must therefore be a theological handing over that makes John the real forerunner of Jesus, particularly in his sacrifice and death23. This Markan idea is accepted by Matt 4,12, but rejected by Luke 4,14, who stresses that Jesus, once the Devil is gone, “goes back to Galilee in the power of the Spirit”, on his own. It is not the Devil, then, but the Spirit that accompanies Jesus out of the temple and to Galilee, without any connection with the incarceration of the Baptist24. 21. Matt 4,5. As in Rev 11,2, the “City” can be called “holy” precisely when it is polluted. 22. The Matthean Infancy narrative, then, seems to be the result of a projection of this Markan historical-theological perspective onto the early phases of Jesus’s life. 23. Since the verb is passive, I suppose it is a passivum divinum, meaning that it is God who hands over his champion to the forces of evil, for the extreme sacrifice. This is, therefore, a sort of prophetic anticipation of Jesus’s death, both executions being constitutive steps in the history of salvation, as planned by God, and not victories for the enemies or the Enemy. 24. Jesus’s movements are therefore independent from the events in John’s life. According to Luke and the Fourth Gospel, John the Baptist’s and Jesus’s preaching and possibly baptizing do overlap for some time (compare Luke 7,18-23 and John 3,22-30). Not so in Mark nor Matthew (besides the passages commented here, compare Matt 11,2-6, where John sends his disciples while in Herod’s prison, creating some problems of historicity). What we have here is the result of what I called the “Theology of the Succession” between the Baptist and Jesus, meaning that John precedes Jesus in everything, and that Jesus can act only when John is imprisoned and executed. In this way John is a real forerunner of Jesus, and even his death is a prophecy of Jesus’s death. This is Mark’s perspective, which is accepted by Matthew. According to Luke, however, no one can really precede Jesus in his salvific death. John’s execution is not only not described, but his incarceration is presented with strong lexical analogies with that of Peter in Acts, transforming John in some sort of disciple ante litteram (compare Luke 3,19-20 with Acts 12,3: both imprisonments are an “addition” by a Herod to a series of crimes). Coherently, Luke seems to disconnect John from Elijah (no description of John’s vest like in Mark 1,6), while connecting Jesus to the biblical narratives on Elijah and Elisha (resurrection of the son of the widow in Nain, multiplication of bread and fish, even the Ascension…). For sure, Luke is convinced of Jesus’s distinction from and superiority to Elijah (compare Luke 9,30-33 [Mark 9,4-6] and esp. Luke 9,54-55 against 2 Kgs 1,10.12) and possibly for this reason the Gospel does not mention Elijah during the crucifixion (against Mark 15,34ff. and Matt 27,46f.49). On all this see already H. PETER, Johannes der Täufer in der urchristlichen Überlieferung, Marburg a.L., Bauer, 1911, pp. 12f.; J.A.T. ROBINSON, Elijah, John and Jesus, in NTS 4 (1958) 263-281;

110

E. LUPIERI

At this point, Luke brings into his narrative the Rejection at Nazareth, called Ναζαρά (4,16-30), adapting it and moving it here from Mark 6,1-6. Matthew is much more faithful to the content and structure of Mark, leaving the scene in a later part of his Gospel (Matt 13,54-58). Nevertheless, after Matthew says that Jesus ἀνεχώρησεν εἰς τὴν Γαλιλαίαν25, he still feels the need to explain here that Jesus “left Ναζαρά26 and went to live in Capernaum” (Matt 4,13). Also in this passage, then, Matthew seems to add some details from the Lukan narrative into the framework of the Markan structure. V. SOME REFLECTIONS Generally speaking, all these discussions come to the somehow obvious yet discouraging conclusion that each version of the Temptation narrative corresponds perfectly well to the theological intention of each Evangelist. If the Matthean version is unacceptable for Luke and the Lukan unacceptable for Matthew, and even if we accept the idea that one reacts to the other, would we still ever be able to convincingly decide who reacts to whom? The point I wish to make here is that it is not absurd to think that Matthew reacts to Luke, even if or when the Matthean text seems less theologically developed or more “primitive” than Luke’s. History does not develop with chronological linearity from the simple to the complex, nor necessarily moves from a more archaic ideological position to a more “modern” one. It is not necessary to think that Luke, having found in Matthew or in any other source for the Temptation narrative the sequence desert – Temple – mountain, and out of his theological love for the Temple and disdain for mountains, eliminates this geographical detail and puts the Temple in the closing position, so that the action would culminate with it. We can also hypothesize that Matthew, having found the sequence desert – W. WINK, John the Baptist in the Gospel Tradition (SNTS MS, 7), Cambridge, Cambridge University Press, 1968, pp. 43ff.; M.M. FEIERSTEIN, Why Do the Scribes Say that Elijah Must Come First?, in JBL 100 (1981) 75-86. 25. Matt 4,12, with a verb he loves, especially in the Infancy narrative. Matthew uses it ten times in his Gospel, Mark only once (like John), while it does not appear in Luke nor in Paul. 26. Here in Matthew, this is probably the most ancient version of the name, exactly as in Luke 4,16 (while in Matt 2,23 the most ancient spelling was probably Ναζαρέτ/Ναζαρέθ [although 𝔓70’s Ναζαρά should be explained], as always in the Lukan Infancy narrative [Luke 1,26; 2,4.39.51]). It may not be a coincidence that in the only passage where Luke changes the spelling Matthew also does the same, apparently.

WHAT IF?

111

somewhere high – Temple, creates a better sequence and builds a different crescendo, from the desert to Jerusalem’s Temple in Zion, to culminate with an apocalyptical mountain that connects Jesus to the Sinai and Moses27. With this, he finds a way to narrate that Satan seems to already have some sort of control over the sacred space of the “holy City”, creating one of the many narrative connections with the Passion and the ending of the Gospel. It is my hypothesis that Matthew reacts to the Gospel of Luke, rewriting Mark and adding to it whatever he finds acceptable in Luke, but avoiding or modifying whatever he deems unacceptable. This process begins with the Infancy narrative, which, similar to the one in Luke, is an ideological pamphlet illustrating the Matthean version of salvation history28. We are used to bringing together the magi and the shepherds, the manger and the star, the imperial census and the Herodian massacre, but these are actually incompatible. A small but important set of elements that are not found in Mark appear in both Luke and Matthew: Jesus is born in Bethlehem, during the reign of King Herod, to Mary, a virgin betrothed to Joseph, of the house of David, and an “angel of the Lord” is in charge of at least some of the communications between the divine and human realms. Almost everything else, though, is radically different in the two narratives. In Matthew, Joseph is not the obedient subject of the Roman Empire who travels from Nazareth of Galilee, where he lives, to Bethlehem of Judea with his pregnant spouse, to obey some historically fragile dispositions of a Roman census29, with the result that Jesus is born in Bethlehem thanks to an imperial order. Rather, Joseph simply lives in Bethlehem because he has his house there, and there also lives Mary, for whom there is no need to be of a priestly family, given the fact that there is no contact whatsoever, much less a maternal lineage, with John the Baptist until the baptism of Jesus30. And the angel does not spend his time talking to a priest in the Temple or 27. In this way, Matthew eliminates the geographical incertitude of the second Lukan temptation. 28. For this reading of the Infancy narrative, see E. LUPIERI, Due diversi racconti sulla nascita di Gesù, in La famiglia di Gesù. “Ecco di fuori tua madre e i tuoi fratelli” (Atti del Seminario invernale di BIBLIA), Settimello, Biblia, 2005, 73-91, and ID., La figura di Gesù di fronte al potere politico, a partire dai testi evangelici, in E. MANICARDI – L. MAZZINGHI (eds.), Il potere politico: bisogno e rifiuto dell’autorità. XXXVIII Settimana Biblica Nazionale, Roma 6-10 settembre 2004 (RSB, 18), Bologna, EDB, 2006, 165-182. 29. Neither Galilee nor Judaea were under Roman jurisdiction when Herod was king. The Romans never asked their subjects to move from their place of residence to their place of origin for a “census”. 30. It seems to me that given the constant opposition in Matthew of the Sadducees to both John and Jesus, this is not the Gospel where we can imagine finding a blood connection with the lineage of Aaron for any of them, even if only on the maternal side for Jesus.

112

E. LUPIERI

to a woman in Galilee: he goes straight to Joseph in Judea and tells him to name the child Jesus. Joseph, the man, is the hero of the narrative. And the good news is not brought to a Jewish priest or to a group of characters representing a category of Jewish impurity, the shepherds, but rather to a group of pagan priests who, through their own pagan knowledge of the stars, determine when the “newborn king of the Jews” (Matt 2,2) is born. Even more, the Jewish king and the Jewish religious authorities plan to destroy the newborn “Christ” (Matt 2,4), so that Joseph, far from taking his wife and child to the Temple of Jerusalem where the infant would have been killed like all the other children of Bethlehem and its surroundings, is instead obliged to take refuge first in the pagan land of Egypt and then in the semi-pagan Galilee. No Jewish prophet recognizes Jesus in the Temple, but the old prophecies are consulted to find a way to kill him. Under no circumstances would the Matthean Jesus go every year on a pilgrimage to Jerusalem. Rather, as we have seen, it is the Devil who takes him in and out of the Temple, apparently for the first time in Jesus’s life. This goes on throughout the entire narrative. If, in Luke, Jesus is the scion of David through a cadet line, in Matthew he is the direct descendant of all the real kings31. And at the end of the Gospel, if in Luke (and in Acts) the followers of Jesus are ordered to stay in Jerusalem until the descent of the Spirit, while the resurrected Lord leaves them by ascending to heaven (from a nearby locality), and they spend most of their time praying in the Temple, so that the beginning of the new religious reality originates from the geographical heart of Judaism, in Matthew (following Mark) the followers are ordered to immediately go to Galilee. Actually, they go to (another?) mountain where the Eleven32 meet with the resurrected Lord. There, they “worship” him33 and receive the so-called Great Commission, with which the Lord not only denies his Ascension and the descent of the Spirit, since he will remain with them without going away until the end of the eon, but

31. The two genealogies probably reflect a different numerology, the one in Luke being based on seven times twelve and the one in Matthew on seven times seven. On this and other aspects of the genealogies, see E. LUPIERI, Dodici, sette, undici, ventiquattro: numeri, chiese e fine del mondo, in Annali di Storia dell’Esegesi 22 (2005) 357-371. 32. Their number is stressed by Matthew, and there is no need to reconstitute the “Jewish” number of Twelve, contrary to what is told in Acts 1,15-26. 33. For the interpretive problem of Matt 28,17, regarding those who have doubts in that crucial moment and the possibility of a Matthean criticism of the group of the “brothers” of Jesus, see E. LUPIERI, La comunità di Matteo e il gruppo dei ‘fratelli’ di Gesù, in S. GUIJARRO (ed.), Los comienzos del Cristianismo. IV Simposio Internacional del Grupo Europeo de Investigación Interdisciplinar sobre los Orígenes del Cristianismo (G.E.R.I.C.O.) (Bibliotheca Salmanticensis, 284), Salamanca, Universidad Pontificia de Salamanca, 2006, 171-180.

WHAT IF?

113

with one sentence orders a mission for the benefit of “all the peoples” and introduces the Christian baptism as a conversion ritual, comprising a mature trinitarian formula, bypassing completely the whole narrative of the ending of the Gospel of Luke and of many chapters of Acts. No Roman centurions, no human discussions in Jerusalem, no celestial visions in Joppa are needed: a simple word of the resurrected Lord sets the Matthean church on march, beginning not from Jerusalem or the Temple, of which there will not remain a stone upon a stone, but from the mountain of Galilee. If we believe that the chronology of our texts follows logical and theological or ecclesiological evolutions, it seems to me that the conclusion of Matthew is more “modern” and “developed” than the ending of Luke and the narrative in Acts. What I mean, particularly if we accept the historical presupposition that theological ideas develop chronologically following a logical order, is that here Matthew not only “answers” Luke with a different narrative, but it is (ideo)logically more developed, and, therefore, should be understood as posterior. VI. MORE HYPOTHESES Can we imagine a historically realistic scenario behind all this? The way I see it, somewhere in the Roman province of Syria (the mention of Syria in Matt 4,24 has no parallel in any Gospel) there was a group of followers of Jesus who used the Gospel of Mark as their foundational text, together with other Scriptures. It was probably some large city, and the group was able, or was trying, to attract new proselytes or converts from the local pagan population, possibly in antagonism with Pharisaic proselytism (Matthew 23, esp. 23,15, which also has no parallel in the other Gospels). Even if this group was antagonistic with the Jews who had not accepted the good news of Jesus, it was nevertheless strongly connected with their way of thinking and living. I also believe that with the fall of Jerusalem there had been new diasporic waves of Palestinian Jews towards other regions of the empire, with different levels of cultural Hellenization and sympathy or antipathy for the Roman hegemony. Our group would also have considered itself different from the groups of followers of Jesus who had accepted the good news brought by the Pauline missionaries. Then it happened that in a group of believers tied to the Pauline mission, maybe living just a couple of insulae down the road, a person or a group whom we traditionally call Luke, collecting various documents and traditions, apparently written and oral, felt the need to re-write Mark, to give

114

E. LUPIERI

a more solid foundation to their faith and to the faith of the newly converted who were mostly pagans, being more Hellenized and tied to the Greco-Roman way of life. The Gospel of Luke and Acts of the Apostles were gradually constructed and put into circulation. When these new texts and/or their contents became known to the first group, someone whom we traditionally call Matthew (I still believe in an individual person, possibly with the cooperation of a scribal group) thought that, while some of the novelties were acceptable, the general tone and many details were not. Therefore, a new and more faithful rewriting of Mark took place, a rewriting which would have incorporated whatever was deemed acceptable from the Lukan work and corrected the rest. Even if they cannot be excluded, I do not think it necessary to hypothesize new or different written sources in order to explain the difference of the Matthean work from the Lukan one. Inspired reflections on the scriptural lore and authorial interventions on particular traditions, together with literary creativity, can explain most discrepancies between Matthew and Luke. Matthew must have been successful among the churches34 with ties to, yet antagonistic with the Pharisaic traditions, churches which at the same time aimed at distinguishing themselves from the Pauline missions. I would frame its success with a wave of new proselytism which in the past we would have defined as Jewish-Christian. It possibly accompanied the diffusion of the Book of Revelation, with which Matthew shares many apocalyptic ideas: from the persecution of the newborn messianic figure to the devilish dimension of religious and political powers, both Jewish and Roman35. If all this makes sense, the last quarter of the 1st century CE would be the ideal timeframe for Matthew’s composition36. Loyola University Chicago United States of America

Edmondo LUPIERI [email protected]

34. We should be authorized to use the word corresponding to ἐκκλησίαι, since Matthew notoriously uses it at 16,18, which is fully redactional. 35. See on this L. PAINCHAUD – E. LUPIERI (eds.), Who Is Sitting on Which Beast? (Judaïsme Ancien et Origines du Christianisme, 29), Turnhout, Brepols, forthcoming. 36. I am fully aware of the hypothetical aspects of this reconstruction. I present it here to explain why I think that Matthean Posteriority is plausible. Nevertheless, I will welcome all criticisms for an open discussion. As I have changed my opinion in the past, I believe there is always space for reconsidering one’s scholarly position.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS: COHERENCE AND DISTINCTNESS?

Q researchers plough a hypothetical field, and not everyone in the scientific community is persuaded by their answer to the synoptic question. In my view, there is a path forward in systematically testing the coherence and distinctness of the double tradition materials over against Mark and the Sondergut materials of Matthew and Luke. For analytical reasons, I use the term “double tradition material” (TD, traditio duplex). This textual corpus which Matthew and Luke have in common over against Mark can be analysed as any other textual corpus. Conclusions can be drawn which might support Q but do not presuppose it. The many cases where Matt and Luke coincide offer solid ground for observation. If characteristic threads are found in different parts of the TD, this may be one reason to conclude that it stems from some kind of a coherent document. These threads can be systematically compared to the other corpora of synoptic material: Mark, Matthean and Lukan Sondergut. As a result of this comparison, it can be determined not only whether the TD is coherent, but also whether it is distinctive. From the time of Heinz Eduard Tödt’s study on the Son of Man, many scholars have recognized that Q, or the TD, has a distinct profile. As, e.g., Arland Jacobson already observed, there are features which run exclusively through the TD1. But the limits of probing coherence and distinctness are quite clear. First of all, we have to work with, at best, fragments of an assumed document. We cannot draw any precise limits as to its extent. Then, there may be erratic blocks in the TD that cannot be connected to other texts through one or another parameter. Finally, it can prove difficult 1. Cf. H.E. TÖDT, Der Menschensohn in der synoptischen Überlieferung, Gütersloh, Gütersloher Verlagshaus, 21963; A.D. JACOBSON, The Literary Unity of Q, in JBL 101 (1982) 365-389; cf. H. SCHERER, Königsvolk und Gotteskinder: Der Entwurf der sozialen Welt im Material der Traditio duplex (BBB, 180), Göttingen, Bonn University Press – V&R Unipress, 2016, pp. 72, 74 n. 274.

116

H. SCHERER

to profile the TD against Matthew’s/Luke’s redaction and Sondergut, because the redactional process obviously preferred the “Luke-pleasing” (to use Austin M. Farrer’s term2) or “Matthew-pleasing” material. But if the TD material differs from Mark and shows traces of diachronic growth, one might at least say that Mark and another older stream of tradition converged in Matthew. Moreover, if alternating primitivity can be demonstrated, one can plausibly assume Luke’s independent access to the TD corpus. And then we are talking about Q. In this essay, I confine myself to the first step of this procedure, the search for threads running through the TD materials and the comparison of the profiles. To develop the distinct profile of a textual corpus, one needs parameters that run through as much of the material as possible. In my Habilitationsschrift, I used the parameter of social categories3. Here, I would like to share reflections on another parameter which I think is worth studying, i.e., marked references to biblical traditions. Over the course of Q research, it has been observed that the treatment of biblical texts in Q might exhibit certain characteristics. In J. Schlosser’s overview of biblical traits in Q, which already differentiates between different categories of reference, he states, in a way similar to others before him, that “[l’] évocation sommaire mais évidente” is typical for Q and is hardly found in other materials4. But this need not be the only way to characterize the phenomenon in question or its salient characteristics. Therefore, a systematic analysis calibrated to prioritize the precision of the description and comparative evidence is in order. To that end, I would first like to clarify what I understand by “biblical” traditions and what type of references can be safely identified for this analysis. Then, I will compare the use of explicit references (marked by introduction formulae) and onomastic references to significant proper nouns in the TD, Mark, Matt and Luke.

2. A.M. FARRER, On Dispensing with Q, in D.E. NINEHAM (ed.), Studies in the Gospels: Studies in Memory of R.H. Lightfoot, Oxford, Blackwell, 1955, 55-88, p. 57. 3. SCHERER, Königsvolk und Gotteskinder (n. 1); EAD., Coherence and Distinctness: Exploring the Social Matrix of the Double Tradition, in M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018, 185-200. 4. J. SCHLOSSER, L’utilisation des écritures dans la source Q, in A. MARCHADOUR (ed.), L’évangile exploré: mélanges offerts à Simon Légasse à l’occasion de ses soixante-dix ans (LD, 166), Paris, Éditions du Cerf, 1996, 123-146, p. 132. Cf., e.g., already D. LÜHRMANN, Die Redaktion der Logienquelle (WMANT, 33), Neukirchen, Neukirchener Verlag, 1969, pp. 98-99, who also mentions the singularity of the temptation story and the fact that only one further quotation with a formula appears in the TD, supposedly due to the lack of Streit- and Schulgespräche.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

I. MAKING USE

OF

117

TEXTS: TAXONOMY

Before addressing the synoptic texts, some preliminary reflections are at order: Firstly, what are biblical traditions? Secondly, and more importantly, when is a reference a reference5? Because texts can be interwoven in different ways, a taxonomy is needed6. 1. Biblical Traditions Apart from the general concept of intertextuality as the interwovenness of all texts available to a culture, this study focusses on the relations of texts to older reference texts7 from “biblical” traditions. One can anticipate that a possible objection is that the concept of “biblical” text is biased by an anachronistic perspective. The establishment of the biblical canon may be seen as still underway in the 1st century AD. Even if they were part of the canon, the texts on scrolls were not evenly distributed or available to all in the manner that a modern printed codex is8. Furthermore, oral transmission may have creatively adapted the written texts. Therefore, the texts qualified as “biblical” need to be characterized more clearly9: I use this term for texts which reasonably can be assumed to have been written down, distributed 5. Literary theories have coined different terms for (a) texts referring to other texts and (b) texts being refered to. In my ‘minimalist’ terminology, I follow J. HELBIG, Intertextualität und Markierung: Untersuchungen zur Systematik und Funktion der Signalisierung von Intertextualität (BNL DF, 141), Heidelberg, Winter, 1996, pp. 76-78, who, after compiling and critically commenting on different terms, opts for “Text”/text (with a fitting adjective, e.g., “manifester Text”, on p. 80) and “Referenztext”/reference text (p. 78). 6. In order to minimize terminological considerations, I refrain from using the term “intertextuality”. I follow R.L. MEEK, Intertextuality, Inner-Biblical Exegesis, and Inner-Biblical Allusion: The Ethics of a Methodology, in Bib 95 (2014) 280-291, who critiques the use of this term out of concerns that are neither synchronic nor reader-oriented nor open, as the concept of intertextuality implies in its original form. For identifying the foregoing texts where intentional references are present, he recommends the terms “inner-biblical exegesis” (filling gaps in the forerunner texts) or “inner-biblical allusion” (without any intention to modify the meaning of the older text). Unfortunately, Meek does not problematize the term “biblical”. 7. Cf. the discussion and definition of W.J.C. WEREN, Intertextuality: Theories and Practices, in ID. (ed.), Studies in Matthew’s Gospel: Literary Design, Intertextuality, and Social Setting (BIS, 130), Leiden – Boston, MA, Brill, 2014, 91-106, p. 94. Weren insists on comparing textual units. 8. Cf. C.D. STANLEY, Rhetoric of Quotations, in B.J. OROPEZA – S. MOYISE (eds.), Exploring Intertextuality: Diverse Strategies for New Testament Interpretation of Texts, Eugene, OR, Cascade, 2016, 42-62, p. 53, who mentions the techniques of copying and memory. 9. Cf. C.M. TUCKETT, Scripture and Q, in ID., From the Sayings to the Gospels (WUNT, 328), Tübingen, Mohr Siebeck, 2014, 196-218, p. 200, referencing Q 11,51 for the limitations of the canon. However, the issue is still under debate. Cf. SCHERER, Königsvolk und Gotteskinder (n. 1), pp. 232f. n. 846 for the discussion. A. LANGE – M. WEIGOLD, Biblical Quotations and Allusions in Second Temple Jewish Literature (SupplJAJ, 5), Göttingen, Vandenhoeck & Ruprecht, 2011, pp. 35f., limit the search to the Hebrew bible based on interpretive interests.

118

H. SCHERER

and regarded as canonical in the Hebrew language in the 1st century A.D. To be sure, the texts that meet this definition and are engaged in this paper are not very controversial. The Tora and the Prophets are rather fixed. Among the Ketubim, only the Psalms and Chronicles are engaged, which may safely be presupposed10. 2. Types of References References to already existing texts are marked according to literary conventions that depend upon time and culture11. Since in antiquity there were no standardized quotation marks12, and hardly any paratexts13 such as, e. g., footnotes, references could be introduced by a formula. Therefore, the most evident class of textual reference is (1) a reference marked by an introduction formula14, e. g. “as is written”; “by the prophet Isaiah”. This is what J. Helbig calls in his helpful systematization15 “Einschreibung” or “Intertextuelle Spur”16. Apart from that, there are also more subtle ways of alluding to older texts, such as: (2) A reference may be clearly recognizable even if an introduction and source are lacking, whenever literary characters or authors are introduced by their proper name. This is what Helbig calls

10. Cf. E. ZENGER – C. FREVEL, Heilige Schrift der Juden und der Christen, in C. FREVEL (ed.), Einleitung in das Alte Testament (KST, 1/1), Stuttgart, Kohlhammer, 92016, 11-36, p. 22. A. LANGE, Canonical History of the Hebrew Bible, in ID. – E. TOV (eds.), Textual History of the Bible, vol. 1A: The Hebrew Bible (Overview Articles), Leiden – Boston, MA, Brill, 2016, 35-48, p. 43, observes an understanding of the Jewish canon in Herod the Great’s time, after centuries of development. 11. Cf. HELBIG, Intertextualität und Markierung (n. 5), pp. 83-84, who limits his discussion to Anglo-American texts of the 19th and 20th centuries. In Helbig’s view, following significant changes, the socio-cultural conditions of reception became rather homogenous during this period. U. LUZ, Intertexts in the Gospel of Matthew, in HTR 97 (2004) 119-137, pp. 126-127, notes that quotations also depend on the canonical status of the reference text and on its metatext (in Genette’s categories), i.e., the genre of the reference text. 12. Other markers that might be visible in the text’s arrangement are also lacking during this time period, e.g., indentation, italics, colon, etc. Cf. HELBIG, Intertextualität und Markierung (n. 5), pp. 121-126. To the best of my knowledge, in scriptio continua and the New Testament manuscript corpus there is no evidence of such conventions. 13. A term coined by Gérard Genette. Cf. HELBIG, Intertextualität und Markierung (n. 5), p. 106, who names titles, motti, dedications, prefaces, headlines, and footnotes. 14. Which correlates to Helbig’s “Potenzierungsstufe”, where the reference text is identified or even the act of reference is in some way discussed in the text. Cf. ibid., pp. 131-137. LUZ, Intertexts (n. 11), p. 135: quotation formula as “Jewish custom”. 15. HELBIG, Intertextualität und Markierung (n. 5), used by, e.g., M.T. PLONER, Die Schriften Israels als Auslegungshorizont der Jesusgeschichte: Eine narrative und intertextuelle Analyse von Mt 1–2 (SBB, 66), Stuttgart, Katholisches Bibelwerk, 2011, pp. 52-56. I owe thanks to Sr. M.M. Gächter, TH Chur, who used Helbig’s work in her inspiring master thesis. 16. HELBIG, Intertextualität und Markierung (n. 5), pp. 79-80, for a sign or a chain of signs belonging to both texts (both phrases are set in italics in the original), which can be subject to change (p. 97).

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

119

“Vollstufe”17. Helbig also uses this term when there is a rupture in the linguistic shape of the reference over against its surroundings. A minimal case would consist of a reference to only one singular yet identifiable word18. If these explicit markers are lacking, Helbig lists a further set of phenomena that could signal a reference is present19. To avoid speculation, I limit the discussion to the aforementioned cases. Biblical exegesis has identified myriads of interplays between the synoptic texts and Scripture. These interplays are not explicitly intended to be read as references, to bear rhetorical weight, or to perform multiple functions in discourse20. They might be overlooked by readers who are less familiar with a particular text or with Scripture writ large21. Thus, I confine myself to the most obvious and secure classes of references: firstly, those marked in the text as references; secondly, Helbig’s category of onomastic markers, which on his definition include “re-used characters” and “re-used authors”22. One last clarification: in what follows I confine myself to characters inhabiting the human world, because discourse on them can be limited – in contrast to superhuman characters such as God, Satan, angels or Wisdom. By their very nature, the latter group takes up so much space in the religious literature of Israel that it cannot be handled here. Applied to the synoptic texts, these parameters will provide enough data to draw conclusions. 17. Cf. ibid., pp. 111-121: “Vollstufe”, or explicitly marked references; but he relativizes the phenomenon by demonstrating that even these references can fail. Unfortunately, Helbig limits “onomastic” references to characters and authors; in my opinion, all proper nouns have the potential to mark references. 18. Cf. ibid., p. 103: e.g., characteristic phrases delivered by literary characters or changes in the code of a text; p. 104: even morphological references. 19. Cf. ibid., pp. 98-106: frequency, proportion (= quantity); position as implicit markers/“Reduktionsstufe”. 20. Cf. STANLEY, Rhetoric (n. 8), pp. 42-45, 56 for Paul’s use of quotations, which might also be applied to the evangelists. 21. Cf. HELBIG, Intertextualität und Markierung (n. 5), pp. 96-97: The more famous a particular text is, the more often it is detected as a reference – but familiarity depends on subjective factors and contexts. Moreover, not all parts of a work are remembered in the same way, e.g., titles are more obvious; they are remembered while the rest can be forgotten. Any changes in the quotation can blur its recognizability. Helbig calls “Nullstufe” (ibid., p. 87) or “[u]nmarkiert” (p. 88) that which is hidden from the readers who do not know the referenced text or do not activate their knowledge (p. 89). 22. Ibid., pp. 113-117: outstanding among these figures are protagonists and figures appearing in the title of a work. I will not take up here Helbig’s category of markers by linguistic code (pp. 117-121). As far as I can see, this category contains the most difficult type of markers to decipher. The link to the reference text must be very strong and distinctive for the reference not to be “textimmanent erklärbar” (p. 120); according to Helbig, it is often visible only to an experienced reader (“dem geübten Leser”, p. 119). Helbig reminds us that the use of characters or authors might not be detected by the reader, e.g., if the names are common or if the reference is loose. Thus, we have to ask whether the figures are of prominence, and, above all, whether the text would make sense without engaging the biblical stories represented by those names.

120

H. SCHERER

II. DATA

AND

ANALYSIS

1. Explicit References Explicit references to Scripture can be analyzed and compared in different respects, e.g., the mode of introduction, the potential textual base of quotations23 or the mode and range of alterations. As a necessary limitation of our scope, only the thematic contexts of the references will be selected here. The TD’s use of explicit references24 is limited to Q 4,3-8 and Q 7,2725. The first one is an example of a reference in an apophthegmatic or dialogue mode; and the second one is used for characterization. Both modes, the apophthegmatic and characterization-oriented one, will now be analysed in the different textual corpora. (1) References in an Apophthegmatic/Dialogue Positioning The sequence of three apophthegms in Q 4,3-826 constitutes a dense cluster of explicit references to Scripture27. These references are exchanged 23. Cf. the short summary in WEREN, Intertextuality (n. 7), pp. 100-103. 24. One may wish to differentiate between verbatim quotations and less precise allusions, cf. LANGE – WEIGOLD, Biblical Quotations and Allusions (n. 9), pp. 25-27. Because I do not regard it as productive, I do not note this difference: the material base of the TD is too small to draw conclusions from the mode of reference. LANGE – WEIGOLD (ibid., pp. 27-28) use the term reference in a different way, “as a referral to another text in a posterior text without a quotation of it or an allusion to it” (p. 27). In the end, they do not differentiate between quotations and allusions (pp. 28-29). 25. In quoting the TD I adopt the Q convention for practical reasons, using the siglum “Q” and following the Lukan versification. 26. In analysing the TD I use the text of the Critical Edition of Q (P. HOFFMANN – C. HEIL, Die Spruchquelle Q: Studienausgabe Griechisch und Deutsch, Darmstadt, Wissenschaftliche Buchgesellschaft, 32009), except for the uncertain parts placed in double brackets. The exact sequence of the verses (a well-known problem in Q 4,1-13) is not of consequence for our purposes here. Portions of the text identical in Luke and Matthew are marked in bold print: Q 4,3 καὶ εἶπεν αὐτῷ ὁ διάβολος· εἰ υἱὸς εἶ τοῦ θεοῦ, εἰπὲ ἵνα οἱ λίθοι οὗτοι ἄρτοι γένωνται. Q 4,4 καὶ ἀπεκρίθη ὁ Ἰησοῦς γέγραπται· οὐκ ἐπ᾽ ἄρτῳ μόνῳ ζήσεται ὁ ἄνθρωπος. Q 4,9 Παραλαμβάνει αὐτὸν εἰς Ἰερουσαλὴμ καὶ ἔστησεν αὐτὸν ἐπὶ τὸ πτερύγιον τοῦ ἱεροῦ καὶ εἶπεν αὐτῷ· εἰ υἱὸς εἶ τοῦ θεοῦ, βάλε σεαυτὸν κάτω· Q 4,10 γέγραπται γὰρ ὅτι τοῖς ἀγγέλοις αὐτοῦ ἐντελεῖται περὶ σοῦ Q 4,11 καὶ ἐπὶ χειρῶν ἀροῦσίν σε, μήποτε προσκόψῃς πρὸς λίθον τὸν πόδα σου. Q 4,12 Καὶ εἶπεν αὐτῷ ὁ Ἰησοῦς· γέγραπται· οὐκ ἐκπειράσεις κύριον τὸν θεόν σου. Q 4,5 Καὶ παραλαμβάνει αὐτὸν ὁ διάβολος εἰς ὄρος, καὶ δείκνυσιν αὐτῷ πάσας τὰς βασιλείας τοῦ κόσμου καὶ τὴν δόξαν αὐτῶν, Q 4,6 καὶ εἶπεν αὐτῷ, ταῦτά σοι πάντα δώσω Q 4,7 ἐὰν προσκυνήσῃς μοι. Q 4,8 ὁ Ἰησοῦς εἶπεν αὐτῷ, γέγραπται Κύριον τὸν θεόν σου προσκυνήσεις καὶ αὐτῷ μόνῳ λατρεύσεις. 27. For the biblical textual foundation cf. T. HIEKE, Schriftgelehrsamkeit in der Logienquelle: Die alttestamentlichen Zitate in der Versuchungsgeschichte Q 4,1-13, in J.M. ASGEIRSSON

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

121

in the framework of an interaction between Jesus and the mythical character of Satan. Q

Quotation Marker

Content

4,4

γέγραπται· γέγραπται

Deut 8,3

γέγραπται28 γέγραπται

Deut 6,16 LXX (Isa 7,12)

4,10 4,12 4,8

Ps 91,11-12 Deut 5,9; 6,13; 10,20

The overarching topic of the apophthegmatic cluster Q 4 is the position of Jesus as a human being before God. By virtue of this theological topic and dense clustering, Q 4 stands out as singular29 vis-à-vis biblical references in the TD. But in these respects it differs also from Markan material. When the Markan Jesus explicitly points to biblical references in apophthegms or dialogues30, they are placed in discussions with humans and substantiate ethical and exegetical positions. Mark

Quotation Marker

Content

Topic

2,25-26

οὐδέποτε ἀνέγνωτε

Episode of David eating bread from the temple

legitimacy of plucking grain ears on Sabbath

7,10

ὡς γέγραπται Μωϋσῆς γὰρ εἶπεν·

Decalogue

hypocrisy of purity rites/halachic practices

– K. DE TROYER – M.W. MEYER (eds.), From Quest to Q: Festschrift James M. Robinson (BETL, 146), Leuven, Leuven University Press – Peeters, 2000, 43-71, pp. 46-63. 28. In Luke 4,12 the quotation formula is εἴρηται. 29. Cf. M. FRENSCHKOWSKI, Nicht wie die Schriftgelehrten: Nichtschriftgelehrte Rezeption alttestamentlich-jüdischer Traditionen in der Logienquelle und im Koran, in D.T. ROTH – R. ZIMMERMANN – M. LABAHN (eds.), Metaphor, Narrative, and Parables in Q (WUNT, 315), Tübingen, Mohr Siebeck, 2014, 221-252, p. 224. 30. For our purposes, I think it would be more helpful to use this broad category instead of differentiating between Schul- and Streitgespräch. E.-M. BECKER, Die markinischen ‘Streitgespräche’ im Plan des Evangeliums: Eine kritische relecture der formgeschichtlichen Methode, in O. WISCHMEYER – L. SCORNAIENCHI (eds.), Polemik in der frühchristlichen Literatur: Texte und Kontexte (BZNW, 170), Berlin – New York, De Gruyter, 2011, 433-463, focuses on the formal criterion of Streitgespräche and notes (ibid., p. 457) that these are absent from Q and Matthean Sondergut (whereas Luke introduces three new instances and leaves out two), despite Matthew’s awareness of conflict and polemic (albeit in other forms). Ibid., p. 458. I exclude the reference in Mark 1,44 because it does not belong to apophthegmata. Mark 9,12; 12,10-11; 14,27 are grouped with the references used for characterization.

122

H. SCHERER

Mark

Quotation Marker

Content

Topic

10,3-4

τί ὑμῖν ἐνετείλατο Μωϋσῆς;31

Deut 24,1.3

question regarding divorce

10,19

τὰς ἐντολὰς οἶδας·

Decalogue

question regarding conditions of eternal life

11,17

οὐ γέγραπται

Isa 56,7

justification of temple action

12,19

Μωϋσῆς ἔγραψεν ἡμῖν

Deut 25,5-6

resurrection and marriage partners

12,26

οὐκ ἀνέγνωτε ἐν τῇ βίβλῳ Μωϋσέως

Exod 3,6

self-presentation of God

Deut 6,4; Lev 19,18

first commandment

12,28-31 πρώτη (ἐντολὴ) ἐστίν· δευτέρα αὕτη 12,36

αὐτὸς Δαυὶδ εἶπεν ἐν Ps 110,1 τῷ πνεύματι τῷ ἁγίῳ·

Davidic sonship

To be sure, questions about ethics and exegetical discourse are also dealt with in the TD. There, we also encounter individuals who question Jesus. But in contrast to Mark, the TD never uses marked references to tackle them. Ethical questions are solved by means of the woes in Q 11,39-47 and the statements on adultery in Q 16,16. In Q 11,17-23 Jesus justifies himself against the accusation of acting through the power of Beelzebul by means of a wisdom discourse. In Matthew, both threads converge: the use of explicit references in apophthegmata for theological purposes, on the one hand, and ethos/ halachah, on the other. Matthew actually expands the number of explicit references for ethical argumentation (Matt 9,13; 12,5.7) in Markan contexts32, but he has no further theological apophthegmata. In Luke, the traditions seem to converge as well. He reuses six of the nine ethical and exegetical references that he finds in the Markan apopththegmata. Luke skips the three items concerning purity and marriage located in Mark 7 and 10. Nevertheless, he integrates the purity woes (Luke 11) and the marriage cluster (Luke 16,16-18) from the TD.

31. Jesus asks for the quotation which is then presented by his dialogue partners. 32. Rephrasing Mark 12,26 as God´s word. The halachic statements of Matt 5,21-48 have an introduction formula, but it refers to the oral exposition of “the elders”. Cf. M. KONRADT, Das Evangelium nach Matthäus (NTD, 1), Göttingen, Vandenhoeck & Ruprecht, 2015, pp. 78-79. Luke has the sapiential answers in different contexts, but offers no introductory statements.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

123

Against the background of all this halachic discourse, Q 4 remains singular33 in the TD, the Markan material, and also in the Sondergut of Matthew and Luke. (2) Characterization by Reference Explicit references are used for the characterization of narrative figures, regardless of the theological, ethical or exegetical positions. This category might overlap with the wider category of apophthegmatic/dialogue settings, but it is marked off by the special intention of the references. They are not used as arguments to obtain a decision in open questions, but as links between characters and biblical “prophecies”. Mark uses this type of characterization for John (1,2-3) and also three times for Jesus. Mark 9,12-13 uses a quotation formula to introduce a free paraphrase, stating that the son of man should suffer much and be killed and that Elijah was treated badly. Mark 12,10f. quotes Ps 117,22-23 LXX which presents Jesus as the cornerstone. Mark 14,27 has Jesus interpreting the events of the passion with the help of Zech 13,7. On one occasion, in Mark 7,6, the dialogue partners of Jesus are characterized by him when he applies to them Isaiah’s description of hypocrisy (Isa 29,13). Matthew recapitulates Mark 12,10-11 and Mark 7,6 with slightly different introduction formulae and Mark 14,27 with a slightly altered quotation. The negative characterization of those who do not understand parables is applied by means of a reference to Scripture once more in Matt 13,14-15. It makes the unmarked allusion of the Markan Vorlage explicit. Matt 21,16 similarly contains an approving characterization of Jesus by means of an explicit reference. Only when Matthew points to the suffering son of man and Elijah in his parallel to Mark 9,12-13 are explicit references to Scripture left out. In his Sondergut, Matthew uses the strategy of characterization most extensively in his fulfilment citations34 (with a christological relation: 1,22-23; 2,5-6.15.23; 4,14-16; 8,17; 12,17-21; 21,4-5). As a result, we can see an increase in the use of the characterizing reference type as we go from Mark to Matthew. On his part, Luke skips Mark 14,27, which is placed in the second gospel on the way to the Mount of Olives. But he inserts a different reference 33. Cf. FRENSCHKOWSKI, Nicht wie die Schriftgelehrten (n. 29), p. 224: the use of Scripture is untypical und singular in early Christianity. 34. Cf. PLONER, Schriften (n. 15), p. 304, for characterization by reference in Matt 1–2: “wesentliche Elemente bei der Charakterisierung von Figuren und Orten”.

124

H. SCHERER

pointing to the passion in the last words of Jesus before leaving for the Mount of Olives (Luke 22,37). He recapitulates Mark 12,11 with a different introduction formula that features new material. He expands John’s characterization from Mark 1,2 and creates a new reference at the outset of Jesus’ ministry (Luke 4,17-19). To be sure, he skips the question concerning Elijah (Mark 9,12-13) completely. But unlike Matthew, who in his characterization and fulfillment quotations avoids all references to suffering, Luke inserts an explicit reference to suffering in Luke 18,31-33 over against Mark. Scriptural prognoses of Jesus’ suffering are not only part of his Sondergut in Luke 24,46 but can also be found in Acts 3,18; 17,2-3; 26,22-23. Further characterization quotations referencing the life and death of Jesus appear in Acts 3,22-23 and 8,32-33. None of Luke’s quotations are the same as in Matthew, apart from Acts 2,34-35, which stems from the Markan tradition. Thus, Luke is highly conservative in his treatment of references made for the purpose of characterization. While characterization quotations flourish in Mark and the Sondergut of Matthew and Luke, characterization by means of reference is found only once in the TD, i.e., in the characterization of John (Q 7,27). Although this passage overlaps with Mark 1,2, there are reasons to attribute it to the TD35. The explicit introduction formula notwithstanding, the wording of what is presented here as a reference has no proper equivalent in the MT, the LXX, or any other Greek translation that we know of36. If the text is classified as part of the TD, characterization by reference is used only here for John and not at all for Jesus. If it is classified as triple tradition, then the TD is void of this type of reference. To sum up, explicit references with formulae are rare in TD material in spite of the availability of ethical and exegetical issues or occasions for characterization. The situation is different in Mark. This might be one indicator of the TD as a distinct and coherent stream of tradition which Matthew integrated. This is not to say that the TD lacked deep roots in the tradition of Israel37. One can only prove that the distinct tradition streams used different strategies in pointing to Scripture. These strategies include onomastic references. 35. Matt and Luke insert it in a more extensive passage, where it fits properly, and both feature ἔμπροσθέν σου in agreement against Mark. 36. If it had reached Mark and the TD in so similar a wording, it must have been an established traditional formula in Greek. FRENSCHKOWSKI, Nicht wie die Schriftgelehrten (n. 29), p. 225, regards it as a quotation from a free-standing baptist tradition (“freies Legitimationswort für den Täufer”), not a “biblical” quotation in its proper sense. TUCKETT, Scripture and Q (n. 9), p. 202, calls it “an earlier, not a later stage”. 37. This is the conclusion of Tuckett (ibid., p. 218), after analysing an unmarked allusion.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

125

2. Onomastic References The phenomenon of onomastic reference was already noted, e.g., by Christopher Tuckett38 and investigated by Marco Frenschkowski. Now this phenomenon must be further situated in relation to the TD. The TD contains proper nouns which typically appear in texts known today as Scripture and which probably were regarded as such also during the period of canonical development in the 1st century. When names from the past are used in TD material, one does not find any names that do not appear in Scripture or biblical tradition. The names are as follows: Q 3,8; 13,28

Abraham, Isaac and Jacob

Q 10,12-15

Cities/inhabitants of Sodom, Tyre and Sidon

Q 11,30-32

Jonah, the Ninevites

Q 11,31; 12,27

Solomon, the Queen of the South

Q 11,51

Abel

Q 11,51

Zachariah

Q 17,26-30

Noah

With the exception of the Patriarchs (Q 3,8; 13,28), Tyre/Sidon (Q 10,1314), and Solomon (Q 12,27), the characters are part of narrated episodes in their reference texts. These episodes are meant to be engaged by the audience, or otherwise the logia would not make sense. In some cases, however, the allusion points to events or details which are not mentioned in the biblical text: • The “sign” of Jonah has no clear reference in the book of Jonah. • In the biblical reference texts 1 Kgs 10,1-13 and 2 Chr 9,1-12, the “Queen of the South” is the Queen of Sheba, and the TD places her in the “corners of the earth”, perhaps to make her journey more exceptional. However, according to the traditional texts, the Queen of the South did not simply come to “listen”, but to challenge Solomon (1 Kgs 10,1). • Q 11,51 locates the death of Zachariah with some precision “between the altar and the house”39, whereas 2 Chr 24,21 only knows of the court. • The biblical accounts are not interested in Solomon’s clothing, whereas Q 12,27 presupposes that Solomon was famous for wearing splendid garments.

38. Ibid., p. 207: required for “meaning to be understood”; doubtful Q 11,30/Jonah. 39. “House” is Luke’s wording. Matthew reads “between the temple and the altar”.

126

H. SCHERER

• Q 17,26-30 describes the generation of Noah living their daily lives unconcerned. That is not what the biblical account tells. The generation of Noah is characterized by evil and violence (Gen 6,5.11-12). In sum, this gives the impression that the biblical characters have developed independent afterlives. They seem to belong to living traditions, with details changed and inserted into the main plot of the stories, as Frenschkowski has already observed40. The main reason for adapting the stories is clear. The biblical characters provide historical examples for the relevant points of Jesus’ teaching: metanoia, clothing, lack of concern, or the search for wisdom. Moreover, the stories of some of these characters of old are expanded into the present and above all into the future, as Frenschkowski has also noticed41. They live on and participate in present or eschatological settings: • Sodom, Tyre and Sidon will face trial at the eschaton (Q 10,12; despite Sodom having already been punished by immediate destruction) • The Queen of the South and the Ninevites will raise their voices at the eschatological trial (Q 11,31-32) • Abraham, Isaac and Jacob are shown to be celebrating in the basileia (Q 13,28-29) Both phenomena, the creative rereading of biblical texts as examples and the projection of characters into the future, reappear several times between Q 10 and Q 17. Accordingly, they can be seen as traits running through at least this portion of the TD. Onomastic references in Mark look rather different. Apart from the patriarchs Abraham, Isaac and Jacob, the Markan characters from the biblical tradition include: 1,2; 7,6

Isaiah

1,44; 7,10; 9,4-5; 10,3-4; 12,19.26

Moses

2,25-26; 10,48; 11,10; 12,35-36

David

2,25-26

Abiathar

6,15; 8,28; 9,4f.11-13; 15,35

Elijah

12,16

Abraham, Isaac, Jacob

40. Cf. FRENSCHKOWSKI, Nicht wie die Schriftgelehrten (n. 29), p. 230: “Gestalten volkstümlicher Imagination”/characters of popular imagination, also active in the eschaton. E.g., Abraham (Q 3,7-8; 13,28). Cf. ibid., p. 232: “einprägsame, schlichte, farbige Gestalten – Kindergottesdienststoff”. 41. Ibid.: “sozusagen vorhistorisches Zeitverständnis”; cf. p. 240 for the non-scribal employment of the texts.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

127

Here, it is not only the dramatis personae that differ, but also the types of reference. In the ensuing analysis, I will therefore begin with the modes employed by Mark to refer to these biblical characters. This will make it easier to show how Matt and Luke42 deal with them and enable a comparison with the TD. (1) Renarrated Biblical Characters Renarrated examples are not limited to the TD. They also appear in Mark 2, in the story of Abiathar and David, which is not an exact match of the biblical reference text. Another instance would be Luke’s Elijah and Elisha example in his Sondergut passage Luke 4,25-27. Luke also duplicates and expands the Noah motif with the references to Lot (Luke 17,28-29) and Lot’s wife (Luke 17,32), while Matthew duplicates the Sodom reference in Matt 11,23-24. All of these characters appear in Jesus’ speech in apothegmatic settings. A comparison of the materials reveals a higher frequency of this mode of reference in the TD. (2) Characters with an Imagined Afterlife The biblical characters inhabiting Mark’s gospel and Luke’s Sondergut also have an imagined afterlife: the patriarchs in Mark 12 are counted among the living; Abraham participates in the basileia according to Luke 16,23-3143. As regards the afterlife, especially noteworthy is the character of Elijah, who is mentioned five times in Mark. First, some suggest that John or Jesus be identified with the (returning) Elijah (Mark 6,15; 8,28). The biblical background of this expectation is Mal 3,23-2444, where God promises to send Elijah before the Day of the Lord. Moses and Elijah are alive and present at the transfiguration (Mark 9,4-5). This is followed by Mark 9,11-13, a reference to Elijah redivivus, whereby Jesus himself identifies John with the prophet of old. Finally, in Mark 15,35-36 some of the bystanders make 42. The observations here are limited to the Jesus traditions, thus Acts is excluded. 43. Jeremiah as inserted by Matthew in Mark 8,28 is not counted here on account of the lack of originality. 44. Also Sir 48,10. These reference texts are quite common, as is the Elijah redivivus theory. But cf. (for both), e.g., R.S. NOTLEY – J.P. GARCÍA, The Hebrew Scriptures in the Third Gospel, in C.A. EVANS – J.J. JOHNSTON (eds.), Searching the Scriptures: Studies in Context and Intertextuality (LNTS, 543), London – New York, Bloomsbury T&T Clark, 2015, 128-147, pp. 137f. Both Moses and Elijah expectations are represented in the synoptic material (ibid., p. 138).

128

H. SCHERER

a misguided comment regarding Jesus as calling on Elijah and mockingly expect Elijah to take him down from the cross, events that have no counterpart in the key quotation from Malachi. Nevertheless, all of the Elijah references depend on the concept of Elijah being taken up to heaven in order to play a special role, either as Elijah redivivus or as a special person of reference for Jesus. Matthean Sondergut also plays on the Elijah redivivus motif with regard to John (Matt 11,13-14), although Matthew skips Mark 6,15. Luke inserts the Elijah motif at Luke 1,17 in relation to John, but he skips Mark 9,11-13 and Mark 15,35. In contrast to Mark, there is no explicit reference to Elijah in the TD. Matthew makes the reference explicit in his version of Q 7,27 in Matt 11,14, but Luke does not. (3) Onomastic References to Authors In Mark’s gospel, Isaiah (1,2; 7,6), Moses (1,44; 7,10; 10,3-4; 12,19.26), and David (12,36-37) are often named as authors of biblical texts, and thus assume the function of a prophet, lawgiver or inspired poet. When Matthew and Luke use Markan material45, they utilize this pattern. Interestingly, Matthew on two occasions rephrases the Markan references to Moses with a reference to “God” having “said/commanded” (Matt 15,4 for Mark 7,10; Matt 22,31 for Mark 12,26). In his Sondergut, Matthew uses this pattern extensively in the fulfilment citations: Matt 2,17; 27,9 (the latter wrongly attributed to Jeremiah instead of Zechariah), 4,14-16; 8,17; 12,17-21; 13,14 (Isaiah). In Luke’s Sondergut, Moses and Isaiah are mentioned as biblical authors (Luke 2,22; 4,17; 16,29.31; 24,44). In contrast to the above, the TD nowhere references the authors of biblical texts by name. (4) References to Genealogy Another Markan concern in using biblical onomastic references appears to be genealogical. David is referenced as a father twice: he is the father of Jesus (Mark 10,47-48) and also of the group expecting his basileia (Mark 11,10). In Matthew and Luke, we find extensive use of this pattern in the two versions of Jesus’ genealogy. In the words of U. Luz, Matthew “narrates here the history of Israel in the condensed form of a genealogy”46. In both cases, the names of Jesus’ forefathers sooner or later find no analogue in a biblical text. 45. Luke skips Mark 7 and Mark 10,1-12 completely. 46. LUZ, Intertexts (n. 11), p. 128.

TESTING BIBLICAL TRADITIONS IN THE SYNOPTIC MATERIALS

129

Davidic sonship is relevant to Matthew (1,1.20; 9,27; 12,23; 15,22; 21,9), although he skips Mark 11,10. Luke also highlights Davidic sonship in his infancy narrative (1,27.32.68; 2,4; 3,31). Beyond this, he also mentions the biblical genealogies of Zechariah, Elisabeth, and Hanna (Abijah and Aaron in Luke 1,5; Asher in Luke 2,36), and Jacob (1,33). Moreover, Luke develops the motif of Abrahamic kinship (1,55.73; 13,16; 16,24; 19,9). By contrast, in the TD, the references to biblical characters as forefathers are limited to Abraham in the Baptist’s speech, which occurs in a polemical context (3,8). III. CONCLUSION In comparison to other synoptic materials, the TD lacks explicitly marked biblical references. Q 4 with its marked theological apophthegmata remains singular in the synoptic tradition, where ethical and halachic uses of biblical references are often found. Despite the fact that the characterization of narrative figures by marked reference is a common strategy in other synoptic materials, it is only found once in the TD (Q 7,27). As regards the onomastic reference type, the TD exhibits a predilection for creatively re-narrated characters, while lacking the patterns detected in other pertinent materials. There is no Elijah reference, no onomastic reference to the biblical authors and no reference to the forefathers. The contrast between the TD and Mark is striking. As regards explicit and onomastic references, Matthew (and Luke) seems to have integrated materials with distinct profiles. Our preliminary sketch on the question of coherence and distinctness ends here. There is much that remains to be explored, including the study of the less overtly marked references and the investigation of the milieus where the patterns of reference appear. One could inquire into the modes of paraphrasing or the affinities to the Masoretic text and the extant Greek translations47. Obviously, the scope of this study is limited to one particular pattern of circumstantial evidence for the TD’s coherence and distinctness. Its results need to be supplemented and substantiated by other observations. The recognition of the distinct profile of a text is a matter of detail that requires one to ponder modes and degrees of similarities and differences. A profile can only be detected by naming precisely the differences. One 47. Helpful categories are also provided by, e.g., M. Pfister. Cf. LUZ, Intertexts (n. 11), pp. 123-124.

130

H. SCHERER

might ask where the line between profile and coincidence is to be drawn. I would like to conclude with a quotation from an expert on criminal profiling, Agatha Christie: “‘Any coincidence’, said Miss Marple to herself, ‘is always worth noticing. You can throw it away later if it is only a coincidence’”48. Universität Duisburg-Essen Essen, Germany

Hildegard SCHERER [email protected]

48. A. CHRISTIE, Nemesis, quoted from the Ebook edition: London, Harper Collins, 2016, 78/position 1021.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022: PROSPECTS IN RETROSPECT

In a recent conference on the Synoptic Problem, Hildegard Scherer, a fellow proponent of the Two-Document Hypothesis (2DH/2ST), observed that “[t]he synoptic question is far from resolved”1. Every decade seems to structure this perennial conversation in a new way, and the 2010s have been no exception. Reflecting the tone set by the 2008 Oxford Conference, in the ensuing decade the landscape of this debate remained contested, with the long-time standard of the field, the 2DH, coming under fire from a new combination of directions2. The arguments made by its critics over this 1. H. SCHERER, Coherence and Distinctness: Exploring the Social Matrix of the Double Tradition, in M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018, 185-200, p. 185. I retain the designation 2ST (Two-Source Theory) in referencing the scholars who use(d) it. On the documentary character of Q (hence, 2DH), see, most recently, J.S. KLOPPENBORG, Oral and Literate Contexts for the Sayings Gospel Q, in C. HEIL – G. HARB – D.A. SMITH (eds.), Built on Rock or Sand? Q Studies: Retrospects, Introspects and Prospects (BiTS, 34), Leuven – Paris – Bristol, CT, Peeters, 2018, 49-72. Cf. infra, n. 58. Of course, most scholars who use the designation 2ST presuppose the documentary character of Q as well. See already in G. STRECKER (ed.), Minor Agreements: Symposium Göttingen 1991 (GTA, 50), Göttingen, Vandenhoeck & Ruprecht, 1993, pp. 21, 241. 2. The proceedings of the 2008 Oxford Conference were published in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011. Since 2010, in support of the 2DH/2ST see ibid. – esp. the essays by Tuckett, Hedrick, Loader, Kirk, and Downing; G. HÄFNER, Das Matthäus-Evangelium und seine Quellen, in D. SENIOR (ed.), The Gospel of Matthew at the Crossroads of Early Christianity (BETL, 243), Leuven – Paris – Walpole, MA, Peeters, 2011, 25-71; J.S. KLOPPENBORG, A Response, in J.C. POIRIER – J. PETERSON (eds.), Marcan Priority without Q: Explorations in the Farrer Hypothesis (LNTS, 455), London – New York, Bloomsbury T&T Clark, 2015, 226-244; C.A. EVANS, The Two Source Hypothesis, in S.E. PORTER – B.R. DYER (eds.), The Synoptic Problem: Four Views, Grand Rapids, MI, Baker, 2016, 27-45; ID., Two Source Hypothesis Response, ibid., 113-125; P. FOSTER, The Doublets in Matthew: What Are They Good For?, in J. VERHEYDEN – G. VAN BELLE (eds.), An Early Reader of Mark and Q (BiTS, 21), Leuven – Paris – Bristol, CT, Peeters, 2016, 109-138; H. SCHERER, Königsvolk und Kinder: Der Entwurf der sozialen Welt im Material der Traditio duplex (BBB, 180), Göttingen, Bonn University Press – V&R Unipress, 2016; EAD., Coherence (n. 1);

132

O. ANDREJEVS

timespan – most notably, from the positions of the Farrer (FH)3 and Matthean Posteriority hypotheses (MPH)4 – invite a renewed treatment of A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016; ID., The Synoptic Problem, Ancient Media, and the Historical Jesus, in JSHJ 15 (2017) 234-259; C.J. MONAGHAN, Why Have We Reduced the Oral Tradition to Silence? The Oral Tradition’s Role in the Formation of the Minor Agreements, in ABR 65 (2017) 1-16; C.M. TUCKETT, Watson, Q and ‘L/M’, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 115-138; H. WOJTKOWIAK, Tradition und Redaktion im Matthäusevangelium: Formale und inhaltliche Charakteristika matthäischer Redaktionspraxis (BZNW, 245), Berlin – New York, De Gruyter, 2021; D.C. ALLISON, Luke Rewriting Matthew? The Case of the Sermon on the Plain, in M.N. PIERCE et al. (eds.), Gospel Reading and Reception in Early Christian Literature, Cambridge, University Press, 2022, 94-121; O. ANDREJEVS, Reception of the Twelve in Matthew and Luke: Comparing the Current Synoptic Hypotheses, in ExpT 133 (2022) 233-244. 3. FOSTER et al. (eds.), Studies (n. 2) – the essays by Poirier, Goodacre, and Eve (see now also Poirier and Eve in this volume); S. ALKIER, Neues Testament, Tübingen, Francke, 2010, pp. 126-127; W. KAHL, Erhebliche matthäisch-lukanische Übereinstimmungen gegen das Markusevangelium in der Triple-Tradition, in ZNW 103 (2012) 20-46; J.C. POIRIER, The Roll, the Codex, the Wax Tablet and the Synoptic Problem, in JSNT 35 (2012) 3-30; F. WATSON, Gospel Writing: A Canonical Perspective, Grand Rapids, MI – Cambridge, Eerdmans, 2013 (esp. pp. 117-285); ID., The Archaeology of the Q Hypothesis: The Case of H. J. Holtzmann, in J.S. KLOPPENBORG – J. VERHEYDEN (eds.), Theological and Theoretical Issues in the Synoptic Problem (LNTS, 618), London – New York, Bloomsbury T&T Clark, 2020, 37-52; POIRIER – PETERSON (eds.), Marcan Priority without Q (n. 2); M. GOODACRE, The Farrer Hypothesis, in PORTER – DYER (eds.), Synoptic Problem (n. 2), 47-66; ID., Farrer Hypothesis Response, ibid., 127-138; M. MÜLLER – J.T. NIELSEN (eds.), Luke’s Literary Creativity (LNTS, 550), London – New York, Bloomsbury T&T Clark, 2016; MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1) – the essays by Müller, Watson, Goodacre, and Kahl; E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021; J. ARCHER, Ancient Bioi and Luke’s Modifications of Matthew’s Longer Discourses, in NTS 68 (2022) 76-88. Although a detailed analysis of M. Klinghardt’s Urevangelium and the Siegert-Wittkowsky hypotheses is not possible here, the most pertinent aspect of their theories for our purposes – canonical Luke’s use of pre-canonical Matthew – can be included in the context of my analysis of the FH. Cf. F. SIEGERT – V. WITTKOWSKY, Von der Zwei- zur Vier-Quellen-Hypothese: Vorschlag für ein vollständiges Stemma der Evangelienüberlieferungen, in HEIL et al. (eds.), Built on Rock or Sand (n. 1), 517-548; M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021 (and his contributions in this volume). 4. R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1992) 1-22; B. ADAMCZEWSKI, Q or not Q? The So-Called Triple, Double, and Single Traditions in the Synoptic Gospels, Frankfurt am Main, Peter Lang, 2010; R.K. MACEWEN, Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS, 501), London – New York, Bloomsbury T&T Clark, 2015; A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226; ID., Plausibility, Probability, and Synoptic Hypotheses: A Response to F. Gerald Downing, in ETL 96 (2020) 131-137. See the contributions by Garrow, Huggins, Lupieri, MacEwen, Saulina, and Tripp in this volume. With the emergence of the MPH, the claim of WATSON, Archaeology (n. 3), p. 38, that the 2DH’s “most vocal rivals … are conventionally named after individual scholars and thus identified with minority views” no longer holds.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

133

certain enduring questions. With all three literary-critical solutions presupposing the now widely accepted hypothesis of Markan priority, some of the basic issues that one can identify in the current Synoptic Problem (SP) discussion include: (1) study and utilization of the Minor Agreements (MAs); (2) feasibility of the synoptic double (DT) and triple (TT) traditions as accurate categories of data5; (3) validity of Mark-Q overlaps as a category of data6; (4) relevance of the alternating primitivity phenomenon7; (5) rejection of congenial materials by FH Luke and MPH Matthew8; and (6) study and utilization of the TT/DT doublets9. A number of other issues could be identified, but because the above six are either foundational to the 2DH or constitute its traditional strong points, I would like to primarily direct the reader’s attention to these. It is in the spirit of collegial dialogue 5. These categories are called into question by ADAMCZEWSKI, Q (n. 4), p. 20; W. KAHL, Inclusive and Exclusive Agreements – Towards a Neutral Comparison of the Synoptic Gospels, or: Minor Agreements as Misleading Category, in MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 3), 44-78, pp. 47, 65. The calibration of the TT by MACEWEN, Posteriority (n. 4), pp. 52-53, n. 89 (criterion No. 6), and the objection recently made by M. GOODACRE, Taking our Leave of Mark-Q Overlaps: Major Agreements and the Farrer Theory, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 201-222, p. 203, n. 6 (“count[ing] elements within pericopae separately”), belong to the conversation as well. For my response to Goodacre’s objection, see O. ANDREJEVS, ‘Major Agreements’? The Synoptic Problem and Mark Goodacre’s Redescription of the ‘Mark-Q Overlaps’, in CBQ 84 (2022) 231-251, pp. 238-241. Cf. infra, Section II. 6. On this, see esp. GOODACRE, Overlaps (n. 5), and my response in ANDREJEVS, Overlaps (n. 5). Cf. infra, Section II. 7. Called into question by ADAMCZEWSKI, Q (n. 4), pp. 93-95; GARROW, Streeter’s ‘Other’ (n. 4), pp. 208-209; GOODACRE, Hypothesis (n. 3), p. 64; EVE, Relating (n. 3), pp. 120-126; A.T. FARNES, The Synoptic Problem and Lectio Brevior Potior: Does Brevity Suggest an Earlier Reading?, in ID. – S.D. MACKIE – D. RUNIA (eds.), Ancient Texts, Papyri, and Manuscripts: Studies in Honor of James R. Royse (NTTSD, 64), Leiden – Boston, MA, Brill, 2021, 47-76. Cf. infra, Section III. 8. Recent attempts to explain this phenomenon for the FH include M. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, pp. 49-52; J.C. POIRIER, The Composition of Luke in Source-Critical Perspective, in FOSTER et al. (eds.), Studies (n. 2), 209-226, p. 222; WATSON, Writing (n. 3), pp. 174-175, n. 32; EVE, Relating (n. 3), pp. 126-131. Cf. my Reception (n. 2) and infra, Section III. 9. Problematized by W. GRÜNSTÄUDL, Luke’s Doublets and the Synoptic Problem, in NTS 68 (2022) 13-25, p. 24. The 19th-century foundation of the modern study of the synoptic doublets is summarized in G. VAN OYEN, The Doublets in 19th-Century Gospel Study, in ETL 73 (1997) 277-306. Cf. F. NEIRYNCK, The Two-Source Hypothesis, in ID., Evangelica III: 1992-2000. Collected Essays (BETL, 150), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001, 343-362, pp. 351-353; H.T. FLEDDERMANN, Q: A Reconstruction and Commentary (BiTS, 1), Leuven – Paris – Dudley, MA, Peeters, 2005, pp. 54-60; FOSTER, Doublets (n. 2) (and the literature there); W. GRÜNSTÄUDL, Wie viele Dubletten finden sich im Lukasevangelium? Forschungsgeschichtliche und methodologische Klärungen, in ETL 97 (2021) 191-222; ANDREJEVS, Overlaps (n. 5). Cf. infra, Section III.

134

O. ANDREJEVS

with the 2DH’s current critics and appreciation of their input that I revisit here the methodological presuppositions of this hypothesis and attempt to develop the implications of the synoptic interrelations model that it presents.

I. CALIBRATION

AND

CLARIFICATION: WHAT THE 2DH IS

AND IS

NOT

An essay centered on the 2DH needs a clarification of its presuppositions, since the hypothesis remains “a very large house with many dwellingplaces”10. We can start our discussion with the canon formulated by Frans Neirynck thirty years ago: “our first hypothesis is the priority of Mark … [t]he Q source is a second and in fact secondary hypothesis, designed to explain the Matthew-Luke agreements in the double-tradition passages”11. If the FH/MPH were found acceptable, there would be no need for Q as the secondary hypothesis. The 2DH’s foundation, then, is a Markanpriority-based analysis of the TT, DT, and Sondergut with all of their phenomena, which leads to a negative assessment of the FH/MPH and, consequently, to the independent use of Mark and Q12. This procedure takes full stock of the MAs13 and – in light of the availability of multiple possible explanations 10. NEIRYNCK, Hypothesis (n. 9), p. 345. 11. ID., Literary Criticism, Old and New, in ID., Evangelica III (n. 9), 65-92, p. 88. 12. ANDREJEVS, Overlaps (n. 5), pp. 236-237. Cf. KIRK, Synoptic Problem (n. 2), p. 251: “Q is the corollary of the analytical postulate that Luke does not know Matthew”. 13. Otherwise, the founding 2ST scholars must be viewed as either so inept that they ignored the MAs or as minimizing the phenomenon’s significance. The expertise of such early 2ST advocates as, e.g., B. Weiss, H.J. Holtzmann, and Sir J.C. Hawkins casts doubt on both options. Of late, the FH theorists have concentrated on the 2ST founders’ supposed ideological motivations, which presumably reveal the superior objectivity of the FH. E.g., KAHL, Agreements (n. 5), p. 65. His claim that the 2ST “has been a function of theological interests rooted in the nineteenth-century quest for the ‘historical Jesus’”, however, paints with a rather broad brush. And is it clear that the recent opposition to the 2DH is itself free of ideological and/or cultural biases? See S.J. Joseph’s contribution in this volume. Cf. C.M. TUCKETT, The Reception of Q Studies in the UK: No Room at the Inn?, in M. TIWALD (ed.), The Q Hypothesis Unveiled (BWANT, 225), Stuttgart, Kohlhammer, 2020, 62-85, p. 70 and notes 38-40. Kahl has also asserted that “Christian Hermann Weisse’s ‘solution’ to the synoptic problem in the form of the [2ST] in 1838 was not the result of an objective investigation”. Even if one granted this assertion (is Weisse’s argument for Markan priority also sub-scientific?), would this then mean that those who subsequently developed Weisse’s “solution” equally lacked objectivity? W. KAHL, The Gospel of Luke as Narratological Improvement of Synoptic Pre-Texts, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 223-244, p. 224. For a more balanced approach, see now Daniel A. Smith: “the obvious role that theological concerns played in the rise and reception of various Synoptic solutions … does not mean that these hypotheses depend on such agendas, for obviously they also include the consideration of ‘serious and substantial literary-critical observations’”. D.A. SMITH,

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

135

– finds them insufficient to require an adjustment of the above conclusion. Mark Goodacre’s description of the 2DH’s procedure does not spell out but probably presupposes the step italicized earlier: “if Matthew and Luke independently edited Mark in the triple tradition, the double tradition necessitates the existence of a common source”14. The negative assessment of the FH/MPH, however, must be emphasized. Its key role was recognized at least as early as Bernhard Weiss and has been restated frequently since then15. It is what calls for an independent redaction of Mark and a mediated solution for the DT. The 2DH theorists are Markan priorists who reject the FH and MPH. Given such emphasis on the negative assessment of the alternatives, it may be appropriate to insert here a brief non-methodological aside and note that the 2DH presents a fundamentally optimistic hypothesis, as far as its reconstruction of the historical events is concerned. Namely, it enables a view of the second and third synoptists as fellow custodians of the tradition rather than, potentially, theological rivals or opponents16. In its lack of a deconstructive component, the 2DH’s view of the synoptic tradition differs “No Weapon but That of Analysis”: Issues at Stake in the Rise and Reception of the TwoDocument Hypothesis, in KLOPPENBORG – VERHEYDEN (eds.), Theological and Theoretical Issues in the Synoptic Problem (n. 3), 113-134, p. 134 (quoting Kloppenborg). 14. Case (n. 8), p. 167. 15. B. WEISS, Lehrbuch der Einleitung in das Neue Testament, Berlin, Hertz, 21889, p. 486, n. 1. The negative assessment of the alternatives appears on the very first page of Weiss’s discussion of Q (which he calls “die älteste Quelle”) and addresses the MPH and FH (in that order), far from the leading challengers to the 2ST in the 1880s. Weiss references the phenomenon of alternating primitivity. Similarly, H.J. HOLTZMANN, Die Synoptiker (HCNT, 1,1), Tübingen, J.C.B. Mohr, 31901, p. 14 (on Watson’s recent selective presentation of this very argument by Holtzmann, see infra, n. 198). Later in his discussion, Weiss (Einleitung2, p. 541) also references what one may describe as rejection of congenial materials by FH Luke. Cf. infra, Section III. More recently, on opposite sides of the recent discussion, C.M. TUCKETT, Q and the History of Early Christianity: Studies on Q, Edinburgh, T&T Clark, 1996, p. 4; F. WATSON, Q as Hypothesis: A Study in Methodology, in NTS 55 (2009) 397-415, p. 398. 16. GOODACRE, Case (n. 8), p. 132: “perhaps Luke is like Martin Scorsese and the third Gospel like The Last Temptation of Christ … both derivative and subversive, influenced by and yet critical of those Jesus films that came before it … a literary artist at work”. For a less romantic assessment of the same phenomena, see V. WITTKOWSKY, Luke Uses/Rewrites Matthew: A Survey of the Nineteenth-Century Research, in MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 3), 3-25, p. 4: it is not “unimaginable that Lukan theological and literary strategy was rather contradictory to that of Matthew … [and] naturally deconstructive”; there is “‘corroborative evidence’ of serious tensions and ‘animosity’ between the Synoptic Gospels … unfortunately widely ignored nowadays”. There has been some acceptance: KAHL, Luke (n. 13), p. 230, seeks to rename the FH as SIM (Synoptic Improvement Model). WATSON, Archaeology (n. 3), p. 38, notes that “Luke’s familiarity with Matthew may … imply that Luke is critical of Matthew” (author emphases in Wittkowsky and Watson). Cf. infra, n. 31.

136

O. ANDREJEVS

from those produced by a number of its current critics. One may include this struggle over competing visions with the conceptual stakes of the discussion. Returning to methodology, it remains, in my view, unnecessary to adopt Deuteromarcus or a similar recensional variant of the 2DH17. These hypotheses are typically utilized to account for the challenge posed by the significant MAs18. Werner Kahl correctly objects to Deuteromarcus’s role of an additional hypothetical source (less convincingly, he uses it as an argument against the 2DH’s plausibility)19. The harmonizing trajectory of our textual tradition problematizes the hypothesis of such a Markan recension20. As importantly, its construction is a product of the 2DH’s occasional surrender to the demand for a global explanation of the significant MAs21. We will not be so pessimistic. Finally, turning to Q, the use of the term Two-Document Hypothesis conveys the presupposition of the singular and written character of the synoptic source22. There may have been QMatt / QLk recensions, although there does not appear to be sufficient evidence for extensive differences between them. I accept the possibility of reconstructing Q’s approximate

17. Popularized by W. SANDAY, The Conditions under Which the Gospels Were Written, in Their Bearing upon Some Difficulties of the Synoptic Problem, in ID. (ed.), Studies in the Synoptic Problem by Members of the University of Oxford, Oxford, Clarendon, 1911, 3-26, pp. 21-24. Cf. infra, n. 18. T.A. FRIEDRICHSEN, The Matthew-Luke Agreements against Mark 1974-1991, Ph.D. dissertation, Leuven, 1992, pp. 164-177, distinguishes between Deuteromarcus (“a complete later redaction of canonical Mark” – p. 169) and “some recension[s] of Mark” (p. 164). 18. Recently, e.g.: H. KLEIN, Das Lukasevangelium (KEK, I/3), Göttingen, Vandenhoeck & Ruprecht, 102006, p. 45; M. WOLTER, Das Lukasevangelium (HNT, 5), Tübingen, Mohr Siebeck, 2008, p. 12. The definitive resources for the Deuteromarcus hypothesis remain A. ENNULAT, Die “Minor Agreements”: Untersuchungen zu einer offenen Frage des synoptischen Problems (WUNT, II/62), Tübingen, Mohr Siebeck, 1994; A. FUCHS, Spuren von Deuteromarkus I-V (SNTU NF, 1-5), Münster, LIT, 2004-2007. For analysis and critique of the hypothesis, see F. NEIRYNCK, Deuteromarcus et les accords Matthieu-Luc, in ID., Evangelica: Gospel Studies. Collected Essays, ed. F. VAN SEGBROECK (BETL, 60), Leuven, Leuven University Press – Peeters, 1982, 769-780; FRIEDRICHSEN, Agreements (n. 17), pp. 164-177; W. SCHENK, Zur Frage einer vierten Version der Seesturm-Erzählung in einer Mt/Lk-Agreement-RedaktionsSchicht („Dt-Mk“): Versuch einer textsemiotischen Geltungsprüfung von A. Fuchs, in STRECKER (ed.), Minor Agreements (n. 1), 93-118; HÄFNER, Matthäus-Evangelium (n. 2), p. 70; M. MCLOUGHLIN, Listing the Minor Agreements, in ETL 87 (2011) 201-228, pp. 201, 205-206. 19. Agreements (n. 5), p. 45. 20. KLEIN, Lukasevangelium (n. 18), p. 45, n. 8 offers an interesting theory that nevertheless contains too many untestable assumptions. 21. The principal difference between the Deuteromarcus hypothesis and the Q hypothesis is that the former reacts to the (significant) MAs. In contrast to this, the Q hypothesis reacts to a variety of phenomena. It is not a ‘one issue’ hypothesis. 22. Supra, n. 1; infra, n. 58.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

137

extent and sequence23. Those less comfortable with Q’s word-level reconstruction may be referred to Neirynck’s Q-Parallels24. With Neirynck, I exclude from Q’s reconstruction (as found with a varying degree of confidence in its current baseline, the Critical Edition)25 the following verses: Q 3,2b-3a.[[21-22]]26; 7,[[29-30]]; 11,17a.[[21-22]]27; 15,[[8-10]]; 17,2. [[20-21]].?28-29?.3328; also with Neirynck, I include Q 14,529. *  *  * Having calibrated the 2DH within the range of parameters provided by its theorists, we can now take up three important claims promulgated by its critics that have the capacity to affect its perception. First, there is the juxtaposition, formulated by the FH theorists, between the “narrowly source-critical focus of Streeter and those in his legacy” and Luke’s presumable literary artistry30. This juxtaposition paints the 2DH as restrictive. It is, 23. The principal argument advanced against this in recent discussion has been the Reconstructed Mark (=MarkR) experiment. There are a number of problems, however, with accepting this experiment as analogous to Q’s reconstruction. See my The Reconstructed Mark and the Reconstruction of Q: A Valid Analogy?, in BTB 50/2 (2020) 35-43 (and the literature there). Cf. SCHERER, Coherence (n. 1), pp. 186-187. 24. Q-Parallels: Q-Synopsis and IQP/CritEd Parallels (SNTA, 20), Leuven, Leuven University Press – Peeters, 2001. All in-verse delimitations in this essay are based on this resource (e.g., Luke 3,16b-f; 11,15bc). 25. J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000. I retain the CritEd’s sigla (if present). 26. F. NEIRYNCK, The First Synoptic Pericope: The Appearance of John the Baptist in Q?, in ID., Evangelica III (n. 9), 209-244; ID., The Reconstruction of Q and IQP / CritEd Parallels, in A. LINDEMANN (ed.), The Sayings Source Q and the Historical Jesus (BETL, 158), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001, 53-147, pp. 78-82. 27. ID., Mt 12,25a / Lc 11,17a et la rédaction des évangiles, in ID., Evangelica II: 19821991. Collected Essays, ed. F. VAN SEGBROECK (BETL, 99), Leuven, Leuven University Press – Peeters, 1991, 481-492; ID., Mark and Q: Assessment, in ID., Evangelica III (n. 9), 505-545, pp. 513-517. 28. ID., The Minor Agreements and Q, ibid., 245-266, pp. 252-254; ID., Saving/Losing One’s Life: Luke 17,33 (Q?) and Mark 8,35, ibid., 480-499. 29. ID., Luke 14,1-6: Lukan Composition and Q Saying, in ID., Evangelica II (n. 27), 183-204. See ID., Literary Criticism (n. 11), p. 88: it is not warranted “to exclude from Q dispersed sayings common to Matthew and Luke”. 30. GOODACRE, Case (n. 8), p. 104. See also, just recently, M. MÜLLER, Were the Gospel Authors Really ‘Simple Christians without Literary Gift’ (Albert Schweitzer)? Arguments for the Quest for Sources behind the Gospels, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 79-96. According to Müller, the “two main tendencies in today’s discussion of the mutual relations between the synoptic gospels” either attribute to the synoptists “the least possible independent creativity” or view them as “independent and creative authors” (ibid., p. 79). There is a more moderate view, one where Matthew and Luke are fellow creative custodians of the tradition.

138

O. ANDREJEVS

however, less than accurate; 2DH Matthew and Luke have been shown to exhibit considerable finesse in weaving together Mark, Q, and Sondergut31. Moreover, 2DH and FH Luke’s creativity in contextualizing the received DT materials does not change materially based on the hypothesis. The second popular claim is that the 2DH’s critics are “able to make good sense of all the data”32 (the implication here appears to be that the 2DH does not fare as well or, at least, is not as efficient). While one can appreciate the optimism of such assertions, there surely must be a difference between testable, partially testable, and, in some cases, entirely untestable explanations. Any of these can help one “make sense” of the synoptic data; the premium in scientific analysis should probably be placed on explanations that can be tested. Typically, the 2DH’s critics point to a promising methodological asset in the MAs33. These are factual albeit raw data (to which I will return in Section II) and they are testable34. Whether the results actually favor the 2DH’s critics is debatable, as we shall see later. The theoretical frameworks of the FH/MPH, in any case, supplement their interpretation of these data with explanations of other synoptic phenomena that tend to run much lower on the scale of testability. Thus, one encounters references to the third synoptist’s hypothetical creativity35, vagaries of 31. E.g., NEIRYNCK, Mt 12,25a / Lc 11,17a (n. 27); ID., The Minor Agreements and the Two-Source Theory, in ID., Evangelica II (n. 27), 3-42; KIRK, Q in Matthew (n. 2). Somewhat perplexing is the view of J.T. NIELSEN, Introduction, in MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 3), xiii-xxvii, whose FH Luke possessed “the necessary creativity” to rewrite Matthew (p. xxvii) but whose 2DH Matthew and Luke were not “writers in their own right” (p. xxi). Are not essentially the same DT materials utilized by Luke on both the 2DH and FH? It appears, then, that FH Luke’s distinctive creativity is of a negative, discriminating, deconstructive sort (frequently, though not always, dislodging the DT materials from their Matthean contexts – in which regard FH Luke indeed differs from [any] 2DH Luke) rather than a positive, constructive one (contextualizing the received DT materials, which the FH and [conventional] 2DH Luke do similarly). By “conventional” 2DH I refer to Q as largely limited to the DT. A degree of negativity of a somewhat different – theological rather than compositional – sort is explored by WITTKOWSKY, Luke (n. 16), pp. 3-25: Matthew is a “negative pretext to the Gospel of Luke” (p. 24); the latter is written “as a direct antithesis to Matthew, i.e., [as] a theological opponent of the First Gospel” (p. 9, summarizing the opinion of E. Zeller). Cf. supra, n. 16. 32. GOODACRE, Response (n. 3), p. 132. Similarly (for the MPH), GARROW, Streeter’s ‘Other’ (n. 4), p. 226. 33. E.g., GOODACRE, Case (n. 8), pp. 152-169; MACEWEN, Posteriority (n. 4), pp. 61-62; GARROW, Streeter’s ‘Other’ (n. 4), p. 222; KAHL, Agreements (n. 5); SIEGERT – WITTKOWSKY, Hypothese (n. 3), pp. 522-523, 544; EVE, Relating (n. 3), pp. 75-85, 114-115; KLINGHARDT, Gospel (n. 3), pp. 250-251. 34. As noted by M.E. BORING, The “Minor Agreements” and Their Bearing on the Synoptic Problem, in FOSTER et al. (eds.), Studies (n. 2), 227-251, p. 232: the MAs “exist, because Matthew, Mark, and Luke exist”. 35. GOODACRE, Case (n. 8), pp. 104, 132; KAHL, Agreements (n. 5), p. 46; F. WATSON, Seven Theses on the Synoptic Problem, in Disagreement with Christopher Tuckett, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 139-147, pp. 145-146.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

139

his memory36, unknowability of his mind37, sufficiency of his prerogative38, or some combination thereof. There is now a concession of the “data we do not and can never have” and “factors we cannot reconstruct with any hope of precision”39. While this concession is likely to be true with regard to the SP generally and can be equally claimed by all available solutions, the frequency with which specific problems are redescribed40 should factor into the overall plausibility of each available model. Redescriptions may give the appearance of making sense of the data, but if their testability is limited, they can ultimately only produce limited evidential value. Given the recent proliferation of the above claims, one wonders if they are running the risk of becoming their own evidence base. To be sure, both the 2DH and its current rivals are forced to turn to untestable explanations somewhere. What is important to recognize is that they do so at different points in their theoretical frameworks. The 2DH as calibrated here turns to untestable explanations for a small percentage of the MAs, while offering a testable theory for the vast majority of them and for Matthew’s and Luke’s source utilization. The FH/MPH specialists will need to turn to rationalizations such as listed above in explaining what the 2DH does not struggle with – the third synoptist’s source utilization. Since it is generally accepted that unfalsifiability undermines the legitimacy of a hypothesis, it is especially surprising to see the persistent appeal to something like an epistemic gap between the modern scholar’s mind and that of FH Luke. The apparent need to turn to this appeal reveals stress fractures at the key juncture in the theory. The appeal itself is hardly bulletproof, of course. If one can consistently understand Luke’s utilization of Mark, it should not be beyond one’s cognitive ability to establish similar consistency in making reasonable surmises about FH Luke’s or MPH Matthew’s expressed intent when utilizing the other’s gospel. The SP then essentially boils down to this: one has to show that the third synoptist acted logically 36. GOODACRE, Hypothesis (n. 3), p. 57 (revisiting his earlier argument from ID., Fatigue in the Synoptics, in NTS 44 [1998] 45-58, pp. 55-56 [and see further examples on pp. 54-56]); KAHL, Agreements (n. 5), p. 46 (distractedness); EVE, Relating (n. 3), p. 132, n. 57 (on Luke 11,15bc). Against Goodacre’s utilization of the hypothesis of editorial fatigue, see T. HÄGERLAND, Editorial Fatigue and the Existence of Q, in NTS 65 (2019) 190-206. 37. KAHL, Agreements (n. 5), p. 46; ID., Luke (n. 13), p. 233 and n. 46. Cf. EVE, Relating (n. 3), p. 194: “The proposals … outlined above cannot be proved beyond all reasonable doubt, since they postulate what cannot strictly be known, namely what associations Matthew’s text may have triggered in Luke’s mind”. 38. FH: WATSON, Writing (n. 3), p. 169; ID., Seven Theses (n. 35), p. 146. MPH: GARROW, Streeter’s ‘Other’ (n. 4), p. 223; ID., Plausibility (n. 4), p. 131, n. 4. 39. EVE, Relating (n. 3), p. 205. Cf. KAHL, Luke (n. 13), p. 233. 40. See the discussion of “renaming the problem” as a type of editorial rationalization in J.S. KLOPPENBORG, Conceptual Stakes in the Synoptic Problem, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 1), 13-42, pp. 22-24.

140

O. ANDREJEVS

and consistently, and that his hypothetical actions reflect the contemporary compositional practices. If a lack of logic or consistency is detected, a hypothesis that seeks to make a compelling case cannot remain above offering a testable explanation. The testability of one of the FH’s most important theoretical components – FH Luke’s presumable creativity – can be briefly illustrated by surveying the recent studies by Alan Kirk (2DH) and Eric Eve (FH)41. Kirk observes that the FH has a tendency “to dismiss the need to give an account of Luke’s use of Matthew on source-utilization grounds”, deferring instead “to authorial creativity and the literary imagination”42. In response, Eve acknowledges the need to refine the argument, moving the conversation past the shopworn ‘hypothetical vs. manuscript-based’ binary43. His engagement of Virgil’s Homeric imitation44, however, while aimed at Kirk’s challenge45, does not substantially advance the FH’s cause, given that the Aeneid offers a continuation of the events of the Trojan War. By contrast, FH Luke’s imitation of Matthew forms a remake rather than a sequel46. FH Luke and Virgil, therefore, would have faced different types of editorial challenges and decisions, something that Eve’s proposed analogy cannot reflect47. Nor do the instances of a potential imitation of one or multiple episodes from an otherwise different narrative offer much clarity48. While ancient authors certainly could imitate and emulate one another, the principal question – again – is whether we possess other examples of remakes in antiquity whose contents range from highly fluid literary imitation to preserving doublets of sayings. Eve’s suggestion that only “one” example of 41. Now further developed by each author in this volume. I limit my observations to Kirk’s and Eve’s earlier work. 42. Synoptic Problem (n. 2), pp. 253-254. See the broader discussion in his Q in Matthew (n. 2). 43. For MÜLLER, Without Literary Gift (n. 30), p. 79, the turn to authorial creativity is still a product of “operating primarily with the actual texts”. Similarly, KAHL, Luke (n. 13), pp. 230-231, 234. See, however, KIRK, Synoptic Problem (n. 2), pp. 251-252: “[s]cientists confidently posit entities not directly observed but evident from their effects, for example, planets in distant solar systems through the effects of their motion on the light signature of their stars”. 44. Relating (n. 3), pp. 55-73, 195-196. 45. Ibid., p. 65: “imitation cannot be reduced to a neat set of mechanical operations”. 46. Ibid., pp. 197-198. 47. For example, because Virgil is not remaking the Iliad, he is not confronted with a complex set of decisions regarding his depiction of its main protagonists. See my Reception (n. 2) and infra, Section III. 48. EVE, Relating (n. 3), pp. 58-59, 195-196, references Tobit’s potential imitation of Homer and Genesis. These potential imitations are distant and fragmentary compared to FH Luke’s and MPH Matthew’s projects.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

141

Luke imitating Homer would be sufficient “to show that imitation is a technique Luke employed”49 does not yet constitute proof of Luke’s imitation of Matthew. When it comes to the mechanical operations of the FH/2DH evangelists, Kirk is able to limit 2DH Matthew’s source-utilization to one particular set of editorial techniques50. Moreover, all of the case studies examined by Kirk share the objective proposed for 2DH Matthew, which is to organize his sources into a coherent work with minimal losses51. Thereby Kirk’s model passes the test, with multiple attestations. Eve’s FH Luke counterpart, in the meantime, is not based on the same analogy as his earlier appeal to imitation. Instead, Eve turns to Philo of Alexandria’s utilization of the Pentateuch52. But – not to belabor the point – Philo’s intention is not to remake the Pentateuch (contrast FH Luke vis-à-vis Mark and Matthew). In his amalgam of Virgil and Philo, seeking to testably ground FH Luke in antiquity, Eve therefore ends up combining two vastly different analogies, neither of which shares FH Luke’s apparent objective53. While something 49. Ibid., p. 69. 50. Q in Matthew (n. 2), pp. 73-92: Apollonius Sophista (Lexicon Homericum); Antonius (Melissa); Melissa Augustana (= Melissa Barocciana; Florilegium Monacense; Florilegium Baroccianum); Joannes Stobaeus (Anthologion). See now his essay in this volume. 51. Q in Matthew (n. 2), p. 89. 52. EVE, Relating (n. 3), pp. 46-55. 53. The amalgam is also less than consistent. The Philonic-Virgilian tour de force whereby Eve’s FH Luke exercises his virtuoso memory command of Matthew to dissect, rearrange, and imitate it forms a rather sharp contrast with the blocks where the same author works with Mark. EVE, Relating (n. 3), p. 198, references M. GOODACRE, Re-walking the “Way of the Lord”: Luke’s Use of Mark and His Reaction to Matthew, in MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 3), 26-43, to correctly observe that “Luke’s use of Mark also encompasses instances of more creative imitation”. But these instances do not materially affect FH Luke’s default manner of utilizing Mark (copy and paraphrase) and the enormous contrast this sets up with FH Luke’s utilization of Matthew. Goodacre’s own solution (Case [n. 8], pp. 88-90), whereby FH Luke knew Mark longer than Matthew, is neither probative nor particularly evident (cf. HÄFNER, Matthäus-Evangelium [n. 2], p. 41, on MPH Matthew’s presumable favoritism of Mark: “Die Annahme einer Höherbewertung des MkEv ist … eine Konsequenz … die aber aus den Texten selbst nicht wahrscheinlich zu machen ist”). It is disrupted by the many minute MAs that require substantial memory expertise in Matthew – that is, if one follows M.D. GOULDER, The Order of a Crank, in C.M. TUCKETT (ed.), Synoptic Studies: The Ampleforth Conferences of 1982 and 1983 (SupplJSNT, 7), Sheffield, JSOT, 1984, 111-130, p. 113: “My Luke has probably a cramped writing table with space for his own scroll and the one he is using as his base-of-the-moment. Mark and Matthew take turns to go on the floor. Where there are overlaps and minor agreements and such things, it is from reminiscence of a familiar parallel text”. For an assessment of this procedure, see R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, pp. 190-200. To be sure, Goodacre could turn to a strictly visual model (both sources are consulted simultaneously). But this has its own challenges, such as the rejection of congenial materials and the

142

O. ANDREJEVS

similar to Philo’s fluency in the Pentateuch can be detected in FH Luke’s Matthean proficiency (although it remains to be seen whether Philo’s fluency is as detached from the Pentateuch’s compositional schematic as FH Luke is from Matthew’s), the model as such remains sui generis. 2DH Matthew fits an established literary pattern exhibited in antiquity – in terms of its objective and its execution – whereas FH Luke does not. The third popular claim made by the 2DH’s critics today is to define it as requiring complete independence of Matthew and Luke54. This framing of the discussion, however – insofar as it tends to axiomatically exclude (or just deny the 2DH) any potential role played by orality – does not seem very adequate55. The frequent inference whereby if Luke was aware of Matthew’s gospel he had to have used it56 calls for some nuancing57. While the matter admittedly has not been helped by those 2DH theorists who have adopted similarly Gutenbergian frameworks (Deuteromarcus, Luke’s subsidiary use of Matthew, etc.), such constructs are only necessary in a bookish imaginarium that hardly seems supported by what we know of the Greco-Roman antiquity. I am, of course, not calling for a revival of James

unsustained MAs (see the discussion of both below). Goodacre’s other suggestion (Case [n. 8], p. 89), that harmonizing Mark and Matthew would have been “anachronistic”, is already somewhat problematized in his own footnote (ibid., p. 89, n. 28). N. PERRIN, What Justin’s Gospels Can Tell Us about Tatian, in M.R. CRAWFORD – N.J. ZOLA (eds.), The Gospel of Tatian: Exploring the Nature and Text of the Diatessaron (RJT, 3), London – New York, Bloomsbury T&T Clark, 2019, 93-109, p. 93, now draws attention to “the genre of gospel harmonies … reaching as far back as 2 Clement”. 54. E.g., M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, p. 6; WATSON, Writing (n. 3), p. 120; GOODACRE, Hypothesis (n. 3), p. 51; KLINGHARDT, Gospel (n. 3), pp. 9, 188, 528; etc. 55. Cf. KLINGHARDT, Gospel (n. 3), p. 10: “orality” is “principally unprovable”, does not explain the “minute analogies in the formulations” (cf. pp. 251, 375). And yet, “[r]ecognition of the pervasiveness of orality in the ancient world has rightly brought into disrepute the so-called literary paradigm that posits a closed textual universe of reading, copying, editing, and redacting”. KIRK, Q in Matthew (n. 2), p. 13. Cf. ibid., pp. 1-28. The FH-2DH debate is less antagonistic. The FH as calibrated by Goodacre has ample room for oral tradition. Surely what is good for the goose is good also for the gander? 56. E.g., GOODACRE, Hypothesis (n. 3), p. 48 (referencing A. Farrer); EVE, Relating (n. 3), p. 205. See also the back cover of MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 3): “[i]f it can be shown that Luke knew Matthew’s Gospel the Q-hypothesis is superfluous”. 57. E.g., W. ARNAL, The Synoptic Problem and the Historical Jesus, in FOSTER et al. (eds.), Studies (n. 2), 371-432, p. 381: Luke could have “heard Matthew recited and was inspired in some of his own literary choices by that knowledge” or known “of Matthew, with some degree of influence, but did not use Matthew as a source”. The second option in particular offers a sufficient explanation of the structural similarities: the third synoptist would have been aware of the existence and rough outline of the middle synoptist’s gospel. Watson’s (Writing [n. 3], p. 120) “total independence”, then, is hardly a necessary requirement.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

143

D.G. Dunn’s expansive role of orality58. But there may be room here for a rapprochement with Goodacre and some space for “a reassessment of – and perhaps ultimately a more nuanced role for – oral tradition in Synoptic relationships”59. In other words, awareness and utilization should probably not be merged indiscriminately. Some degree of awareness may be suggested by the third synoptist endeavoring a project structurally similar to the second synoptist’s60. Utilization, however, remains to be demonstrated.

II. RECENT ARGUMENTS AGAINST THE 2DH’S METHODOLOGICAL FOUNDATION Twenty years after Goodacre bid adieu to Q61, the 2DH continues to hold its position. To analyze the most recent arguments against its methodological foundation, I have organized the following discussion to move from the principal (TT; DT) to subsidiary categories of synoptic data (MAs; Mark-Q overlaps). Arguing that the TT “has been necessitated by the 2ST”, Kahl defines it as “passages that all three synoptic gospels have in common”62. As we shall see, this definition needs more precision. But let us begin with Kahl’s claim of a pro-2ST bias. In fact, the TT is a direct heir of Eusebius’ canon-tables and the 2ST was hardly on the fourth-century agenda. For approximately a millennium the Eusebian apparatus facilitated cross-referencing of the fourfold gospel in the Christian East and West, with canons I-II offering the equivalent of the synoptic TT63. Of course, these and the remaining canontables were not calibrated with the desiderata of modern source criticism in mind. They simply map similar (or, in the case of canon X, unique) passages, not concerned with the possible direction of utilization. The very idea of such utilization seems to have no place in Eusebius’ theory of the gospels’ 58. J.D.G. DUNN, Altering the Default Setting: Re-envisaging the Early Transmission of the Jesus Tradition, in NTS 49 (2003) 139-175. See the critical analysis of Dunn’s and his student’s T.C. Mournet’s proposals in J.S. KLOPPENBORG, Variation in the Reproduction of the Double Tradition and an Oral Q?, in ETL 83 (2007) 53-80. Cf. A. KIRK, Orality, Writing, and Phantom Sources: Appeals to Ancient Media in Some Recent Challenges to the Two Document Hypothesis, in NTS 58 (2011) 1-22, pp. 2-3, 16-17. 59. Case (n. 8), p. 188. Cf. ID., Hypothesis (n. 3), p. 48. 60. Ibid., pp. 53-54. Cf. supra, n. 57. 61. Case (n. 8), p. 189. 62. Agreements (n. 5), p. 47. 63. Epistula ad Carpianum (NA28, pp. 89*-94*). One needs to subtract John from canon I but this hardly affects most of the synoptic parallels.

144

O. ANDREJEVS

origins64. His canon-tables, however, are designed to produce categories of data. Accordingly, they may be worth revisiting in the context of the current SP discussion as an independent consultant not yet far removed from the NT era, even as one must remember that these tables “contain a wide range of both precise and less precise parallels”65 and should factor in Markan priority for the hypotheses that are based on it. In anticipation of a larger discussion, I limit myself to five observations. First, on theories presupposing Markan priority, the TT can be defined more precisely as utilization of the Markan text by both later synoptists, which is a shared point of departure for the FH/MPH/2DH66. There is no need to avoid this rather obvious conclusion for those who reject the Griesbach Hypothesis (which, notably, includes all the participants in this volume). Then, it can be recognized that Matthew’s and Luke’s versions of Markan passages feature expansions containing texts either shared by both later synoptists (DT) or unique to one of them (Sondergut). The hypothesized origin of the DT texts (with the second synoptist or Q) is immaterial to the fact that on Markan priority the third synoptist edits Mark with their help. Presumably calibrated without any awareness of the modern Markan priority hypothesis, the Eusebian apparatus is fully compatible with the theoretical foundation just described, differentiating between the categories of in quo tres, in quo duo, and in quo [evangelist X] proprie. Second, on Markan priority, it is possible to add precision by defining the TT as any number of unaltered or redacted sentences and phrases that can be identified as derived from Mark’s gospel67. Within a Markan passage of Matthew and Luke, the DT/Sondergut materials with which Matthew 64. Hist. eccl. II,15; III,4,7; III,24; III,39,15-16. 65. C. NORDENFALK, The Eusebian Canon-Tables: Some Textual Problems, in JTS 35 (1984) 96-104, p. 103. 66. ADAMCZEWSKI, Q (n. 4), p. 20, correctly notes that “the conventional definition” of the TT calls for further precision. However, once Markan priority is presupposed (as Adamczewski himself does – ibid., p. 443), the extent of what constitutes the TT emerges rather naturally if one thinks in terms of source utilization. 67. E.g., in Luke 3,1-18, according to the Eusebian apparatus (NA28): Luke §6-11 (canons III [DT+John: 7/6/2; 7/6/25]; I [TT+John: 8/2/7/10]; V [DT: 10/8]; X [Sondergut: 9]; I [TT+John: 11/4/10/6 et al.]; V [DT: 12/11]). From the perspective of Markan priority, I would correct Luke 3,1-2 to TT (canon I in the Eusebian system). I would furthermore include Luke 3,16b-f in the DT (see the fourth observation above). In all of this, I admittedly go beyond Eusebius who does not predicate his system on Markan-priority-based source utilization. But if Markan priority can be presupposed (and this presupposition today appears sufficiently secure), we can move one step further. My immediate point, in any case, is to show that Eusebius’ equivalent of TT material is found in Luke 3,3-6.16b-f. Kahl’s (Agreements [n. 5], p. 50) disagreement is therefore not only with the 2DH theorists whom he accuses of “an arbitrary pre-decision” in limiting the MAs to “a few words” and attributing

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

145

and/or Luke may have expanded a Markan passage should be excluded from the TT68. Once again, we see this frequently done independently of the modern discussion by Eusebius69. Third, when Matthew and Luke feature a doublet of a Markan text, this doublet, provided that it exhibits closer contact between Matthew and Luke than between the synoptist who features it and Mark, should be classified as DT70. Robert K. MacEwen’s recent statistics rely on his decision to “[r]egard all forms of ‘Mark-Q overlap’ sayings as TA [triply attested] material”71. On this approach, all that matters is whether a text has a Markan parallel; the phenomenon of doublets is of secondary relevance. But this runs the risk of rendering irrelevant whether a text is attested in the synoptic tradition three, four, or five times72. Eusebius, who did not operate on the presupposition of the synoptists utilizing each other’s work, already either affirms what modern scholarship sees as the DT or categorizes it as Sondergut in a number of Mark-Q overlaps73. Fourth, on Markan priority, the material found in a Mark-Q (MarkMatthew [FH]; Mark-Luke [MPH]) overlap that lacks a DT doublet should also be considered DT74. Here, it is significant that the contexts in which Matt 11,10//Luke 7,27 (par. Mark 1,2); Matt 3,11//Luke 3,16b-f (par. Mark 1,7-8); Matt 12,24.25b-26//Luke 11,15.17b-18 (par. Mark 3,22-26); Matt 7,2//Luke 6,38 (par. Mark 4,24); Matt 5,13//Luke 14,34-35 (par. Mark 9,50) appear are either fully or at least partly non-Markan. These

longer agreements to Q. It is also with Eusebius, who already assigns Luke 3,1-2.7-9.17-18 to his equivalent of the DT. 68. E.g., in Luke 3,1-18: Eusebius’ equivalent of DT/Sondergut materials is 3,1-2.716a.17-18. Cf. supra, n. 67. 69. E.g., note a new section (§10) at what since Stephanus’ fourth edition is Luke 3,16(b-f), offsetting Eusebius’ equivalent of TT from the surrounding DT/Sondergut. Cf. supra, n. 67. 70. Several configurations are possible here. See infra, Section III (Table 1). The Markan text should be considered TT even if only preserved by Matthew or Luke. 71. Posteriority (n. 4), pp. 52-53 n. 89. 72. Infra, Section III (Table 1); for MacEwen all of these remain triply attested. See his Posteriority (n. 4), pp. 219-228 (Table 2.15, following MacEwen’s order; the superscripted number shows the actual number of a saying’s occurrences in the synoptic tradition): his TA pericopes No. 364.264[p. 221].474.405.265[p. 223].674.324.214.434. There is, factually speaking, not one triply attested saying on this list. Cf. my Overlaps (n. 5), pp. 234-235. 73. Infra, Section III (Table 1). 74. If the TT is defined as pericopes, the overlap texts lacking DT doublets become TT. Accordingly, all such Mark-Q overlaps can be found in F. NEIRYNCK, The Minor Agreements of Matthew and Luke against Mark, with a Cumulative List (BETL, 37), Leuven, Leuven University Press – Peeters, 1974 (here cited: MA74), and ID., The Minor Agreements in a Horizontal-Line Synopsis (SNTA, 15), Leuven, Leuven University Press – Peeters, 1991 (here cited: MA91). See infra, notes 227-229.

146

O. ANDREJEVS

contexts tend to be either DT or DT/Sondergut, as recognized already by Eusebius75. Accordingly, on Markan priority, it is difficult to avoid the impression that Matthew and Luke are utilizing in these instances a source other than Mark (namely, Q or the middle synoptist). Fifth, guided by its overarching framework, the FH may occasionally be inclined to identify a Luke-Mark parallel as Luke-Matthew. The following cases are not insignificant: Luke 3,2b-4; 11,21-22; 12,1; 17,2.3176. A Markanpriority-based neutral analysis (i.e., analysis neutral with respect to the Markan-priority-based theories) will yield a more reserved assessment77. Notably, the Eusebian apparatus tabulates the aforementioned texts analogously to the modern TT78. If Markan priority is the default neutral position of (the overwhelming majority of) the current discussion, then “in tripletradition passages where Matthew’s and Luke’s independent redactions provide a satisfactory explanation of their agreements against Mark there is no need to suggest the existence of a second [non-Markan] source”79. Far from being necessitated by the 2ST, the TT supplies this hypothesis and its current rivals with a shared neutral foundation. Similar to the DT and Sondergut, it is an ancient category of data, entirely independent of the modern hypotheses. Turning to the DT, we can now address the recently identified phenomenon of Strings of Verbatim Agreement (SVAs)80. MacEwen, Goodacre, and Garrow are among the scholars who propose that the higher SVAs in the DT are best explained in terms of a direct relationship, even as they obviously disagree on its editorial direction81. As summarized by Garrow, because Matthew and Luke “never achieve the same levels of compound 75. The Mark-Q overlap is underlined. (1) Luke 7,18-35 = Luke §69-73 (canons V [DT]; II [TT]; V [DT]; X [Sondergut]; V [DT]); (2) Luke 3,1-18 = Luke §6-11 (supra, n. 67); (3) Luke 11,14-32 = Luke §126-132 (canons V [DT]; II [TT]; V [DT]; II [TT]; V [DT]; X [Sondergut]; V [DT]); (4) Luke 6,32-39 = Luke §55-57 (canons V [DT]; II [TT]; V [DT]); (5) Luke 14,16–15,7 = Luke §181-187 (V [DT]; V [DT]; X [Sondergut]; V [DT]; II [TT]; II [TT]; V [DT]). The situation is similar in Matthew. The Beelzebul pericope, while situated by Matthew in the Markan sequence, nevertheless is surrounded by (and contains) enough DT material to suggest that MPH Matthew’s main source here was Luke. 76. E.g., WATSON, Writing (n. 3), pp. 186-187 (M26, M28, M40). 77. Supra, notes 26-28. 78. Canons I (TT+John) and II (TT): Luke §7 (8/2/7/10), §129 (122/33/129), §144 (164/79/144), §197 (179/99/197), §209/253 (248/143/209; 248/143/253). 79. NEIRYNCK, The Minor Agreements and Q (n. 28), p. 249. 80. The term first appeared in MacEwen’s Posteriority (n. 4) as SVA*4 (the lowest number of consecutive words that MacEwen studied was 4 – see his Table 2.15 on pp. 219-228). 81. Ibid., pp. 50-73; M. GOODACRE, Too Good to Be Q: High Verbatim Agreement in the Double Tradition, in POIRIER – PETERSON (eds.), Marcan Priority without Q (n. 2), 82-100; ID., Hypothesis (n. 3), pp. 51-53; GARROW, Streeter’s ‘Other’ (n. 4), pp. 212-213.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

147

faithfulness to Mark [as they do] to Q”, the DT passages with the high level of verbatim agreement “are most likely the product of direct, rather than indirect, copying”82. This argument requires closer attention. The participants in the discussion can agree that the DT mostly consists of the sayings of Jesus. For MacEwen, who anticipates this observation, the high SVAs in the DT are nevertheless “not satisfactorily explained” by this: “[t]here are plenty of sayings of Jesus in the TA [triply attested] material, yet Matthew and Luke are unable to achieve the highest levels of sustained, word-for-word agreement” in the TT83. MacEwen’s numbers may be bound to the text of Nestle-Aland a bit too closely for comfort84. But whatever one may think of counting word sequences in reconstructed texts, even if one limits the discussion to the NA28 SVAs, the high end of the TT SVAs is still loaded with Markan sayings material preserved by both Matthew and Luke85. Elsewhere I have highlighted that “omissions of Markan sayings” by both Matthew and Luke “are very rare”, and when they occur, “cannot be characterized as frivolous”86. The later synoptists “are more interested in the preservation of the Markan sayings of Jesus than Mark’s narrative”87. It is only in a number of Mark-Q overlaps (most of which are TT/DT doublets) that the Q (DT) “sayings of Jesus are favored by one or both of the evangelists over their Markan counterparts”88.

82. Ibid., p. 213. 83. MACEWEN, Posteriority (n. 4), pp. 62-63. 84. One may consider Mark 8,35 (ibid., p. 219, Table 2.15), with an NA28 SVA of 16 words in Matthew and Luke. Here, NA28 Luke 9,24 replacing NA28 Mark 8,35’s ἐάν with ἄν (effectively, removing one letter) is all that stands in the way of MacEwen’s new longest TA (=modified TT) SVA, adding six more words to the above total. If NA28 had gone with Luke 9,24 ‫ א‬C G S Y Γ Ψ (ἐάν), all three synoptists would read ἐάν and the resulting TT SVA would come in at 22 words: καὶ ἀκολουθείτω μοι. ὃς γὰρ ἐάν θέλῃ τὴν ψυχὴν αὐτοῦ σῶσαι ἀπολέσει αὐτήν· ὃς δ᾽ ἂν ἀπολέσῃ τὴν ψυχὴν αὐτοῦ ἕνεκεν ἐμου. Can one then really say, based on this one particle replaced on Markan priority by NA28 Luke (which may or may not reflect the reality of what happened), that the highest TT SVA lags substantially behind the highest DT SVA of 27 words? And how much does this particle (a difference of one letter character) actually mar the retention of the TT text by Luke in this instance? 85. GARROW, Streeter’s ‘Other’ (n. 4), p. 213: Mark 8,34–9,1; 13,28-32. Cf. MACEWEN, Posteriority (n. 4), p. 219 (Table 2.15). MacEwen (ibid., pp. 54-55, 219) lists Mark 1,40-42 as tied for the longest Matthew-Luke TA (=modified TT) SVA, with 18 words. Cf. supra, n. 84. GARROW, Streeter’s ‘Other’ (n. 4), p. 213, is interested in “compound faithfulness to Mark”, that is, in accumulations of SVA*4s. While his SVAs do not reach the peaks of MacEwen’s SVAs, they capture something that MacEwen’s data do not, namely, Matthew’s and Luke’s faithfulness to Mark’s sayings material below the SVA peak values. 86. Reconstructed Mark (n. 23), p. 86. 87. Ibid., p. 85. 88. Ibid., p. 87.

148

O. ANDREJEVS

The following picture emerges. The highest degree of verbatim agreement in the synoptic tradition is found in sayings material, both TT and DT. On the FH/MPH, the third synoptist encountered sayings material in Mark and the second synoptist. As indicated by the third synoptist’s selective utilization of the second one’s gospel, Mark’s antiquity was apparently taken seriously89. Yet, the third synoptist on the FH/MPH exhibits superior faithfulness to the sayings material recorded by the second synoptist. At the same time, the third synoptist’s utilization of the second one’s sayings material also features a high degree of variability90. The FH/MPH, then, both produce a third synoptist who, when it comes to the sayings of Jesus, which is the material most faithfully reproduced in the synoptic tradition, is neither impressed by the presumable antiquity of Mark nor remains consistently faithful to the second synoptist, while nevertheless favoring the latter over Mark in terms of the SVAs91. Perhaps, some will write this off as competitive textualization. For others, the increase in the SVA levels of sayings material from the TT to the DT may receive a more satisfactory explanation in 2DH Matthew’s and Luke’s dependence on Q, a document that appears to predate Mark92 and to have been “normative” for Matthew and Luke93. As a result, one can expect that it would have been edited by them less liberally. The instances of low SVAs in the DT would then derive naturally from the evangelists’ independent utilization of Q. Precisely here one can already see that excessive rigidity is not property of the 2DH: “less liberally” cannot imply with “the least possible independent creativity”94. *  *  * 89. This, however, does not demonstrate the third synoptist’s acquaintance with Mark’s gospel for a longer period of time (cf. supra, n. 53). One can acquire (or be introduced to) a newer book and a vintage book at the same time. 90. MACEWEN, Posteriority (n. 4), p. 218 (Table 2.14). 91. Hence, WOJTKOWIAK, Tradition (n. 2), p. 466: “Bei [die FH] wäre dieses Phänomen … konstanter zu erwarten”. One could say the same about the MPH. 92. This is the presupposition of the vast majority of 2DH scholars, emerging logically from the analysis of the reconstructed text. For a response to objections against the feasibility of Q’s reconstruction, see supra, n. 23. 93. KIRK, Q in Matthew (n. 2), p. 305. This would make even more sense if Q were a document (in possession) of the Jerusalem Church. S.J. JOSEPH, The Quest for the “Community” of Q: Mapping Q Within the Social, Scribal, and Textual Landscape(s) of Second Temple Judaism, in HTR 111 (2018) 90-114, makes a sustained argument against limiting the precise geographical location of Q to Galilee (see now also his contribution in this volume). Joseph calls attention to Q’s Jewishness and explores the available possibilities, which include Jerusalem. He finds “the idea that social and scribal networks in Jerusalem served to support and inform the earliest post-30 Jesus movement” to be “an economical and explanatorily powerful hypothesis” (ibid., p. 111). One may recall that WEISS, Einleitung2 (n. 15), rather logically thought of Q as an “Apostelschrift” (e.g., pp. 497-500). 94. Cf. supra, n. 30.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

149

Turning to the subsidiary categories, virtually all of the 2DH’s critics campaign heavily on the MAs, for which various statistics have been given95, as the 2DH’s “fundamental problem”96. Some call into question the category itself, with Kahl reminding us that the agreements “are not few, insignificant or occasional”97. He does not forget about the other warhorse of the 2DH’s critics: the segregation of the MAs from the DT98. One can sympathize with the criticism of the term; some of these agreements are certainly not “minor” in the sense of being “insignificant”99. The objection to the category’s existence, however, needs to be understood for what it is: a rejection of the TT’s validity. The MAs are “exclusively a triple-tradition phenomenon”100. Kahl’s characterization of the distinction between the MAs and the DT as “arbitrary”101 goes hand-in-hand with his rejection of the “so-called” TT which, as we saw earlier, he thinks “has been necessitated by the 2ST”102. It is for this reason that above we discussed the antiquity of the TT and updated its definition to reflect Markan-priority-based source utilization. The same author rather casually dismisses two of the important explanations of the phenomenon, referring to “coincidence” and “text critical references”103. He believes that these “strategies” are “invalidated by the clear evidence of the large number of MA of a high quality”104, a monologic assessment that relies on the pre-judged significance of the overall number 95. E.g., M. MCLOUGHLIN, Using the Minor Agreements, in ETL 94 (2018) 93-114, p. 97 (about 350); EVE, Relating (n. 3), p. 79 (at least 750). 96. KAHL, Agreements (n. 5), p. 45. 97. Ibid., p. 47. 98. Ibid., pp. 50-52. Cf. GOODACRE, Overlaps (n. 5), p. 210. 99. Kahl’s (Agreements [n. 5], p. 52) complaint that the MAs “can never exceed their status as minor and insignificant” is puzzling, however, in light of Neirynck’s frequent and thorough analysis of the significant MAs and explicit rebuke of scholars (Theory [n. 31], p. 6) who consider the MAs “a minor and neglectable problem”. 100. ANDREJEVS, Overlaps (n. 5), p. 237. Elided in Goodacre’s (Overlaps [n. 5], p. 210) “continuum” of agreements is that in the MAs Matthew and Luke work primarily with Mark rather than with each other. To emphasize this is not to insist on evaluating the TT and DT phenomena separately for the sake of the categories. It is, rather, to take seriously the implications of the TT on Markan priority. The MAs are Matthew-Luke agreements that occur under conditions other than direct utilization (unless one argues that FH Luke or MPH Matthew utilized multiple sources, Mark and the middle synoptist, simultaneously). See infra, n. 102. 101. Luke (n. 13), p. 236. 102. Agreements (n. 5), p. 47. The question of utilization retains a pivotal role as well. If Kahl presupposes that FH Luke utilized Mark and Matthew simultaneously (with the help of scribes; utilizing wax tablets; laying the scrolls side by side; etc.), he can object to the separate study of the MAs. But this presupposition should be stated clearly. See supra, n. 100. 103. Ibid., p. 48. 104. Ibid., p. 49.

150

O. ANDREJEVS

of agreements. He references Neirynck’s contribution to the 1991 Göttingen symposium “Die Minor Agreements in den Synoptischen Evangelien”105 in a footnote under “coincidence”106. There is, however, no further engagement. Whether or not Kahl is happy with Neirynck’s refutation of the arguments he now champions as “academically sound”107, does he actually contest “that some of these agreements can be explained by independent redaction on the basis of the text of Mark”108? Since no engagement with this line of inquiry is found, one surmises that Kahl’s methodology rests on the Goulder-Goodacre canon: “the theory of Matthew’s and Luke’s independent use of Mark is untenable given the presence of the minor agreements”109. But this assessment is predicated on raw unprocessed data. Since Kahl’s essay conveniently summarizes the current treatment of the MAs by the vast majority of the FH/MPH theorists, two further observations may be useful. My first remark regards what Kahl characterizes as “coincidence”. In fact, this item includes a number of “strategies”, among them the study of Matthew’s and Luke’s redactional tendencies such as (a) stylistic characteristics110; (b) tendency to correct “the frequent theological and literary inelegancies of Mark” (=Mk instigator)111; (c) awareness of the nearby Markan context112; and (d) awareness of remote Markan contexts (=expertise in Mark, which may have suggested to Matthew and Luke words and phrases from remote Markan episodes appropriate to their rewriting of the episode in question)113. All four types of explanations can 105. NEIRYNCK, Theory (n. 31). 106. Agreements (n. 5), p. 48 and n. 13. 107. Ibid., p. 65. 108. NEIRYNCK, Theory (n. 31), p. 5. 109. GOODACRE, Case (n. 8), p. 167. 110. E.g., J.C. HAWKINS, Horae Synopticae: Contributions to the Study of the Synoptic Problem, Oxford, Clarendon, 21909, p. 209; B. DE SOLAGES, La composition des évangiles de Luc et de Matthieu et leurs sources, Leiden, Brill, 1973, pp. 102-152; W.D. DAVIES – D.C. ALLISON, A Critical and Exegetical Commentary on the Gospel according to Saint Matthew (ICC), Edinburgh, T&T Clark, 1988-97, vol. 1, pp. 112-113 and throughout the commentary; NEIRYNCK, Hypothesis (n. 9), p. 350, throughout the Evangelica I-III volumes, and in MA74 (n. 74), pp. 199-288; FRIEDRICHSEN, Agreements (n. 17), pp. 207-209; J.S. KLOPPENBORG, Excavating Q: The History and Setting of the Sayings Gospel, Minneapolis, MN, Fortress, 2000, p. 35. 111. MCLOUGHLIN, Using (n. 95), p. 94. Cf. HAWKINS, Horae Synopticae (n. 110), pp. 117-125, 209; B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, pp. 295-305; NEIRYNCK, Hypothesis (n. 9), p. 350; FRIEDRICHSEN, Agreements (n. 17), p. 207; KLOPPENBORG, Excavating (n. 110), p. 34. 112. E.g., NEIRYNCK, The Matthew-Luke Agreements in Mt 14,13-14 / Lk 9,10-11 (par. Mk 6,30-34): A Response to M.-É. Boismard, in ID., Evangelica II (n. 27), 75-94, pp. 85-88; ID., Theory (n. 31), pp. 24-27. 113. E.g., ibid., pp. 12-16; ID., Mt 12,25a / Lc 11,17a (n. 27); ID., Hypothesis (n. 9), p. 351 (“in light of the whole gospel”); MCLOUGHLIN, Listing (n. 18), p. 208.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

151

be tested and form different aspects of the independent redaction hypothesis114. Even before they are supplemented with text-critical solutions (only “in a few exceptional cases”)115, these “strategies” can account for a large number of MAs of a high quality116. It is certainly possible also to argue for the utility of secondary orality117. My preference would be to retain these supplemental strategies in a strictly auxiliary role, similar to the FH/MPH’s additional sources and/or traditions118. It is worth noting that the few MAs that demand untestable explanations can “stand in the text as a service” to the 2DH, providing falsifiability and reminding us that history consists of “unrepeatable particularities”119. Of course, this only works if these MAs are few, an assessment whose verification requires engagement with the independent redaction hypothesis. My second remark therefore circles back to Kahl’s aforementioned appeal to a “large number of MA of a high quality”120, apparently intended as incontrovertible proof of the FH’s case. It is unclear how many MAs Kahl’s “large number” actually represents. Does he allow the existence of MAs of a low quality, and if so, how many? It has been recognized for some time 114. ALKIER, Neues Testament (n. 3), p. 126; GOODACRE, Hypothesis (n. 3), p. 56; SIEGERT – WITTKOWSKY, Hypothese (n. 3), pp. 522-523; and KLINGHARDT, Gospel (n. 3), p. 9, bypass this hypothesis with all of its components. But is it not “the real enemy”? Cf. GOULDER, Luke (n. 54), p. 47. In his critique of the 2ST, Klinghardt assures the reader that, on the 2ST, the MAs “should not exist at all” (Gospel [n. 3], pp. 9, 188, 528); calls them “unresolved” (ibid., p. 20); and pronounces them “fatal” (ibid., pp. 188, 332). A perusal of Neirynck’s Evangelica I-III volumes and MA74 (n. 74), pp. 199-288, indicates that the rumors of the 2DH’s demise may have been exaggerated. If the MAs should not exist at all, then 2ST Matthew and Luke are not allowed to share any linguistic, stylistic, and theological preferences. Cf. DAVIES – ALLISON, Matthew (n. 110), p. 113: “[t]here are only so many ways to revise or correct a Greek sentence”. 115. MA74 (n. 74), p. 37. 116. N. PERRIN, The Limits of a Reconstructed Q, in M. GOODACRE – N. PERRIN (eds.), Questioning Q: A Multidimensional Critique, Downers Grove, IL, InterVarsity Press; London, SPCK, 2004, 71-88, p. 77, appears to misunderstand the independent redaction hypothesis, which does not attempt to explain all the MAs. It therefore does not press the argument “to the point of breaking”. E.g., NEIRYNCK, Hypothesis (n. 9), p. 351 (“not a reasonable expectation”); FRIEDRICHSEN, Agreements (n. 17), p. 213 (“[i]n addition to the coincidental redactions … [there] are auxiliary hypotheses”); MCLOUGHLIN, Listing (n. 18), p. 203. 117. See the definition in A. GREGORY, What is Literary Dependence?, in FOSTER et al. (eds.), Studies (n. 2), 87-114, pp. 91-92. I distinguish this from a more vague appeal to shared oral tradition. E.g., MONAGHAN, Tradition (n. 2). His No. 5 on p. 6 can easily be modified to secondary orality, which seems to me more probable. Cf. supra, n. 57. 118. E.g., MACEWEN, Posteriority (n. 4), pp. 130, 135; GARROW, Streeter’s ‘Other’ (n. 4), pp. 223-224, 226; GOODACRE, Hypothesis (n. 3), pp. 64-66; EVE, Relating (n. 3), p. 103. See P. FOSTER, Is It Possible to Dispense with Q?, in NT 45 (2003) 313-337, pp. 321-322; ARNAL, Synoptic Problem (n. 57), p. 426. 119. S.D. BLACK, One Really Striking Minor Agreement: TIΣ EΣTIN O ΠAIΣAΣ ΣE in Matthew 26:68 and Luke 22:64, in NT 52 (2010) 313-333, p. 333. 120. Agreements (n. 5), p. 49.

152

O. ANDREJEVS

that some MAs are more significant than others121. I do not suppose that Kahl or other critics of the 2DH can accept that to adopt Kahl’s solution is to “set the pyramide on its apex”122. Neirynck’s vintage observation may therefore still serve as a noteworthy reminder: The word “atomization” has been used with reference to the various explanations of the agreements, but there is also the atomization of the evidence by concentrating on one passage and collecting all sorts of agreements without studying each type of agreement together with all other similar changes of Mark elsewhere in the Gospel123.

Because a proper discussion of the MAs cannot be conducted in the abstract, I would like to concentrate on the examples of MAs handpicked by three recent critics of the 2DH: Kahl (2016), Eve (2021), and Klinghardt (2021)124. Below, I reproduce Kahl’s clusters of Exclusive Agreements of Matthew and Luke (EA)125, with the bold highlights and underline identifying the verses containing the significant MAs noted by Hawkins, DeWitt Burton, Lagrange, and de Solages (tabulated by Neirynck and more recently updated by McLoughlin [who adds de Solages])126. The first set of parentheses gives Kahl’s count of his EA clusters for each chapter of Mark. The second set of parentheses gives the count based on the number of clusters containing the significant MAs (Neirynck-McLoughlin, underlined, emboldened). Square brackets contain additions to Kahl’s list, when NeirynckMcLoughlin tabulate a significant MA omitted by Kahl.

121. E.g., HAWKINS, Horae Synopticae (n. 110), pp. 208-212; E. DEWITT BURTON, Some Principles of Literary Criticism and Their Application to the Synoptic Problem, in The Decennial Publications of the University of Chicago, vol. 1, Chicago, IL, University of Chicago Press, 1904, pp. 194-264 (= pp. 1-72); M.-J. LAGRANGE, Évangile selon saint Luc (EB), Paris, Gabalda, 1921, pp. LXX-LXXIII; B. DE SOLAGES, A Greek Synopsis of the Gospels: A New Way of Solving the Synoptic Problem, trans. J. Baissus, Leiden, Brill; Toulouse, Institut Catholique, 1959; DAVIES – ALLISON, Matthew (n. 110), vol. 1, pp. 109-114; NEIRYNCK, Theory (n. 31), pp. 10-12; ID., Hypothesis (n. 9), p. 351; FRIEDRICHSEN, Agreements (n. 17), p. 209; MCLOUGHLIN, Listing (n. 18); ID., Using (n. 95), p. 96; EVE, Relating (n. 3), p. 80. 122. DEWITT BURTON, Principles (n. 121), p. 245 (= p. 53). 123. Hypothesis (n. 9), p. 351. Cf. FRIEDRICHSEN, Agreements (n. 17), pp. 208-209; KLOPPENBORG, Excavating (n. 110), p. 35. 124. KAHL, Agreements (n. 5), p. 63 n. 35-37; EVE, Relating (n. 3), pp. 75-85; KLINGHARDT, Gospel (n. 3), pp. 240-251. 125. Agreements (n. 5), p. 63. Cf. EVE, Relating (n. 3), p. 77. These are essentially MAs under a more tendentious name. I retain the standard designation. 126. NEIRYNCK, Theory (n. 31), pp. 10-11; MA91 (n. 74), pp. 101-102; MCLOUGHLIN, Using (n. 95), p. 98. Cf. supra, n. 121.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

153

Mark 1 par.

vv. 4.6b-7a.7.8.10.12-13.[39].40-42

(7)

(1)

Mark 2 par.

vv. 3.10.12bis.16.[21].22.[23].24.26

(7)

(4)

Mark 3 par.

vv. [1].[7b].[10].14-16.[18].21b-22a 22.23.24.26.27b-28a.28b-29a.30b-31a.32

(10)

(4)

vv. 3-4.[10].11.15.20.21.22.30-31.32 32b-33a.35.37.38.41

(13)

(2)

Mark 5 par.

vv. 22-23.27.36

(3)

(2)

Mark 6 par.

vv. 7.10.11.14.32-33.34.36.38.43

(9)

(3)

Mark 7 par.

-----------

Mark 8 par.

vv. 11-12.14-15.29.31.35-36

(5)

(1)

Mark 9 par.

vv. 2-4.7.14-17.18-19.27.29.31.37.42b.49

(10)

(3)

Mark 10 par.

vv. 20-21.22.23.26-28.29.[30].34

(6)

(1)

Mark 11 par.

vv. 1-3.4.8-9.[11b].27-28a.29.31-32

(6)

(2)

Mark 12 par.

vv. 1.6-7.9-10.11-12.14.17.18-19.22-24 27b-28a.28-29.30.37

(12)

(0)

Mark 13 par.

vv. 1-2.3-5.11.15-16.19.21.23b.25-26.34

(9)

(1)

Mark 14 par.

vv. 10-11.20.35-36.37-38.45.47.54 61b-62bis.65.66-67.69-70.72

(12)

(5)

Mark 15 par.

vv. 1.2.12.15.22.[30].39.40.43.46

(9)

(3)

Mark 16 par.

vv. [1].5.6.8B

(2?)

(3)

Mark 4 par.

The emerging picture is quite different from what Kahl suggests, and it should be noted that even the significant MAs are not all of equal strength127. The majority of Kahl’s list (unhighlighted entries) does not feature significant examples but rather assembles all sorts of agreements. Eve now notes that some of Kahl’s clusters feature “collocation[s] of minimally impressive EA”128. The examples selected by Kahl to illustrate his position can be illuminating, given that none of them are on the Neirynck-McLoughlin list of significant MAs. Kahl’s choice of illustration for his category ‘a’ (Matt 27,58a// Luke 23,52a [par. Mark 15,43c] οὗτος προσελθὼν τῷ Πιλάτῳ) sets the tone129. To define the Markan subject is typical Matthean and Lukan redaction130; both are quite capable of supplying οὗτος in such scenarios 127. See already HAWKINS, Horae Synopticae (n. 110), pp. 210-211, Nos. 6.10.12.18. 128. EVE, Relating (n. 3), p. 80. 129. Agreements, p. 63, n. 35. 130. MA74 (n. 74), No. 22, pp. 261-266.

154

O. ANDREJEVS

independently131. The change of Mark’s εἰσῆλθεν to προσελθών can be attributed to Matthew’s and Luke’s predilection for the participle of προσέρχομαι: 28 of 51 occurrences of the verb προσέρχομαι in Matthew (14 redactional); 9 of 10 in Luke (6 redactional)132. Presumably, FH theorists would view this predilection in Luke 23,52 as provoked by Matt 27,58a. It remains unclear, however, why FH Luke is provoked by Matthew’s participle of προσέρχομαι at Mark 4,38; 5,27; 15,43 but always refuses it in the DT when he actually works with Matthew133. We may designate this phenomenon ‘unsustained MAs’134. These are MAs whose linguistic influence is either entirely or in the vast majority of cases not sustained in the DT, which are precisely the situations where the middle synoptist would presumably have exerted the strongest influence on the third one. In Luke 23,52a, this influence was apparently congenial to FH Luke and dovetails neatly with his own habit of introducing the participle of προσέρχομαι into the Markan narrative. What about the DT? While some degree of rewriting could be expected, universal omission is curious (including three instances in the span of five verses; Matt 25,20.22.24). For his category ‘b’, Kahl gives three examples135. The first one is Matt 12,47-48a//Luke 8,20-21 (par. Mark 3,32-33a). Here one finds the ubiquitous correction of a historic present136 and the change of Mark’s στήκοντες to ἑστήκασιν. Matt 12,47, however, is probably a harmonizing addition137. With its subtraction, the MA featuring ἵστημι + ἔξω becomes somewhat less impressive138. Kahl’s second example focuses on Matthew’s 131. Matt 12,24; 27,47; Luke 8,13; 9,24; 22,59; 23,22. Luke uses οὗτος as a nominative subject for Joseph of Arimathea twice, in 23,51 and 23,52; cf. Acts 1,18; 7,19.36; 10,6; 13,7; 14,9; 18,25-26 (the occasional questioning of Luke’s authorship of Acts remains, in my view, bereft of demonstration). Accordingly, if one still deems the coincidental use of οὗτος unlikely, the argument is stronger for the MPH. But then, it becomes necessary to explain MPH Matthew’s omission of Luke 23,51 (one of MPH Matthew’s many omissions of uniquely Lukan additions to the passion narrative). 132. NEIRYNCK, Theory (n. 31), p. 14. Redactional: Matt 8,2.25; 9,20; 13,10; 14,12; 15,12; 16,1; 17,19; 19,16; 26,50.60bis.73; 27,58; Luke 8,24.44; 9,42; 20,27; 23,36.52. 133. Participle of προσέρχομαι: Matt 4,3 (diff. Luke 4,3); 8,19 (diff. Luke 9,57); 25,20.22.24 (diff. Luke 19,16.18.20), a 0:5 ratio for FH Luke. See the Appendix to this essay (Table 2). 134. To the best of my knowledge, this phenomenon has not been discussed before. 135. Agreements, p. 63, n. 36. 136. MA74 (n. 74), No. 10, pp. 224-228. 137. Matt 12,47 omitted: ‫ *א‬B L Γ ff1 k sys.c sa. KLINGHARDT, Gospel (n. 3), p. 241, seems to agree. Cf. J. WEISS, Das Matthäus-Evangelium (Die Schriften des Neuen Testaments, 1), Göttingen, Vandenhoeck & Ruprecht, 1906-1907, 212-378, p. 308; G.S. PAULSON, Scribal Habits and Singular Readings in Codex Sinaiticus, Vaticanus, Ephraemi, Bezae, and Washingtonianus in the Gospel of Matthew (GHDS, 5), Wilmore, KY, GlossaHouse, 2018, p. 36, n. 52. 138. Matt 12,46 (εἱστήκεισαν) // Luke 8,20 (ἑστήκασιν).

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

155

and Luke’s coincidental use of λέγων and ἔφη139. Here, the introduction of the participle of λέγω before direct discourse illustrates another characteristic feature of Lukan and Matthean redaction140. Less expected is the near-universal omission of the formula in the DT, now both by FH Luke141 and MPH Matthew142. One may thus distinguish between the singly and doubly unsustained MAs (that is, observed in one or both later synoptists). Singly unsustained MAs are here illustrated by FH Luke’s rejection of φημί in the DT143. The quest to explain Luke’s agreement with Matt 27,11 probably needs to look no further than Luke’s own nearby context (Luke 22,70; diff. Matt 26,64). On his part, when Matthew uses φημί, in 15 of 16 (possibly, 17) instances it is to introduce a response to a previous statement144. 5 (possibly, 6)145 of these instances are redactional, rendering a source-critical explanation at Matt 27,11 unnecessary. The bigger picture, once again, is that yet another example on Kahl’s list is compromised by the reluctance of FH Luke (and, in one case, MPH Matthew) to sustain these presumably congenial TT influences in the DT. Kahl’s third example in his second category146 features an accumulation of incomplete MAs (Matt θέλετε; ἀπολύσω; λέγουσιν // Luke θέλων; ἀπολῦσαι; λέγοντες), matching names (τὸν Βαραββᾶν; Ἰησοῦν), and a word order change (ἐποίησεν κακὸν to κακὸν ἐποίησεν). Matt 27,21’s θέλετε … ἀπολύσω can be traced back without difficulty to Matt 27,17//Mark 15,9 (θέλετε ἀπολύσω); Luke 23,20’s θέλων ἀπολῦσαι to Mark 15,9. The overlapping use of key names from the Markan episode (τὸν Βαραββᾶν; [τὸν] Ἰησοῦν) is hardly surprising and their context / sequence matches Mark’s147. 139. Matt 27,11//Luke 23,3 (par. Mark 15,2). 140. MA74, No. 15, pp. 246-249; F. NEIRYNCK, Minor Agreements Matthew – Luke in the Transfiguration Story, in ID., Evangelica (n. 18), 797-810, p. 808, n. 58. 141. Participle of λέγω before direct discourse: Matt 5,2 (diff. Luke 6,20); 6,31 (diff. Luke 12,29); 7,21 (diff. Luke 6,46); 8,6 (diff. Luke 7,3); 9,33 (diff. Luke 11,14); 10,7 (diff. Luke 10,9); 12,38 (diff. Luke 11,16); 22,1 (diff. Luke 14,16a); 25,11 (par. Luke 13,25); 25,20 (par. Luke 19,16), a 2:8 ratio for FH Luke. See the Appendix to this essay (Table 3). 142. Participle of λέγω before direct discourse: Luke 7,6 (diff. Matt 8,8); 7,19 (diff. Matt 11,3); 13,25 (par. Matt 25,11); 13,27 (diff. Matt 7,23); 17,4 (diff. Matt 18,21-22); 19,16 (par. Matt 25,20); 19,18.20 (diff. Matt 25,22.24), a 2:6 ratio for MPH Matthew. See the Appendix to this essay (Table 3). 143. φημί: Matt 4,7 (diff. Luke 4,12); 8,8 (diff. Luke 7,6); 25,21.23 (diff. Luke 19,17.19), a 0:4 ratio for FH Luke. See the Appendix to this essay (Table 4). 144. Matt 4,7; 8,8; 13,28.29; 14,8*; 17,26; 19,21*; 21,27*; 22,37*; 25,21.23; 26,34*; 27,11.23.65 (asterisk indicates Matthew’s redaction of Mark). If Matt 19,18 is included (𝔓71 B f 13), another redactional instance can be added. The exception: Matt 26,61. 145. Matt 19,18 (𝔓71 B f 13). 146. Matt 27,21-23a//Luke 23,(17)18-22a (par. Mark 15,12-14a). 147. τὸν Βαραββᾶν: Matt 27,21//Luke 23,18 (par. Mark 15,11). (τὸν) Ἰησοῦν: Matt 27,22//Luke 23,20 (par. Mark 15,12).

156

O. ANDREJEVS

Of the five examples handpicked by Kahl, this particular accumulation of MAs (Mark 15,12-14a) is instructive in illustrating how consideration of the later synoptists’ utilization of Mark – specifically, their awareness of the nearby Markan context – can alone sometimes be sufficient to account for a large portion of the agreements. The word order change (verb-object in Mark to object-verb in Matthew/Luke) is not so unusual for either evangelist as to require a source-critical solution148. Kahl’s fifth and final example represents his category ‘c’ (Matt 16,21b// Luke 9,22b [par. Mark 8,31b])149. Eve references this “significant” example as well, but grants that Goulder150 may have “overestimated [underestimated?] the odds of its occurring by chance”151, with a reference to Neirynck and Friedrichsen152. We may likewise refer to these publications, since they have not yet received a response, and address Eve’s other examples. He follows Goodacre in revisiting the latter’s six single-word MAs, characterized as “certainly significant”153. Compare Neirynck’s assessment of the same list in 1997154: “can [Goodacre] really call them ‘striking’ examples?”155. This time, no reference to Neirynck’s analysis is given. Two decades later, it therefore remains “strange that statements such as ‘foreign to Luke’s style’ continue to be quoted … without informing the reader how they can and have been answered”156. Eve’s next “particularly striking” example is in the well-known cluster of MAs in Matt 9,1-8//Luke 5,17-26 (par. Mark 2,1-12)157.

148. MA74 (n. 74), No. 21A, pp. 257-259, esp. both later synoptists at Mark 1,8; 8,12.36; 11,17; 15,15 and the Lukan instances on pp. 258-259. For some curiously unsustained MAs, see the Appendix to this essay (Tables 5 and 6). 149. Agreements, p. 63 n. 37. 150. GOULDER, Luke (n. 54), pp. 48-50. 151. Relating (n. 3), p. 84. 152. F. NEIRYNCK – T.A. FRIEDRICHSEN, Note on Luke 9,22: A Response to M. D. Goulder, in NEIRYNCK, Evangelica II (n. 27), 43-48; F. NEIRYNCK, Luke 9,22 and 10,25-28: The Case for Independent Redaction, in ID., Evangelica III (n. 9), 295-306, pp. 295-300. Is this example still “significant”, then? 153. EVE, Relating (n. 3), p. 81. 154. (1) Matt 12,15//Luke 6,19 (par. Mark 3,10); (2) Matt 13,54//Luke 4,15-16 (par. Mark 6,2); (3) Matt 14,13//Luke 9,11 (par. Mark 6,33); (4) Matt 8,27//Luke 8,25 (par. Mark 4,41); (5) Matt 22,27//Luke 20,32 (par. Mark 12,22); (6) Matt 26,47//Luke 22,47 (par. Mark 14,43) (listed in the order of their discussion by Neirynck, with the significant MA emboldened). Cf. F. NEIRYNCK, Goulder and the Minor Agreements, in ID., Evangelica III (n. 9), 307-318. For No. 3, see also his Theory (n. 31), pp. 29-34; Matthew-Luke Agreements (n. 112). For No. 5, cf. HÄFNER, Matthäus-Evangelium (n. 2), p. 52. Unsustained MA: (No. 5) ὕστερον in Matt 4,2 (diff. Luke 4,2); 25,11 (diff. Luke 13,25), a 0:2 ratio for FH Luke. 155. Goulder (n. 152), p. 308. 156. Ibid., p. 311, n. 8. 157. Relating (n. 3), p. 83, n. 38. Cf. KLINGHARDT, Gospel (n. 3), pp. 243-244.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

157

But here too Neirynck awaits a rejoinder, now for almost fifty years158. As could be expected, Eve also references τίς ἐστιν ὁ παίσας σε (Matt 26,68// Luke 22,64 [par. Mark 14,65]) and ἐξελθὼν ἔξω ἔκλαυσεν πικρῶς (Matt 26,75//Luke 22,62 [par. Mark 14,72])159, both significant MAs160. It is not clear, however, that these MAs are a slam dunk for the FH/MPH, at least if a certain amount of awkwardness in Matt 26,67-68 and Luke 22,61-63 is recognized161. Eve closes his discussion by stating that the FH succeeds in “many such” MAs where “the 2DH struggles to offer any explanation at all”162. But is it a real ‘success’ if a number of the 2DH theorists’ arguments have simply been ignored? This leaves us with Klinghardt’s recent list of fourteen MAs (and MA clusters), divided into texts derived by (a) Luke and *Matthew from *Ev; (b) Luke from *Matthew163. I have already referenced four of his examples (of which two are significant MAs)164. Another example (Luke 13,18f.) is a Mark-Q overlap and I address these at the end of this section. Luke 9,7 (ὁ τετραάρχης), a significant MA, finds parallels in Luke 3,1.19165. It is not clear on what basis one should deny Matthew and Luke independent 158. Les accords mineurs et la rédaction des évangiles: L’épisode du paralytique, in NEIRYNCK, Evangelica (n. 18), 781-796. Cf. HÄFNER, Matthäus-Evangelium (n. 2), pp. 52-53. KLINGHARDT, Gospel (n. 3), p. 243, asserts that at Luke 5,17-26 “the 2ST provides no satisfactory explanation”. He references Neirynck in n. 12, but does not engage with the arguments. 159. Relating (n. 3), pp. 78-79. Cf. KLINGHARDT, Gospel (n. 3), pp. 243, 245. 160. Supra, n. 126. 161. F. NEIRYNCK, TIΣ EΣTIN O ΠAIΣAΣ ΣE, in ID., Evangelica II (n. 27), 95-138 (see p. 109, n. 85 subsidiarily on Luke 22,62); BLACK, TIΣ EΣTIN (n. 119), pp. 329-330; HÄFNER, Matthäus-Evangelium (n. 2), pp. 53-54. Cf. WOJTKOWIAK, Tradition (n. 2), p. 247 (on Matt 26,75 and the MPH). 162. Relating (n. 3), p. 85. 163. Gospel (n. 3), pp. 240-251. Klinghardt’s list is not intended to be exhaustive (ibid., pp. 250-251). However, his examples are apparently all “problematic for the 2ST” (ibid.). In Klinghardt’s theory, *Matthew represents pre-canonical Matthew; *Ev represents the oldest gospel used by all four canonical gospels (essentially, a proto-Luke). 164. Supra, notes 136-138, 157-161. Of these, two MAs belong to Klinghardt’s group ‘a’ (derived by Matthew and Luke from *Ev; Mark changes *Ev), including τίς ἐστιν ὁ παίσας σε. The supposition that Mark (Klinghard: [pre-canonical] *Mark) deleted these words would be very difficult to demonstrate. The existence of this clarifying phrase (and its context) in Matthew and Luke has persuaded a clear majority of scholars (2DH, FH, MPH, and 2GH!) of its secondary character. It is but one of a number of MAs (e.g., MA91 [n. 74], §38.25 τὸ περισσεῦσαν[-ον] [D W περίσσευμα]; §51.3 το[ῦ] πρόσωπον[-ου] αὐτου; §54.4 μέλλει; §77.1 νομικός + [ἐκ]πειράζων) that present a problem for Klinghardt’s hypothesis of Markan posteriority to *Ev. If these MAs can be reasonably assessed as secondary to *Mark’s text (cf. ENNULAT, Minor Agreements [n. 18], pp. 178, 195, 207, 287, 378-381; FRIEDRICHSEN, Agreements (n. 17), p. 214), then the hypothesis of *Mark removing them based on their attestation in *Ev becomes a circular argument. 165. Gospel (n. 3), p. 244.

158

O. ANDREJEVS

awareness of Herod’s correct title166. The remaining eight examples are Luke 5,12-14; 6,1-5; 6,6-11; 8,22-25; 9,[41]-47; 19,36-40; 20,1-8; and 22,39-46 (these include a further eight significant MAs)167. It should be apparent that Klinghardt’s examples are significant MA-heavy. But does this emphasis accurately reflect how small the latter group is in comparison to the rest of the MAs? While Klinghardt references Neirynck’s Minor Agreements (1974), his interest appears to wane with the conclusion of the Cumulative List. Consulting the Classification of Stylistic Agreements located in the same study suggests that further discussion of the phenomenon remains warranted, in light of the explicability of most MAs168. It may be appropriate to revisit Neirynck’s quote: “our first hypothesis is the priority of Mark”169. The MAs constitute a phenomenon that has to do with the utilization of Mark’s gospel. Even a cursory examination of Matthew’s and Luke’s redactional and linguistic characteristics reveals that their independent redactions of Mark can explain the vast majority of the MAs without the help of subsidiary (text-critical; secondary orality) solutions. We can see this illustrated in Kahl’s five handpicked cases and all but the significant cases where Neirynck-McLoughlin, Eve, and Klinghardt overlap. Does the remainder of the MAs offer a sufficient basis for positing a direct Matthew-Luke relationship? Given that alternative solutions are available, the value of these remaining MAs should be assessed in light of the 2DH’s strengths and the direct dependence theories’ weaknesses. The 2DH’s critics have focused the discussion on the existence and concentrations of the MAs, while leaving these raw data rather under-examined. But assuming the significance of numbers is hardly a good reason to recoil 166. E.g., J.A. FITZMYER, The Gospel according to Luke I–IX (AB, 28), Garden City, NY, Doubleday, 1981, p. 758; DAVIES – ALLISON, Matthew (n. 110), vol. 2, p. 463; WOJTKOWIAK, Tradition (n. 2), pp. 216-217. The fact that Matt 14,9 subsequently uses βασιλεύς may convey interchangeability (contra, e.g., GOODACRE, Hypothesis [n. 3], pp. 50-51). Herod’s position, while technically that of a tetrarch, was functionally, i.e., in his local jurisdiction – as the characters in the story come to experience – akin to a king (e.g., he had the authority to take a life). The fact that his famous father also was one (Matt 2,1.3) further blurs the line between the titles. Cf. I. GOMÁ CIVIT, El evangelio según san Mateo. Volumen Segundo (14–28) (Comentario al Nuevo Testamento, 3), Madrid, Marova, 1976, p. 19: Mark uses the “lenguaje del pueblo”. Why could Matthew not have done the same after he had introduced Herod by his correct title? Cf. R.H. GUNDRY, Matthew, Grand Rapids, MI, Eerdmans, 21994, p. 285. 167. Luke 5,12bis; 6,1.4.6; 8,25; 9,41; 20,1. Cf. supra, n. 126. 168. MA74 (n. 74). One can reference a number of other publications not engaged by Klinghardt, most importantly, NEIRYNCK, Deuteromarcus (n. 18), pp. 773-774; ID., Theory (n. 31), pp. 16-23; SCHENK, Frage (n. 18), pp. 95-113; HÄFNER, Matthäus-Evangelium (n. 2), pp. 50-55. See now also (after the publication of Klinghardt’s Gospel) WOJTKOWIAK, Tradition (n. 2) and my A Source-Critical Analysis of the Lord’s Prayer, in ETL 96 (2020) 661-679, pp. 677-679. 169. Literary Criticism (n. 11), p. 88.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

159

from further study of the phenomenon170. It can lead to something of an echo chamber. As we have seen, it would be especially helpful to establish to what degree MPH Matthew and FH Luke sustain the influences presumably exhibited in the MAs when one of them utilizes the other in the DT. That is, it would be necessary to explain why writer X, if he is indirectly influenced by writer Y’s stylistic characteristics when utilizing writer Z, refuses those very characteristics when he utilizes writer Y directly. These data seem to suggest independence. Examining them should provide important context for the phenomenon of the MAs. At the very least, until such data are properly assessed, it seems reasonable to see them as compromising the Goulder-Goodacre canon and its emphasis on the existence and overall number of the MAs171. Finally, Kahl’s list, Eve’s discussion, and Klinghardt’s examples also incorporate some Mark-Q overlaps172. These texts have long been in the crosshairs of the 2DH’s critics. In 2018, Goodacre published an essay on the subject173. His intention, as before174, is to suggest that “Matthew and Luke [do not] only agree in ‘minor’ ways in the triple-tradition material”175. I may quote the following statement from my response: “Goodacre’s nomenclature [major agreements] only works when the category of Mark-Q overlaps is reduced to a few triple-tradition pericopae”176. Describing the overlap texts as major agreements therefore “does not adequately characterize the entire body of data encompassed by the overlaps in question”177. Moreover, it is possible to define the overlap category as in virtually its entirety consisting of DT material178. Viewed in this light, Goodacre’s position becomes akin to an explorer making a determination about an island’s shape while viewing its coastline from only one angle. To put it simply, the Mark-Q overlap category is closely related to the phenomenon of the TT/DT doublets and should be studied in conjunction with the latter, through the lens of source-utilization179. 170. NEIRYNCK, Theory (n. 31), p. 40: “[s]tatistical evaluation … is far less promising. In fact, it is hardly conceivable that the total number of explained agreements could become unexplainable”. Cf. ID., Hypothesis (n. 9), pp. 350-351. 171. GOODACRE, Case (n. 8), pp. 165-169. See the Appendix to this essay. 172. Supra, n. 124. 173. GOODACRE, Overlaps (n. 5). 174. Case (n. 8), pp. 163-165. 175. GOODACRE, Overlaps (n. 5), p. 209. 176. ANDREJEVS, Overlaps (n. 5), p. 238. Of course, it is also possible (and, I think, necessary) to further object to defining the TT as pericopes. Cf. supra, Section II. 177. Ibid., p. 238. Cf. MCLOUGHLIN, Listing (n. 18), pp. 209-210. 178. Cf. supra, Section II. 179. Cf. infra, Section III.

160

O. ANDREJEVS

Eve’s recent objections to the feasibility of Mark-Q overlaps are consistent with Goodacre’s: the category is “an ad hoc epicycle created to preserve the 2DH”180. But the overlaps referenced by Eve remain with the small group of texts explored by Goodacre181 and are similarly based on (an abridged reading of) Streeter182, whose opinion has been nuanced more recently183. Once again, then, the majority of the category as it has been defined in more recent studies is not analyzed. Eve also finds difficult “the need to postulate that Mark and Q happen to agree on including versions of just these passages” and “the extent of the verbal agreement between Mark and Q”184. This does not seem to be as much of a problem for Goodacre185. III. THE 2DH’S TRADITIONAL STRENGTHS IN RECENT DISCUSSION In the remaining portion of this essay, I would like to briefly revisit three phenomena that problematize the FH/MPH and/or favor the 2DH186. These phenomena have been established since the 19th century, perhaps indicating that the founding 2ST scholars were guided by more than just alleged ideological concerns187. Not surprisingly, all three phenomena have recently come under various degrees of scrutiny from the 2DH’s critics. 1. Alternating Primitivity Even if this phenomenon cannot by itself decisively demonstrate the independence of Matthew and Luke, it has been rightly referenced as favoring 180. EVE, Relating (n. 3), p. 76. Similarly, KLINGHARDT, Gospel (n. 3), p. 191. 181. EVE, Relating (n. 3), p. 76. Cf. GOODACRE, Overlaps (n. 5), pp. 203-204. 182. See the discussion in my Overlaps (n. 5), pp. 238-241. 183. Cf. R. LAUFEN, Die Doppelüberlieferungen der Logienquelle und des Markusevangeliums (BBB, 54), Bonn, Hanstein, 1980; H.T. FLEDDERMANN, Mark and Q: A Study of the Overlap Texts (BETL, 122), Leuven, Leuven University Press – Peeters, 1995; NEIRYNCK, Assessment (n. 27). The occasional tendency to engage the contemporary 2DH through Streeter seems outdated. How many 2DH scholars today engage the FH through A. Farrer and bypass the more recent nuancing of his ideas? A similar reservation, in fact, forms the basis of Goodacre’s (Case [n. 8], p. 14) objection against retaining “Goulder” in the name of the hypothesis (which thereby has now morphed into FH/Farrer Hypothesis as opposed to FGH/Farrer-Goulder Hypothesis in previous scholarship). 184. Relating (n. 3), p. 76. 185. GOODACRE, Overlaps (n. 5), pp. 208-209. 186. These, of course, are not the only arguments that support the 2DH. Other important phenomena include those of order (e.g., NEIRYNCK, Hypothesis [n. 9], p. 347; TUCKETT, Q and the History [n. 15], pp. 8-9; KLOPPENBORG, Excavating [n. 110], pp. 18-33; FLEDDERMANN, Q: A Reconstruction and Commentary [n. 9], pp. 46-54 [and the additional literature there]) and the duality in Mark (NEIRYNCK, Hypothesis [n. 9], pp. 348-349). 187. Cf. supra, n. 13.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

161

the 2DH188. The recent objections can be divided into two categories: those that turn to additional sources and all other objections. Goodacre and Garrow represent the former group. As noted earlier, I am sympathetic toward Goodacre’s attentiveness to oral tradition. However, his appeal to it to explain alternating primitivity189 scarcely renders the latter less of a favorable argument for the 2DH. At issue is not only that Goodacre must sacrifice his appeal to the FH’s superior simplicity, but also the literariness of the DT as demonstrated by a number of scholars190. Contrary to Goodacre’s assessment191, the argument from alternating primitivity is based not on a “narrowly literary” paradigm but on the degree to which the DT can be demonstrably correlated to written sources. It seems to me that to make a compelling case Goodacre must further develop his argumentation, closely engaging with the work already done on this subject. On his part, Garrow seeks to close the discussion before it really begins: “[i]nstances of Alternating Primitivity, even if very numerous, only suggest the use of a shared source or sources – the presence of which does not preclude the further possibility that Matthew also used Luke, or vice versa”192. I cannot of course object to the existence of shared sources, but even if there should only be one non-Markan shared source (a form of Q?), the MPH sacrifices its claim to superior simplicity. The Q source is then simply replaced by some other source(s) or tradition(s), whenever alternating primitivity happens to pose a difficulty. More recently, Garrow has upgraded his assessment of what has been accomplished. He now claims to have demonstrated “the irrelevance of Alternating Primitivity in relation to the question of direct contact between Matthew and Luke”193. But the necessity of multiple criteria hardly invalidates the relevance of any one criterion. Due consideration should be given to the key feature of the phenomenon: alternating primitivity is found in thirty-six of the thirty-eight 188. E.g., WEISS, Einleitung2 (n. 15), p. 486, n. 1; HOLTZMANN, Synoptiker3 (n. 15), p. 14; STREETER, The Four Gospels (n. 111), p. 183; DAVIES – ALLISON, Matthew (n. 110), vol. 1, p. 116; TUCKETT, Q and the History (n. 15), p. 10; FOSTER, Q (n. 118), pp. 321, 336; J.S. KLOPPENBORG, On Dispensing with Q? Goodacre on the Relation of Luke to Matthew, in NTS 49 (2003) 210-236, pp. 223-225; FLEDDERMANN, Q: A Reconstruction and Commentary (n. 9), pp. 60-65 (and the additional literature there); SCHERER, Coherence (n. 1), p. 199; KLINGHARDT, Gospel (n. 3), pp. 20, 189-190; ALLISON, Rewriting (n. 2), pp. 118-119. 189. Hypothesis (n. 3), pp. 64-66. 190. Cf. A. KIRK, The Composition of the Sayings Source: Genre, Synchrony, and Wisdom Redaction in Q (SupplNT, 91), Leiden – Boston, MA – Köln, Brill, 1998; ID., Orality, Writing, and Phantom Sources (n. 58); KLOPPENBORG, Oral and Literate Contexts (n. 1); ID., Variation (n. 58); SCHERER, Königsvolk und Kinder (n. 2), and see her contribution in this volume. 191. Hypothesis (n. 3), p. 64. 192. Streeter’s ‘Other’ (n. 4), p. 209 (emphasis mine). 193. Plausibility (n. 4), p. 131, n. 4 (emphasis mine).

162

O. ANDREJEVS

DT pericopes194. This does not mean that in some of the thirty-six DT pericopes primitivity favors Matthew while in the rest it favors Luke195. Rather, in each of the thirty-six pericopes there are components pulling in both directions196. A data-oriented discussion would be required to show that this is irrelevant. Turning now to authors who do not appeal to extra sources, another objection against alternating primitivity has recently been made by Alan Taylor Farnes197. Contra Farnes, however, the logic of the phenomenon is not limited to the application of lectio brevior. The argument from length is at best tertiary to the linguistic/stylistic and statistical considerations198. Finally, Eve’s claim that the alternating primitivity argument “must be able to demonstrate a number of cases where Luke’s transformation of Matthew would be positively implausible” essentially seeks to render communis opinio

194. FLEDDERMANN, Q: A Reconstruction and Commentary (n. 9), p. 64. 195. This would be the case in the two remaining DT pericopes. Ibid., n. 28. 196. Ibid., pp. 60-65 and throughout the commentary. For an illustration, see infra, n. 198. 197. Synoptic Problem (n. 7). 198. See the scholars listed supra, n. 188, e.g., ALLISON, Rewriting (n. 2), p. 119: “word statistics [and] redactional tendencies”. Cf. FLEDDERMANN, Q: A Reconstruction and Commentary (n. 9), pp. 60-65. These considerations are neglected by, e.g., ADAMCZEWSKI, Q (n. 4), p. 93. WATSON, Archaeology (n. 3), p. 50, now reduces Holtzmann’s (Synoptiker3 [n. 15], p. 14) argument to being “based … largely on omissions”. In fact, however, “modern Q advocates” can accept a good portion of “Holtzmann’s version of [the alternating primitivity] claim” (contra ibid.). HOLTZMANN, Synoptiker3 (n. 15), p. 14, references Luke’s μετεωρίζεσθε (12,29); (τὰ ἔθνη) τοῦ κόσμου (12,30); and πλήν (12,31) – all absent from Matt 6,31-33 – along with an argument in support of the greater primitivity of ravens (Luke 12,22 diff. Matt 6,26). Cf. ibid., pp. 221, 372. These observations have nothing to do with the potential Q origin of Luke 12,13-21 (presupposed by Holtzmann and engaged by Watson). They belong to the statistical and redactional argumentation. Cf., e.g., S. SCHULZ, Q: Die Spruchquelle der Evangelisten, Zürich, TVZ, 1972, pp. 149-152 (see the previous scholarship in his footnotes); J. JEREMIAS, Die Sprache des Lukasevangeliums: Redaktion und Tradition im Nicht-Markusstoff des dritten Evangeliums (KEK, Sonderband), Göttingen, Vandenhoeck & Ruprecht, 1980, pp. 209-210, 217-218; J.A. FITZMYER, The Gospel according to Luke X–XXIV (AB, 28A), Garden City, NY, Doubleday, 1985, pp. 978-980; J. GNILKA, Das Matthäusevangelium. Erster Teil: Kommentar zu Kap. 1,1–13,58 (HTKNT, 1/1), Freiburg i.B. – Basel – Wien, Herder, 1986, p. 246; A. SAND, Das Evangelium nach Matthäus (RNT), Regensburg, Pustet, 1986, p. 140; DAVIES – ALLISON, Matthew (n. 110), vol. 1, pp. 648, 650, 654, 657-658, 660; GUNDRY, Matthew2 (n. 166), pp. 116-119; CritEd (n. 25), pp. 334-353; FLEDDERMANN, Q: A Reconstruction and Commentary (n. 9), pp. 594-604, 894-896; J. NOLLAND, The Gospel of Matthew (NIGTC), Grand Rapids, MI, Eerdmans, 2005, p. 308; U. LUZ, Matthew 1–7 (Hermeneia), trans. James E. Crouch, Minneapolis, MN, Fortress, 2007, p. 339; A. DENAUX – R. CORSTJENS – H. MARDAGA, The Vocabulary of Luke: An Alphabetical Presentation and a Survey of Characteristic and Noteworthy Words and Word Groups in Luke’s Gospel (BiTS, 10), Leuven – Paris – Walpole, MA, Peeters, 2009, pp. 215, 268-269, 334-335, 395, 509; F. BOVON, Luke 2: A Commentary on the Gospel of Luke 9:51–19:27 (Hermeneia), trans. Donald S. Deer, Minneapolis, MN, Fortress, 2013, p. 211.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

163

irrelevant199. The fact that one cannot demonstrate something to be positively implausible does not, of course, negate the value of high or even relative implausibility. Hypotheses seldom win universal assent; a majority typically carries the day. Two of Eve’s cases – Luke 11,2-4 and 11,20 – illustrate the situation; they may indeed feature some dissenting voices where arguments have occasionally been made for Lukan redaction. But such voices remain – in both cases – in a sufficiently clear minority to indicate that their arguments have not yet been found plausible by the larger academic community. The onus probandi remains on said minority to persuade a majority of the guild of their arguments’ strength. 2. Rejection of Congenial Materials The second important phenomenon may be described as rejection of congenial materials by FH Luke and MPH Matthew. At the heart of this is the third synoptist’s (in twentieth-century discussion, mostly FH Luke’s) frequent omission of the second one’s additions to the text of Mark200. Some of these additions can be quite congenial to the third synoptist. This is now recognized by Eve (who quotes Craig A. Evans): “[a]nother problem for the [2GH and FH] is having to account for Luke’s omission of Matthean material that otherwise accords well with Luke’s theology”201. Eve acknowledges the insufficiency of Goodacre’s appeal to Luke-pleasingness202. His own solution is a return to Goulder’s ‘block policy’. On this solution, FH Luke works with one source at a time and stubbornly disregards the other source’s parallel accounts203. But why should FH Luke have adopted such a policy? There is also the following problem: the MAs call for FH Luke’s awareness of and willingness to use at least some of Matthew’s additions to Mark, breaking with the aforementioned policy204. It is, then, difficult to escape from “some type of a deliberate rejection of Matthew’s addition[s] to Mark”205. 199. Relating (n. 3), p. 126. 200. E.g., WEISS, Einleitung2 (n. 15), p. 541; FITZMYER, Luke I–IX (n. 166), pp. 73-74; DAVIES – ALLISON, Matthew (n. 110), vol. 1, p. 116; TUCKETT, Q and the History (n. 15), pp. 7-8; KLOPPENBORG, Dispensing (n. 188), p. 219; FOSTER, Q (n. 118), pp. 326-328, 336. More recently, including assessments of the MPH, see HÄFNER, Matthäus-Evangelium (n. 2), pp. 41-45; ANDREJEVS, Reception (n. 2). 201. EVANS, Hypothesis (n. 2), p. 44. Quoted in EVE, Relating (n. 3), p. 127. 202. Ibid. Cf. GOODACRE, Case (n. 8), p. 51. 203. Relating (n. 3), p. 127. Cf. GOULDER, Luke (n. 54), pp. 39, 44. 204. HÄFNER, Matthäus-Evangelium (n. 2), pp. 54-55; ANDREJEVS, Reception (n. 2), p. 236. 205. Ibid.

164

O. ANDREJEVS

The same phenomenon, in reverse, occurs on the MPH. Something like a ‘block policy’ could be proposed for the passion narrative, where MPH Matthew stubbornly steers clear of Luke’s substantive – and in many instances congenial to Matthew – additions to Mark. Except that, again, Matthew reproduces the MAs. If Kahl’s opinion represents the general position of the FH (and, adjusted, of the MPH), namely, that “Luke read the gospel of Mark critically through the Matthean re-lecture of it”206, rejection of congenial materials continues to pose a challenge. There is simply not much for FH Luke and MPH Matthew to be critical of when it comes to, for example, something like the second synoptist’s depiction of Peter and the apostles207. Nor is the phenomenon limited to omission of congenial episodes (or in-episode developments). As with alternating primitivity, the analysis must also engage with linguistic/ statistical considerations. This includes omission of congenial terminology, such as, for example, Matthew’s κύριε by FH Luke in contexts where this title would be fully compatible with Luke’s own usage208. The third synoptist’s unsustained MAs would certainly also be a part of the conversation209. Appealing to the sufficiency of the third synoptist’s prerogative (which is what the ‘block policy’ essentially is) to explain all of this hardly goes beyond redescribing the problem210. The question is not what we think the third synoptist must or must not have done. It is simply “whether the modern [Markan-priority-based] synoptic problem solutions have done proper justice to the evangelists’ visible redactional tendencies”211. 3. TT/DT Doublets This phenomenon has frequently been studied from a particular angle: the existence of doublets212. However, the TT/DT doublets (which mostly consist of sayings), also comprise the majority of the 2DH’s Mark-Q overlaps. Therefore, as noted earlier, the two phenomena are related. Moreover, 206. Luke (n. 13), p. 237. 207. For details, see WOJTKOWIAK, Tradition (n. 2), pp. 371-372 and my Reception (n. 2). 208. Ibid., p. 239. 209. Supra, Section II and see the Appendix to this essay. 210. Supra, n. 40. 211. ANDREJEVS, Reception (n. 2), p. 243. In Reception I did not engage the argument made by POIRIER, Composition (n. 8), p. 222. He recognizes the problem and attempts to explain FH Luke’s omission of Matt 16,16b-19. Whether or not his explanation holds up, he has numerous other passages to address, along with FH Luke’s omission of congenial terms and the phenomenon of unsustained MAs. 212. The centrality of the sayings doublets to the emergence of the 2ST is documented in VAN OYEN, Doublets (n. 9).

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

165

as shown in Table 1 below, a third phenomenon can be brought into the discussion as well: FH Luke’s so-called unpicking of his Matthean source, separating Matthew’s non-Markan material from its Markan context213. As a result, the TT/DT doublets should be viewed as a juncture of several key phenomena in the discussion. Our interest, then, lies with the doublets whose text is found in the DT, with a TT parallel in Matthew, Luke, or both214. Table 1 contains the TT/ DT doublets listed by Geert Van Oyen (Matthew and Luke), Wolfgang Grünstäudl (Luke, following Van Oyen’s numeration), and myself (Matthew and Luke)215. Van Oyen’s / Grünstäudl’s numeration forms the basis of the table and is found in column VO/G. My numeration is found in column A216. Van Oyen’s list is narrowed to the TT/DT doublets and occasionally adjusted217. The Matthean and Lukan TT/Mk columns contain the TT texts (based on the definition given earlier) including the Markan material omitted by the other evangelist. In Table 1, bold highlights indicate that the DT text is found in Eusebius’ canons III, V (DT), or X (Sondergut), with the Eusebian paragraph number superscripted and placed in square brackets (to distinguish from footnotes)218. 213. Unpicking is essentially the famous “crank” argument, influentially (if perhaps too strongly) stated by STREETER, The Four Gospels (n. 111), p. 183. This phenomenon is more complex than some of the recent discussion allows. For details, see my FH Luke’s ‘Unpicking’: Some Observations on Francis Watson’s Recent Analysis (2018) and the Extent of the Phenomenon, in JSNT 45 (2022) 3-22 (and the literature there). See also O. ANDREJEVS – D.B. SLOAN, Matthean Conflations and Luke’s Utilization Procedure under the Farrer Hypothesis: In Defense of F. Gerald Downing, in JBL 141 (2022) 737-759. 214. GRÜNSTÄUDL, Doublets (n. 9), p. 25, urges that “reduced lists of doublets that contain only supposed ‘source’ doublets” be avoided. However, the DT can be utilized to study a particular set of data – DT/TT doublets – without prejudging the included doublets’ redactional or source origin. 215. VAN OYEN, Doublets (n. 9), pp. 283-285; GRÜNSTÄUDL, Dubletten (n. 9), pp. 219221; ANDREJEVS, Overlaps (n. 5), pp. 232-235. 216. Ibid. 217. Van Oyen’s list “is mainly compiled on the basis of doublets referred to by Weisse (1838, 1856), Holtzmann (1863) and Wernle (1899)” (Doublets, p. 282, n. 20). A century later, it is possible to make some adjustments: Van Oyen’s No. 2 (TT/DT parallels adjusted); Nos. 14 and 44 (given their own “special case” sections [2 and 5 in my Table 1]); No. 17 (relocated to my section 1; parentheses around Luke 10,16 removed); No. 22 (relocated to my section 1; TT/DT parallels adjusted). Omissions: Van Oyen’s No. 6 (Mark 13,9-10.12-13 does not have a DT parallel); Nos. 8.11.13.15.20 (not doublets); Nos. 16.37-38.44.46 (no DT text); No. 35 (redundancy; see Van Oyen’s No. 6). Additions: Mark 9,40 = No. 19 in the A column (my section 4). 218. NA28, pp. 92*-94*. The significance of the canons III, V, and X is that these texts in Table 1 were not categorized by Eusebius as what a modern specialist would describe as TT. The significance of the canons II, IV, and VI is that in these Table 1 texts Eusebius separates the two halves of the doublet into his equivalent of the TT (or Mark+Matthew) and either DT or Sondergut.

166

O. ANDREJEVS

Asterisks indicate the eight synoptic doublets whose utilization by FH Luke involves unpicking of Matthew’s TT material219. In potential source doublets, the texts forming the doublet pair are derived from two separate sources220. In potential redactional doublets, one of the two texts forming the doublet pair can be seen as an evangelist’s creation221. Table 1 VO/G

A

Mark

Mt TT/Mk

Mt DT

Lk TT/Mk

Lk DT

1 TT + DT in Matthew and Luke 2DH: (potential) source doublets in Matthew and Luke FH/MPH: (potential) source doublets in the third, redactional or source doublets in the second synoptist

 1

22

4,25

 2

 9

8,11-12[77 IV; 78 VI] 16,1.4 [161 IV; 163 VI] 12,38-39[127 V; 128 V] (11,16) [128 V]

13,12

25,29

8,18

 3

23

8,34[85 II]

16,24[170 II]

 4

24

[85 II]

8,35

16,25

[170 II]

10,39

 5

25

8,38

16,27

10,33

17

26

9,37

18,5

10,40

9,48

10,16

 7

27

13,12[141 II]

10,21[88 II]

10,35-36[95 V]

21,16[251 II]

12,52-53[160 V]

22

28

13,35

24,42

24,44

21,36

12,40

10,38[96 V] [97 III]

9,23[96 II]

19,26 11,29[132 V] 14,27[182 V] *

9,24

17,33[211 III]

9,26

12,9

[96 II]

2 (special case) TT and DT conflated in Matthew and Luke; duplication of the DT text in Matthew 2DH: Matthew and Luke conflate Mark and Q; a redactional or source doublet in Matthew FH: Luke copies Matthew; a redactional or source doublet in Matthew MPH: Matthew copies Luke; a redactional doublet in Matthew

9,32-34a[75 X]

14 3,22c

9,34b [12,22] [119 V] 12,22-24a

3,22c-26

12,24b-26

[11,14] [126 V] 11,14-15a 11,15bc.17-18

219. Supra, n. 213. See esp. my Unpicking (n. 213), pp. 12-14 (Table 2). Cf. WATSON, Writing (n. 3), pp. 185-187; ID., Seven Theses (n. 35), pp. 143-144 and notes 6-8. 220. GRÜNSTÄUDL, Doublets (n. 9), p. 19, notes that “it is very difficult” to distinguish between redactional and source doublets “without implicitly presupposing a certain solution” to the SP. Accordingly, in Table 1 this distinction is explicitly coordinated with specific solutions to the SP. 221. Cf. FOSTER, Doublets (n. 2), pp. 112-116.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

VO/G

A

Mark

Mt TT/Mk

Mt DT

Lk TT/Mk

167 Lk DT

3 Markan parallel in Matthew only 2DH/MPH: (potential) source doublets in Matthew FH: redactional or source doublets in Matthew

19

10

10,11-12[105 II]

19,9[190 II]

5,32[37 X]

16,18[195 II] *

21

12

11,23[124 VI]

21,21[215 VI]

17,20[175 V]

17,6[200 V]

*

29

 8

3,28

12,31

12,32

12,10

*

32

11

10,31[111 II]

19,30[199 II]

20,16[200 X]

13,30[173 II]

34

13

11,24[125 IV]

21,22[216 IV]

7,7-8[53 V]

11,9-10[125 V]

24,26

17,23[204 II] *

36

14

[148 II]

13,21

[253 II]

[255 V]

24,23

4 Markan parallel in Luke only 2DH/FH: (potential) source doublets in Luke MPH: redactional or source doublets in Luke

39

15

4,21

5,15

8,16

11,33

40

16

4,22[40 II]

10,26[92 II]

8,17[80 II]

12,2[145 V]

10,9-14 [10,12-13] [84 V]

9,4-5

10,4-11 * [10,5-6][111 V] 11,23

41

17/18 6,8-11 19

9,40

12,30

9,50

42

20

12,38-39

23,6-7a

20,46

11,43

43

21

13,11

10,19-20

21,14-15

12,11-12

*

5 (special case) Markan parallel in Matthew and Luke; no DT on the 2DH FH: Luke copies Mark and Matthew MPH: Matthew copies Mark

44

13,15-16

24,17-18

21,21 (from Mark)

17,31 (FH) * (from Matt)

As noted by Van Oyen, “Matthew’s and Luke’s use of Mark and Q cannot be used [by the 2DH] as an interpretive key for all sayings doublets”; each synoptic doublet “should be put to the test of possible redactional duplication”222. However, if the FH/MPH theorists should adopt Goulder’s hypothesis of the middle synoptist (Matthew on the FH, Luke on the MPH) engaging in a substantial project of manufacturing doublets223, the redactional origin of the DT texts in that synoptist’s doublet pairs needs to be 222. Doublets (n. 9), p. 305. E.g., Luke 17,33 (cf. supra, n. 28). 223. GOULDER, Luke (n. 54), pp. 34-35.

168

O. ANDREJEVS

demonstrated224. Until such a demonstration becomes available, the FH/ MPH are left with the following identical argument: the second synoptist potentially inherited doublets of Markan sayings from some unspecified source(s), but the third synoptist inherited them from the second225. Furthermore, the third synoptist will have largely treated the second one as a sayings source (and that selectively), frequently tapping him for duplicate sayings, while mostly avoiding non-Markan narrative matter (occasionally with surgical precision, as the phenomenon of unpicking shows). This historical reconstruction hardly closes the door on the alternative presented by the 2DH, on which Matthew and Luke independently used a shared sayings collection that contained some materials overlapping with Mark. Our list of doublets (twenty-two) supports Paul Foster’s recent assessment: “perhaps more than twenty examples”226. Of these, the Eusebian apparatus recognizes what a modern specialist will describe as nine DT doublets227, in a further three instances assigns Matthew’s DT doublet to Sondergut228, and elsewhere recognizes a Lukan TT/DT doublet, marking Matthew as TT229. Even if the last example is omitted, in twelve of twentytwo instances the doublets are recognized as something other than the TT. Of course, modern scholarship is not bound to follow Eusebius’ conclusions. Nevertheless, one can see that already in Late Antiquity the TT classification of all synoptic doublets with a TT parallel, as now adopted for statistical purposes by MacEwen, was found unwarranted230. 224. GRÜNSTÄUDL, Doublets (n. 9), pp. 19, 25, observes that this procedure may be challenging. 225. Cf. FOSTER, Doublets (n. 2), p. 136. The problematic feature for the FH/MPH here is the presence of TT/DT doublets already in the middle synoptist. Explaining FH Luke’s or MPH Matthew’s treatment of the doublets does nothing to address this. The argument of NEIRYNCK, Hypothesis (n. 9), p. 352, which he directs against the 2GH, applies here: “questions must arise about the repetitions in Matthew” (one can substitute “Luke” on the MPH). Accordingly, the assessment of GRÜNSTÄUDL, Doublets (n. 9), p. 24, that the argument from doublets “is not to be considered an independent argument in favour of Q” remains, I think, somewhat premature. Perhaps, Q skeptics can find some refuge in KAHL, Luke (n. 13), p. 233 and n. 46: the “limitations of our possible understanding of Luke’s mind” apply “even more to Matthew’s mind” (Luke’s mind on the MPH)! 226. FOSTER, Doublets (n. 2), p. 136. 227. Eusebian canons III and V. Table 1 (VO/G): Nos. 2-4.7.14.21.34.36.41. For No. 36, see Eusebian §202 (Luke 17,21), which is a common point of disagreement. Cf. my Unpicking (n. 213), pp. 6-7, n. 13. 228. Eusebian canon X. Table 1 (VO/G): Nos. 14.19.32. 229. Table 1 (VO/G): No. 40. Eusebius recognizes DT material in Luke 12,2-8// Matt 10,27-32. Apparently guided by the absence of a doublet in Matthew, he places Matt 10,26 in canon II. But because Eusebius has already assigned Luke 8,17 to canon II (in parallel with Matt 10,26//Mark 4,22) he does not do the same at Luke 12,2. As a result, Luke 12,2, while assigned to canon V (DT), is left without a DT parallel. This is a clear illustration of a logic that runs counter to MacEwen’s policy (supra, n. 72). 230. Supra, n. 72.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

169

IV. CONCLUSION: OCKHAM’S RAZOR (AGAIN) I stop here to draw the conclusions. A key claim of the 2DH’s recent critics is the alleged superior economy of their solutions231. This claim undergirds the entire enterprise of the Mark-without-Q hypotheses. One can spot it lurking behind the calls to dispense with Q and the appeals to the cumulative number of the MAs. With a deft rhetorical touch, the claim to economy can help downplay the difficulties encountered in the third synoptist’s presumable utilization of the middle one: yes, it is not clear why Luke does X with Matthew (or vice versa), but at least we are dispensing with Q! But, as has been emphasized by a number of authors, “the simplicity of the theories of direct utilization of one gospel in another is only a fictitious one, while, in such a theory, hypotheses have to be multiplied to explain omissions, transpositions, semitisms, etc.”232. Earlier I noted that alternating primitivity can be particularly damaging to the Mark-without-Q hypotheses’ claims to superior economy. In addition to this, there are the curious compositional strategies produced by the SVAs and the TT/DT doublets. In all of this, the 2DH’s current crop of critics have yet to demonstrate that their theories require fewer untestable components233. Explanations may well come to an end somewhere234, but does it mean that a case has been demonstrated? It can also mean that a particular theory has run out of answers. It is quite possible that the SP will never fully be solved. Yet some hypotheses seem to endure while others fade into the pages of history. It is no secret that progress is paid for by repeated rigorous testing. The 2DH has been rightly subjected to just such testing since essentially its inception. Consequently, it seems only fair that the arguments from the testability and explanatory capacity of the FH/MPH (and, in earlier discussion, 2GH) should belong to the 2DH’s foundation235. In 2022, it is not an empty claim to observe that, despite over a century of critical scrutiny, the traditional strengths of the 2DH remain valid. They include its compatibility with the main categories of synoptic data and established utilization patterns 231. E.g., ALKIER, Neues Testament (n. 3), p. 127; NIELSEN, Introduction (n. 31), p. xxi; GOODACRE, Hypothesis (n. 3), p. 48; EVE, Relating (n. 3), p. 205; KLINGHARDT, Gospel (n. 3), p. 438. 232. F. NEIRYNCK, The Gospel of Matthew and Literary Criticism: A Critical Analysis of A. Gaboury’s Hypothesis, in ID., Evangelica (n. 18), 691-723, p. 722 (quoting A. GABOURY, La structure des évangiles synoptiques: La structure-type à l’origine des Synoptiques [SupplNT, 22], Leiden, Brill, 1970, p. 13). Cf. KIRK, Q in Matthew (n. 2), p. 308. 233. Supra, notes 35-38. 234. WATSON, Seven Theses (n. 35), p. 146: “explanations must come to an end somewhere”. On the contrary, one hopes that the discussion is only now hitting its stride! 235. Supra, notes 12 and 15.

170

O. ANDREJEVS

exhibited in antiquity; logical reconstruction of historical events; and capacity to be calibrated in a testable manner. As long as this alignment is maintained, the 2DH should retain an advantage over the current alternatives, offsetting the argument from Ockham if not reversing it in its favor. Appendix: Unsustained Minor Agreements In this appendix, I provide tabulation of the examples of unsustained MAs offered above. Due to the small sample size offered in this essay, the examples discussed in Section II of the essay mostly problematize the FH (Tables 2-4 below). Only one example from Section II also problematizes the MPH (Table 3). Accordingly, to show that the phenomenon is equally problematic for both Mark-withoutQ hypotheses, it seemed necessary to supply another example of unsustained MAs that could challenge the MPH. Such an example is offered below in Tables 5 and 6 (it is, in fact, equally problematic for both the FH and MPH). I hope to offer a more thorough tabulation and analysis of this phenomenon in a separate publication. In each of the following tables, the MAs (TT) and agreements in the DT are emboldened. With the exception of Table 3, the examples tabulated below were presented and analyzed at the 2022 General Meeting of the Catholic Biblical Association, in Santa Clara, CA236. Table 2: Participle of προσέρχομαι TT Matt

DT Matt

DT Luke

TT Luke

4,3  προσελθὼν 8,2  προσελθὼν 8,19  προσελθὼν 8,25 προσελθόντες 9,20 προσελθοῦσα

8,24  προσελθόντες 8,44 προσελθοῦσα 9,12  προσελθόντεςMK 6,35

13,10  προσελθόντες

236. Research report: “Unsustained Minor Agreements in the Synoptic Double Tradition” (Sunday, July 31).

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

TT Matt

DT Matt

DT Luke

171

TT Luke

19,16  προσελθὼν 20,27  προσελθόντες 25,20  προσελθὼν 25,22  προσελθὼν 25,24  προσελθὼν 26,49  προσελθὼνMK 14,45 26,73  προσελθόντες 27,58 προσελθὼν

23,52 προσελθὼν

Table 3: Participle of λέγω before direct discourse TT Matt

DT Matt

DT Luke

TT Luke

 1

3,2

 1

 2

3,3

 2

 3

3,17

 3

 4

5,2

 4

 5

6,31

 5

 6

7,21

 7

 6 13,27

MK 1,40

 7 MK 1,40

 8

8,2

5,12

 8

 9

8,3

5,13

 9

10

8,6

11

10 7,6

11

12

8,25

8,24

12

13

8,27

8,25

13

14

8,29

14

15

8,31MK 5,12

15

16

5,21

16

17

5,26

17

18

5,30

18

172

O. ANDREJEVS

TT Matt

DT Matt

DT Luke

TT Luke

19

9,14

19

20

9,18MK 5,23

20

21

9,29

21

22 23

9,33

22

10,7

24

10,5

23

24 25 26

7,19 12,10

27

25 14,3

26

12,38

27

28

13,3

28

29

13,31

29

30

14,15

30 MK 8,27

MK 8,27

31

16,13

9,18

31

32

17,5

9,35

32

33

17,15

9,38

34

18,1

33 34

35

17,4

35

36

18,16

36

37

18,18

37

MK 10,26

38

38

19,25

39

20,30 (9,27)

18,38

39

40

21,2

19,30

40

41

21,9

42 43

21,23 MK 11,31

44

21,25

45

21,37MK 12,6

41 42

20,2

43

MK 11,31

20,5

20,14

47

46

22,1 22,16 MK 12,18

49

22,24

50

22,31MK 12,26

51

44 45

46 48

19,38 19,46

47 20,21 MK 12,18

20,28

48 49 50

10,25

51

52

22,42

52

53

23,2

53

54

24,3

21,7

54

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

TT Matt

DT Matt

DT Luke

MK 13,6

55

TT Luke 21,8MK 13,6

24,5

173

55

56

25,11

13,25

56

57

25,20

19,16

57

19,18

58

58 59

19,20

60

26,8

61

26,17

59 7,39

60 61

62

22,19

62

63

26,27

22,20

63

64

26,39

22,42

64

22,67

65

65 66

26,68

67

26,69

68

26,70MK 14,68

22,64

69 27,11

70 71 72

66 67

22,57MK 14,68

68

22,59

69

23,3

70

23,21

71

27,23

72

73 27,54

74

23,35

73

23,47

74

Table 4: φημί TT Matt

DT Matt

DT Luke

TT Luke

4,7 8,8 19,21 21,27 22,37 25,21 25,23 7,44 26,34 22,58bis 27,11 27,23

23,3

174

O. ANDREJEVS

Note on Table 4: φημί occurs 25 times in Acts237, more than in any other NT document. There are 33 total instances of the verb in Luke-Acts, more than in any other NT author238. Unless the Lukan authorship of Acts is contested, this is a congenial term for Luke. Table 5: Word Order: Verb-Object (dir. and indir.) in Mark; Object-Verb in Matthew and/or Luke (TT); Object-Verb (dir. and indir.) in Matthew and/or Luke (DT) TT Matt  1

3,11A

 2

3,11B

 3

DT Matt

DT Luke

TT Luke 3,16

4,9

 4

 3 11,33

 5

 1  2  4

5,25

 5

 6

12,59

 7

6,27

 8

5,39

 9

5,40A

 6  7  8  9

10

6,29C

10

11

6,42

11

12

11,11

13

7,10bis

14

7,18bis

15 16 17

14 6,44

15

7,9

16

8,27

17

18

10,2

19

18

10,8

19

20

9,5

21

10,29

22

12,6

16,4

20 21

12,51

23 24

12 13

22 8,13

23

11,29

24

237. Acts 2,38; 7,2; 8,36; 10,28.30-31; 16,30.37; 17,22; 19,35; 21,37; 22,2.27-28; 23,5.17-18.35; 25,5.22.24; 26,1.24-25.32. 238. Matthew is in second place (16 occurrences); Paul in third (7 occurrences, in undisputed epistles only). There are 8 occurrences of φημί in Luke’s gospel: 7,40.44; 15,17; 22,58bis.70; 23,3.40.

THE TWO-DOCUMENT HYPOTHESIS AND ITS MAIN ALTERNATIVES IN 2022

TT Matt 25

DT Matt

DT Luke

16,26

26

TT Luke 9,25

25

9,35

26

19,46

27

28

20,11

28

29

20,13

29

30

20,22

30

27

21,13

175

31

20,44

32 33

32

22,30

34

23,39

34

31

11,39

33

35

(22,48)

35

36

27,23

23,22

36

37

27,26

23,25

37

Table 6: Simplified presentation of the DT data from Table 5  1

3,11A

 2

B

3,16

 1  2

3,11

 3

X

 Y

 3

 4

 Y

X

 4

 5

X

 Y

 5

 6

 Y

X

 6

 7

 Y

X

 7

 8

X

 Y

 8

 9

X

 Y

 9

10

 Y

X

10

11

 Y

X

11

12

 Y

X

12

13

Xbis

 Ybis

13

14

Xbis

 Ybis

14

15

 Y

X

15

 Y

X

16

16 17

8,27

17

18

 Y

X

18

19

X

 Y

19

20

9,5

20

176

O. ANDREJEVS

21

X

X

21

22

 Y

X

22

23 24

16,4

25

16,26

8,13

23

11,29

24

9,25

25

9,35

26

19,46

27

28

20,11

28

29

20,13

29

30

20,22

30

31

20,44

31

26 27

21,13

32

 Y

X

32

33

X

 Y

33

34

 Y

X

35

34 (22,48)

35

36

27,23

23,22

36

37

27,26

23,25

37

Notes on Table 6: X = object-verb; Y = verb-object. The existence of only one match in the DT (Matt 10,29//Luke 12,6) requires that FH Luke or MPH Matthew reverse all but one of the middle synoptist’s object-verb sequences, while supplying a roughly similar number of their own object-verb sequences (reversing the middle synoptist’s verb-object sequences).

Department of Theology, Loyola University Chicago United States of America

Olegs ANDREJEVS [email protected]

ANCIENT COMPOSITIONAL PRACTICES, MEDIA, AND THE SYNOPTIC PROBLEM

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE? SOME RECENT STUDIES AND THEIR IMPLICATIONS FOR THE SYNOPTIC PROBLEM

It has long been recognised that the Gospel of Mark, when compared to the other canonical Gospels, is “rougher”, “less-polished”, and “more-unfinished”. We see such descriptions as early as Papias, who, according to Eusebius, described the Gospel of Mark as lacking “order” (τάξις)1. Modern commentators have noted the “simple and popular” style of Mark’s Gospel2, seen in the frequency of the historical present, parataxis, pleonasms, anacolutha, and so on. The Gospel’s seemingly abrupt beginning and abrupt ending (again, when compared to the Gospels of Matthew and Luke) also contribute to this sense of incompleteness and roughness3. Afterall, the manuscript tradition of the Gospel of Mark indicates at least three later attempts of provide some resolution to Mark’s narrative in the additional endings for the Gospel. And of course, understanding Mark as “less-polished” and “more-unfinished” is 1. Eusebius, Hist. eccl. III,39,15. 2. V. TAYLOR, The Gospel according to St. Mark, London, Macmillan, 1952, p. 52. Taylor states: “Mark’s Gospel is written in a relatively simply and popular form of Greek which has striking affinities with the spoken language of everyday life as it is revealed to us in the papyri and inscriptions. In this respect it differs profoundly from the masterpieces of Attic prose and even from the more cultured language of certain parts of Luke and Acts…”. 3. One wonders if we left out the apparently “incomplete” ending of the Gospel from our reckoning of Mark, would we still see Mark as “incomplete” or even possibly “unfinished”? M.D. HOOKER’s (The Gospel according to St. Mark [BNTC, 2], London, A&C Black, 1991, p. 389) argument still has relevance. Hooker notes that the later scribal attempts to conclude Mark do little to “complete” the Gospel – the “Longer Ending” “does not attempt to deal with problems caused by Mark’s abrupt ending – the women’s silence and the unfulfilled promise to the disciples that they would see Jesus in Galilee – and it shows no reliance on vv. 1-8”. Thus, Hooker makes the case that the Gospel ends abruptly because that is what the author intended – the “sense that Mark is incomplete may be misleading,” for even the “inelegant” 16,8 is “complete” and consistent with Mark’s “notoriously rough” style (ibid., p. 391). She continues: “If we are unhappy with the end of Mark’s gospel as he has left it, it is perhaps because we expect him to ‘round it off’ with an appearance of the risen Lord. … [So perhaps] Mark is inviting us [the reader] to make our own response…” (ibid., p. 393).

180

R.A. DERRENBACKER

seen by some as compelling evidence for the author as a subordinate “literary worker”4, as well as an indication of Markan priority5. Within this larger perspective on the Gospel of Mark, one could argue that the Deutero-Markan hypothesis takes seriously the notion that canonical Mark was succeeded by a secondary, later edition that was accessed by both Matthew and Luke, and can account, at least in part, for the Minor Agreements (agreements in wording between Matthew and Luke against Mark in the Triple Tradition), a lingering problem for the Two-Document Hypothesis (2DH). The Deutero-Markan hypothesis, along with its related but distinctively different Urmarkus, both capture the various editions and versions which clearly must have been in circulation in the late first century when the authors of the Gospels of Matthew and Luke accessed them. On the one hand, the proto-Markan (or Urmarkus) hypothesis posits that a primitive, early form of Mark (now lost) was the chief literary source for Matthew and Luke and was the basis for canonical Mark6. On the other hand, advocates of Deutero-Mark posit a later (“second”) edition of Mark (now lost) that is based on the earlier (canonical) version of Mark, and is used by Matthew and Luke7. In particular, the Deutero-Markan hypothesis 4. See, for example, C. MOSS, Fashioning Mark: Early Christian Discussions about the Scribe and Status of the Second Gospel, in NTS 67 (2021) 181-204, p. 202. Moss writes: The “designation of Mark as Peter’s interpreter [in Papias] identifies him not only as a subordinate but also a literary worker. The role he plays in textualising Peter’s memories was one that ordinarily would have been performed by a literate slave, servile worker or local scribe for hire. The description of Mark’s resulting text as ‘without order,’ a characteristic shared with ancient notes or unpolished drafts, only further cements the idea that Mark and thereby his text, are of lower status”. But see a competing view by H. BOND, The First Biography of Jesus: Genre and Meaning in Mark’s Gospel, Grand Rapids, MI, Eerdmans, 2020, pp. 88-89: “Overall, Mark’s writing is perfectly proficient. … While we should certainly not overplay Mark’s literary abilities, he was clearly a competent and reasonably skilled writer who was perfectly able to convey his ideas in the literary form of a bios”. 5. See, for example, B. WITHERINGTON III, The Gospel of Mark: A Socio-Rhetorical Commentary, Grand Rapids, MI, Eerdmans, 2001, p. 19: “When one takes this evidence together, and especially when one compares how Matthew and Luke deal with the same material, the argument for Markan priority becomes as close to a certainty as one can imagine in scholarly discourse. Matthew and Luke deliberately smooth out the harsh, rough edges of Mark’s grammar and syntax and vocabulary and, of course, some of the content … of his Gospel as well. … One cannot very readily imagine the process working the other way around”. See also M.F. BIRD, The Gospel of the Lord: How the Early Church Wrote the Story of Jesus, Grand Rapids, MI, Eerdmans, 2014, p. 161: “The priority of Mark is … supported by improvements to Mark’s grammar and editorial adjustments to his narrative made by Matthew and Luke”. 6. For a thorough-going defense of the Proto-Mark (Urmarkus) hypothesis, see D. BURKETT, Rethinking the Gospel Sources: From Proto-Mark to Mark, New York – London, T&T Clark, 2004; ID., The Case for Proto-Mark (WUNT, 399), Tübingen, Mohr Siebeck, 2018. 7. Main advocates of this perspective have included A. FUCHS, Defizite der Zweiquellentheorie, Frankfurt am Main, Peter Lang, 2008, and A. ENNULAT, Die “Minor Agreements”:

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

181

implies that this earlier version of Mark (which again approximates our canonical Mark) was somehow less finished and in need of further redaction, which then produced Deutero-Mark – the putative source for Matthew and Luke on this particular variation within the 2DH. However, both of these Markan recensional theories have largely failed to take hold within synoptic source-critical discussion8. Michael F. Bird captures some of the reasons for this: “Theories about Ur-Mark, deutero-Mark, or proto-Luke are problematic: Why would someone bother to reissue a book in only slightly revised form? And such theories are ultimately impossible to prove given the nature of our extant sources”9. More recently, Matthew D.C. Larsen has argued that the Gospel of Mark is an example of an “unfinished” text, not unlike the catalogue of “unfinished” and “less authored” texts that Larsen produces in two important contributions to the study of the Gospels10. Cataloguing “unfinished”, “less authored”, and “accidentally published” literary works from antiquity, Larsen explores the generic category of hypomnēmata as the likely descriptor of the Gospel of Mark – an example of an unfinished draft, a collection of notes, not yet a “book” but a literary text that both the Gospels of Matthew and Luke bring to completion in their respective Gospel “books”. Larsen’s argument will be explored in detail below. But the general assertion that the Gospel of Mark circulated and was utilized (appropriated?) by both Matthew and Luke raises a number of questions for the Synoptic Problem and synoptic source criticism more generally, including the following: • If Matthew and Luke regarded the Gospel of Mark as an unfinished text – a text in need of completion – how does this complexify the Synoptic Problem? • On each of the synoptic theories that presuppose Markan priority (i.e., the 2DH, the Farrer Hypothesis [FH], and the Matthean Posteriority Hypothesis [MPH]), how is one to imagine, from the standpoint of Roman literary culture, Matthew and Luke each accessing copies of (unfinished) Mark? • In what format or medium are we to envisage Matthew and Luke accessing the (unfinished) text of the Gospel of Mark? Untersuchungen zu einer offenen Frage des synoptischen Problems (WUNT, II/62), Tübingen, Mohr Siebeck, 1994. 8. Burkett has attempted to give a renewed focus to Proto-Mark. See The Case for ProtoMark (n. 6). 9. BIRD, Gospel (n. 5), p. 155. 10. M.D.C. LARSEN, Accidental Publication, Unfinished Texts and the Traditional Goals of New Testament Textual Criticism, in JSNT 39 (2017) 362-387; ID., Gospels before the Book, Oxford, Oxford University Press, 2018.

182

R.A. DERRENBACKER

This paper will attempt to address these (and other) questions while critically engaging with the theory that the Gospel of Mark was an “unfinished” and “incomplete” Gospel “book”, or at least an edition of the Jesus story that was in need of revision and expansion, engaging chiefly with the work of J. Andrew Doole, Matthew Larsen, and Chris Keith. I. THE GOSPEL OF MATTHEW AS A NEW (REPLACEMENT) EDITION OF THE GOSPEL J. ANDREW DOOLE

OF

MARK:

As we think about the source-critical implications of an “unfinished” (or at least “less-finished”) Gospel of Mark being available to Matthew and Luke, we need to consider how the authors of the Gospels of Matthew and Luke potentially viewed their main narrative source-text Mark (at least on the synoptic theories that posit Markan priority – the 2DH, the FH, and MPH). Matthew, of course, makes no explicit reference to the Gospel of Mark and instead absorbs most of Mark (sometimes verbatim, or nearverbatim) into a new Gospel narrative. J. Andrew Doole tackled this question in his 2013 monograph What Was Mark for Matthew?11. Presupposing the 2DH, and seeing “M” as “a hodge-podge of unique Jesus material” that “is not a single, written entity”12, Doole builds his case through an analysis of Matthew’s use of Mark, Q and the “M” material, where Matthew builds the “independent material [Q and ‘M’] into his gospel which reveals the dominance of the Markan framework; despite the appendage of brief pro- and epilogues, Mark is followed from start to finish, with only minor omissions and slight rearrangement [in Matthew 3–11] which, however, soon abates. Matthew’s story is that of Mark”13. Thus, Matthew’s “foremost proximity” is to Mark and not to Q; as such, “Matthew is a new edition of Mark”14. With Matthew as a “new edition” of Mark, and “in his almost complete reproduction of his sources in one single, coherent whole [narrative]”, causes his source texts Mark and Q to become “redundant”15. As a result, Q moves into obscurity and ultimately becomes a lost text, perhaps due in part to its

11. J.A. DOOLE, What Was Mark for Matthew? (WUNT, II/344), Tübingen, Mohr Siebeck, 2013. 12. Ibid., p. 36. 13. Ibid., p. 79. 14. Ibid. 15. Ibid., p. 174.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

183

potential medium as a codex16. But what about the Gospel of Mark – why does it survive? Doole suggests that Mark may have been maintained by Matthew’s “community”, or even “forwarded to other groups at the same time as Matthew’s gospel was spreading throughout the area”17. Doole, then, concludes that author of the Gospel of Matthew saw his literary work as a “replacement” for the Gospel of Mark: “Matthew has replaced Mark, the authoritative tradition with which he has long been familiar, with a gospel which both agrees with Mark and develops Mark in the same direction that the Jesus story had been taking”18. And while both Matthew and Luke “rework the gospel of Mark” and engaged in “rewriting the life of Jesus”, only Luke explicitly “leaves open the possibility of future accounts” by “acknowledging his antecessors” in Luke 1,119. The Gospel of Matthew, on the other hand, is “a direct development of Mark”20, a “new edition” of the story of Jesus21. II. THE GOSPEL OF MATTHEW COMPLETION OF AN UNFINISHED (AND OPEN) DRAFT: MATTHEW LARSEN AND THE GOSPEL OF MARK AS HYPOMNĒMATA AS A

A thesis similar to Doole’s has been advocated by Matthew Larsen, but with less of focus on what the Gospels of Matthew and Luke are in relation to Mark, and more emphasis placed on what sort of text Mark was to the two other synoptic evangelists22. As stated above, Larsen has catalogued a number of “unfinished”, “less-authored”, and “accidentally published” works from antiquity23. In his article (2017) that preceded his more extensive treatment of the question in his monograph (2018), Larsen concludes the following: The prevalence of accentual publication, stolen texts and author variants simultaneously identifies and destabilizes one of the foundational assumptions of traditional textual criticism: without the assumption of a text existing in a 16. Ibid. 17. Ibid. 18. Ibid., p. 193. 19. Ibid., p. 194. 20. Ibid. 21. Ibid., p. 195. 22. It is worth noting that despite the seemingly natural affinities between Doole’s monograph (2013) and Larsen’s article (2017) and monograph (2018), Larsen only mentions Doole’s contribution once in passing (see LARSEN, Gospels [n. 10], p. 183, n. 27). 23. See LARSEN, Accidental Publication (n. 10), pp. 368-376; and ID., Gospels (n. 10), pp. 11-78.

184

R.A. DERRENBACKER

final form, the boundaries between text, form, and redaction criticism fall apart. Ancient writing practices and the prevalence of textual fluidity invite us to rethink some foundational categories and ideas of the discipline24.

As such, Larsen argues that text of Mark, at least as appropriated by the later synoptic writers Matthew and Luke, “was unfinished raw material, ὑπομνήματα, notes, memoirs, a draft”, with the Gospel of Matthew “as a continuation of the … mushrooming textual tradition of the gospel [of Mark]”25. This is the thesis that Larsen more fully develops in his 2018 monograph Gospels Before the Book. Larsen begins Gospels Before the Book by lamenting the fact that most scholars have approached the Gospels with rather “bookish” assumptions, including a “standard set of historical questions” that “are brought to bear on the text”26. Larsen states that these assumptions are both “modern and ‘bookish’”, ones that “presuppose an author producing some sort of finished book at a particular time and for a specific situation”27, often presupposing rather anachronistic notions around “publication”. The problem here, according to Larsen, is that there is “no evidence of someone regarding the gospel as a discrete, stable, finished book with an attributed author until the end of the second century CE”, with the idea of a “gospel” as a “discrete[ly] authored book” appearing not until the third century CE at the earliest28. Larsen writes: “That is, though gospel became textualized in the first and second centuries, there is no evidence of the idea of gospel as a gospel book with an author until much later”29. Larsen argues that the evidence instead “suggests a first- or secondcentury reader of the texts we now call the Gospel according to Matthew and the Gospel according to Mark would not have thought of them as two separate books by two different authors. Rather, they would have regarded them as the same open-ended, unfished, and living work: the gospel – textualized”30. Thus, the implications for the traditional historical-critical methods of Gospels study are significant, particularly for source and redaction critics. For Larsen, an appropriate ancient literary genre that better represents the textualized Gospels in first two-to-three centuries of the Common Era 24. ID., Accidental Publication (n. 10), p. 376. 25. Ibid., pp. 377-378. 26. ID., Gospels (n. 10), p. 1. 27. Ibid. 28. Ibid., pp. 1-2. 29. Ibid., p. 2. 30. Ibid., p. 4.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

185

is the category of hypomnēmata. Often translated as “notes”, “memoranda”, or “drafts” (and often found in the medium of a notebook), Larsen states that hypomnēmata “are not literature per se, at least not yet. They are inchoate, provisional, and often exist specifically for the creation of other texts”31. Larsen zeros in on Mark’s Gospel, and sees it as rough and unfinished, an entity that is not viewed as a “gospel book” until the third century CE at the earliest. Larsen describes Mark as follows: “It remained fluid and open to revision. It was a text without being a book. It was written without being authored – at least in any strict sense of the term. It had a readership, but it was not formally ‘published,’ finished, or otherwise closed off”32. Larsen notes that Mark’s Gospel “lacks many elements expected of ancient bioi”33. Thus, while it could be argued that the other Gospels may reflect the literary genre of a modified biography, what is absent from the Gospel of Mark leads Larsen to conclude that it is certainly not example of a bios of Jesus. Instead, the Gospel of Mark should be seen as hypomnēmata – that is, “as less finished, less authored, and less bookish than other gospel constellations”, and instead, “a fluid, unfinished, and relatively open constellation of textual objects”34. Larsen builds his case for seeing Mark’s Gospel as hypomnēmata in terms of how the Gospel was utilized by its first readers and users, including the author of Luke’s Gospel. In the prologue in Luke 1,1-4, the author of Luke’s Gospel makes reference to other “textualized” accounts of the life of Jesus, giving the impression, argues Larsen, that these “previous accounts” (1,1) were found to be “dissatisfying”35. As a result, Larsen maintains that the prologue indicates that the author of Luke’s Gospel “does not take the Gospel according to Mark (or the saying source document called Q, or any other sources to which he may have

31. Ibid., p. 11. See also the definition of the related term commentarii in S. HORN– A. SPAWFORTH (eds.), The Oxford Classical Dictionary, Oxford, Oxford University Press, 31996, p. 373. 32. LARSEN, Gospels (n. 10), p. 81. 33. Ibid.: “Most notably, [Mark] lacks stories about Jesus’s family lineage, birth, childhood, ethical focus, and things after his death”. 34. Ibid., p. 83. 35. Ibid., p. 84. See Larsen’s translation of Luke 1,1-4: “Since lots of other people have attempted to set in order the story about the things that have been fulfilled among us, just as the witnesses from the beginning and the servants of the story handed it down to us, it also seemed good for me, since I had followed closely (pareœkoloutheœkoti) carefully (akriboœs) everything from the beginning (anoœthen), to write [the story] for you, most excellent Theophilus, in order that you might know the convincing nature (asphaleian) of the things you were taught”. BLOWER

186

R.A. DERRENBACKER

had access) to be sufficiently careful (akribeœs) [1,3]. It was not a wellordered narrative that closely follows from the beginning, thus necessitating the creation of the Gospel according to Luke”36. Thus, Larsen maintains, the author of Luke “reworks” Mark’s Gospel because he “finds it to be poorly ordered”, “less finished” and “less accurate”, and supplements Mark’s Gospel extensively37. Thus, Mark’s Gospel looked less like a “book” and more like hypomnēmata to the author of Luke’s Gospel. Larsen then turns to a discussion of other “earliest users” of Mark’s Gospel, and focuses in on the Matthew’s use of Mark. Again, Larsen questions the status quo understanding that viewed Mark as a “finished” and “closed” book by the mid-to-late first century of the Common Era. The various traditional “solutions” to the Synoptic Problem, seen most readily in sourcecritical stemma diagrams, reveal this presupposition, Larsen argues38. Thus, Larsen offers a different and helpful means of visualizing the relationship between the Gospels of Mark and Matthew, proposing that a proportional Venn diagram that illustrates the overlap between the two Gospels better presents the “degree of similarity between the textual traditions” between the two Gospels than do the traditional stemma diagrams39. Again, if the first-century Gospel of Mark was encountered by the author of the Gospel of Matthew in the form of hypomnēmata, the latter was engaged in “finishing, continuing, or otherwise altering the same unfinished and still fluid textual tradition” of the former40. Using several examples from Matthew’s Gospel, Larsen illustrates how the author of Matthew’s Gospel “seems to correct missing elements [from Mark’s Gospel], as well as continue and ‘finish’ the gospel as found in the text we now call the Gospel according to Mark, rather than write a ‘new and separate’ book”, as well as “narrow [its] ambiguities”41. Here, Larsen argues that Matthew’s Gospel “continues the same fluid textual tradition found in” Mark’s Gospel “by removing potential problems and ambiguities”42.

36. Ibid., p. 85. 37. Ibid., p. 86. 38. “In all these graphic depictions, each constellation of textualized gospel tradition is represented as its own discrete unit, bounded by lines within a box or a circle or some other shape, with arrows indicating the direction of source relationship and redaction. All of this, however … serves to reinforce the third-century and subsequent gospel textuality and authorship discourse, reifying each gospel as an enclosed, separate text with its own unique author” (ibid., p. 102). 39. Ibid., p. 104. 40. Ibid., p. 106. 41. Ibid., pp. 108, 111. 42. Ibid., p. 113.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

187

The Gospels of Mark and Matthew “should not be viewed as two separate books”. Instead, Larsen argues, we should see the Gospel of Matthew “continuing the same unfinished tradition of ‘the gospel’ more broadly understood, adding stories to a textual tradition that help that tradition conform better to ancient readers’ expectations about what should be in a story about an individual”43. This same tendency can be found in the textual history of Mark’s Gospel, and its various endings within its manuscript tradition. These various endings within the Markan manuscript tradition illustrate several things, including revealing that ancient readers found the Gospel of Mark “as incomplete and lacking a sufficient ending”, as well as being “unfinished but also as open, and therefore able to be further reworked”44. With all this in view, Larsen is able to conclude the following about Mark’s Gospel: The interaction between the textual traditions we now call the Gospels according to Matthew and Mark, as well as the variety of ways of ending the Gospel according to Mark, are concrete data points that demonstrate early readers’ attempts to revise and polish existing hypomnēmata, thus improving the Gospel according to Mark’s rough yet powerful text. That is to say, perhaps there is a way to understand the text we now call the Gospel according to Mark as a powerful text that was not “finished” until the tradition we now call the Gospel according to Matthew took it up and continued it by reworking it, until the various longer endings of the Gospel according to Mark added proper resurrection narratives, until later prologues were added, until the different gospel traditions were gathered together into a fourfold gospel, until commentators made meaning of the text, until modern narrative critics supply readers with nuanced readings, and so on45.

As a result, then, Larsen urges contemporary readers to see Mark’s Gospel as ancient readers would have seen it – as “less finished, more rough, open, revisable, unpublished, and not very ‘bookish’”, as “unfinished notes”, as hypomnēmata46. There is much that commends Larsen’s thesis, not least of which is that he is attempting to get Gospels scholars to think more realistically, more accurately, and less anachronistically about the literary cultures behind the canonical Gospels. But Larsen’s thesis also raises a few implications for the Synoptic Problem, along with a number of questions that will be addressed below. 43. Ibid., 44. Ibid., 45. Ibid., 46. Ibid.,

p. p. p. p.

114. 118. 120. 122.

188

R.A. DERRENBACKER

III. MATTHEW AND LUKE AS “COMPETITIVE TEXTUALIZERS”: CHRIS KEITH In his book The Gospel as Manuscript, Chris Keith utilizes the phrase “competitive textualization”, a term that captures “the various ways in which the tradents of the Jesus tradition drew upon its material form – and parasitically upon prior instances of the Jesus tradition in material form – in order to assert a particular position within a reception history, which was often characterized by claims of superiority”47. And in seeing Matthew and Luke as “competitive textualizers” of the Gospel of Mark, Keith does interact with Larsen’s work on Mark as hypomnēmata. Keith agrees with Larsen that “the Gospels were part of an open-ended tradition process”48. but takes issue with the characterization of the Gospel of Mark as hypomnēmata, since there is “no evidence” that “first- and early second-century readers of the Gospels could have conceptualized them as such”49. But even if the Gospel of Mark was not seen as hypomnēmata by its earliest readers, Keith argues that they still may “have viewed it is unfinished”50. Even “unfinished books” are still books, so Keith “retain[s] the term ‘book,’ not in reference to finalized texts or collections but in reference to artifacts like bookrolls and codices that were important factors in the transmission of streams of tradition”51. Keith argues, then, that the term and concept of “book” can have a broad conceptualization – anywhere between “unrestrained stream of tradition” or “associat[ion] with the security of restrained tradition”52. Focusing on the “physical space” that Matthew creates when it calls itself a “book” (βίβλος; Matt 1,1), Keith argues that the Gospel of Matthew, as a βίβλος (like the Gospel of Mark), has, in fact, “usurped” Mark: “The author of Matthew’s Gospel … attempts to occupy not only the same narrative space as Mark’s Gospel but also the same physical space in the hands of ‘the reader’ (Mark 13:14//Matt 24:15)”53. While “Matthew’s Gospel does not outright reject Mark’s Gospel … Matthew’s Gospel cannibalizes Torah and Mark’s Gospel – their content but also their media format as 47. C. KEITH, The Gospel as Manuscript: An Early History of the Jesus Tradition as Material Artifact, Oxford, Oxford University Press, 2020, p. 8. Keith first coined the phrase in The Competitive Textualization of the Jesus Tradition in John 20:30-31 and 21:24-25, in CBQ 78 (2016) 321-337. 48. KEITH, Gospel (n. 47), p. 51. 49. Ibid., p. 52. 50. Ibid., p. 53. 51. Ibid., pp. 59-60. 52. Ibid., p. 61. 53. Ibid., pp. 122-123.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

189

authoritative tradition”54. Thus, even though “Matthew’s Gospel would proceed to become substantially more popular than Mark’s Gospel in the early church …it did so harnessing the technology and the gospel format that Mark first introduced to the Jesus tradition”55. When it comes to the Gospel of Luke, Keith rightly points to the Prologue (Luke 1,1-4) as key to understanding Luke’s perspective on his sources (including Mark), “the most textually self-conscious statement yet among Jesus followers”56. Can the reader of Luke’s Prologue discern what Luke’s view of his chief source-text was? Keith surveys a variety of answers to this question57, and concludes that while he might not have been “anti- his predecessors”58, he did engage in “competitive textualization” and indicates in the Prologue “his sense of superiority over, and improvement upon, those prior attempts”, including the Gospel of Mark59. What, then, is the Gospel of Mark for Matthew and Luke, according to Keith? The Gospel of Mark, he suggests, was an innovation in that it created a “written Gospel”, a “manuscript” that was the earliest “textualization of the Jesus tradition”60. Instead of seeing the Gospel of Mark “supplanted” or “supplemented” by the Gospels of Matthew and Luke, the Gospel of Mark functioned as a “manuscript to the Jesus tradition”, for it caused a “subsequent explosion of written Gospels”61. For “Mark’s Gospel introduces ‘textual self-consciousness’ to the Jesus tradition … [It] self-consciously reflects its cultural status as a material artifact, a book that requires a reader to vocalize its script”62.

54. Ibid., p. 123. 55. Ibid. 56. Ibid., p. 124. 57. E.g., F. BOVON, Luke 1: A Commentary on the Gospel of Luke 1:1–9:50 (Hermeneia), trans. Christine M. Thomas, Minneapolis, MN, Fortress, 2002, p. 19: Luke “refutes” his literary predecessor Mark and “introduces his own product as better and more reliable”; contra M. WOLTER, The Gospel According to Luke, Volume 1 (Luke 1–9:50) (BMSEC), trans. Wayne Coppins and Christoph Heilig, Waco, TX, Baylor University Press, 2016, p. 45: “…Luke does not devalue his predecessor’s works with a single word”. Keith also notes (Gospel [n. 47], pp. 127-128) Loveday Alexander’s description of a more “neutral” perspective by Luke on his predecessors. L. ALEXANDER, The Preface to Luke’s Gospel: Literary Convention and Social Context in Luke 1.1-4 and Acts 1.1 (SNTS MS, 78), Cambridge, University Press, 1993, esp. pp. 133-135. 58. KEITH, Gospel (n. 47), p. 128. 59. Ibid., p. 129. 60. Ibid., p. 96. 61. Ibid., p. 98. 62. Ibid., p. 98. Keith references Mark 13,14 as evidence that Mark saw his Gospel as a “material artifact” – “Let the reader (ὁ ἀναγινώσκων) understand”.

190

R.A. DERRENBACKER

IV. IMPLICATIONS FOR

THE

SYNOPTIC PROBLEM

While Doole, Larsen, and Keith differ in a number of areas, each sees the Gospels of Matthew and Luke “usurping” or “replacing” the Gospel of Mark, a text that is, at the very least, “less-finished” than the two later Gospels. Such a perspective has a number of implications for the Synoptic Problem, particularly when these implications are conceived in light of the physical realia of first-century textual production and literary compositional conventions. In Gospels before the Book, Larsen does not delve much into the Synoptic Problem, nor does he necessarily defend one source-critical “solution” over another, other than work with a particular assumption of (a form of) Markan priority – that is, a textualized form of Mark’s Gospel was used by both the authors of the Gospels of Matthew and Luke. The hypothetical sayings document Q, which is posited within the 2DH, is mentioned just a few times in the book63. However, Larsen’s thesis has a number of important implications to the Synoptic Problem. While Larsen’s focus is on Mark’s Gospel as hypomnēmata, there could be merit to applying the same literary category to Q on the 2DH, with Q seen as an unfinished, unpolished, rough draft-like written collection of sayings of Jesus (plus a few narrative stories) that was a literary source for Matthew’s and Luke’s Gospels. After all, this has been suggested before by George Kennedy in 197864, and more recently discussed by Alan Kirk in his book on Matthew’s use of Q65. Of course, the biggest weakness to Matthew’s and Luke’s use of Q is the lack of ancient manuscript evidence for the putative document. But if Q were hypomnēmata, its absence from any extant ancient manuscript tradition is more easily explainable and understandable. As Larsen notes: “A modern reader does well to keep in mind that many texts in the ancient world, perhaps even most, were considered not ‘literature’ but, rather, textual objects and unfinished [including hypomnēmata]. Yet this vast sea of unfished textual objects has not survived to the present. Rather, they have been lost to time”66. In other words, if Q were hypomnēmata, its disappearance is explainable, understandable, and, indeed, expected. 63. E.g., LARSEN, Gospels (n. 10), pp. 85, 102. 64. G. KENNEDY, Classical and Christian Source Criticism, in W.O. WALKER, Jr. (ed.), The Relationships Among the Gospels: An Interdisciplinary Dialogue, San Antonio, TX, Trinity University Press, 1978, 125-155, pp. 136-137. 65. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, pp. 44-50. 66. LARSEN, Gospels (n. 10), p. 36.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

191

All synoptic solutions that assume Markan priority (i.e., the 2DH, FH and MPH) need to imagine both Matthew and Luke independently accessing their copies of the Gospel of Mark. This means that at least two copies of the Gospel of Mark were in some type of literary circulation, and therefore independently accessible by Matthew and Luke; in other words, both Matthew and Luke had copies of Mark on their (hypothetical) bookshelves. If the Gospel of Mark were in some unfinished format (like hypomnēmata, according to Larsen), it seems less likely that Matthew and Luke would have independent access to that pre-publication format of the text of Mark, which one would assume is less likely to circulate. Could one imagine multiple copies of hypomnēmata circulating in such a way that both Matthew and Luke had access to them? How and why is it that a “less-finished” Gospel of Mark survived, in the end? Keith answers that question by arguing that (canonical) Mark saw itself as a book (as defined broadly by Keith [see above]), a distinctive “literary artifact” that was ground zero for “the explosion of Gospel literature in the early church”, with “Mark’s employment of textuality … overwhelmingly successful in creation not only cultural memory but a mark for this kind of cultural memory”67. Larsen, on the other hand, argues that the gospels of Mark and Matthew “should not be viewed as separate books”68. But of course, the Gospel of Mark has survived as a “book” distinct from the Gospel of Matthew. Of the four Gospels, portions of Mark are only preserved in (currently) four papyri (cf. Matthew [24], Luke [11], and John [32]), indicating perhaps a preference away from Mark at least in north Africa (where the papyri were preserved and discovered) in the first few centuries of early Christianity. As well, the canonical sequencing of the four Gospels in uncial manuscripts also betrays a similar perspective in the early Christian scribal tradition (i.e., Mark as the “least favourite” will, at times, turn up as the fourth Gospel in the canonical sequence, with preference typically given to Matthew and John)69. But why preserve Mark at all if Mark were hypomnēmata? Would not hypomnēmata typically vanish from ancient manuscript traditions? Why, for example, would Mark with its “Shorter Ending” be preserved in the rather “bookish” Siniaticus and Vaticanus? Why would multiple copies of Mark be preserved in the scribal tradition if Mark in its first and second century forms was hypomnēmata? 67. KEITH, Gospel (n. 47), p. 96. 68. LARSEN, Gospels (n. 10), p. 114. 69. See, for example, the sequence of the Gospels in Codex Bezae: Matthew, John, Luke, and Mark – the so-called “Western order”.

192

R.A. DERRENBACKER

If the Gospel of Mark was unfinished to the point of being no more than an unfinished collection of notes (e.g., hypomnēmata), why does Matthew so often abbreviate the material he sees in Mark, typically eliminating key details (and sophisticated literary techniques) that existed in his source-text? Take for example Matthew’s use of Mark’s version of the raising of Jairus’ daughter and the healing of the hemorrhaging woman (Mark 5,21-43 par.). There, Mark’s 23 verses become nine in Matthew. The number “twelve” that linked the hemorrhaging woman and the leader’s daughter is eliminated by Matthew, as is the dramatic bit about Jesus sensing power “going forth” from him (v. 30) and the ensuing rebuke by the disciples (v. 31). As well, many of the details at the house of the leader (including Jesus’ words “Talitha cum” [v. 41]) are dismissed by Matthew, details that make the story engaging and interesting. As is well known, this is typical for Matthew in his use of Mark’s Gospel – most Matthean pericopes in the Triple Tradition are much briefer than those in his source-text Mark. If Mark was hypomnēmata, an unfished and unpolished text written for someone else to finish, why does Matthew so frequently eliminate the details from Mark’s stories that make Mark so interesting? Why are the later versions of the stories found in Matthew so often “blander” than the versions found in Mark? Here, I am not disagreeing with Larsen’s understanding of Matthew’s use of Mark, and the problems that Matthew has with the “gaps” in Mark’s text. Instead, it may be that the Gospel according to Mark is actually a bit more “bookish” as Doole and Keith argue, and perhaps also somewhat more “finished” than Larsen allows. Some synoptic source critics have attempted to wrestle with the question of the potential media of the source(s) for Matthew and Luke, in part to explain the resequencing of source material in the later Gospel. For example, Ulrich Luz has argued that Q may have been utilized by Matthew and Luke as “a large notebook bound at the edge with twine”, which “permitted an insertion of new leaves at any time”, while the Gospel of Mark was accessed by Matthew (and Luke) as “a firmly bound codex (or a scroll?) and thus [was] a literary work that for this reason alone continued to be passed on after Matthew had expanded it”70. Similarly, I suggested briefly in 2005 that a potential medium for Mark’s Gospel as accessed by Matthew and

70. U. LUZ, Matthew 1–7 (Hermeneia), trans. James E. Crouch, Minneapolis, MN, Fortress, 2007, p. 19. I developed this suggestion in my essay ‘The Medium is the Message’: What Q’s Content Tells us about its Medium, in D.T. ROTH – R. ZIMMERMANN – M. LABAHN (eds.), Narrative, Metaphor, and Parables in Q (WUNT, 315), Tübingen, Mohr Siebeck, 2014, pp. 207-220.

“UNFINISHED” MARK “REPLACED” BY MATTHEW AND LUKE?

193

Luke could be an early form of a codex71, a medium that is consistent with an “unfinished” (or at least “less-finished”) Gospel of Mark. Going forward, further exploration of a “less-finished” Mark should also lead to further exploration of the potential media of Gospel sources. Finally, the assumption of an “unfinished” Mark also raises some questions around Luke’s use of Matthew on the FH. With this assumption, it is clear that the author of Luke’s Gospel sees and treats his textualized source – Mark – as unfinished, potentially as hypomnēmata (again, Larsen argues that the Prologue in Luke 1,1-4 established that Mark’s Gospel was not “sufficiently careful” or “well-ordered”, and thus “[necessitated] the creation of the Gospel according to Luke”)72. But what about Luke’s other sourcetext (Matthew) on the Farrer Theory? Could it be argued that Matthew’s Gospel, as Luke encounters it, was also not “sufficiently careful” or “wellordered”? Larsen’s work yet raises the question again as to why (for example) Luke would dismantle and restructure Matthew’s Sermon on the Mount the way that he does on the Farrer Theory. As Michael Goulder acknowledges: “For Luke’s warmest admirers must admit that if he has revised Matthew’s Sermon, his result is not really to be compared with his coevangelist’s”73. In other words, it would appear that, on the FH, Luke treats Matthew as he treats Mark (because Mark is “unfinished”, perhaps even hypomnēmata). However, Matthew’s Gospel is clearly not an example of an “unfinished” text (let alone hypomnēmata), as it is polished and sophisticatedly structured, looking more like a “book” (and perhaps even an ancient biographical book). Thus, if, for the sake of argument, one were to dispense with Q and maintain Markan priority, Matthew’s use of Luke (as advocated by supporters of the MPH) actually makes the best sense of the material in a Mark-without-Q scenario. V. CONCLUSION Doole, Larsen, and Keith have all helped us to more realistically conceive of how the Synoptic Gospels (at least the Gospels of Matthew and Luke) came to be, and how the writers of each may have conceived of the Gospel of Mark. In the depictions put forward of the Synoptic Gospels, Doole, 71. R.A. DERRENBACKER, Jr., Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, p. 254. 72. LARSEN, Gospels (n. 10), p. 85. 73. M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, p. 346.

194

R.A. DERRENBACKER

Larsen, and Keith each describe varying degrees “incompleteness” and “unfinishedness” for the Gospel of Mark. At the same time, they each offer various descriptions of the later Synoptic Gospels – Matthew and Luke – in how they eclipse and replace their source text, the Gospel of Mark. At the same time, that same Gospel of Mark survived and endured, perhaps due to its inherent “bookishness” (see Doole and Keith) and potential genre as a biographical treatment of Jesus74. As heuristic devices, the stemma diagrams map out the literary relationships between the Synoptic Gospels by helpfully illustrating the various “solutions” to the Synoptic Problem. But as Doole, Larsen, and Keith have shown us, there is much complexity that lies behind these stemma diagrams, with the exploration of this complexity reflective of the maturing and evolving discussions on the Synoptic Problem. Trinity College Theological School, University of Divinity Melbourne, Australia

Robert A. DERRENBACKER [email protected]

74. In addition to the (now classic) argument for Mark (and the rest of the Gospels) as examples of (modified) ancient bioi in R.A. BURRIDGE, What Are the Gospels? A Comparison with Graeco-Roman Biography, Waco, TX, Baylor University Press, 32018, see BOND, The First Biography of Jesus (n. 4), and C.S. KEENER, Christobiography: Memory, History, and the Reliability of the Gospels, Grand Rapids, MI, Eerdmans, 2019.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

I. INTRODUCTION One of the main objections to the Farrer Hypothesis (FH) is its alleged difficulty in accounting for Luke’s ordering of the double tradition, especially in Luke’s central section. It is felt unclear both why Luke would have reordered Matthean material as he would need to have done, and how he could have done so given the limitations of ancient writing technology and the working practices of other ancient authors. A possible answer to the question how is that Luke possessed good memory command of all or part of Matthew’s Gospel. The present essay will explore how this might have worked through comparison with the ordering of source material in the Gospel of Thomas and the allegorical commentaries of Philo of Alexandria1. Although Matthew and Luke often place their double tradition material in different contexts, much of it occurs in (more or less) the same order in the two Gospels; this is one of the reasons the Two-Document Hypothesis (2DH) gives for postulating a documentary Q. That this commonly ordered material often appears in different contexts presents no mechanical difficulty for the FH, since it could result from FH Luke using Mark and Matthew in alternating blocks. The difficulty arises from the substantial amount of Matthean material (such as much of Matthew’s Sermon on the Mount) that FH Luke would have employed out of its Matthean sequence, some of it in blocks, some of it just a verse at a time. If Luke were reliant on eye-contact with his written sources, we should have to imagine him winding and unwinding a scroll of Matthew to pick out a verse here and a passage there to weave into his own composition. Luke’s possession of a codex of Matthew (conceivable though by no means assured) might alleviate 1. I discuss the issue of FH Luke’s (re-)ordering of source material in substantially greater length in E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021, pp. 159-204. The present paper aims to provide further support for the use of associative cueing (in memory) that is suggested there.

196

E. EVE

the difficulty, but would hardly eliminate it, since leafing back and forth through a codex to extract so much material out of order would still be laborious. The difficulty is greatly reduced if Luke worked from his memory of Matthew, at least when using Matthean material out of its Matthean sequence. The bulk of this out-of-sequence material is taken from four of Matthew’s five discourses, Matthew chapters 5–7, 10, 18, 23–252, so while FH Luke’s memory command of Matthew could well have extended beyond those chapters, a good memory command of just these parts of Matthew would have enabled him to access most of the double tradition material he needed to re-order. The question is how readily Luke would have been able to retrieve material from these sections in memory to reuse them in the quite different contexts and sequences in which they occur in his Gospel (with varying degrees of resemblance to Matthew’s wording)3. Some previous work on the use of memory in ancient composition suggests that this should not have been particularly problematic4. But a caution has been sounded by Alan Kirk, who has addressed this question in considerable detail through a discussion based in part on David C. Rubin’s work on serial cueing and in part on Kirk’s analysis of the reordering of material by ancient scholarly writers5. In brief, Kirk argues that the fusion of memory and manuscript in a written text would not permit it to be dissolved back into orality willy-nilly, since memory command of the written text would require following the structured (and hence memory-guiding) sequencing of the text. In Kirk’s view, this does not mean that no reordering of source material is possible, but that rather that it would be subject to certain restrictions. Kirk’s examples of ancient scholarly authors (whose use 2. See Table 6.8 in ibid., pp. 199-204. 3. There is some tendency for Luke’s wording to be closer to Matthew’s when they are both employing double tradition material in the same sequence, but this is far from being a universal rule. See EVE, Relating (n. 1), p. 26, engaging with R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, p. 238. 4. See M. CARRUTHERS, The Book of Memory: A Study of Memory in Medieval Culture (CSML), Cambridge: Cambridge University Press, 22008, pp. 21-22; A. GREGORY, What Is Literary Dependence?, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 87-114, pp. 95-103; E. EVE, Writing the Gospels: Composition and Memory, London, SPCK, 2016, pp. 81-102. 5. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, pp. 29-150; D.C. RUBIN, Memory in Oral Tradition: The Cognitive Psychology of Epics, Ballads, and Counting-Out Rhymes, Oxford, Oxford University Press, 1995, pp. 39-193.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

197

of source material Kirk takes to be analogous to Matthew’s use of Q) suggests the use of a particular index and search strategy (first locate a particular topic in memory, then scan the arrangement of material under that topic) for the retrieval of material from a source text to be employed in a partially different sequence in the target text. On this model, the new text is still constrained to move broadly forward through the material in its source and to broadly preserve the relative order of material taken from the source. In Kirk’s view, this procedure accounts neatly for 2DH Matthew’s use of Q but not so well for FH Luke’s of Matthew6. The counterargument in Relating the Gospels employs the notion of associative cueing, meaning the ability of memory to follow cues in a non-linear fashion, on the basis of verbal, topical, or other conceptual links, to allow for a freer retrieval of source material than Kirk’s proposal readily permits7. In support of this I cite Paul’s catena of biblical quotations at Rom 3,10-18, the principles of rabbinic exegesis (some of which would appear to rely upon associative cueing to operate) and Philo’s drawing on material from across the Pentateuch in the Migration of Abraham and his reordering of biblical source material in the Life of Moses and the Special Laws, but these Philonic examples were tabulated without much analysis or discussion8. Here I shall explore one of them (Philo’s allegorical exegesis) in more depth to see how Philo’s retrieval of material from across the Pentateuch would have been facilitated by associative memory cueing. This exploration will aim to support proposals for the role of associative cueing in FH Luke’s ordering not only of Matthean material, but also of some of the L material with which it is interspersed in his central section9. I previously suggested that Luke’s selection both of some L material and of some Matthean material taken out of its Matthean sequence was sometimes prompted by material Luke encountered (but did not directly employ) in its Matthean sequence, so that Luke combined direct use of Matthew (double tradition appearing in the same relative order in both Gospels) with indirect use (generated by imitation or associative use of Matthew) to form a more or less continuous sequence through both Gospels. This means that Luke’s 6. KIRK, Q in Matthew (n. 5), pp. 184-297, 306-309; ID., Memory, Scribal Media, and the Synoptic Problem, in FOSTER et al. (eds.), New Studies in the Synoptic Problem (n. 4), 459-482, pp. 473-479. 7. EVE, Relating (n. 1), pp. 11-15, 42-55, 145-162, in part depending on A. BADDELEY – M.W. EYSENCK – M.C. ANDERSON, Memory, Hove, Psychology Press, 2009, pp. 165-80; CARRUTHERS, Book of Memory (n. 4), pp. 21-22. 8. EVE, Relating (n. 1), pp. 48-54. 9. Ibid., pp. 167-195. By L material I mean material that is regarded as unique to Luke, without thereby implying anything about its origins.

198

E. EVE

order is rather more dependent on Matthew’s (as well as Mark’s) than is commonly recognized. For example, I suggest that Luke’s account of the rejection of Jesus by the Samaritan village at Luke 9,52-56 was prompted by the command to go nowhere among the Samaritans at Matt 10,5 and that the Parable of the Good Samaritan and its introduction at Luke 10,25-37 was prompted by the Matthean additions to the pericope about plucking grain on the Sabbath at Matt 12,1-7 (more on which below)10. Luke’s order then arises from complex interactions between the topics he is concerned to address, the construction of his own narrative and topical sequences, and the material he finds in his sources. The plausibility of this account will be strengthened if we can find anything analogous occurring in Philo. Rather than rely on a single parallel, however, I shall begin with a different one, namely the Gospel of Thomas. This is relevantly similar to FH Luke’s use of sources in that both Luke and GTh deploy synoptic (or synoptic-like) material along with other material in their own compositions. Neither GTh nor Philo’s allegorical commentary provides a precise parallel to Luke’s use of sources, but together they straddle Luke (potentially in date as well as in compositional aims and methods) so that what can be learned from them jointly can reasonably be applied to Luke as well. II. THE GOSPEL

OF

THOMAS

The Gospel of Thomas (henceforth GTh) comprises a list of sayings attributed to Jesus, introduced with the incipit “These are the obscure sayings that the living Jesus uttered and which Didymus Jude Thomas wrote down”11. GTh is unlike the canonical gospels in that there is no narrative and little obvious structure. Some of the sayings in GTh have clear synoptic parallels, many do not, and some lie midway between, that is they echo or partly parallel a synoptic saying, but if the synoptic saying is the source, it must have been substantially adapted12. If GTh is dependent on the canonical gospels (especially Matthew and Luke) for a substantial portion of its material, there are a number of relevant parallels with FH Luke’s of Matthew and Mark. Both FH Luke and 10. Ibid., pp. 175, 177. 11. The translation is that of B. LAYTON, The Gnostic Scriptures: A New Translation with Annotations and Introductions, London, SCM, 1987, p. 380. 12. S.J. PATTERSON, The Gospel of Thomas and Jesus, Salem, OR, Polebridge, 1993, pp. 95-97, lists 88 sayings with some kind of synoptic parallel, of which 58 are “twins” and 18 are “siblings” (for this terminology see ibid., pp. 17-18).

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

199

GTh would have adapted and reordered material from their sources to fit agendas that differed from those of their sources. GTh’s ordering would have been more extensive than FH Luke’s, as would have been the extent of his reshaping of his source material to fit a viewpoint that was further from those of the Synoptic Gospels than any of the Synoptic Gospels were from one another. But then, a fortiori, what was possible for GTh should also be possible for FH Luke. The case for GTh’s dependence on the Synoptic Gospels has recently been argued by Simon Gathercole and Mark Goodacre on the basis of the traces of synoptic redaction appearing in GTh and the degree of verbatim agreement between several synoptic sayings and their GTh parallels13. But many others have argued that GTh was wholly independent of the canonical gospels, and that the parallels between GTh and the synoptics arose mainly from the former’s independent use of an early form of the synoptic oral tradition together with the occasional influence from the canonical gospels on the subsequent textual tradition of GTh14. While some scholars have been persuaded by Gathercole’s and Goodacre’s arguments, several others have not15. This debate cannot be settled here, but it may be noted that several of the arguments deployed for GTh’s independence rely on assumptions that many, including many supporters of the FH, would reject. For example, Patterson calls on Q to provide a parallel for another earlier sayings collection and for minimizing the evidence for Lukan or Matthean redaction

13. S. GATHERCOLE, The Composition of the Gospel of Thomas: Original Language and Influences (SNTS MS, 151), Cambridge, Cambridge University Press, 2012, pp. 129-224; M. GOODACRE, Thomas and the Gospels: The Making of an Apocryphal Text, London, SPCK, 2012. Cf. earlier (and more cautiously), C.M. TUCKETT, Thomas and the Synoptics, in NT 30 (1988) 132-157. 14. See, e.g., PATTERSON, Gospel of Thomas (n. 12), pp. 9-110; H. KOESTER, Ancient Christian Gospels: Their History and Development, London, SCM, 1990, pp. 84-113; cf. GOODACRE, Thomas (n. 13), pp. 1-9. 15. C.M. TUCKETT, The Gospel of Thomas: Gathercole and Goodacre, in SJT 66 (2013) 221-229 and N.D. LEWIS, A New Gnosticism: Why Simon Gathercole and Mark Goodacre on the Gospel of Thomas Change the Field, in JSNT 36 (2014) 240-250, are both broadly sympathetic. S.J. PATTERSON, Twice More – Thomas and the Synoptics: A Reply to Simon Gathercole, The Composition of the Gospel of Thomas, and Mark Goodacre, Thomas and the Gospels, in JSNT 36 (2014) 251-261 and J.S. KLOPPENBORG, A New Synoptic Problem: Mark Goodacre and Simon Gathercole on Thomas, in JSNT 36 (2014) 199-239, are more critical. For responses to criticisms see M. GOODACRE, Did Thomas Know the Synoptic Gospels? A Response to Denzey Lewis, Kloppenborg and Patterson, in JSNT 36 (2014) 282-293 and S. GATHERCOLE, Thomas Revisited: A Rejoinder to Denzey Lewis, Kloppenborg and Patterson, in JSNT 36 (2014) 262-281.

200

E. EVE

appearing in GTh16. Thomasine independence is also said to be supported by the alleged primitiveness of many GTh sayings compared with their synoptic counterparts, but this relies on methods of detecting relative primitivity (for example, through comparative length or simplicity) that many would dispute17. It is also argued that the differences in order between the parallel material in GTh and the Synoptic Gospels counts strongly against the latter’s dependence on the former, but the question of order is precisely the point at issue here18. Finally, Patterson argues that John’s independence from the synoptics parallels that of GTh, but not everyone would agree that John is independent of the synoptics19. Many of the perceptions underlying the plausibility or otherwise of GTh’s use of the canonical gospels are thus similar to those underlying that of Luke’s use of Matthew, not least whether one gives more weight to similarities or differences between texts in determining the likelihood of a literary relationship between them20. It may therefore seem perilous to rely on GTh’s use of the canonical gospels in developing a supporting argument for the FH, even if a good case has been made for the former. We shall therefore proceed by enquiring first what may be learned about how reordering might be facilitated by memory on the basis that GTh was dependent on the Synoptic Gospels, and then ask what changes if GTh is an independent composition21. 16. PATTERSON, Gospel of Thomas (n. 12), pp. 12-13 (acceptance of the 2DH), 20-70 (frequent appeals in one way or another to Q), 102-108 (Q as a support for the genre of GTh). For critiques see GATHERCOLE, Composition (n. 13), pp. 8-9; GOODACRE, Thomas (n. 13), pp. 9-14. 17. PATTERSON, Gospel of Thomas (n. 12), pp. 22-24, 31, 42, 47-49, 59, 66-71. Against this see GOODACRE, Thomas (n. 13), pp. 17-20, 109-127, 145-150; GATHERCOLE, Composition (n. 13), pp. 5-6, 132-138; TUCKETT, Thomas and the Synoptics (n. 13), p. 156. 18. PATTERSON, Gospel of Thomas (n. 12), pp. 94-99. Against this see GOODACRE, Thomas (n. 13), pp. 14-17; GATHERCOLE, Composition (n. 13), pp. 130-132; TUCKETT, Thomas and the Synoptics (n. 13), pp. 139-140. 19. For a history of the debate on the question, see H.W. ATTRIDGE, John and Mark in the History of Research, in E.-M. BECKER – H.K. BOND – C.H. WILLIAMS (eds.), John’s Transformation of Mark, London, T&T Clark, 2021, 9-22. See also the other essays in the same volume that make the case for John’s use of Mark. 20. This is the issue behind what GOODACRE, Thomas (n. 13), pp. 54-57, dubs the “plagiarist’s charter”. On which see further PATTERSON, Twice More (n. 15), pp. 255-256; KLOPPENBORG, New Synoptic Problem (n. 15), pp. 202-203; and GOODACRE, Did Thomas Know (n. 15), p. 285. 21. To be sure these are not the only two possibilities, since GTh could be partially dependent on the synoptics for its synoptic-like sayings while deriving others via another route, as it must have done for its non-synoptic material; see, e.g., the caution in TUCKETT, Thomas and the Synoptics (n. 13), pp. 156-157, but for the purpose of this paper reckoning with the two extremes will have to suffice.

201

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

Roughly half of the sayings in GTh have parallels in one or more of the Synoptic Gospels, but there is virtually no commonality in their order22. This may be illustrated by setting out how the parables of Matthew 13 are dispersed in GTh (Table 1) and then listing the synoptic parallels to a small section of GTh (Table 2). Table 1 Matthew

Thomas

13,7-8

  9

13,24-30

 57

13,31-32

 20

13,33

 96

13,44

109

13,45-46

 76

13,47-48

  8

Table 2 Thomas

Matthew

Mark

Luke

 4

19,30

10,31

13,30

 5

10,26

4,22

8,17; 12,2

 6

6,2-16?

8,5-8

 7  8

13,47-48

 9

13,2-8

4,3-9

24,35?

13,31?

10 11

12,49 21,33?

12 13

16,13-22?

14

6,1-18; 15,11

10,8

10,34

12,49-53

15 16

22. See PATTERSON, Gospel of Thomas (n. 12), pp. 17, 94-97; GATHERCOLE, Composition (n. 13), pp. 152-153.

202

E. EVE

It is sometimes unclear which of the Synoptic Gospels would be the most likely source for GTh, or even whether the GTh saying in question is an adaptation of any of them or a wholly independent saying. Moreover, several of the GTh sayings that are shown as having synoptic parallels also contain material not obviously paralleled in the synoptics. Three of the GTh sayings in Table 2 (7, 12, 15) have no clear canonical parallel at all. Sayings 8 and 9 are both taken from Matthew 13, but from opposite ends of the chapter and in reverse order. The apparent contiguity of material from Luke 8,17 in sayings 5 and 6 is a result of the repetition of the substance of the second half of saying 5 (about there being nothing hidden that will not be revealed) within saying 6. Together, these two tables illustrate how GTh selects material paralleled in the canonical gospels but arranges it in a way that owes little or nothing to the canonical arrangement23. This is also true of GTh as a whole, except that overall GTh has a lower proportion of material paralleled in the Synoptic Gospels than the example sequences shown above. On the assumption that GTh did derive its synoptic-like material from the canonical gospels, this raises the question of how its author (or compilers) went about deploying it. It is not totally inconceivable that the author of GTh scrolled (or leafed) back and forth through manuscripts of Matthew and Luke picking out a saying here and a parable there to adapt for their own composition, but this seems highly unlikely. One would have expected such a procedure to result in more of the order of the sources (as well, perhaps, as more of the wording) being preserved in GTh. It seems far more likely, then, that the author(s) of GTh accessed their sources primarily through memory24. Given the apparent (and possibly quite deliberate) lack of any obvious narrative, logical or rhetorical sequence in GTh and the equally apparent freedom of GTh to select material from anywhere in their sources, this invites the question how the compiler’s memory worked. 23. Although there are occasional traces of common order between GTh and one of the Synoptic Gospels. E.g., GOODACRE, Thomas (n. 13), p. 151, suggests that sequences like GTh 63–66 might indicate the author of GTh working through Luke in sequence, while PATTERSON, Gospel of Thomas (n. 12), p. 96, identifies brief common sequences at GTh 43,3–45,4; 47,3-5; and 65–66. 24. So GATHERCOLE, Composition (n. 13), pp. 220-221, with the qualified approval of PATTERSON, Twice More (n. 15), p. 258, although both Gathercole and Patterson talk in terms of “secondary orality” and thus envisage some kind of oral memory. GOODACRE, Thomas (n. 13), pp. 150-151 argues for GTh’s use of the Synoptic Gospels but suggests this was probably often through memory of the source text. For memory as a possible mechanism of literary dependence see GREGORY, Literary Dependence (n. 4), pp. 95-103. For GTh as self-consciously textual, see C. KEITH, The Gospel as Manuscript: An Early History of the Jesus Tradition as Material Artifact, Oxford, Oxford University Press, 2020, pp. 154-157.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

203

One can occasionally trace some contiguity of image or topic within individual sayings and, more rarely, between consecutive sayings. So, for example, sayings 5 and 6 both talk about what is hidden becoming revealed, although in saying 5 this is related to knowing both what is obvious and what is hidden and in saying 6 to not lying or doing what you hate. Sayings 8 and 9 both contain parables taken from Matthew 13 (the dragnet and the sower), and while their application within GTh may be less than immediately apparent, it may be that their juxtaposition is intended to aid or at least provoke interpretation. The collocation of material possibly derived from Matt 6,1-18 (here negated), Luke 10,8 and Matt 15,11 in saying 14 progresses from a rejection of typical acts of Jewish piety to a rejection of Jewish concerns about proper food25. It would not be impossible to see the disciples’ question about the end in saying 18 as having been prompted by Jesus’ promise to give them something never before conceived in saying 17. Sayings 32 (based on Matt 5,14b?) and 33 (based on Matt 10,27) are connected by the notion of something being or becoming apparent from a height (hill/rooftop). Saying 36 urges lack of worry over clothing, while in Saying 37 Jesus tells the disciples they will see him when they strip off their clothes. Saying 54 pronounces a blessing on the poor (Luke 6,20; Matt 5,3) while Saying 55 urges renouncing one’s family and carrying one’s cross (which might imply embracing poverty) to be worthy of Jesus (Luke 9,23; 14,26-27). The theme is arguably continued into Saying 56, which likens the world to a carcass (cf. Luke 9,24-25). Sayings 68–69 contain a series of macarisms with some loose resemblance to Matt 5,11 and Luke 6,21. This suggests that some form of associative cueing (retrieving material from memory on the basis of image or theme) may well have been at work over short sections of GTh, though hardly at the level of macrostructure or even meso-structure26. More prominent is the use of catchwords connecting one saying to the next27. As Goodacre observes, this would be another form of associative cueing28. It may be questioned whether such catchwords functioned as aids 25. Although the listing of prayer, fasting and almsgiving as typical of Jewish piety may be too common to indicate any dependence on Matthew here; see R. URO, Thomas and Oral Gospel Tradition, in ID. (ed.), Thomas at the Crossroads: Essays on the Gospel of Thomas, Edinburgh, T&T Clark, 1998, 8-32, p. 22. 26. See also GATHERCOLE, Thomas Revisited (n. 15), p. 268, on thematic links between and generic groupings of GTh sayings. 27. PATTERSON, Gospel of Thomas (n. 12), pp. 99-102; KOESTER, Ancient Christian Gospels (n. 14), pp. 80-82. 28. GOODACRE, Thomas (n. 13), p. 151: “memory is associative, and many of Thomas’s links look as if they could be the result of simple associations in memory”.

204

E. EVE

for the composition of GTh or for subsequent users of GTh, especially those who wished to commit the material to memory29. But these possibilities are by no means mutually exclusive, and so it seems probable that catchword association was one of the mnemonic techniques employed by the author of GTh to select material both from the canonical gospels and his own traditions. The lack of any clear overarching macrostructure could well be a deliberate decision by the author of GTh30. Saying 1 suggests that the meaning of the collection is not meant to be immediately clear but needs to be puzzled over in order to avoid tasting death31. Whether one regards GTh’s position as primarily Gnostic, subversive wisdom, ascetic, Middle Platonic, or some combination of these, it is plainly intended to convey a different view of Jesus from that of the Synoptic Gospels32. Removing the synoptic-like material from anything resembling its synoptic contexts and redeploying it in a manner far removed from that of the synoptic discourses that may have been familiar to at least part of GTh’s target audience may be one way of establishing that distance from the canonical viewpoint, constituting an instance of what Chris Keith calls “competitive textualization”33. Goodacre sees the use of this synoptic material as a way of borrowing the authority of the Synoptic Gospels to lend credence to the non-synoptic materials that GTh mixes in with them (and hence to the overall project of GTh)34. It could also be that GTh’s use of this material is intended to subvert the synoptic presentation of Jesus. GTh could be adapting synoptic material and mixing it with material of its own to simultaneously borrow and subvert the prestige of the synoptic presentation, by making it serve GTh’s alternative vision35. 29. PATTERSON, Gospel of Thomas (n. 12), p. 102, allows that both are possible. 30. GOODACRE, Thomas (n. 13), p. 16; PATTERSON, Gospel of Thomas (n. 12), p. 102, argues against this. 31. GOODACRE, Thomas (n. 13), p. 16, describes GTh as a gospel that “announces itself as an enigma from the beginning”. 32. For discussions of various possibilities see PATTERSON, Gospel of Thomas (n. 12), pp. 154-155, 199-214; ID., The Gospel of Thomas and Christian Origins: Essays on the Fifth Gospel, Leiden – Boston, MA, Brill, 2013, pp. 9, 11-19, 33-59, 134-136, 173-196; KOESTER, Ancient Christian Gospels (n. 14), pp. 124-128; A. MARJANEN, Is Thomas a Gnostic Gospel?, in URO (ed.), Thomas at the Crossroads (n. 25), 107-139; R. URO, Is Thomas an Encratic Gospel?, ibid., 140-162. 33. KEITH, Gospel as Manuscript (n. 24), pp. 8-9, 154-157. 34. GOODACRE, Thomas (n. 13), pp. 178-182. 35. Cf. C. KEITH, ‘If John Knew Mark’: Critical Inheritance and Johannine Disagreements with Mark, in BECKER et al. (eds.), John’s Transformation of Mark (n. 19), 31-49, p. 40: “the Gospel of Thomas eliminates the contextualized narrative for Jesus’ sayings in order to reproduce a Jesus who is simultaneously similar to and dissimilar to the Synoptic Jesus”.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

205

There is no definitive way to tell whether GTh used the Synoptic Gospels purely from memory or through consulting manuscripts or from some combination of the two (for example, one might refresh one’s memory from a manuscript and then compose on the basis of one’s short-term memory or newly reinforced long-term memory of the source without maintaining eye-contact with it when actually composing)36. Ancient authors did typically adapt their source material37. Ancient authors also typically relied on memory of their sources38, and there seems to be nothing in GTh’s use of synoptic material that is incompatible with its being employed primarily from memory. That the precise wording of a synoptic parallel is sometimes reflected in GTh’s version could be the result of that wording have stuck in the compiler’s memory, particularly if GTh is reworking material that has become familiar from repeated use. Many of the substantial variations from synoptic parallels could equally well be due to a combination of GTh relying on the memory of the gist of the synoptic parallel and a conscious decision to give it a Thomasine twist. But what difference would it make if GTh were substantially independent of the Synoptic Gospels? On this scenario, GTh would presumably have obtained its synoptic-like material from the pre-gospel oral tradition. It may then be that GTh has to do less work in adapting this material, since in many cases it might have resembled GTh’s version more closely than the canonical versions. But GTh still has to be able to select a range of material, and it is not immediately clear why this should be any easier from one’s memory of oral tradition than from one’s memory of manuscripts, which may include one’s aural memory of hearing manuscripts read aloud39. 36. EVE, Relating (n. 1), pp. 25-27. This raises further questions about whether GTh’s putative knowledge of the synoptics is mediated through direct eye contact with a source, (possibly repeated) aural reception of a source (hearing it read aloud), or so-called “secondary orality” (better termed “feedback” or “re-oralization”) whereby the oral performance of a written text influences subsequent oral tradition. See GATHERCOLE, Composition (n. 13), pp. 155-159, 214-221; URO, Thomas and Oral Gospel Tradition (n. 25); and note the warnings at GOODACRE, Thomas (n. 13), pp. 137-142, against setting up a sharp oral-written dichotomy and the need for a more nuanced oral-written interface model. 37. See, e.g., EVE, Relating (n. 1), pp. 29-73; J.S. KLOPPENBORG, Variation and Reproduction of the Double Tradition and an Oral Q?, in ETL 83 (2007) 53-80; C.B.R. PELLING, Plutarch’s Adaptation of His Source-Material, in JHS 100 (1980) 127-140. 38. EVE, Writing (n. 4), pp. 58, 81-89; ID., Relating (n. 1), pp. 11-14; C.B.R. PELLING, Plutarch’s Method of Working in the Roman Lives, in JHS 99 (1979) 74-96, pp. 92-94; Pliny the Younger, Letters 9,36,1-3. 39. On reading in antiquity see EVE, Writing (n. 4), pp. 4-10; W.V. HARRIS, Ancient Literacy, Cambridge, MA, Harvard University Press, 1989, pp. 25-28, 226-228, 231-233; H.Y. GAMBLE, Books and Readers in the Early Church, New Haven, CT, Yale University Press, 1995, pp. 4-10; L.W. HURTADO, Oral Fixation and New Testament Studies? ‘Orality’, ‘Performance’ and Reading Texts in Early Christianity, in NTS 60 (2014) 321-340.

206

E. EVE

The term “oral tradition” can sometimes be used in an unhelpfully vague manner40. What must be meant here is not the sequence of oral performances that in fact constitute the process of oral tradition, but the capacity to reproduce from memory the content of what has previously been heard. But absent knowing what form these putative oral performances took, one is left trying to compare known written texts with unknowable oral texts performed in a range of hypothetical situations and then simply to assume that it is easier to randomly access material from the latter than it is from the former41. If this were correct, it would lead to the seemingly paradoxical conclusion that knowing material from a written text made it harder to manipulate it in memory than knowing it from oral performances. This seems strongly counterintuitive not least because there are many people today who can cite individual parables or sayings of Jesus out of context despite having read their bibles, and there is no obvious reason to suppose that people’s memories were any less able to do this in antiquity. If it be countered that such persons may also be familiar with hearing Jesus material outside its wider literary context in the course of Christian worship, this is also likely to be true of Christians in antiquity who would have been even more reliant on oral performance for their familiarity with written texts. Conversely, it is simply an assumption that oral performances regularly presented Jesus material in atomized form so the that the hearers’ memories would have been stocked with atomized material readily deployable in any order simply because it was atomized. For all we know, oral performances of the Jesus tradition may already, at least on occasion, have delivered it in the form of connected discourses, and presumably must have made some attempt to render the tradition memorable. To have a stock of Jesus material readily deployable from oral tradition surely requires one to have developed some sort of mental schema to access it from memory. The assumption that familiarity with written texts would make one less able to access material in any desired order derives its force from the stable structure a written text imposes on its material, which structure is then taken to constrain the way the material is accessed in memory. But this arguably relies on too sharp a distinction between the memory of written and of orally delivered material as they may often be encountered in practice. That being so, it seems reasonable to suppose that what GTh could do 40. For various ways in which oral tradition might be envisaged, see E. EVE, Behind the Gospels: Understanding the Oral Tradition, London, SPCK, 2013. 41. See note 5 above.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

207

with oral tradition GTh could also do with written texts. To be sure, the case being made here is stronger if GTh used the Synoptic Gospels, but even if GTh were wholly independent of the Synoptic Gospels it still illustrates how memory access to synoptic (or synoptic-like) material via associative cueing may facilitate its arrangement into a fresh sequence. III. PHILO’S ALLEGORICAL COMMENTARY The dependence of Philo’s allegorical commentaries upon written sources is scarcely in doubt. The reason these commentaries are of interest here is that in the course of presenting his verse-by-verse exegesis of the text he is primarily commenting on, Philo frequently quotes or alludes to material taken from widely scattered locations elsewhere in the Pentateuch. For example, in moving from Gen 6,3 to Gen 6,4 at Gig. 20-58, Philo draws in material from Gen 1,2; Exod 31,2-3; Num 11,7; Lev 18,6; Num 14,44; Deut 5,31; Exod 18,14; Lev 16,2.34; Exod 33,7; Exod 20,21 and Deut 34,7 (in that order)42. Similarly, after quoting the command to Abraham to leave his land, kindred and father’s house at Gen 12,1-3 at Migr. 1, at Migr. 3-54 Philo sets out on his own tortuous migration through Gen 3,19; Gen 28,17; Exod 24,12; Gen 12,1; Gen 13,9; Exod 2,23; Gen 50,24-25?; Gen 39,7; Gen 40,8; Gen 41,41; Gen 40,15; Gen 39,8-9.12; Gen 42,18; Gen 45,28; Gen 50,19; Gen 45,7-8; Exod 12,11; Gen 31,3; Gen 26,2; Exod 15,25; Gen 2,9; 1 Sam 9,9; Gen 1,31; Gen 12,1; Gen 15,6; Deut 34,4; Exod 20,18.22; Deut 4,12; Exod 1,9 and so on through many other parts of the Pentateuch43. Philo is thus able both to range freely over his primarily text (here Genesis) and to bring in material from diverse locations in a number of other texts. Although he is notionally working forwards verse-by-verse through small sections of Genesis in both these examples, he calls on a good deal of material that is unrelated to this sequence. It seems most likely that he would do so from memory. The structure of Philo’s allegorical commentaries is not always wholly apparent. Sometimes he proceeds by question and answer, identifying a potential difficulty in the text and then providing a solution to it44. Sometimes he proceeds by elaborating on the text. Sometimes he appears to 42. D.T. RUNIA, Further Observations on the Structure of Philo’s Allegorical Treatises, in VC 41 (1987) 105-138, pp. 133-134. 43. EVE, Relating (n. 1), pp. 53-54. 44. E.g., at Gig. 1–5 (on Gen 6,1) and Deus 1–4; see P. BORGEN, Philo of Alexandria: An Exegete for his Time, Leiden – Boston, MA – Köln, Brill, 1997, pp. 104-105, 113.

208

E. EVE

wander off in substantial digressions. Most of the time he is engaged in a spiritualizing, or Platonizing, exegesis of the text that aims to delve beneath its surface, historical meaning to a deeper sense that the passages brought in from elsewhere may be called upon to support. The most helpful work on the structure of Philo’s allegorical commentary for our purposes is that of David Runia, based on a close study of De Gigantibus and Quod Deus sit immutabilis, partially in response to earlier studies45. Among the issues Runia considers is the principle underlying Philo’s selection of secondary texts, the passages he brings in from elsewhere in the Pentateuch in the course of explicating the primary text he is covering seriatim. Runia observes that Philo employs various means to “break open” what Philo takes to be the hidden meaning of the biblical text. Among these is the employment of “outside material”, so that “[a]spects of the biblical lemma under discussion can be illumined by invoking other biblical texts containing parallel words, phrases or ideas”46. Philo carries out both “primary exegesis, which concentrates on direct exegesis of the main biblical lemma, and … secondary exegesis … of the subordinate biblical lemmata” that are invoked “in order to cast more light on the main text”47. For Runia, in order to understand the overall structure of Philo’s allegorical treatises, it is important to discern the principles underlying his selection of the subordinate biblical lemmata (secondary texts). On occasion, “the sequence of the biblical texts used for illustratory purposes” is mainly determined by the “diaretic schema” with which he begins his explanation of the main biblical lemma or by “the analysis of a particular term.” Runia is particularly concerned with what he calls two basic types of “modes of transition”. The first “is a verbal mode [in which a] word or phrase in the main [or subordinate] biblical lemma catches Philo’s attention and prompts him to recall another biblical passage where that same word or phrase also occurs”48. Runia subsequently observes that this “enchaining of biblical texts, especially on the basis of verbal parallels, clearly has its roots in Jewish exegesis”49. The second “is a thematic mode” where “[a] theme or topic raised in the main [or subordinate] biblical lemma causes Philo to think of another 45. D.T. RUNIA, The Structure of Philo’s Allegorical Treatises: A Review of Two Recent Studies, in VC 38 (1984) 209-256, and ID., Further Observations (n. 42). 46. RUNIA, Structure (n. 45), p. 237. 47. Ibid., p. 238. 48. Ibid., p. 239. Cf. K. SCHENCK, Old Testament Context: Insights from Philo, in D. ALLEN – S. SMITH (eds.), Methodology in the Use of the Old Testament in the New: Context and Criteria (LNTS, 549), London – New York, Bloomsbury T&T Clark, 2020, 103-114, pp. 110-112. 49. RUNIA, Further Observations (n. 42), p. 120.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

209

biblical text which contains the same theme or can be used to illustrate that theme”50. In the portion of Quod Deus Runia examines, he finds both types of transition employed equally. As an example of the first (verbal) type, Runia cites the comparison of Noah with Moses and Joseph at Deus 109–116. Here the main biblical lemma being expounded is Gen 6,8, cited at Deus 104 as Νῶε εὗρε χάριν παρὰ τῷ θεῷ although previously at Deus 86 Philo’s wording ἐναντίον κυρίου τοῦ θεοῦ more accurately reflects the biblical original. Philo connects this to Moses via the verbal links in Exod 33,17 (εὕρηκας γὰρ χάριν ἐνώπιόν μου) and to Joseph by those in Gen 39,21 (ἔδωκεν αὐτῷ χάριν ἐναντίον τοῦ ἀρχιδεσμοφύλακος)51. Here one might also see a thematic link (Runia’s second type) in the notion of receiving grace, or of different degrees of worthiness52. Another example of a thematic link is afforded by the transition to material concerning the Nazirite vow from Num 6,2-12 at Deus 86–90. In the course of focusing on the word εὗρεν at Gen 6,8 Philo proceeds to contrast discovery and rediscovery (a defiled Nazarite has to “rediscover” the spiritual commitment of his vow)53. The Migration of Abraham begins by citing God’s command to Abraham to leave his land (ἐκ τῆς γῆς σου), his kindred and his father’s house (Gen 12,1-3), which Philo goes on to interpret as meaning removal from body, sense perception and speech. Philo proceeds to support the first of these equations by citing Gen 3,19 to show that the body took its substance from the earth or land: γῆ εἶ καὶ εἰς γῆν ἀπελεύσῃ. Once again here there is a verbal link, via γῆ, but this clearly is not some random association in Philo’s mind; Gen 3,19 has been carefully selected to underpin Philo’s assertion that ‘land’ is a symbol for ‘body’. His choice of subordinate lemma has been guided by the schema he has chosen for expounding the primary lemma. Philo’s next secondary lemma, Gen 28,17 (οὐκ ἔστιν τοῦτο ἀλλ’ ἢ οἴκος θεοῦ) cited at Migr. 5, follows the same principle, this time via the word οἶκος, to support the association of father’s house with speech, Philo’s point here being that ‘this’, the ostensible world of sense perception around Jacob, was not the house of God, which is invisible and can only be apprehended by the soul. Philo is here contrasting the human mind which has speech (λόγος) for its house with God’s mind, which has his Word (λόγος) for his house. 50. ID., Structure (n. 45), p. 240. 51. Ibid. 52. BORGEN, Philo (n. 44), pp. 117-118. 53. Ibid., p. 117; RUNIA, Structure (n. 45), p. 240.

210

E. EVE

The third secondary lemma, “pay attention to yourself” (πρόσεχε σεαυτῷ, Exod 24,12, cited at Migr. 8) is employed in support of Philo’s point that leaving body, sense perception and speech does not imply complete severance, since this would mean death, but rather rising above them and not allowing oneself to be ruled by them, thereby coming to know yourself (γίνωσκε σεαυτόν). The verbal and conceptual links here are not with the biblical text Philo is seeking to exegete but with his own exegetical explanation. Our final example from Migration occurs at Migr. 13, where Philo cites Gen 13,9, “separate yourself from me”, said by Abraham to Lot. Philo takes it as read that among the Hebrews, Lot stands for the soul that inclines towards sense-perception. The link here appears to be a purely conceptual one. In the main lemma, Abraham is commanded to depart from his land, his kindred, and his father’s house, the second of these taken by Philo to symbolize sense-perception. Now here Abraham commands Lot, symbolizing an inclination to sense-perception, to separate himself from him. What Philo has done is to find another secondary lemma that fits his interpretive scheme. In sum, Runia’s first mode is akin to the catchword association identified by Patterson as a structuring device in GTh, while his second resembles the continuity of image or topic that occasionally links material in GTh. But there are differences. Philo’s overall structuring is far more elaborate than that of GTh. Philo seeks to bring out hidden meanings he sees lying beneath the surface of the text, while GTh leaves its audience to discern these hidden meanings for themselves, and so lacks anything corresponding to Philo’s extensive explanations. The catchword and topic connections in GTh seem to be little more than mnemonic conveniences, whereas the verbal and conceptual connections in Philo serve his larger design. Despite these differences, roughly similar kinds of associative cueing appear to have been at work in these two quite different kinds of writing, and if GTh is likely to have relied on memory, the same seems true of Philo as well. Runia entertains the possibility that Philo employed a concordance to locate verbal parallels, but he nevertheless believes that “Philo’s method of structuring his treatises … requires at the very least an intimate knowledge of the Pentateuch and a formidable memory”54. Or as Runia says in his later article, “Philo appears to have possessed a highly associative mind, whose thought is quickly ‘triggered’ by parallels (verbal and thematic) and analogies”, although he at once goes on to add, “Nevertheless 54. Ibid., p. 245.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

211

I am persuaded that the chief reason for Philo’s concatenative tendency is literary, and is probably derived from literary theory and practice”55. Philo illustrates associative cueing even more clearly than GTh. It will hardly be news that Philo employs exegetical techniques that are common elsewhere in Second Temple and Rabbinic Judaism, or that these in turn frequently employ thematic and verbal links56. The point here is to illustrate how such mnemonic associations could contribute to the structuring of extended compositions, and that what I have called associative cueing is by no means a purely passive process of following whatever links may happen to come to mind but may be a highly active one of seeking out material to suit one’s purpose. IV. APPLICATION TO LUKE Neither GTh nor Philo provides an exact parallel with FH Luke’s source utilization, but each of them has enough in common with Luke that together they bracket FH Luke’s proposed use of associative memory to construct sequences that depart from those of his sources. The source material employed by GTh is similar in kind to that used by Luke, and if GTh is indeed dependent on the Synoptic Gospels, in some cases it may be the very same source material. Where Luke adapts his sources, his adaptations are often similar in kind to those made in GTh, although GTh’s adaptations are often more radical. Luke transforms the order, content and wording of his source material to give his own slant on the material he uses, while presenting a figure who is recognizably the same Jesus as the figure depicted in Matthew and Mark, who might thus reasonably be expected to be acceptable to audiences for whom Mark and Matthew are authoritative57. GTh appears to be both borrowing and subverting the authority of his synoptic source material to present a Jesus who bears sufficient resemblance to his synoptic counterpart to be taken as the same figure but who nevertheless takes on a radically different significance. If (as suggested above) GTh is dependent on the synoptics, then it has deliberately decontextualized its synoptic source material while interweaving it with other material of its own. There is some parallel here with the way in 55. ID., Further Observations (n. 42), p. 130. 56. See, e.g., G. STEMBERGER, Introduction to the Talmud and Midrash, Edinburgh, T&T Clark, 21996, pp. 18-20. 57. See further EVE, Relating (n. 1), pp. 195-198.

212

E. EVE

which FH Luke will have decontextualized material taken from Matthean discourses (and occasionally other Matthean contexts) and placed it in his own arrangement alongside his own special material (notably in his central section). That Luke retains more of Matthew’s relative ordering of the double tradition material than GTh retains of its source material does not detract from this parallel. If GTh can carry out its reordering of synoptic material through memory, then so can FH Luke. ‘Reordering’ may, however, be a misleading term in both instances since neither FH Luke nor GTh attempts to incorporate the entirety of their source material in a different sequence. Instead, both use a relatively small selection (in relation to the whole) of their source material out of the sequence in which it appears in their sources, together with some additional material peculiar to themselves. In the case of GTh this appears to have been facilitated by the use of catchwords, and occasionally common themes or images, to both select and arrange the material. Both types of linkage constitute associative cueing. Philo’s allegorical commentaries are a substantially different kind of writing from either the Gospel of Luke or the Gospel of Thomas58. The biggest difference is that of genre: Philo is writing a commentary on a sacred text whereas Luke writes a narrative based on source material he is adapting. A major consequence of this is that a far higher proportion of Philo’s prose is his own, as opposed to material taken from a source. This distinction would be diminished to some extent if Luke composed the bulk of his L material and GTh the bulk of the material peculiar to itself, but even if this were the case (and it may be well not be)59 the distinction would be far from eliminated. A second difference is that Philo tends to either explicitly quote or paraphrase his source material to comment on it or appeal to it for support rather than reshaping it for use in a composition of the same kind from which it was taken without explicit attribution. Philo’s employment of source material is thus different in kind from that of Luke or GTh. That said, Luke and Philo share some features that are not found (or not so clearly found) in GTh. For one thing, Luke and Philo are both far more concerned than GTh to order their material in a manner that serves their themes, although both can occasionally go off in directions that seem puzzling. Philo’s sequence is controlled first of all by the text on which he commenting, so that 58. Philo’s Life of Moses might provide a closer parallel to Luke’s Gospel, for which see ibid., pp. 46-50, but his allegorical commentaries are more instructive for the issues of memory and order being examined here. 59. In the case of Luke, see M. GOODACRE, Goulder and the Gospels: An Examination of a New Paradigm (SupplJSNT, 133), Sheffield, Sheffield Academic Press, pp. 132-191, esp. 281-291, and EVE, Relating (n. 1), p. 174.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

213

his primary lemmas are cited in the order in which they appear in Genesis, and second by a combination of the interpretive scheme Philo wishes to apply in his interpretation and the structure of the primary lemma being interpreted (so that, for example, his commentary at Migr. 1–13 is controlled by his discussion of the symbolic meaning of the terms “land”, “father’s house”, and “kindred” in his primary lemma). In an analogous manner, FH Luke’s order is primarily controlled by that of the source he is currently following (alternately Mark or Matthew), and just as Philo always eventually returns to the sequence of his primary text, so in his central section FH Luke sooner or later always returns to Matthew’s sequence60. And just as the order of the material (his own and his secondary lemmas) Philo deploys between his citations of his primary lemmas is at least in part controlled by his interpretive schema (and hence the points he wants to get across), so Luke’s sequences of (L and reordered Matthean) material in his central section is at least in part controlled by the particular themes he wishes to cover, such as prayer, riches and repentance. While Luke’s current themes controls which material he deploys, the order of material in Matthew may in turn influence which theme he turns to next61. This leads to another respect in which Luke may be closer to Philo than to GTh, namely in the sophistication of his associative cueing. Whereas the catchwords employed by GTh appear to be little more than a mnemonic convenience for sequencing otherwise unstructured material, both Luke and Philo employ verbal and thematic cues to retrieve material that has some direct relevance or appropriateness to their designs. This may be illustrated with a selection of examples. We shall begin with the most straightforward, FH Luke’s use of Matt 6,9-13 and 7,7-11 at Luke 11,1-4.9-13. In the wider context Luke appears to be constructing a sequence concerning three fundamental aspects of discipleship: love of neighbour, listening to the Lord, and prayer62. At Luke 11,1-13 he comes to the last of these, his short discourse on prayer, comprised of his version of the Lord’s Prayer (taken from Matt 6,9-13) and the Encouragement to Pray (taken from Matt 7,7-11) surrounding the L Parable of the Importunate Friend at Midnight (Luke 11,5-8). Here the links are thematic: Luke selects two items from Matthew’s Sermon on the Mount related to prayer and adds a third, which he either composes himself for the purpose, or, more likely, takes from the tradition available to him. Compared with the 60. Ibid., pp. 171-172. 61. Ibid., pp. 170-171. 62. Ibid., pp. 176-178.

214

E. EVE

way Philo gleans thematically relevant material from across the Pentateuch (and occasionally beyond), this would hardly seem to place any implausible strain on Luke’s memory. A second example is Luke’s deployment of two more items from the Sermon on the Mount, the Light on the Stand (Luke 11,33//Matt 5,13) and the Sound Eye (Luke 11,34-36//Matt 6,22) following Luke’s version of the sequence incorporating the Beelzebul Controversy and the Sign of Jonah (in which Luke has broadly used Matt 12,22-45 at the point he has reached following Matthew’s overall sequence). The broad connection of thought here is that Jesus constitutes a sign that people ought to be able perceive but which many do not63. In addition to this overall thematic connection, the two passages taken from the Sermon on the Mount are linked verbally through the words light and lamp (φῶς/φωτεινός and λύχνος). Luke’s technique here thus corresponds to both kinds of transition Runia identifies in Philo, and presumably relies on the same kind of associative memory. This leads straight into our third example, namely FH Luke’s bringing forward his reworking of Matthew’s Woes on the Scribes and Pharisees (Matt 23,24-26) to Luke 11,39-52. As noted above, FH Luke has just been employing material from Matthew 12 in its Matthean sequence, but he has not yet made use of Matt 12,33-37, in which Jesus condemns his audience (identified at Matt 12,24.38 as Pharisees) for their hypocrisy: “how can you speak good, when you are evil?” Luke 11,39-40 and Matt 12,33-35 both thus contrast outward show with inner depravity. In what we might term an indirect or hidden parallel FH Luke has here seemingly taken one passage in Matthew as a prompt for introducing material from another64. The link here would be a purely thematic one. A similar process attends Luke’s employment of Matthew 13. Luke 13,18-21 employs Matt 13,31-33 directly (in its Matthean sequence) in the Parables of the Mustard Seed and the Leaven, but it is notable how much of the surrounding Lukan material (in Luke 12,1–13,29) echoes themes and language from the surrounding Matthean material (in Matt 13,1-50). Thematically,

63. Ibid., pp. 178-179. 64. Ibid., p. 179. Cf. A. DAMM, Ancient Rhetoric and the Synoptic Problem: Clarifying Markan Priority (BETL, 252), Leuven – Paris – Walpole, MA, Peeters, 2013, pp. 272-273 and the suggestion by M.A. MATSON, Luke’s Rewriting of the Sermon on the Mount, in M.S. GOODACRE – N. PERRIN (eds.), Questioning Q: A Multidimensional Critique, Downers Grove, IL, InterVarsity Press; London, SPCK, 2004, 43-70, pp. 55-56, that Luke used the Pharisees and Scribes to represent the dark side of the contrast between light and darkness running through Luke 11,29-54.

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

215

this includes the eschatological thrust of both Matt 13,30.36-42 and Luke 12,35–13,9; 13,22-30. A striking linguistic link is provided by the uncharacteristic (for Luke) “There you will weep and gnash your teeth” at Luke 13,28, which suggests a verbal prompt from the “weeping and gnashing of teeth” at Matt 13,50 to Luke’s use of Matt 8,11-12. The admonition against anxiety at Luke 12,22-32 parallels the similar passage at Matt 6,25-34 but shares with the Parable of the Wheat and Tares at Matt 13,24-30 the image of burning unwanted vegetation, raising the possibility of an associative link via common imagery65. Our final example is one noted in the introduction to this paper, namely the Parable of the Good Samaritan and Luke’s introduction to it at Luke 10,25-37. If we picture Luke working forwards through Matthew in his central section, then this could be the point at which Luke reaches Matthew’s version of Plucking Grain on the Sabbath at Matt 11,1-766. He has already used Mark’s version of this in its Markan sequence at Luke 6,1-5, but Matthew has made a couple of additions to Mark’s version which seem to be echoed in this section of Luke. One of these is Matthew’s citation of Hos 6,6, “I desire mercy and not sacrifice” (Ἔλεος θέλω καὶ οὐ θυσίαν). Not only is the keyword ‘mercy’ echoed in the reply Luke’s Jesus elicits from the lawyer that the person who acted as neighbour was the one who showed mercy (ὁ ποιήσας τὸ ἔλεος) but the entire parable could be seen as exemplifying the quotation from Hosea (even if Luke ostensibly uses it to make a different point)67. Moreover, while the lawyer’s question that introduces the parable in Luke may be loosely based on Mark 12,28-31, the counter-question Luke alone assigns to Jesus at Luke 10,26 (ἐν τῷ νόμῳ τί γέγραπται; πῶς ἀναγινώσκεις;) resembles the addition by Matthew at Matt 12,5 of Jesus’ question ἢ οὐκ ἀνέγνωτε ἐν τῷ νόμῳ; in both cases Jesus questions what his interlocutor has read in the Law68. To the extent that Luke has employed pre-existing material here, this would be analogous to Philo’s use of verbal 65. EVE, Relating (n. 1), pp. 181-182. On the wailing and gnashing of teeth here, see also M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, pp. 574-575. 66. Luke has just been in direct sequential parallel with Matthew at Luke 10,21-24// Matt 11,25-27 and will next come into direct sequential parallel at Luke 11,14-32// Matt 12,22-45. 67. If it seemed at all probable that Luke might have known something like the Hebrew version of the wider context of this verse from Hosea, it would be tempting to argue that the Parable of the Good Samaritan had been further prompted by Hos 6,1 (binding up the traveller’s wounds) and Hos 6,9 (the robbers who attack him), but this does not work nearly so well with the LXX of Hos 6,9. 68. EVE, Relating (n. 1), p. 177; cf. GOULDER, Luke (n. 65), pp. 485-488.

216

E. EVE

and thematic links to select secondary lemmas. To the extent that Luke is composing (or adapting) his own material, it is somewhat analogous to the way in which Philo picks up words and themes from the texts he cites in the commentary he provides for it. In a sense, though, Luke and Philo would be working in opposite directions. Whereas Philo draws out what he takes to be the deeper meaning of a narrative text in a series of propositional statements, Luke either employs or constructs a brief narrative to illustrate the meaning of a key saying. But both authors nevertheless use similar processes of associative cueing. At this point it may be objected that Luke’s indirect (or hidden) parallels are unlike anything encountered in GTh and Philo. The catchwords and occasional thematic links between sayings in GTh are, necessarily, links between sayings we can see in the text. Similarly, in Philo the verbal and thematic links that apparently prompt the selection of secondary lemmas are visibly connected to the primary lemma Philo cites or to other features of his commentary. Several of the Lukan examples just discussed, however, rely not on verbal or thematic links with something else that appears in Luke’s text but on something Luke supposedly finds in one of his source texts (Matthew) but does not directly quote or adapt in his own. Yet while this means that we may not be able to prove that Luke was prompted by this or that text in Matthew, it does not materially affect the kind of associative memory cueing Luke would be employing if he were working from that Matthean text, which is the point being argued here. That an underlying Matthean text is not directly employed by Luke does not mean that it could not form the basis of memory cueing of the sort employed by Philo (and to a lesser extent GTh) from the material they visibly cite. The point has been to show, or at least to illustrate, that the memory operations FH Luke would need to have employed to use Matthean material out of its Matthean context and sequence are by no means implausible in the light of the kinds of memory cueing Philo and GTh most probably required to employ their source material out of sequence. V. CONCLUSION This paper has had a limited aim. It does not aim either to demonstrate Luke’s use of Matthew or to account for Luke’s ordering of double tradition material on the assumption that Luke derived it from Matthew, but merely to show that Luke’s use of a limited selection of Matthean material (mainly from Matthew’s discourses) out of its Matthean sequence would have been

MEMORY AND ORDER IN THOMAS, PHILO, AND LUKE

217

perfectly possible using the kinds of associative memory cueing employed by other authors at the same time who similarly incorporated source material out of its original sequence. Neither GTh nor Philo provides a precise parallel to FH Luke, but each has sufficient in common with Luke that together they bracket what FH Luke would need to have done with Matthew (and may also have done with his L material) to arrive at the order he did. It thus appears that Luke’s reordering of Matthean material (and deployment of L material) based on a combination of verbal and thematic memory cues would have been perfectly feasible. Harris Manchester College, Oxford United Kingdom

Eric EVE [email protected]

CAN MEMORY RESCUE THE FARRER HYPOTHESIS? THE TEMPLE SCROLL AND FH LUKE’S MATTHEW-UTILIZATION

The perennial difficulty for hypotheses that posit Luke’s use of Matthew has been accounting for Luke’s rather large-scale rearrangements of his Matthean materials, especially in the central travel section. In Farrer Hypothesis (FH) scholarship this has typically been treated as task for redaction-critical ingenuity, but more recently, as our awareness of ancient media realities has grown, the difficulties of the actual source-utilization operations involved have moved to the forefront. These lie not just in the search and location operations involved, but also in the seeming randomness of FH Luke’s utilization movements around his Matthew scroll. My interest in the FH problem comes from my effort to address its counterpart in the Two-Document Hypothesis (2DH): Matthew’s rearrangement of the Q materials. I argued that in line with well-attested ancient practices Q’s topical sequence was exploited by Matthew as his mnemonic for navigating around the source in a targeted search for suitable materials, guided by his own topical schema. As it did generally in the ancient world, memory-based utilization practices helped overcome the difficulties of so-called random access to materials in Matthew’s source scroll1. I have long wondered whether memory-utilization practices might help the Farrer Hypothesis finally crack the problem of Luke’s use of Matthew. Farrer advocates have suggested as much. “Sauce … for the goose”, Mark Goodacre observes, “is sauce for the gander”2. The quest is to find an analogy to what seems Luke’s rather unsystematic movements around Matthew (in contrast to Matthew’s more systematic movements around the hypothesized Q source). The scale of the challenge that this presents cannot be 1. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016. 2. M. GOODACRE, Q, Memory, and Matthew: A Response to Alan Kirk, in JSHJ 15 (2017) 224-233, p. 228.

220

A. KIRK

better illustrated than by Michael Goulder’s rather convoluted scrolling account, which notwithstanding its disavowal by contemporary Farrer scholars constituted a serious attempt to come to grips with the sourceutilization problematic3. The Temple Scroll, a prominent Qumran work dating from around the middle of the second century B.C., seems to offer just such a promising analogy to FH Luke’s use of Matthew. Its composition involves wide ranging movements, primarily within the Pentateuch but across the biblical corpus, without any particular reliance on the arrangement of the biblical works that it exploits with such sovereign freedom in the pursuit of its own literary and hermeneutical purposes (different 2DH Matthew with Q). To all appearances this utilization is random. That the composer’s expert source memory is instrumental in all of this is as good as certain. The Temple Scroll has been the subject of a number of compositional studies. We will explore the source utilization techniques employed in the Temple Scroll, and then see how these line up with FH Luke’s use of Matthew.

I. MEMORY-BASED COMPOSITION IN

THE

TEMPLE SCROLL

Though the Temple Scroll was once thought to be a redactional conglomerate of earlier sources, contemporary scholarship has increasingly come to view it as a unified composition, though doubtless the product of a long tradition of intra-Torah legal exegesis4. Like other Qumran scribes (and highly trained Second Temple scribes in general) its author exercised a thorough memory competence in the biblical sources. This competence operated at the level of an active compositional capability: scribes could re-express the written memory tradition fluently; it existed as a register for re-enactment5. In Dwight Swanson’s words, “We are in fact presented in the [Temple] 3. Chiefly in M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989. We will have occasion to discuss Goulder’s appeal to memory later in this essay. Goodacre for his part is generally content to subsume the utilization question to Luke’s narrative art, to the intelligibility of Luke’s literary arrangements (see, e.g., GOODACRE, Memory [n. 2], p. 229). That is, he continues to treat it as a question for redaction criticism. 4. M.M. ZAHN, 4QReworked Pentateuch C and the Literary Sources of the Temple Scroll: A New (Old) Proposal, in DSD 19 (2012) 133-158, pp. 137 [and n. 13], 155); ID., Rethinking Rewritten Scripture: Composition and Exegesis in the 4QReworked Pentateuch Manuscripts (STDJ, 95), Leiden – Boston, MA, Brill, 2011, p. 180, n. 2. 5. S. MILLER, Dead Sea Media: Orality, Textuality, and Memory in the Scrolls from the Judean Desert (STDJ, 129), Leiden – Boston, MA, Brill, 2019, pp. 230, 256-260, 265-266; M.M. ZAHN, Genres of Rewriting in Second Temple Judaism: Scribal Composition and Transmission, Cambridge, Cambridge University Press, 2020, pp. 32-33.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

221

Scroll with an example of the consummate skill of the ancient scribes in drawing from their deep and intimate knowledge of the scriptures to weave together the various strands of each topic in a new unity which both respects the tradition drawn upon and draws out new interpretations of it”6. This brings us to our probe question of how the TS composer executed his project of drawing materials at will from across the biblical corpus and recombining them. He did so by means of a clearly configured compositional schema. The TS takes as its compositional baseline the material that falls between Exodus 34 and Deut 23,1. It “reorders and ignores” the materials along this baseline “according to the requirements of its own plan”7. The materials in this stretch of the Pentateuch are the orientation points from which the author draws in materials from other parts of the biblical corpus. This is not “willy-nilly” but highly systematic. His principal mnemonic technique for pulling elements together from disparate source locations into a fluent compositional unity is topical 8. This is supported by a vividly drawn spatial-visual concentric schematic. Again Zahn: [F]or the most part [the TS] is organized spatially, proceeding outward from the Temple: first instructions are given for building the Temple itself and the altar. After the construction of the altar comes the list of festivals … Thereafter come instructions for the courts of the Temple, followed by rules ensuring the purity of the holy city in which the Temple is located and then governing the purity of the rest of the land of Israel9.

The topical technique and the visual-spatial schematic function together in tight mnemonic tandem. In his treatment of the Yom Kippur festival, for instance, “the author gathers and systematizes all the relevant pieces of the Torah that deal with this festival and creates a more legally coherent set of prescriptions”10. The baseline for the Festival Law section itself (Columns 3-49) is the festival sequence in Numbers 28–29, and it takes the biblical order as its spine: “Tamid, Sabbath offering, first of the month, first of the year, Passover, Unleavened Bread, the Omer, First-fruits, first

6. D.D. SWANSON, The Temple Scroll and the Bible: The Methodology of the 11QT (STDJ, 14), Leiden – New York – Köln, Brill, 1995, p. 15. 7. Ibid., p. 5. 8. ZAHN, Genres (n. 5), p. 188; J. VROOM, The Authority of Law in the Hebrew Bible and Early Judaism (SuppJSJ, 187), Leiden – Boston, MA, Brill, 2018, pp. 105-107. 9. ZAHN, Scripture (n. 4), pp. 227-228; similarly, H. STEGEMANN, The Literary Composition of the Scroll, in G.J. BROOKE (ed.), Temple Scroll Studies: Papers Presented at the International Symposium on the Temple Scroll Manchester, December 1987 (SuppJSP, 7), Sheffield, Sheffield Academic Press, 1989, 123-148, p. 134. 10. VROOM, Authority of Law (n. 8), p. 101.

222

A. KIRK

of the seventh month, Day of Atonement, and the Feast of Booths”11. The Law of the King (Columns 56-59) is a section of halakhic innovations that as such stands outside the schematic framework that structures the bulk of the TS. However, it has as a base text Deut 17,14-20, and from this fulcrum the author pulls materials topically (“kingship”) from various other scriptural locations, extending these halakhically through application of his legal hermeneutic. These sub-topics are themselves distinct and arranged coherently, e.g., law on the royal guard, on the council, on the queen12. Supplementary texts feeding into the sub-topical composition – indicated, Swanson says, “by the use of word-form, or by phraseology, and not only by word-association (or, key-word link)” – are “uniformly pertinent to the subject of the section in which they appear”13. He describes the sub-topic, “meeting an invading enemy”, as “a mosaic of biblical contexts which have to do with battle”14. To the base text of the law of the queen, Deut 17,17, the author draws together thematically pertinent materials from Genesis 24, Proverbs 31, and for its critique of polygamy and intermarriage apropos connections to David and Solomon15. The facility with which the author conflates textual elements from diverse biblical sources into a new text is so fluent that scholars can sometimes disagree in their identification of the scriptural elements that are feeding into the composition. Column 59,5b-7a (from the Law of the King) runs as follows: ‫ומזעיקים מפני עול כבד וקראו‬/‫והמה בארצות אויביהמה מתאנחים‬ ‫אותמה מפני רוע מעלליהמה‬/‫ולוא אשמע וזעקו ולוא אענה‬, which Swanson analyzes as follows: The opening phrase is from Lev 26:36 ‫ … בארצח איביהם‬and the final phrase in line 7 returns to Deuteronomy 28 at verse 20 ‫ מּפני רע מעלליך‬merged with Jer 44:22 (etc.). In between is a most skillful interweaving of texts built on the base of Exod 2:23, which picks up the link from Lam 1:11, ‘groan’ ‫ אנח‬and in turn provides the verb “cry out” ‫ זעק‬as the key-link word. Lines 5-7 contain these key words (highlighted text): “And in the land of their enemies they will groan and cry out because of a heavy yoke and they will call, and I will not hear; and they will cry out and I will not answer them.” Use of Exod 2:23 11. SWANSON, Temple Scroll (n. 6), p. 17. 12. P.R. CALLAWAY, Extending Divine Revelation: Micro-Compositional Strategies in the Temple Scroll, in BROOKE (ed.), Temple Scroll Studies (n. 9), 149-162, pp. 158-159; also ZAHN, Genres (n. 5), pp. 190-191. 13. SWANSON, Temple Scroll (n. 6), p. 168. 14. Ibid., p. 146. 15. Ibid., p. 139; see also M.M. ZAHN, Identifying Reuse of Scripture in the Temple Scroll: Some Methodological Reflections, in E.F. MASON (ed.), A Teacher for All Generations: Essays in Honor of James C. VanderKam (SuppJSJ, 153), Leiden – Boston, MA, Brill, 2011, 341-358, p. 357.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

223

evokes the image of Israel in Egypt, “The children of Israel groaned under their bondage and cried out for help…” Three additional texts are woven together: God warns Jeremiah in 11:11, “Though they cry out to me, I will not listen to them”; Zech 7:13 says, “As I called and they would not listen, so they called and I would not listen”; and Micah 3:4, “Then they will cry to the Lord, but he will not answer them.” … No one text contains all the ingredients of lines 5-7, but combined they do. Jeremiah says “they cry out” and “I will not listen”; Zechariah does not have ‘cry’ but adds “they call” to “I will not listen”; Micah does have ‘cry’ and adds “I will not answer them”16.

For her part Zahn identifies as the source texts Lev 26,39 (“they shall rot in their iniquities in the lands of your enemies”), Exod 2,23 (“the Israelites groaned because of their labor, and they cried out”), Zech 7,13 (“and they shall call, but I shall not hear”), Jer 11,11 (“and they shall cry out, but I will not listen to them”), and Jer 21,12//26,3 (“because of the wickedness of their deeds”), so overlapping with but also diverging from Swanson’s determinations17. In any case, one can see how key words are the active factor in cuing these texts and pulling their elements together, with Exod 2,23 acting as a base text. The operative procedure for this here and elsewhere in the TS, however, is anything but spontaneous; it is highly systematized memory technique that exploits the topos principle and the visual-spatial principle. The signal analysis of the textual technique is by Swanson. The TS author lays down a Pentateuchal base text that establishes the subject matter, the topic, of a given section. With this base text is integrated, “by merging, harmonizing, or weaving”, one or more secondary texts. Most often these are other Pentateuchal passages that run topically parallel to the base text. The intent is to bring the one into alignment with the other (and, as Vroom shows, thereby eliminate legal ambiguity). Contributing vocabulary to this emerging compositional entity are topically apposite supplementary texts; these are associated by allusion through topically significant key word links (e.g., as in the above). Though they can also be Pentateuchal it is in the supplementary texts that the author can be found navigating the biblical corpus more widely, with the topos marker of the Pentateuchal base text operating as his mnemonic direction-finder. Though evoked allusively these supplementary texts are of great hermeneutical import, for it is in them that the author’s interpretative stance, his “special concerns”, vis-à-vis the halakhic issues at stake in the combined base and secondary text “is most

16. SWANSON, Temple Scroll (n. 6), pp. 153-154. 17. ZAHN, Genres (n. 5), p. 188.

224

A. KIRK

clearly revealed”18. In the Festival Law section, for example, Leviticus 23 serves as the base text and the Numbers 28 parallel the secondary text. “Leviticus 23 is the text which ‘attracts’ all the other texts”19. In Column 18, lines 2-10, the base text is Lev 23,10-15 (First Fruits), and the secondary text Num 28,26-30. The sequence of the base text in TS 18,2-10 is Lev 23,12; Lev 12,14; Lev 23,13; Lev 23,14; Lev 23,12. Into these elements are woven supplemental textual elements and allusions from Num 28,27; Num 15,11; Num 28,30; Lev 9,3; Num 28,31; Num 15,24; Ezra 8,35; Num 15,4; Exod 29,40; Num 15,5; Lev 16,33; Num 15,25; Lev 14,21; Lev 16,34; Num 15,15; Lev 17,7; Lev 17,820. One notes here the high recurrence of Numbers 15; Leviticus 16; Leviticus 17, with their instructions on sacrifices. The application of this compositional technique is not just a showy display of scribal virtuosity. It issues from media conditions and from the corresponding recruitment of memory for source-utilization. The TS author’s retrieval movements around the biblical sources, moreover, are not extempore but mnemonically directed by the duo-dimensional topical and visualspatial arrangement of his own work. This arrangement is itself the clear reflection of the author’s specific program: to overcome the legal disarray of the Pentateuch, which “scatters laws on similar topics throughout the legal corpora”, by “rearrang[ing] pentateuchal law to create a more topically organized law code”21. The Pentateuchal source materials are accessed and re-indexed by means of the topical progression and concentric spatial layout of the TS. These rubrics harness the powerful associational properties of memory to enable the locating and combining of source materials into the new composition.

II. FH LUKE: GOULDER We now turn to our lead question of whether source-utilization by the author of the Temple Scroll can supply the much-needed analogy to FH Luke’s utilization of Matthew. The place to begin is with Goulder’s detailed response to the source-utilization question. To be clear, we are not attempting here to hang Goulder’s idiosyncratic and widely panned account around the neck of contemporary Farrer scholarship. The point rather is 18. SWANSON, Temple Scroll (n. 6), pp. 14-15, also 228-229; see VROOM, Authority of Law (n. 8), pp. 124-125, 204. 19. Ibid., p. 46. 20. Ibid., p. 27. 21. ZAHN, Scripture (n. 4), p. 20 (emphasis original); VROOM, Authority of Law (n. 8), p. 204.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

225

that Goulder took the full measure of the difficulties and tried to come to grips with them to an extent not attempted again in FH scholarship until Eric Eve’s recent reprise of Goulder’s effort. His account has the great virtue of bringing the source-utilization problematic fully into the light. Of particular relevance to our enquiry, Goulder calls upon memory, understood as “reminiscence”, to help explain FH Luke’s curious utilization movements around Matthew corresponding to his rather dramatic rearrangements of the Matthean materials. Notably, he construes the memory function in Luke’s source-utilization as a spontaneous reflex, triggered by some element or other in the text before Luke at whatever his visual location in his Matthew scroll happens to be at the moment. In composing 12,35-40, Luke’s eye location is Matt 16,27, the Son of Man’s Coming. Luke views this as happening at Passover, and so there “springs to his mind” the Exod 12,11 phrase, “thus shall you eat it, your loins girded”. The underlying idea of “readiness” in turn triggers an association to “lamp” in Matt 6,22: “the eye is the lamp of the body”. Thematically this has zero connection to the “readiness” motif, but it flashes upon Luke because he happens to be in the Matthew 6 neighborhood by memory: visually cued by “what shall a man profit…?” in Matt 16,25, he has just utilized Matt 6,19-21//Luke 12,32-34. Matt 6,22’s role therefore is merely to trigger an association to the bridesmaids’ lamps over in Matt 25,1-13. By this circuitous memory route Luke gets his phrase, “…and your lamps burning” for 12,35b. Thereupon, “other associated phrases from the [Matthew 25] context now leap to mind … ‘Blessed that servant whom his master coming shall find so doing’ (Matt 24,46)”22. Goulder recreates this pinball series of memory cues, which takes Luke back to Matthew 6, forward to Matthew 25, then backward to Matthew 24, because Luke’s visual location in his Matthew scroll at the time is, as noted, Matthew 16. (This bi-locality contrivance will reappear in Eve’s theory of Luke’s “indirect use” of Matthew while composing the travel section). Goulder must nevertheless derive Luke 12,23-40 – somehow – from Matthew, and Luke’s uncontrolled “reminiscence” of these other Matthean passages rationalizes the abruptness of Luke’s corresponding utilization movements. Similarly, in composing 6,44 (Trees and Fruit), though visually fastened at the moment on Matt 7,16a, Luke is drawing the wording by reminiscence from Matt 12,33, involuntarily and unconsciously: because he proceeds immediately to Matt 7,16b, it is Matt 7,16a “which was in Luke’s conscious mind”23.

22. GOULDER, Luke (n. 3), pp. 543-544 (emphasis added). 23. Ibid., p. 371.

226

A. KIRK

This ad hoc procedure of Goulder’s Luke, which does not even rise to the level of a “method”, finds no analogy in the highly systematized memory techniques for source utilization employed by the composer of the Temple Scroll. As with “the eye is the lamp of the body” above, often no associative thematic connection exists between the paired elements cued. The φραγμόν around the vineyard featured in the Wicked Husbandmen parable (Matt 21,33) is Goulder’s cue for the φραγμούς from which the homeless are fetched to the Great Supper (Luke 14,23)24. By what operational compositional conception, moreover, does Luke omit the Matthean materials that he goes back to and picks up later, spontaneously cued by whatever happens to be encountering in the course of following Matthew? To all appearances FH Luke has no well-defined topos sequence of his own that he can use to locate and re-index materials from their various Matthean contexts. This finds no better confirmation than in Goulder’s theory that Luke’s order in the latter part of the travel section is determined by a backward scrolling motion from Matthew 25, to boot a decision that Luke takes on the spot, with no premeditation25. In short, Goulder incorporates randomness and extemporaneity right into his account of Luke’s utilization of Matthew: they are features, not bugs. Aware at some level, however, of the need to find some constructive compositional conception that is guiding Luke and that rationalizes his FH utilization operations, Goulder experiments with the “revised lectionary” theory26. Luke associates the Coming of the Son of Man motif in Matt 16,27 with the “girded loins” of Exodus 12 (triggering the subsequent cascade of memory associations), for instance, because he is composing a Passover lection. After following Matthew for the Sermon and the Healing of the Centurion’s Servant (Matt 5,1–7,28; 8,5-13), it is for lectionary reasons that Luke departs from his procedure thus far of forward progress through Matthew and scrolls ahead to Matt 11,2-13 (Discourse on John the Baptist): he needs a New Year Sunday lection, and his community is accustomed to reflecting on John the Baptist and his fasting practices (i.e., Matt 11,18) at this season. When he thereupon leaves Matthew and rejoins the Markan sequence at Mark 4,1-20 (Parable of the Sower) at Luke 8,1 the pretext is again lectionary: he needs a Tabernacles reading and thus a harvest-themed 24. Ibid., p. 597. Luke then omits the Markan/Matthean φραγμόν in 20,9 (Mark 12,1// Matt 21,33)! 25. ID., The Order of a Crank, in C.M. TUCKETT (ed.), Synoptic Studies: The Ampleforth Conferences of 1982 and 1983 (SupplJSNT, 7), Sheffield, JSOT, 1984, 111-30, p. 121; ID., Luke (n. 3), p. 582. 26. Ibid., p. 170.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

227

pericope27. As he does with “reminiscence”, Goulder invokes Luke’s lectionary project ad hoc, i.e., when he needs a pretext for a particular utilization maneuver vis-à-vis Luke’s sources Mark and Matthew. It all but disappears, for example, in his treatment of Luke’s composition in the travel section, notoriously the graveyard for theories of Luke’s use of Matthew. The lectionary theory, spontaneous reminiscence, and even the reversescrolling expedient are serious attempts to identify a system in Luke’s Matthew-utilization. But even as regards the last one, when one takes a closer look at the actual sequence that constitutes Luke’s reverse scrolling action in the travel section, numerous of the connections between the proposed Matthean cues and the Lukan parallels in question turn out to be far-fetched and contrived, and not really uni-directional at all28. In other words, upon examination the proposed systematic scrolling movement dissolves into a random utilization pattern. Again, the point is not to pin Goulder’s proposals onto the contemporary defenders of the Farrer Hypothesis but to highlight the scope of the problem for the FH, of which Goulder was presciently cognizant. The tight system we have seen to be at work in the Temple Scroll’s source-utilization makes this difficulty for the FH even more acute.

III. FH LUKE: EVE The response of Farrer scholarship post-Goulder, however, has been to do little more than gesture at the problem (if that) until Eric Eve’s recent unflinching confrontation with it in his 2021 volume Relating the Gospels: Memory, Imitation, and the Farrer Hypothesis. Eve brings to the question an expertise in ancient media realities and practices that Goulder lacked, but for that very reason also a Goulder-like urgency about the need to find a fit for the FH within those realities and practices. What gives Eve’s analysis its promise is that in line with ancient practice, he assigns memory a far more expansive role in Luke’s source utilization than Goulder did. His large-scale recourse to Luke’s memory competence in Matthew, unprecedented in FH scholarship, gives our Temple Scroll analogy particular salience in this case. 27. Ibid., pp. 387, 407. 28. See R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, pp. 195-196; ID., Greco-Roman Writing Practices and Luke’s Gospel: Revisiting ‘The Order of a Crank’, in C.A. ROLLSTON (ed.), The Gospels According to Michael Goulder: A North American Response, Harrisburg, PA, Trinity Press International, 2002, 61-83, pp. 78-79.

228

A. KIRK

Eve’s application of Luke’s memory expertise in Matthew to the sourceutilization question allows him to eliminate certain unpalatable features of Goulder’s account, in particular the reverse-scrolling proposal. That said, the extent to which Eve’s account is a reprise of Goulder’s is both noteworthy and revealing. Like Goulder, he has extensive recourse to spontaneous memory associations – buttressed by appeal to research on the cognitive phenomenon of associative memory cuing – to account for a large number of Luke’s rather unsystematic utilization movements. The two sayings that are found juxtaposed in Luke 6,39-40 are a case in point. The first, Blind Leading the Blind (Luke 6,39//Matt 15,14) is found as a Matthean addition embedded the Markan dispute with the Pharisees over handwashing and ritual purity (Matt 15,1-20), i.e., in a narrative with no thematic relation to Matt 7,1-5, Luke’s visual, scroll-based location in Matthew at the moment. The saying, Eve argues, therefore gets associatively “prompted” in Luke’s memory by the speck-/log-in-the-eye images in Matt 7,3-5. In an associative cascade this saying in its turn prompts recollection of the Disciple and Teacher (Matt 10,24-25), found in the Matthew 10 mission segment on persecution, the associative trigger being that “a teacher is a bettersighted guide”29. This solution comes at the cost of allowing quite a bit of spontaneity and lack of system into Luke’s utilization of Matthew. The Temple Scroll likewise combines passages from different source locations on associative principles, but it does so methodically, harnessing the associative powers of memory to a highly systematized compositional schema that is visual, spatial, and thematic. But as Eve acknowledges, a clear compositional schema is exactly what it is difficult to make out in Luke – and in the travel section in particular, ground zero for the FH utilization problem. There is, he admits, “no clear overall structure” to the travel section. It amounts to a half-formed topical menagerie, with materials tending “to cluster around topics such as discipleship, prayer, riches, criticism of the Pharisees, judgement, and the eschaton”30. These themes, though certainly linked, “cannot always be neatly separated out into distinct blocks and are often intertwined”31. In any case, “the rationale behind the resultant Lukan order appears less than immediately apparent”32. The upshot is that different from the author of the Temple Scroll, Luke is without a clearly demarcated schematic cue system for methodical utilization of the Matthean 29. E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021, pp. 161-162. 30. ID., Writing the Gospels: Composition and Memory, London, SPCK, 2016, p. 138. 31. ID., Relating (n. 29), p. 170. 32. Ibid., p. 145.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

229

source. This explains why Eve, like Goulder, adverts to spontaneous memory association and is forced to downplay the importance of structured mnemonic methods for searching out and retrieving requisite source materials33. The Gospel of Matthew of course is famous for its carefully configured arrangements. Beyond their aesthetic virtues these would facilitate its absorption to memory through habituated usage and, conversely, provide the organizational, schematic pathways for its memory-based utilization as a source. But curiously, FH Luke’s use of Matthew does not make use of those pathways. He does not appear to be using Matthew’s arrangement as his navigational search cues. This oddity is sharpened when one realizes that the Gospel of Matthew would have come to exist as a cognitive entity in FH Luke’s memory in the form of those constitutive organizational networks, internalized through its habituated usage. These networks would have been the essential basis – the script – for his activation (i.e., “performance”) of it as a source. Naturally we can assume that FH Luke could rove about the Gospel of Matthew at will, but as we have just seen, he lacks his own compositional schematic that his selective utilization movements around the source might be mapped to34. All he has is a “train of thought”, one that is admittedly “elusive”35. Hence his randomized utilization patterns and the corresponding Eve-Goulder recourse to memory spontaneity.

IV. THE TEMPLE SCROLL AND 2DH MATTHEW If there is an analogy to the memory-based utilization procedures of FH Luke it is unfortunately not to be found in the Temple Scroll36. But there is an evangelist whose procedures line up rather nicely with those of 33. Ibid., pp. 13-14, 151-152. In an earlier reflection on this issue Eve refers to Luke’s “mental jumbling of Matthew’s order,” which he hypothesizes is due to Luke’s reliance upon memory utilization. ID., The Synoptic Problem without Q?, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 551-570, p. 569. No doubt he would eschew the “jumbling” term today, but his underlying model for Luke’s memory use of Matthew is little changed in its essentials. 34. To evade this problem Eve makes a rather questionable distinction between “systematic” utilization of a source and “selective” utilization of a source, the latter supposedly correlating to randomized source access. ID., Relating (n. 29), p. 54. 35. Ibid., pp. 185, 190. 36. Eve puts forward Philo’s Life of Moses and Special Laws as analogies (ibid., pp. 46-50), but in both works Philo is relying on definite compositional schematics to guide his use of the Pentateuchal sources, analogously to the Temple Scroll author.

230

A. KIRK

the Temple Scroll: 2DH Matthew. Unlike FH Luke, 2DH Matthew exploits the topos-sequence and organization of his double tradition source. His utilization patterns accord with how the source would have existed as a cognitive entity within his memory. Moreover, exactly like the author of the Temple Scroll his retrieval movements within the source are directed by his own compositional schematic, a clearly contoured topos-sequence of his own devising37. In composing the Temple Scroll the author combines a “base text” with a “secondary text” at points where their two sequences run parallel. By this means he aligns the two texts with each other and brings them into a compositional unity. In Swanson’s words, it is a matter of “building upon a base text by conflation or merger with a closely related secondary text”38. In the Festival Law Leviticus 23 is the base text and Numbers 28 the parallel text. Leviticus 23 “is the text which ‘attracts’ all other texts”39. With differences in the respective source materials and cultural projects duly controlled for, this is exactly Matthew’s technique: he takes Mark as his base text and aligns Q materials to the Markan narrative at fitting notches in the Markan narrative spine to create a new work, the Gospel of Matthew40. Again, while certainly an impressive display of scribal virtuosity, this technique is dictated to both writers by media conditions and the corresponding need to rely upon memory instrumentality in source utilization. The thematic sequences of their sources in combination with Matthew’s own compositional schematic serve as associative attraction and attachment “operators” for the coherent integration of materials into a new work. In addition to the parallel secondary text, the Temple Scroll author establishes linkages to additional Pentateuchal texts and on occasion to texts from the Former and Latter Prophets. These “supplemental texts” likewise have lexical connections and thematic pertinence to the legal topic under development. Then there is also another group of additions that Swanson labels the author’s “expansions”; these are attributable to no biblical source41. These “supplementary texts” and “expansions” have a special significance: it is in them that the hermeneutical stance and “special concerns” of the author, “his particular viewpoint”, come most clearly to expression. The “expansions” can also be found fulfilling the editorial function of “creating 37. For full discussion see my Q in Matthew (n. 1), chapters 5 and 6. 38. SWANSON, Temple Scroll (n. 6), p. 215. 39. Ibid., pp. 46-47. 40. I oversimplify, since at times the Q sequence triggers realignments of the Markan sequence, principally between Matthew 8–12 (for analysis, see KIRK, Q in Matthew [n. 1]). 41. SWANSON, Temple Scroll (n. 6), p. 215.

CAN MEMORY RESCUE THE FARRER HYPOTHESIS?

231

a smooth-reading text”42. The alert reader will recognize in these “supplementary texts” and “expansions” an analogy to 2DH Matthew’s “M” materials. The M elements are diverse, but very frequently they serve to align the baseline Markan and Q materials to Matthew’s theological perspective. Also, they frequently function as editorial mortar binding Markan and Q materials to each other and into the larger framework of Matthew’s Gospel43. Finally, the remarkable fluency with which the author of Temple Scroll uses the internalized Pentateuchal and biblical traditions as the medium for his own expression, weaving phrases and lexemes into coherent compositional configurations and extending source elements into new tradition formations, finds its counterpart in Matthew’s taking the Q and Markan traditions up into his own language of thought and expression44. FH critics have pointed to overlap in motifs and vocabulary between Matthean and double tradition materials to argue that Q cannot be stylistically or theologically distinguished from Matthew45. The 2DH response to this – the double tradition becomes part of Matthew’s own expressive lexicon – finds good corroboration in the Temple Scroll composition. From a prolific number of cases, we will content ourselves with pointing to the phrase that commences TS 22,8: ‫אשי ריח ניחוח ליהוה‬: “an offering by fire, a pleasing odor to the Lord”, which, Zahn comments, “is used so often in P that its presence here cannot be traced to any particular verse, and probably just reflects the author’s familiarity with the formulaic language of sacrificial texts”46. Analogous expressions in Matthew would include “you of little faith” (Matt 6,30//Q-Luke 12,28; Matt 8,26; 14,31; 16,8; 17,20), “brood of vipers” (Matt 3,7//Q-Luke 3,7; Matt 12,34; Matt 23,33), “weeping and gnashing of teeth” (Matt 8,12//Q-Luke 13,28; Matt 13,42.50; 22,13; 24,51; 25,30), and “blind guides” (Matt 15,14//Q-Luke 6,39; Matt 23,16.24). Incidentally, that the double and triple tradition form Matthew’s internalized expressive language in a manner analogous to the Pentateuchal/biblical tradition for the composer of the Temple Scroll tells us quite a bit about normative status of Matthew’s written tradition.

42. Ibid., pp. 229-231. 43. E.g., Matt 9,27-32. 44. For analysis of the author’s extension of the baseline Pentateuchal/biblical legal tradition into new halakhic formations see ZAHN, Scripture (n. 4), pp. 195-197, 206-207. 45. E.g. GOULDER, Luke (n. 3), pp. 12-14, esp. 14: “[O]f the 89 words in John’s sermon, 18 form part of Matthew’s vocabulary elsewhere … That 20% of an alien text should happen to have been written in Matthew’s preferred phrases would rather strain belief”. 46. ZAHN, Scripture (n. 4), p. 213.

232

A. KIRK

The 2DH therefore can add source-utilization in the Temple Scroll to the lengthening roster of ancient analogies to the compositional procedures that the hypothesis attributes to Matthew. The quest in Farrer scholarship for an analogy to FH Luke’s Matthew-utilization, one that fits with ancient media conditions and practices, will of course doubtless go on. James Madison University United States of America

Alan KIRK [email protected]

COMPETITIVE TRADITIONS: LUKE’S AND MATTHEW’S (CON)TEXTUALIZATION OF THE BEELZEBUL CONTROVERSY

I. DISTINCTIVE COMPOSITIONAL FEATURES AND INTERTEXTUALITY IN THE GOSPELS The conviction that the Gospels were not produced in an intellectual vacuum has led scholars to compare the Gospels’ features and compositional processes with other ancient writings. Over the course of this comparison, increased attention has been paid to the Gospels’ distinctive characteristics. While adapting their sources, ancient authors often take the time to digest the materials and paraphrase them rather than copy their sources verbatim1. Gospel readers, however, can find many verbal agreements and parallels in order between the Synoptic Gospels2. These similarities lead to the question

1. B. SHELLARD, New Light on Luke: Its Purpose, Sources, and Literary Context, Sheffield, Sheffield Academic Press, 2002, p. 62, notes that for ancient classical authors, “imitation and emulation were very tightly bound together”. For surveys of ancient compositional practices, see F.G. DOWNING, Redaction Criticism: Josephus’ Antiquities and the Synoptic Gospels (I), in JSNT 8 (1980) 46-65; ID., Redaction Criticism: Josephus’ Antiquities and the Synoptic Gospels (II), in JSNT 9 (1980) 29-48; ID., Writer’s Use or Abuse of Written Sources, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 523-550; R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, pp. 52-117. See also D.A. RUSSELL, ‘De Imitatione’, in D. WEST – T. WOODMAN (eds.), Creative Imitation and Latin Literature, Cambridge, Cambridge University Press, 1979, 1-16; S.L. MATTILA, A Question Too Often Neglected, in NTS 41 (1995) 199-217. 2. As aptly described by DOWNING, Redaction (II) (n. 1), p. 33: “It is not the divergences among the synoptists (or even between them and John) in parallel contexts, that are remarkable: it is the extraordinary extent of verbal similarities”. Nevertheless, it should be noted that the synoptics also show “the pronounced unevenness in the distribution of verbatim agreement”, which make them once again distinct from other contemporary texts. MATTILA, Question (n. 1), p. 207.

234

C.M. SAULINA

of Gospel interrelations, in particular their literary connections3. For many specialists, these phenomena suggest a literary dependence, either direct or indirect4. The verbal agreements become more pronounced when we consider the sayings rather than the narrative sections in the synoptics, especially Luke and Matthew. This tendency rarely occurs in classical literature. As observed by Sharon Mattila, “Classical authors generally take their greatest liberties when composing speeches, even when relying on source-material rather than on personal speculation (the latter more often being the case)”5. Nonetheless, the close agreements in the synoptics also come with a great degree of inconsistency, “varying from pericope to pericope and within individual pericopae and from the narrative material to the sayings material, [which] appears to be somewhat of an anomaly”6. Some argue that these distinctive features of the Gospels are due to the authors’ distinct relationship with their sources in contrast to the way in which classical authors relate to their sources7. The evangelists show a great respect to their sources and traditions, which influences how they transmit these materials in their writings8. In composing their materials, they are bound by the tradition. Nevertheless, the evangelists’ interaction with their sources does not stop there. They also appear to contribute to their respective faith communities and to the expansion of the Jesus movement by interpreting and cultivating the received traditions9. In this area, the evangelists may have had some freedom to shape their materials for their own theological agenda by constructing them in a certain way for the benefit of their readers or communities. In some ways, they act as creative authors. 3. Literary dependence itself needs to be defined. See A. GREGORY, What is Literary Dependence?, in FOSTER et al. (eds.), Studies (n. 1), 87-114. 4. There is no consensus, however, in how scholars determine the level of verbal and sequential agreement to postulate literary dependence. Ibid., p. 89. 5. MATTILA, Question (n. 1), p. 209. 6. Ibid., p. 209. 7. DOWNING, Redaction (II) (n. 1), p. 33, A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, pp. 40-42. 8. L. ALEXANDER, Luke’s Preface in the Context of Greek Preface-Writing, in NT 28 (1986) 48-74, p. 71, notes, however, that some texts taken from an educational context mimic a similar respect for traditions by “direct personal contact” from master to disciple. This process may also include the use of written texts. 9. KIRK, Q in Matthew (n. 7), p. 40. This practice can also be observed in the ancient Jewish community, e.g., the scribes and redactors of rabbinic literature. See M. FISHBANE, Biblical Interpretation in Ancient Israel, Oxford, Clarendon, 1985, p. 37; C. HAYES, Halakhah le-Moshe mi-Simai in Rabbinic Sources: A Methodological Case Study, in S.J.D. COHEN (ed.), The Synoptic Problem in Rabbinic Literature, Providence, RI, Brown University, 2000, 61-118.

COMPETITIVE TRADITIONS

235

These two compositional aspects lead to a unique positioning of the Gospel authors within the spectrum of author-tradent. The positioning of the Gospel authors helps clarify the manner in which each evangelist utilizes their sources10. As tradents, the evangelists’ manner of familiarizing themselves with their materials differs from that of classical authors. As suggested by Alan Kirk, “The evangelists have not picked up their acquaintance with their sources through a program of preliminary reading preparatory to writing. Matthew and his recipient community’s cultural stake in Mark is of a different order from the stake Livy and his social circles have in Polybius’s History”11. Therefore, in composing their writings, the Gospel authors’ memorization and internalization of their sources appears to have played a crucial role. This unique situation provides some background for the phenomenon of a close agreement between the synoptics. Moreover, the manner in which each evangelist composes their Gospel to achieve certain goals, while potentially cultivating and participating in the tradition for the benefit of their community, corresponds to what Chris Keith refer to as “competitive textualization” vis-à-vis the Gospels’ “textual self-consciousness”12. The latter refers to the Gospels being aware of their own status as a written text, in contrast to other traditions in other forms of media, e.g., oral performance, liturgy, or monument. Alongside this awareness is the presence of a “reading community” as the recipients of these traditions. In this area, the Gospels demonstrate a different degree of awareness13. In its correlation with the textual self-consciousness, the Gospels’ competitive textualization highlights the evangelists’ awareness of the presence of other written traditions and the manner in which they position themselves in relation to those materials. As described by Keith, “this relationship is ‘competitive’ only in the sense that the tradition is aware of a prior tradition’s social position and is vying for its own particular position by drawing parasitically upon that predecessor … the term ‘competitive’ neither assumes nor precludes the possibility that a given tradition views the prior tradition 10. KIRK, Q in Matthew (n. 7), p. 40. 11. Ibid., p. 59. 12. C. KEITH, The Gospel as Manuscript: An Early History of the Jesus Tradition as Material Artifact, Oxford, Oxford University Press, 2020, p. 103. 13. Keith (ibid., p. 101) argues that Mark exhibits a soft from of competitive textualization in relation to the Jewish Scripture. This competitive connection with the Scripture is also found in Matthew, who shows textual consciousness by claiming his work as a Βίβλος (Matt 1,1). Luke incorporates both concepts and heightens the competitive textualization not only vis-à-vis the Jewish Scripture but also the previous written traditions.

236

C.M. SAULINA

in a derogatory manner”14. This concept elucidates how the evangelists may have aligned their works with other traditions, such as their predecessors or/ and their sources. Obvious examples of the Gospels’ competitive textualization can be found in Luke’s prologue (1,1-4) or the end of John’s Gospel (20,30-31; 21,24-25). All of the aforementioned elements of the Gospels become vital when we observe how the Gospels connect with and relate to their sources. The scope and degree of the previous (ante-) text’s influence on the later text can be retraced through intertextuality. Intertextuality can be seen as the memory of an earlier text that is embedded by a subsequent author15. As a concept, intertextuality leads to two important implications. First, although the later text transforms their sources when it absorbs them, intertextuality still helps us trace the footprints of the source materials. In intertextuality theory, as Renate Lachmann aptly claims, “Literature is culture’s memory, not as a simple recording device but as a body of commemorative actions. Literature inscribes itself in a memory space into which earlier texts have inscribed themselves. It does not leave these earlier texts as it finds them but transforms them in absorbing them”16. The possibility for retracing the ante-text in the later text is enabled by the second element of intertextuality. As Lachmann further suggests, “Intertextuality, as the term is conceived in literary scholarship, is the semantic interchange, the contact between texts literary and non-literary”17. Thus, in an interchange between two literary texts, intertextuality should not be studied only on the word level, such as verbal agreements. In fact, the obvious parallels between two texts such as those that exist on the word level may be only ‘the tip of the iceberg’ of the intertextuality between a text and its source(s). The influence of source materials on the later text can extend beyond the immediately apparent elements to include those that are below the “surface”. These may include narrative paradigm techniques, theological motifs, thematic parallels, etc.18. 14. Ibid., p. 108. 15. R. LACHMANN, Cultural Memory and the Role of Literature, in European Review 12 (2004) 165-178. 16. Ibid., p. 172. 17. Ibid., p. 173. 18. D.T. ROTH, The Parables in Q (LNTS, 582), London – New York, Bloomsbury T&T Clark, 2018, p. 41. Also, “Jeder Text hat Vorgänger, auf die er referiert, und zwar in formalstruktureller und/oder in inhaltlich-pragmatischer Hinsicht”. G. WEISE, Zur Spezifik der Intertextualität in literarischen Texten, in J. KLEIN – U. FIX (eds.), Textbeziehungen: linguistische und literaturwissenschaftliche Beiträge zur Intertextualität, Tübingen, Stauffenburg, 1997, 39-48, p. 39.

COMPETITIVE TRADITIONS

237

Analyzing the Synoptic Problem through the lens of intertextuality placed within the realm of Gospel competitive textualization and viewing the evangelists within the spectrum of author-tradent may be rewarding. This framework has the capacity to explain not only the anomalies of verbatim agreements and variation within the Gospel texts, but also the theological or ideological motifs behind them. For instance, the symbolic world of the source materials may have influenced their adapters/users. Adopting this approach also requires students of the Synoptic Problem to use a more holistic framework than the mechanical and fragmented examination of the texts through a literary-critical analysis with un-uniform and disputed criteria19. Furthermore, the holistic method enables intra- and intertextuality. As regards the former, it requires interpreters to conduct not only a contextgrounded microanalysis of a Gospel passage, but also a macroanalysis, e.g., observing how the pericope connects with the larger context and the overall plot of the Gospel. As regards intertextuality, this framework enables a comparison of parallel theological motifs and sequences between two or more Gospels. Since all the Synoptic Gospels are written in narrative form (or at least dominated by narrative), it is fitting to apply a narrative approach in identifying and assessing the intertextuality of the texts. In the current study of the Synoptic Problem, however, the utilization of a narrative approach to analyze Gospel parallels has largely been overlooked20. This study will seek to demonstrate that narrative criticism has much to offer to the study of the Synoptic Problem. Here, it is important to note two potential contributions. First, narrative analysis investigates elements such as settings, plots, characters of each pericope or Gospel as a whole; all of these are essential in the construction of each Gospel text. Since the majority of Jesus’s sayings are framed by narrative materials, this method can also be applied to sections consisting of Jesus’s teachings. Second, the narrative framework necessarily fosters a holistic view of intertextuality. 19. M. HENGEL, The Four Gospels and the one Gospel of Jesus Christ: An Investigation of the Collection and Origin of the Canonical Gospels, London, SCM, 2000, p. 174. J.S. KLOPPENBORG, Conceptual Stakes in the Synoptic Problem, in M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018, 13-42, p. 23, also notes that aesthetic and ad hoc compositional arguments in solving the synoptic problem can be problematic. 20. M.S. GOODACRE, The Synoptic Problem: A Way Through the Maze, London, Sheffield Academic Press, 2001. Some scholars have attempted to do this, e.g., ROTH, Parables (n. 18); W. KAHL, The Gospel of Luke as Narratological Improvement of Synoptic Pre-Text the Narrative Introduction to the Jesus Story (Mark 1.1-8 Parr), in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 19), 223-244.

238

C.M. SAULINA

It is important to note that what I propose in this study is a broader view of narrative criticism, as presented in Kari Syreeni’s tripartite model21. This approach moves beyond the occasional boundaries of narrative criticism by including the world outside the text, such as the symbolic world and the ‘real world’. The former enables readers to observe parallels not only in the texts but also in motifs and theological or ideological agendas. Investigation of the latter, however, will include how the evangelists use their sources (their source and media utilization procedures). In light of this essay’s scope, I will concentrate on the intertextual connections between the Markan, Lukan, and Matthean versions of the Beelzebul controversy and their respective surrounding contexts. I will seek to apply the principles of competitive textualization to inquire into the positioning of the evangelists within the author-tradent spectrum, in an attempt to bring fresh insight to the Synoptic Problem. II. THE BEELZEBUL PASSAGE: ITS IMPORTANCE

AND

COMPLEXITY

The importance of so-called Beelzebul controversy to the Synoptic Problem discussion has been frequently noted. As aptly summarized by M. Eugene Boring, “Participants on all sides of the debate are agreed that the Beelzebul pericope (Matt 12,22-37 and parallels) is a key text for any source hypothesis, because of its difficulty and complexity from the point of view of source analysis”22. The passage appears in each of the synoptics and exhibits similarities and contrasts across its multiple versions. However, it should be noted that how scholars compare the three versions depends on how they determine the scope of the passage23.

21. For a more detailed explanation of this framework, see K. SYREENI, Wonderlands: A Beginner’s Guide to Three Worlds, in SEÅ 64 (1999) 33-46; ID., Peter as Character and Symbol in the Gospel of Matthew, in D.M. RHOADS – K. SYREENI (eds.), Characterization in the Gospels: Reconceiving Narrative Criticism (SupplJSNT, 184), Sheffield, Sheffield Academic Press, 1999, 106-152. 22. M.E. BORING, The Synoptic Problem, “Minor” Agreements, and the Beelzebul Pericope, in F. VAN SEGBROECK – C.M. TUCKETT – G. VAN BELLE – J. VERHEYDEN (eds.), The Four Gospels 1992: FS Frans Neirynck (BETL, 100), Leuven, Leuven University Press – Peeters, 1992, 587-619, p. 600. 23. Downing, for instance, sets the scope of the Beelzebul passages to include the Sin Against the Holy Spirit, the Sign of Jonah, and the Return of the Holy Spirit (Matt 12,22-45; Mark 3,20-29; Luke 11,14-26 [with Luke 12,10; 6,43-45]). F.G. DOWNING, Towards the Rehabilitation of Q, in NTS 11 (1965) 169-181.

COMPETITIVE TRADITIONS

239

Readers can observe the obvious verbatim agreements between Mark and Matthew and between Matthew and Luke while also noticing the small parallels between Mark and Luke24. As we can see from these three sets, Matthew holds the medial position25. It is worth noting that Matthew’s and Luke’s narratives exhibit the greatest number of verbatim agreements 26. According to the Two-Document Hypothesis (2DH), the Beelzebul scene suggests some overlaps between Mark and Q27. Proponents of this hypothesis argue that Matthew conflates Mark and Q, while Luke follows Q alone. According to the Farrer Hypothesis (FH), however, the phenomena presented by this passage in the synoptics can be explained by Luke’s use of both Mark and Matthew. FH Matthew was influenced by Mark in much of the passage, while adding his own editorial work. FH Luke then follows Matthew’s account quite closely. Strikingly, in this scenario, FH Luke seems to disregard Matthew’s adoption and treatment of certain Markan elements in the story; some scholars refer to this phenomenon as FH Luke’s “unpicking” of Matthew’s Markan elements28. In contrast to the FH, the 24. The most obvious one is the use of ἐφʼ ἑαυτήν in Mark 3,24 and Luke 11,17 instead of καθʼ ἑαυτῆς in Matt 12,25. Also, Mark and Matthew use βλασφημέω in the blasphemy against the holy spirit (Mark 3,28-29; cf. Luke 12,10), whereas this verb does not occur in Matthew’s version of the saying. 25. For the definition of this term, see E.P. SANDERS – M. DAVIES, Studying the Synoptic Gospels, London, SCM, 1989, pp. 54, 67-83. 26.  Tyson argues, “At best, Matthew and Mark agree on 52 words, Luke and Mark on 18, and Matthew and Luke on 95”. J.B. TYSON, The Two-Source Hypothesis: A Critical Appraisal, in A.J. BELLINZONI – J.B. TYSON – W.O. WALKER (eds.), The Two-Source Hypothesis: A Critical Appraisal, Macon, GA, Mercer University Press, 1985, 437-452, p. 446. 27. There are no clear criteria to determine where the Mark-Q overlaps occurs; see C.M. TUCKETT, The Revival of the Griesbach Hypothesis: An Analysis and Appraisal (SNTS MS, 44), Cambridge, Cambridge University Press, 1983, p. 78. The 2DH proponents often describe the agreements between Matthew and Luke against Mark in the Mark-Q overlaps as Minor Agreements, distinguishing these from Major Agreements. However, these categories are debatable. E.g., R.T. SIMPSON, The Major Agreements of Matthew and Luke Against Mark, in NTS 12 (1966) 273-284; W. KAHL, Inclusive and Exclusive Agreements - Towards a Neutral Comparisons of the Synoptic Gospels, or: Minor Agreements as Misleading Category, in M. MÜLLER – J.T. NIELSEN (eds.), Luke’s Literary Creativity (LNTS, 550), London – New York, Bloomsbury T&T Clark, 2016, 44-76; M.S. GOODACRE, Taking our Leave of Mark-Q Overlaps: Major Agreements and the Farrer Theory, in MÜLLER – OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (n. 19), 201-222. 28. The term ‘unpicking’ was introduced by Downing. See his Redaction (II) (n. 1); ID., Compositional Conventions and the Synoptic Problem, in JBL 107 (1988) 69-85; ID., A Paradigm Perplex: Luke, Matthew and Mark, in NTS 38 (1992) 15-36. For notable responses to Downing’s argument, see K. OLSON, Unpicking on the Farrer Theory, in M. GOODACRE – N. PERRIN (eds.), Questioning Q: A Multidimensional Critique, Downers Grove, IL, InterVarsity Press; London, SPCK, 2004, 127-150; E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021, pp. 131-143. Recently, O. ANDREJEVS, FH Luke’s ‘Unpicking’: Some Observations on

240

C.M. SAULINA

Griesbach (Two-Gospel) Hypothesis (GH/2GH) explains Matthew’s medial position by theorizing that Luke and Mark adopted Matthew29. In addition to these three hypotheses, the passage has also been employed by proponents of the Matthean Posteriority Hypothesis (MPH), which posits Matthew’s knowledge of Mark and Luke30. However, the MPH theorists have yet to explore the full potential of this passage as they look to bring their hypothesis to the forefront of the Synoptic Problem discussion. By conducting a micro- and macroanalysis of the Beelzebul controversy and its surrounding context in each Synoptic Gospel, this paper seeks to demonstrate that the MPH provides the best possible explanation for the phenomena observed in the Synoptic Gospels, at least, as regards this particular pericope. By employing the narrative and source-utilization analysis to examine the textual, symbolic, and real worlds of Matthew and Luke, this study will seek to show that the intertextual relationship between these two Gospels extends beyond verbatim agreements, thereby increasing the high possibility of a direct literary dependence between the two. The comparison of the textual and symbolic worlds of these two Gospels will suggest the influence of Luke on Matthew rather than vice versa. Due to the spatial limitations, I will simply accept here the widely established hypothesis of Markan priority and focus my arguments on responding to the FH and the 2DH.

Francis Watson’s Recent Analysis (2018) and the Extent of the Phenomenon in JSNT 45 (2022) 3-22, has broadened the scope of this term. The Beelzebul pericope, however, belongs to the discussion of ‘unpicking’ already in Downing’s opening article – see DOWNING, Rehabilitation (n. 23), pp. 172-176. 29. See A.J. MCNICHOL – D.L. DUNGAN – D.B. PEABODY (eds.), Beyond the Q Impasse. Luke’s Use of Matthew: A Demonstration by the Research Team of the International Institute for Gospel Studies, Valley Forge, PA, Trinity Press International, 1998, pp. 176-178. Cf. E. EVE, The Devil in the Detail: Exorcising Q from the Beelzebul Controversy, in J.C. POIRIER – J. PETERSON (eds.), Marcan Priority without Q: Explorations in the Farrer Hypothesis (LNTS, 455), London – New York, Bloomsbury T&T Clark, 2015, 16-43, p. 16; TUCKETT, Revival (n. 27), pp. 85-86. 30. E.g., R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1994) 1-22; HENGEL, Gospels (n. 19), pp. 169-207; G.A. BLAIR, The Synoptic Gospels Compared (SBEC, 55), Lewiston, Edwin Mellen, 2003, pp. 196, 293, 310; B. ADAMCZEWSKI, Q or not Q? The So-Called Triple, Double, and Single Traditions in the Synoptic Gospels, Frankfurt am Main, Peter Lang, 2010; R.K. MACEWEN, Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS, 501), London – New York, Bloomsbury T&T Clark, 2015; A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226. See a thorough survey of earlier scholarship (through 2015) in MACEWEN, Matthean Posteriority, pp. 6-26.

COMPETITIVE TRADITIONS

241

III. THE BEELZEBUL CONTROVERSY IN MARK’S ACCOUNT Mark places the Beelzebul scene between the institution of the twelve (3,13-19a) and the biographical apothegm elucidating the meaning of kinship with Jesus (3,31-35). The latter will be followed by a brief summary passage (4,1-2) and the parable of the sower (4,3-9). Although the thematic links between these episodes may not be obvious at face value, Mark connects all of these scenes chronologically in his narrative world. It can be noted, too, that the future betrayal of Judas – known to a reasonably erudite reader and therefore likely brought to mind with the mention of his name (3,19a) – logically sets up a series of “divergent forms of opposition of Jesus”31. A brief narrative episode involving Jesus’ family helps set up the ensuing controversy32. The sequence begins with Jesus returning to some place of residence after the mountain gathering33. The crowds follow and gather around the house, presumably to listen to his teaching and be healed (cf. Mark 2,4.13; 3,9). This creates a significant obstacle for Jesus’ family to get through. The family intends to restrain Jesus, thinking that he may be out of his mind or possessed (the first accusation that Jesus faces in this series of episodes)34. Mark continues the sequence by describing the arrival of scribes from Jerusalem to investigate Jesus35. These scribes make two further accusations: 1) Jesus is said to be possessed by Beelzebul, which may echo the claim made by Jesus’s family36; 2) Jesus is said to exorcise demons through the power of Beelzebul, the ruler of demons. 31. C. FOCANT, The Gospel according to Mark, trans. Leslie Robert Keylock, Eugene, OR, Pickwick, 2012, p. 137. 32. In both the LXX and contemporary colloquial use, the term οἱ παρʼ αὐτοῦ (Mark 3,20) can refer to family or friends. J. MARCUS, Mark 1–8: A New Translation with Introduction and Commentary (AB, 27), New Haven, CT, Yale University Press, p. 270, concludes that the familial interpretation here “seems to be demanded by the context”. Thus already B. WEISS, Die Evangelien des Markus und Lukas (KEK, 1,2), Göttingen, Vandenhoeck & Ruprecht, 71885, p. 54; H.J. HOLTZMANN, Die Synoptiker (HCNT, 1,1), Tübingen, J.C.B. Mohr, 31901, p. 126. More recently, in recognition of a number of difficulties, R.T. FRANCE, The Gospel of Mark (NIGTC), Grand Rapids, MI, Eerdmans, 2002, p. 166 (“the least unsatisfactory solution”). Cf. A. YARBRO COLLINS, Mark (Hermeneia), Minneapolis, MN, Fortress, 2007, p. 226; FOCANT, Mark (n. 31), pp. 138, 142; G. GUTTENBERGER, Das Evangelium nach Markus (ZB NT, 2), Zürich, TVZ, 2017, p. 89. 33. Mark 3,19b-21.31-35. The words οἶκος and οἰκία become important links connecting the scenes in Mark 3,20-35. FRANCE, Mark (n. 32), p. 165, notes that “it would naturally follow that the same house [as in 2,1] is intended here”. 34. D.C. SIM, Matthew and the Synoptic Problem, in FOSTER et al. (eds.), Studies (n. 1), 187-208, p. 190. 35. According to, e.g., FOCANT, Mark (n. 31), p. 139, this could be an official delegation. Somewhat similarly, YARBRO COLLINS, Mark (n. 32), p. 228. 36. Cf. MARCUS, Mark (n. 32), p. 271.

242

C.M. SAULINA

The passage thus features three separate groups of characters: the crowd, Jesus’s family, and the scribes from Jerusalem. In Mark’s characterization of these groups, they can be divided into two categories based on their response to Jesus. The evangelist signals a positive attitude towards the crowd in 3,34-35. By contrast, Jesus’ family and the scribes are characterized negatively. Accordingly, Jesus’ response in vv. 23-30 rebukes these groups. Jesus responds to them by uttering parables, which can be divided into five parts or sets. Beginning with a rhetorical question (πῶς δύναται σατανᾶς σατανᾶν ἐκβάλλειν;), Jesus proceeds to bolster this rejoinder with metaphors: 1) a divided kingdom (v. 24); 2) a divided household (v. 25)37. These metaphors in turn support the validity of two additional arguments or images (vv. 26-27). After disputing the accusation made by the scribes (and, potentially, the family), Jesus stresses the severe consequences of their accusations. By claiming that Jesus is possessed and exorcises demons through Beelzebul, the scribes are implied to have sinned against the Holy Spirit (3,28-30)38. It is worth noting that the Christological aspect of Jesus’s claim regarding the source of his power is less prominent than what we find in Luke’s and Matthew’s accounts (see the next two sections in this essay). Moreover, although exorcisms have a place in Jesus’ ministry, the significance of his exorcisms vis-à-vis the kingdom of God is less explicit (if not absent) in Mark. By contrast, this is something that Luke and Matthew will underscore in their versions of the passage. Nevertheless, for Mark, Jesus’ confrontations with the forces of evil are highly important. Jesus appoints the apostles to proclaim the message and gives them authority specifically to cast out demons (Mark 3,15; 6,7.13). These confrontations appear to belong to the kingdom’s project39, that is, a cosmic battle between Jesus/his disciples and Satan, between the power of Holy Spirit and the demons40. Furthermore, Mark refers to Satan as

37. These metaphors, in all likelihood, offer examples of the smallest and largest types of social institutions known to the original (intended) reader. 38. MARCUS, Mark (n. 32), p. 283; FRANCE, Mark (n. 32), p. 174; YARBRO COLLINS, Mark (n. 32), pp. 234-235; FOCANT, Mark (n. 31), p. 140; GUTTENBERGER, Markus (n. 32), p. 90. 39. Cf. M.L. HUMPHRIES, Christian Origins and the Language of the Kingdom of God, Carbondale, IL, Southern Illinois University Press, 1999, p. 57. 40. In the burglary scene painted by Mark 3,27, however, there seems to not yet be any indication of a cosmic war between the kingdoms of God and Satan, in contrast to what we will see in Luke 11,21-22. Ibid., pp. 52-53. This interpretation must be inferred from the context. The vast majority of the commentators make this inference, e.g., WEISS, Markus und Lukas7 (n. 32), p. 59; HOLTZMANN, Synoptiker3 (n. 32), p. 127; MARCUS, Mark (n. 32),

COMPETITIVE TRADITIONS

243

the culprit who comes to take away the word (the seed) in the parable of the sower. Not insignificantly, the announcement of Jesus’s ministry in Mark 1,15 indicates the emergence of the eschatological kingdom which comes on the heels of Jesus’s baptism and triumph over temptation41. Mark follows the Beelzebul passage by resuming the sequence in which Jesus’s family are looking for him (3,31)42. Instead of responding to their summons, Jesus uses the opportunity to highlight the true meaning of belonging to his inner circle. The link to the previous passage is apparent insofar as all three scenes share the same setting, in the vicinity of a house (3,19b). Mark may have thought that the family’s accusation (or Jesus’ actions) provides an illustration of a divided household. At the same time, however, the evangelist highlights the manner in which Jesus establishes a new household code predicated on one’s obedience to God rather than one’s family43. The Parable of the Sower (4,1-9), which follows the Beelzebul section, certainly underscores this aspect of discipleship. IV. THE BEELZEBUL CONTROVERSY IN LUKE’S ACCOUNT 1. General Observations Luke locates, frames, and structures the episode differently from Mark. First of all, it appears on the heels of three passages highlighting the importance of prayer (Luke 11,1-4.5-8.9-13). The inclusion of the Beelzebul accusation immediately after the prayer-oriented material seems to be very abrupt, almost as if Luke starts a new subject44. Suddenly, Jesus is no longer alone with the disciples (11,1), but among the crowds. In this public setting, Jesus exorcises a mute demon; the people are amazed by what they see (11,14). The faceless crowd continues to play an important role in the story p. 274; FRANCE, Mark (n. 32), pp. 172-173; YARBRO COLLINS, Mark (n. 32), pp. 233-234; FOCANT, Mark (n. 31), p. 140; GUTTENBERGER, Markus (n. 32), p. 90. 41. J.M. ROBINSON, The Problem of History in Mark, Philadelphia, PA, Fortress, 1957, pp. 79-80. MARCUS, Mark (n. 32), p. 173, references T. Mos. 10,1. 42. The motif of Jesus’ family in a contentious setting is unique to Mark. Luke and Matthew depict sancta familia in an unambiguously positive way. Cf. SIM, Matthew (n. 34), p. 198. 43. Cf. HUMPHRIES, Origins (n. 39), pp. 60-61. 44. E.g., J.A. FITZMYER, The Gospel according to Luke X–XXIV (AB, 28A), Garden City, NY, Doubleday, 1985, p. 917; M. WOLTER, Das Lukasevangelium (HNT, 3), Tübingen, Mohr Siebeck, 2008, p. 416; F. BOVON, Luke 2: A Commentary on the Gospel of Luke 9:51– 19:27 (Hermeneia), trans. Donald S. Deer, Minneapolis, MN, Fortress, 2013, p. 117 and n. 23; EVE, Devil (n. 29), p. 32.

244

C.M. SAULINA

for the next four passages: the parable of the return of the unclean spirit (11,24-26); the contrasting beatitudes (11,27-28); the sign of Jonah (11,29-32); and the parable of the lamp (11,33-36). All of these events occur chronologically (according to Luke’s narrative world), before the same audience. Following the last of these four segments, Luke changes the setting when Jesus enters a Pharisee’s house, which will set the stage for the denunciation of the Pharisees and the lawyers in 11,37-54. The Lukan Jesus heals a mute (κωφός) man by exorcising the demon causing the illness, which prompts a number of different responses from the crowd, including the Beelzebul accusation. The exorcism serves as the catalyst for the ensuing controversy. Based on the crowd’s responses, readers can divide these bystanders into several groups: 1) those who accept Jesus’s work with amazement (11,14); 2) those who accuse Jesus of being an ally of Beelzebul (11,15); and 3) those who demand a sign from heaven (11,16). Jesus responds to each of the aforementioned groups in three separate segments, albeit not presented in an orderly manner: 1) 11,17-23 (to the second group); 2) 11,27-28 (to the first group); and 3) 11,29-36 (to the third group). Thus, in the narrative-critical reading proposed here, the scope of the Lukan Beelzebul controversy includes all of these passages. I will limit my discussion to Luke 11,14-28. 2. Luke 11,14-20: Narrative-Critical Observations The crowd’s accusation in 11,15 disputes not the effectiveness of Jesus’ exorcism but the source of his power; it is said to originate with Beelzebul, the ruler of demons (cf. Matt 12,26-27//Luke 11,18-19). This response highlights the extraordinary power exhibited by Jesus in expelling the mute demon. Casting out a mute spirit (let alone a blind and mute spirit as described in Matthew) would appear to be a particularly challenging type of exorcism45. The unusual power exhibited by Jesus may therefore have led the bystanders to suspect him of acting akin to those miracle workers who utilize the spirits of the dead or demons to do their ‘miracles’46. This suspicion of an ‘illegitimate’ source of Jesus’ power may be further indicated by the earlier rumour that Jesus is John the Baptizer risen from the dead (9,9).

45. In order to gain control over a possessing demon, an exorcist typically needs the demon to speak and to disclose its identity. Cf. C. BONNER, The Technique of Exorcism, in HTR 36 (1943) 39-49; S. SKINNER, Techniques of Graeco-Egyptian Magic, Singapore, Golden Hoard Press, 2014, pp. 53, 96. Cf. Mark 9,17.27-29. 46. S.R. GARRETT, The Demise of the Devil: Magic and the Demonic in Luke’s Writings, Minneapolis, MN, Fortress, 1989, p. 44.

COMPETITIVE TRADITIONS

245

Jesus’s response to this accusation is divided into three parts, each represented by a conditional sentence opening with εἰ δέ (Luke 11,18.19.20). The first part (11,18) revisits the Markan metaphor of a kingdom (Mark 3,24), now modified to explicitly reflect the nature of Satan’s kingdom47. Jesus cannot expel demons with the power of Beelzeebul since, if this were true, Satan’s kingdom would be divided among itself. A divided kingdom would not last and would soon find its end48. In his second response (Luke 11,19-20), Jesus does not refute the aspersion but delivers “…eine Demaskierung der Voreingenommenheit von Jesu Kritikern”49. His answer exposes the flawed logic of the accusation in light of the existence of other exorcists50. Consequently, the charge is superfluous and little more than a smear attempt. 3. Luke 11,14-20: Source-Critical Observations It is important to note that in Luke 11,20 Jesus makes an unusual claim. In Second Temple literature, exorcisms do not seem to convey any eschatological significance51. Consequently, many interpreters have questioned the authenticity of vv. 19-2052. Particularly, verse 20 is often taken to have previously been a part of free-flowing logia subsequently adopted by Q or by one of the synoptics53. Nonetheless, such logia are unknown to Mark. It seems scarcely conceivable that Mark came across something like Luke 11,20 in his sources or tradition and intentionally omitted it54. 47. I agree with the view on which, for Luke, Beelzebul is not identical with Satan. As WOLTER, Lukasevangelium (n. 44), pp. 417-418, observes, “‘Satan’ steht vielmehr hier in metonymischer Weise für alle, durch die er seine Herrschaft ausübt. Im vorliegenden Fall sind das die Damönen und Beelzebul”. But see FITZMYER, Luke X–XXIV (n. 44), pp. 920-921. 48. The phrase ἡ βασιλεία αὐτοῦ in verse 18 is worth noting, especially compared to Mark 3,26. Mark never links ‘kingdom’ and ‘Satan’ in his gospel. One might argue that Satan’s kingdom is implied in Mark 3,23-24. However, looking closer at Mark’s wording of the parables of the divided kingdom and house, these metaphors seem to represent the human society through the angles of two different social institutions. 49. WOLTER, Lukasevangelium (n. 44), p. 418. 50. For Jewish exorcists, see, e.g., Tob 8,1-3; 1 Sam 16,14.23; Josephus, A.J. 8.45-49. See also B.J. 7.185; Acts 29,13. 51. EVE, Devil (n. 29), p. 25. 52. Among those who accept the authenticity: e.g., R. BULTMANN, The History of the Synoptic Tradition, trans. John Marsh, New York, Harper & Row, 21968, p. 162; W.D. DAVIES – D.C. ALLISON, Matthew (ICC), Edinburgh, T&T Clark, 1988-97, vol. 2, p. 339; BOVON, Luke 2 (n. 44), p. 120. 53. HUMPHRIES, Origins (n. 39), p. 27; BULTMANN, History (n. 52), pp. 13-14; J.D. CROSSAN, In Fragments: The Aphorisms of Jesus, San Francisco, CA, Harper & Row, 1983, pp. 184-191. 54. H. RÄISÄNEN, Exorcisms and the Kingdom: Is Q 11:20 a Saying of the Historical Jesus?, in R. URO (ed.), Symbols and Strata: Essays on the Sayings Gospel Q, Göttingen, Vandenhoeck & Ruprecht, 1996, 119-143, p. 134. Some scholars have proposed that Mark knew and

246

C.M. SAULINA

Of course, Mark and Q may have been completely independent, as the majority of Q scholars in fact hold. H. Räisänen persuasively argues that Luke 11,20 fits into the Q Beelzebul accusation now preserved in Luke and Matthew. He suggests that “the ‘I’-form and the direct address (‘upon you’) likewise presuppose a context”55. Jesus’s reference to his exorcistic activity seems to convey an event(s) that triggered the saying. If Luke and Matthew adopted Matt 12,28//Luke 11,20 from their source(s), it must at least have been accompanied by Matt 12,22.25a//Luke 11,14.17a56. Moreover, Räisänen also claims that it is easier to see how the statement in Luke 11,20 presupposes the one in Luke 11,19 rather than vice versa. It may not be necessary, however, to envision Q as the source of Matt 12,28//Luke 11,20. Jesus’ sayings in Luke 11,19-20 fit rather well with the surrounding material. Albeit somewhat redundant, 11,19 follows Jesus’ argument in 11,18 naturally. As noted earlier, 11,19-20 contain the first two parts of a tripartite response to the crowds (εἰ δέ ×3). From the macrostructural perspective, these verses also align with Luke’s overall theological agenda: Jesus’ and the disciples’ successful exorcisms are an important aspect of Jesus’ proclamation of the kingdom of God57. Turning to the FH, I would like to suggest that the view on which Luke adapted 11,19-20 from Matt 12,27-28 is problematic58. To be sure, Matthew’s inclusion of verbatim agreements with Luke 11,19-20 in Matt 12,27-28 has clear source-critical implications, which I will interpret from the position of the MPH59. In the context of Matt 12,27-28//Luke 11,19-20, the high Christological claim conveyed in Luke 11,19-20 would have been too significant to be ignored or rejected. Accordingly, Matthew followed Luke’s wording verbatim, except when he replaced the word “finger” with “spirit” (Matt 12,28//Luke 11,20). used Q. E.g., H.T. FLEDDERMANN, Mark and Q: A Study of the Overlap Texts (BETL, 122), Leuven, Leuven University Press – Peeters, 1995 (and the literature there). However, this view has not persuaded the majority of Q or Markan scholars. E.g., F. NEIRYNCK, Assessment, ibid., 261-307. 55. RÄISÄNEN, Exorcisms (n. 54), p. 129. 56. Ibid. 57. Luke highlights the link between the arrival of God’s kingdom/Satan’s fall and the success of Jesus’ (and the disciples’) exorcisms more than any other synoptist (e.g., Luke 4,3137.41.43; 7,21-22; 10,1-20 [esp. 10,9.11.17-18]). See the discussion in BOVON, Luke 2 (n. 44), pp. 30-31. 58. E.g., M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, p. 504. 59. Besides Mark and Luke, Matthew might have known Luke’s source(s); see GARROW, Streeter’s ‘Other’ (n. 30). Accordingly, Matthew may have taken these sayings from Luke or from what one may designate, as a manner of convenience (without attachment to the term’s use in previous scholarship), as Luke’s L source.

COMPETITIVE TRADITIONS

247

4. Luke 11,20 and the Problem of δάκτυλος θεοῦ Some scholars have proposed that Matthew’s ἐν πνεύματι θεοῦ reflects the saying’s original form while Luke changes it to ἐν δακτύλῳ θεοῦ60. If so, the Lukan replacement could be alluding to Exod 8,15-19, potentially contrasting the Pharaoh’s magicians/οἱ υἱοὶ ὑμῶν and δάκτυλος θεοῦ/Jesus. While the Exodus reference seems valid, the comparison between the two sides is probably less so. For Jesus’ argument in Luke 11,19 to work, the accusing party (the Lukan crowd and Matthean Pharisees) would need to believe that their sons’ exorcisms are powered by a legitimate source. Since Jesus uses this line of thought to justify his exorcisms, both parties are apparently envisioned as being on the same side. If Luke worked with Mark and Matthew (FH), the decision to change πνεῦμα to δάκτυλος seems odd61. Both Mark and Matthew highlight the role 60. E.g., H.T. FLEDDERMANN, Q: A Reconstruction and Commentary (BiTS, 1), Leuven – Paris – Dudley, MA, Peeters, 2005, pp. 483-484; EVE, Devil (n. 29), pp. 36-38. This also appears to be implied by GOULDER, Luke (n. 58), p. 504. 61. Goulder (ibid.) notes that Luke never uses ἐν πνεύματι θεοῦ, but the same argument can be made for ἐν δακτύλῳ θεοῦ. In fact, Luke (1,17; 3,16; Acts 1,5; 11,16; cf. ἐν τῷ πνεύματι Luke 2,27; 4,1; [10,21]) uses the ἐν πνεύματι construction more frequently than Matthew (3,11; 12,28; 22,43). The problem, then, would presumably have been not ἐν πνεύματι but ἐν πνεύματι + θεοῦ. But is it not too rigid to insist that Luke avoided a reference to the Spirit just because it happened to be in combination with θεοῦ? He could have easily modified it to ἐν πνεύματι ἁγίω (Luke 3,16; Acts 11,16). J.A. FITZMYER, The Gospel according to Luke I–IX (AB, 28), Garden City, NY, Doubleday, 1981, p. 227, notes that “[m]ore than any of the other Synoptic evangelists Luke has made the Spirit an important feature of his Gospel and its sequel” (see his detailed analysis in ibid., pp. 227-231). FLEDDERMANN, Q: A Reconstruction and Commentary (n. 60), p. 483, argues that 2DH Luke “does indeed alter his sources to remove a reference to the [holy] spirit” (Mark 12,36 diff. Luke 20,42; Mark 13,11 diff. Luke 21,15). But Luke preserves Mark’ πνεῦμα with this nuance in Mark 1,8.10.12; 3,29 (cf. Luke 3,16.22; 4,1; 12,10). Luke’s 4:2 retention ratio attested for Mark’s πνεῦμα calls Fleddermann’s statistical argument into question. Moreover, Luke is not always consistent with Mark’s unclean spirits. He preserves πνεῦμα with this nuance at Mark 1,23.27; 5,8; 9,17.25 (cf. Luke 4,33.36; 8,29; 9,39.42b), but omits it at Mark 1,26 diff. Luke 4,35; Mark 2,8 diff. Luke 5,22; Mark 3,30 diff. Luke 11,18; Mark 9,20 diff. Luke 9,42a. Does the 5:4 retention ratio mean that Luke deliberately alters his sources to remove references to unclean spirits? The limitations of the statistical argument in this particular case should be apparent. Fleddermann’s second argument (ibid.), that δάκτυλος θεοῦ “forms part of a series of anthropomorphic terms that Luke uses for God”, cannot necessitate a redaction-critical solution; the other references occur in Sondergut passages (Luke 1,51.66; and in Acts). Therefore, we cannot test whether Luke would have been compelled to override source material with an anthropomorphic redactional term. He may well have preserved a congenial term found in his source(s) or tradition. That is to say, Luke’s potential predilection for anthropomorphic terms does not yet prove that he would have omitted Matthew’s (or Q’s) congenial πνεῦμα – other instances of Luke favoring such terminology over his sources (preferably, Mark, to avoid the ambiguity of testing the double tradition) are necessary for Fleddermann’s argument to stand. Fleddermann’s remaining three arguments (ibid., pp. 483-484) assume the 2DH and are therefore not neutral. It is worth

248

C.M. SAULINA

of the Spirit in the Beelzebul pericope. If Luke were to follow Matthew’s wording, the phrase ἐν πνεύματι θεοῦ would have fit perfectly with Jesus’s immediately preceding teaching on the prayer, in which the disciples were encouraged to ask for πνεῦμα ἅγιον (Luke 11,13). As a result, the shift from Luke 11,5-13 to 11,14-36 would have been considerably smoother62. Furthermore, both Mark and Matthew include the sin of blasphemy against the Holy Spirit in their versions of the Beelzebul controversy. Would Luke have ignored all of these explicit references to πνεῦμα in the immediate context of 11,20? Since he utilizes the blasphemy saying elsewhere (Luke 12,10), the exclusion of πνεῦμα in Luke 11,20 should be considered an example of FH Luke’s omission of congenial materials presumably found in his Matthean source. The FH’s problem with δάκτυλος θεοῦ is reinforced by the fact that the Holy Spirit and evil spirits are juxtaposed elsewhere in Luke’s gospel. The Holy Spirit descends upon Jesus in his baptism (3,22), shortly before which, incidentally, one encounters the phrase ἐν πνεύματι ἁγίω (3,16). Returning from his baptism and immediately before encountering the devil in the wilderness, Jesus is said to be πλήρης πνεύματος ἁγίου (4,1). Following the temptation sequence (4,1-14), Jesus returns to Galilee ἐν τῇ δυνάμει τοῦ πνεύματος (4,14). In his work on Luke-Acts, C.A. Evans notes that the Spirit “clarifies why Jesus was successful in facing Satanic temptation, and Adam was not. Jesus’s success was due to his continual reliance upon the noting, however, that Fleddermann’s (ibid., p. 484) appeal to Matthew potentially following Q faithfully in taking over βασιλεία τοῦ θεοῦ does not require that Matthew abstain from making changes elsewhere in 12,28. See DAVIES – ALLISON, Matthew (n. 52), vol. 2, p. 339; U. LUZ, Matthew 8–20 (Hermeneia), trans. James E. Crouch, Minneapolis, MN, Fortress, 2001, p. 204 and n. 64. According to EVE, Devil (n. 29), p. 36, the implication is that “your sons” cast out demons “only by magic” (as opposed to Jesus). See, however, my comment above on the logic of Jesus’ argument. More recently, Eve (Relating [n. 28], p. 122) states that “Luke is sparing with his allusions to the Spirit” from Luke 4 until Acts 2. The reader should see the statistics: 5 instances (Luke 9,55; 10,21; 11,13; 12,10.12) to Matthew’s 7 over the same story-span (post-Temptation) (Matt 10,20; 12,18.28.31-32; 22,43; 28,19), but with 4 occurrences outside the immediate vicinity of the Beelzebul controversy to Matthew’s 3. Of the 5 Lukan occurrences, 1 occurs in special material (Luke 9,55); 2 are typically classified with the double tradition (Luke 12,10.12); and the remaining 2 are redactional on the FH (Luke 10,21 diff. Matt 11,25; Luke 11,13 diff. Matt 7,11). Nothing here suggests avoidance of the term on the FH. In fact, FH Luke takes over Matthean πνεῦμα and redacts Matthew by adding it. Meanwhile, the clustering of πνεῦμα in the immediate vicinity of the Matthean Beelzebul controversy and in the episode itself (Matt 12,18.28.31-32) – the highest accumulation of the term’s use in any 15 verses of Matthew’s gospel (higher than in Matt 1–2 or Matt 3–4, and only slightly behind the 5 combined occurrences in Matt 1–4) – strongly recommends its redactional origin at Matt 12,28. The FH theorists will probably not object to the last statement. They will, however, interpret this accumulation as Matthew’s redactional activity without any knowledge of Luke. 62. See n. 44 above.

COMPETITIVE TRADITIONS

249

power of the Holy Spirit”63. Subsequently, the Spirit also enables the disciples to overcome evil spirits in their mission to liberate the nations from Satan’s power in Acts. The congeniality of Matt 12,28’s πνεῦμα to FH Luke seems difficult to dispute. As a result, his presumable omission of it here appears strange. If we reverse the editorial direction, Matthew has good reasons to replace Luke’s δάκτυλος64. An important part of Matthew’s agenda is to underscore the Markan idea of retribution against those who accuse Jesus of using the unclean spirit in his exorcisms; they commit blasphemy against the Holy Spirit65. By substituting πνεῦμα in 12,28, Matthew creates a clear link between v. 28 and vv. 31-3266. Additionally, Luke’s δάκτυλος θεοῦ may have appeared to Matthew as too enigmatic and even as potentially having a magical ring to it67. 63. C.A. EVANS, Jesus and the Spirit: On the Origin and Ministry of the Second Son of God, in ID. – J. SANDERS (eds.), Luke and Scripture: The Function of Sacred Tradition in Luke-Acts, Minneapolis, MN, Fortress, 1993, 26-45, p. 40. Evans’s work offers a comprehensive survey of the Holy Spirit and Jesus in Luke-Acts. Contra EVE, Devil (n. 29), p. 38, who argues that Luke never associates the Holy Spirit with miracles. 64. Thus, Luke wrote/retained the original form. E.g., C.K. BARRETT, The Holy Spirit and the Gospel Tradition, London, SPCK, 1937, p. 63; S. SCHULZ, Q: Die Spruchquelle der Evangelisten, Zürich, TVZ, 1972, p. 205 (see the previous scholarship in his n. 218); I.H. MARSHALL, The Gospel of Luke (NIGTC), Grand Rapids, MI, Eerdmans, 1978, p. 476; J. JEREMIAS, Die Sprache des Lukasevangeliums: Redaktion und Tradition im Nicht-Markusstoff des dritten Evangeliums (KEK, Sonderband), Göttingen, Vandenhoeck & Ruprecht, 1980, p. 201; R. LAUFEN, Die Doppelüberlieferungen der Logienquelle und des Markusevangeliums (BBB, 54), Bonn, Hanstein, 1980, p. 131; FITZMYER, Luke X–XXIV (n. 44), p. 918; DAVIES – ALLISON, Matthew (n. 52), vol. 2, pp. 339-340; R.H. GUNDRY, Matthew, Grand Rapids, MI, Eerdmans, 21994, p. 235; J.P. MEIER, A Marginal Jew: Rethinking the Historical Jesus (ABRL), New York, Doubleday, 1994, vol. 2, pp. 410-411; J. MARCUS, The Beelzebul Controversy and the Eschatologies of Jesus, in B. CHILTON – C.A. EVANS (eds.), Authenticating the Words of Jesus and Authenticating the Activities of Jesus, vol. 2: Authenticating the Activities of Jesus (NTTSD, 28,2) Leiden – Boston, MA – Köln, Brill, 1998, 247-278; J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000, pp. 232-233; LUZ, Matthew 8–20 (n. 61), p. 200 and n. 24; H. KLEIN, Das Lukasevangelium (KEK, 1,3), Göttingen, Vandenhoeck & Ruprecht, 102006, p. 415, n. 32; BOVON, Luke 2 (n. 44), p. 121; J.R. EDWARDS, The Gospel of Luke (PNTC), Grand Rapids, MI, Eerdmans, 2015, p. 344, n. 11; M. KONRADT, The Gospel according to Matthew, Waco, TX, Baylor University Press, 2020, p. 194; M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021, pp. 833, 1299. 65. The idea of judgement runs through this passage. Cf. B.C. BUTLER, The Originality of St. Matthew: A Critique of the Two-Document Hypothesis, Cambridge, Cambridge University Press, 1951, p. 65. 66. E.g., DAVIES – ALLISON, Matthew (n. 52), vol. 2, p. 340. 67. Ibid. Note too that FH Luke moves Jesus’ statement on blasphemy against the Holy Spirit to a significantly different context, in which the connection with its immediately preceding context (Luke 12,8-9) is difficult. Cf. TUCKETT, Revival (n. 27), p. 88. The blasphemy

250

C.M. SAULINA

For Luke, however, the phrase δάκτυλος θεοῦ fits naturally with his unique version of the parable of the strong man (Luke 11,21-22). The statement in Luke 11,20 not only debunks the false claim regarding Jesus’ source of power, but also functions as a response to those who ask for a sign from heaven to test him (11,16). The phrase δάκτυλος θεοῦ occurs only here in the NT, and only three times in the LXX in two different settings: 1) in Exod 8,19, when the third plague hits the Egyptians, and 2) in Deut 9,10 par. Exod 31,18, when God gives the two covenant tablets to Israel. These contexts are made even more relevant by virtue of their own respective requests for a sign. In the first setting, Moses competes with the Pharaoh’s magicians in performing miraculous works (“sings”). Initially, the rival magicians can imitate Moses’s work. However, at the third plague, they can no longer keep up and declare that the plague has been brought by the finger of God (Exod 8,14-15): δάκτυλος θεοῦ ἐστιν τοῦτο (v. 15). The plague becomes the sign of God’s power that accompanied Moses. By alluding to Exod 8,19 in 11,20, Luke stresses the exclusivity of God as the source of power behind these two events while simultaneously setting up an ironic contrast between the Egyptian magicians and Israel (the crowd). The former group was able to recognize YHWH’s act through Moses, in contrast to the latter group who could not identify God acting through Jesus. Moreover, in the second Pentateuchal event featuring the phrase δάκτυλος θεοῦ, the covenantal tablets are said to be written by the finger of God (ἐν τῷ δακτύλῳ τοῦ θεοῦ). The tablets become the ‘true sign’ of YHWH’s status as Israel’s God. Likely alluding to both events, Luke underscores the importance of Jesus’s exorcisms and the protagonist’s Christological identity. These elements are missing from Mark’s version of the Beelzebul passage. Strikingly, Mark never associates God’s kingdom with Jesus’s miracles68. By contrast, in his version of the pericope, Luke emphasizes the authoritative, revelatory, and, potentially, covenantal aspects of Jesus’s exorcisms. In the Hebrew Bible, the demonstration of God’s salvific power provides the foundation for enacting the law at Sinai (Deuteronomy 9–10). Likewise, in Luke’s gospel, God’s salvific power that manifests in Jesus’s exorcisms/healings precedes God’s establishment of the new order, the coming of the kingdom of God on earth.

saying seems to contradict the preceding two verses. The MPH’s explanation needs not envision a deliberate recontextualization of Matthew’s saying (contrast FH). 68. RÄISÄNEN, Exorcisms (n. 54), p. 134.

COMPETITIVE TRADITIONS

251

5. Luke 11,21-28: Narrative-Critical Observations The Lukan Jesus further elucidates the importance of the coming of the kingdom in 11,21 with the new version of the parable of the strong man. While Luke may have been inspired by the Markan version, he either presents a redactional reworking of Mark 3,2769 aligning with his theological agenda or has adapted Luke 11,21-22 from a source other than Mark70. The differences from Mark’s version are considerable71. Mark and Matthew portray a burglary scene: someone breaks into a strong man’s house (οἰκία), ties him up (δέω), and steals his property (σκεῦος). By contrast, Luke depicts a large-scale military conflict, presumably between Jesus as the agent of God’s kingdom and Satan and his allies72. On this interpretation, Satan is a powerful king living in a fully guarded palace (καθωπλισμένος φυλάσσῃ τὴν ἑαυτοῦ αὐλήν) where all of his belongings are safe (ἐν εἰρήνῃ ἐστὶν τὰ ὑπάρχοντα αὐτοῦ). Jesus (ἰσχυρότερος αὐτοῦ) successfully invades Satan’s territory (ἐπελθὼν νικήσῃ αὐτόν) and distributes the plunder (σκῦλον)73. It is worth noting that by introducing the character of the stronger (ἰσχυρότερος) one, Luke links this parable with John the Baptizer’s words in 3,16 (ἔρχεται δὲ ὁ ἰσχυρότερός μου). Over the course of Luke’s narrative, Jesus already confirms himself as the fulfilment of this prophecy in 7,22. In fact, Luke consistently blurs the line between healings and exorcisms74, heightening the scope of evil spirits’ influence and power over humankind. According to Luke, the defeat of Satan’s kingdom (as manifested in the fall of Satan and exorcisms) is “the epitome of Jesus’s ministry”75. The list in 7,21 also has further significance in the linking of v. 21 to v. 22, whereby Luke associates Jesus’s power, manifested through healings and exorcisms, with the fulfilment of God’s eschatological promise. This idea corresponds to what we find in 11,20-22.

69. NEIRYNCK, Assessment (n. 54), pp. 271-273. 70. E.g., FLEDDERMANN, Q (n. 60), pp. 484-488; BOVON, Luke 2 (n. 44), p. 116. 71. FLEDDERMANN, Q (n. 60), pp. 484-488. 72. Cf. EVE, Devil (n. 29), p. 39. 73. Luke infuses 11,18 and 11,20-22 with the language of kingdom and power, e.g., βασιλεία, ἰσχυρός, καθοπλίζω, ὑπάρχω, and νικάω. Cf. J.B. GREEN, The Gospel of Luke (NICNT), Grand Rapids, MI, Eerdmans, 1997, p. 455. 74. Cf. G. TWELFTREE, Jesus the Miracle Worker, Downers Grove, IL, InterVarsity Press, 1999, p. 176; E. SORENSEN, Possession and Exorcism in the New Testament and Early Christianity (WUNT, II/157), Tübingen, Mohr Siebeck, 2002, p. 135; R.H. BELL, Deliver Us from Evil: Interpreting the Redemption from the Power of Satan in New Testament Theology (WUNT, 216), Tübingen, Mohr Siebeck, 2007, p. 69. 75. Drawing on TWELFTREE, Miracle Worker (n. 74), pp. 275, 347.

252

C.M. SAULINA

Luke’s distinctive ‘stamp’ can also be detected in the wording of 11,21-22. While there are some hapax legomena, Lukan redactional vocabulary dominates76. This incudes such characteristically Lukan words as φυλάσσω, ὑπάρχω, ἐπέρχομαι, πείθω, διαδίδωμι77. The hapaxes include καθοπλίζω and σκῦλον, not encountered elsewhere in the New Testament, and πανοπλία, which appears only in Luke and Eph 6,11.13. Generally speaking, hapaxes do not necessarily suggest the use of a source78. Furthermore, Luke has a tendency to use uncommon words79. While some pre-Lukan influence cannot be fully ruled out, it seems less likely than the alternative. Luke’s inclusion of verse 23 after the parable of the strong man may appear somewhat abrupt. However, this saying fits well with the parable’s bellicose imagery. The parable’s plunderers represent those who have been set free from Satan’s bondage. Witnessing and experiencing Jesus’s liberating ministry therefore requires an urgent response; one must decide whose side they will adopt in the cosmic saga. The implications of choosing poorly are illustrated with the ensuing parable of the return of the unclean spirit. The shift from the palace/kingdom setting to that of a house redirects attention from a cosmic event to personal space. Strikingly, in both parables/settings, this space is initially occupied by demonic entities. In light of the instruction on prayer with its climactic emphasis on asking for the Holy Spirit (11,13), the parable can be interpreted as urging the readers to invite the Holy Spirit to ‘guard’ the ‘house’. Once again, it is made clear that Jesus and Satan cannot be in the same camp. One gathers and builds, while the other destroys. Finally, Luke includes 11,27-28, apparently with Mark 3,31-35 on his mind80. Although Luke has already included a parallel to this Markan material in Luke 8,19-21, the emphasis on a personal decision in 11,23-26 necessitates reiterating the message heard previously in 8,21. However, rather than telling the same story twice (or creating a doublet) – unless, that 76. FLEDDERMANN, Q (n. 60), pp. 486-487; EVE, Devil (n. 29), pp. 38-39. 77. See the authors listed for each term in A. DENAUX – R. CORSTJENS – H. MARDAGA, The Vocabulary of Luke: An Alphabetical Presentation and a Survey of Characteristic and Noteworthy Words and Word Groups in Luke’s Gospel (BiTS, 10), Leuven – Paris – Walpole, MA, Peeters, 2009, pp. 148, 230-231, 491, 617-618, 634-635. Cf. J.C. HAWKINS, Horae Synopticae: Contributions to the Study of the Synoptic Problem, Oxford, Clarendon, 21909, pp. 15-25, 27-29; EVE, Devil (n. 29), pp. 38-39. 78. E.g., R.H. STEIN, The “redaktionsgeschichtlich” Investigation of a Markan Seam (Mk 1,21f.), in ZNW 61 (1970) 70-94, p. 73; GOULDER, Luke (n. 58), pp. 20-22. 79. Ibid., p. 21; EVE, Devil (n. 29), pp. 38-39. 80. E.g., GOULDER, Luke (n. 58), p. 510.

COMPETITIVE TRADITIONS

253

is, 11,27-28 originates in a source or is an adaptation of one81 – Luke changes the narrative setting and supplies the beatitudinal motif of the blessedness of those who hear the word of God and obey it. V. THE BEELZEBUL CONTROVERSY IN MATTHEW’S ACCOUNT: NARRATIVE AND SOURCE-CRITICAL OBSERVATIONS Matthew locates and shapes the Beelzebul controversy differently from the other synoptics. Table 1 below shows the parallels that Matthew has with Mark and Luke. The fact that Matthew has all the rows filled in (except one) suggests his medial position. This impression is bolstered by the verbatim agreements between Mark and Matthew. The correspondences and verbatim agreements in the sections parallel to Luke’s account are also noticeable. As I suggested earlier by way of a preliminary discussion, Matthew may be here following Luke rather than a common source, such as Q82. It seems hard to explain how two independent composers (Luke and Matthew) ended up using Mark and Q in an almost identical manner83. In fact, the parallels extend beyond the passage in question; Matthew and Luke also include parallel units that follow the Beelzebul controversy. Therefore, I propose that Matthew followed Luke’s scope of the Beelzebul passage while adding some materials from Mark and from remote Lukan passages. The following discussion will attempt to demonstrate the feasibility of the MPH as a synoptic solution in this particular pericope. 81. E.g., HOLTZMANN, Synoptiker3 (n. 32), p. 366; BOVON, Luke 2 (n. 44), p. 129. 82. For FH proponents, the editorial direction is reversed. Cf. H.B. GREEN, Matthew 12.22-50 and Parallels: An Alternative to Matthean Conflation, in C.M. TUCKETT (ed.), Synoptic Studies: The Ampleforth Conferences of 1982 and 1983 (SupplJSNT, 7), Sheffield, JSOT, 1984, 157-176. However, the FH and MPH proponents are in agreement in stressing the close parallels between Matthew and Luke, which pose a problem for the 2DH’s independent composition hypothesis. Considering the level of freedom that each evangelist exhibited in paraphrasing, relocating, and reordering their source materials, 2DH theorists need to better explain the accumulation of ‘accidental’ parallels in this pericope. For further arguments against the 2DH, see MACEWEN, Matthean Posteriority (n. 30), pp. 50-73; M. GOODACRE, Too Good to Be Q: High Verbatim Agreement in the Double Tradition, in POIRIER – PETERSON (eds.), Marcan priority without Q (n. 29), 82-100; GARROW, Streeter’s ‘Other’ (n. 30), pp. 212213. On this specific episode, see also OLSON, Unpicking (n. 28), pp. 139-142; EVE, Devil (n. 29); ID., Relating (n. 28), pp. 132-137. 83. Just like in other Mark-Q overlaps, GOODACRE, Overlaps (n. 27), p. 204, rightly observes: “There are agreements in order, where Matthew and Luke agree in placing the same double-tradition material in the same triple-tradition context, something that is not supposed to happen on the Two-Source Theory, and there are major agreements in wording” (emphasis added). Cf. EVE, Relating (n. 28), p. 76; KLINGHARDT, Gospel (n. 64), pp. 190-191.

254

C.M. SAULINA

Matthew adopts Mark’s deployment of this incident in the context of Jesus’s confrontations with religious leaders. In 9,1-17, Matthew presents the first three (of five) confrontation stories that can be found in Mark 2,1– 3,6. The copying of the Markan confrontations is then interrupted, with the last two incidents deferred until Matt 12,1. From this point on Matthew will closely follow Mark’s order with few modifications. Throughout Matthew 12, the evangelist highlights the Pharisees’ hostility and records Jesus’ responses. Matthew includes only two passages in this chapter that do not correlate with this agenda – the fulfilment of Isaiah’s prophecy in vv. 12-21 and the true kindred in vv. 46-49. In terms of the setting of the Beelzebul episode, Matthew’s account resembles Luke’s. As regards the wording, Matthew’s account corresponds to both the Markan and Lukan versions. This resemblance suggests that Matthew may have conflated Mark and Luke. M.E. Boring aptly observes that one half of “a pair contained in Matthew” can be found in Mark, the other half in Luke (see Table 1 below)84. In terms of the verbal agreements between Matthew/ Mark, and Matthew/Luke, the second set exhibits much closer parallels. The differences between Matthew and Luke may therefore suggest Matthean editorial work. Table 185 Sections Mark’s introduction Matthew’s/Luke’s introduction The accusation Jesus knows the opponents’ mind Satan expels Satan Divided kingdom and house Satan divided Your sons’ exorcism The kingdom of God The parable of the strong man With/against Jesus The unforgivable sin No good fruit from a bad tree

Mark

Matthew

Luke

3,20-21

-

-

-

12,22-23

11,14

3,22

12,24

11,15-16

-

12,25a

11,17a

3,23

(12,26a)

-

3,24-25

12,25b

11,17b

3,26

12,26b

11,18

-

12,27

11,19

-

12,28

11,20

3,27

12,29

11,21-22

-

12,30

11,23

3,29

12,31

(12,10)

-

12,32-37

(6,43-45)

84. BORING, Beelzebul (n. 22), p. 612. 85. A modification of Table 5 in EVE, Devil (n. 29), pp. 33-34.

255

COMPETITIVE TRADITIONS

Sections Return of the unclean spirit Doing the will of God/hearing the word and doing it Sign of Jonah

Mark

Matthew

Luke

-

12,43-45

(11,24-26)

3,31-35

12,46-50

(8,19-21; cf. 11,27-28)

(8,11-12)

12,38-42

(11,29-32)

Luke (a)

Mark (b)

Table 286 Matthew (ab) Exorcism as the initial setting

Yes

Specifying the group of opponents Jesus knows the opponents’ mind

Yes Yes

Divided house Your sons’ exorcism/The kingdom of God

Yes Yes

Binding the strong man Sin against the Holy Spirit (part 1) Sin against the Holy Spirit (part 2)

Yes Yes Yes

For instance, in the setting of the Beelzebul accusation (Matt 12,22-24), Matthew differs in describing the physical condition of the demoniac. The Matthean Jesus heals a blind and mute person, while Luke’s account does not feature blindness. This section has a parallel in Matt 9,32-33, which is much closer to Luke 11,14 on the word level. The FH proponents typically argue that in shaping his Beelzebul controversy FH Luke utilizes both sections87. On this theory, FH Luke simplifies the Matthean doublet by using just one account. In the process, FH Luke retains some of the wording of Matt 9,32-34, which actually resembles Mark 3,22. For Ken Olson, FH Luke’s adoption of Matt 9,34’s wording challenges the claim that Luke has unpicked Matthew’s Markan material88. FH Luke’s procedure in this section and Olson’s argument are debatable, however. First, why would FH Luke decline the opportunity to emphasize Jesus’s power, who in Matthew’s version heals two distinct (and difficult) illnesses, blindness and muteness? Alex Damm rightly suggests: “Now it would help Luke to show Jesus’ power if Jesus could cure muteness and blindness – in order words, could make the demoniac “see” (βλέπειν) – 86. A modification of the table in BORING, Beelzebul (n. 22), p. 612. 87. E.g., OLSON, Unpicking (n. 28), pp. 139-140; EVE, Relating (n. 28), pp. 132-133. 88. OLSON, Unpicking (n. 28), pp. 139-140.

256

C.M. SAULINA

especially since ‘blindness’ and ‘sight’ are important for Luke later in the controversy”89. Second, if Olson is right, why would FH Luke adopt a more complicated procedure by visually consulting two different parts of Matthew? One might theorize that Luke incorporated Matt 9,32-33 from memory, but this only raises further questions90. Luke could have followed Matt 12,22-23 without any impact on his overall compositional objective. On the reverse scenario, it is not difficult to explain the phenomena through the MPH’s lens. Matt 9,32-34 certainly shows Luke’s influence, while the doublet in Matt 12,22-24 can be seen as Matthew’s editorial work linking Matthew 12 to Matthew 991. In 12,22, Matthew repeats and reworks the earlier story to create a cross-reference92. With this link, he resumes the theme of an ongoing conflict between Jesus and the religious leaders, interrupted by the materials in Matt 9,35–11,30. Moreover, the demoniac’s dual ailment conveys a typically Matthean redactional feature: “Matthew loves duplicating persons and sometimes also texts”93. In fact, the presence of blindness and dumbness in Matt 12,22 may be a result of “combining reminiscences of the two miracles” of Matt 9,27-34 to reinforce the cross-reference94.

89. A. DAMM, Ancient Rhetoric and the Synoptic Problem: Clarifying Markan Priority (BETL, 252), Leuven – Paris – Walpole, MA, Peeters, 2013, pp. 69-80. 90. EVE, Relating (n. 28), p. 132, n. 57, notes that “It is admittedly far from obvious why Luke should choose to follow Mt. 9.32-34 rather than Mt. 12.12-28 here … Perhaps the best one can suggest is that, for reasons now lost to us, this is how Luke’s memory happened to work at this point” (emphasis added). This is what KLOPPENBORG, Conceptual Stakes (n. 19), pp. 22-24, describes as “renaming the problem”. 91. HAWKINS, Horae Synopticae (n. 77), pp. 95-96, juxtaposes Luke’s version as the “middle” form between the Matthean doublets. The context of Luke 11,14-15 resembles Matt 12,22-24 (the exorcism appears in a “defensive discourse”), while the wording is more similar to Matt 9,32-34, e.g., ἐκβάλλω and the clause καὶ ἐθαύμασαν οἱ ὄχλοι. The wording of Matt 9,32-34 illustrates Luke’s potential influence on Matthew. In Matt 12,22-23, Matthew reworks this material since now his account will be informed by Mark and Luke. Matthean style can be seen in the use of προσφέρω (the word has a different function in Luke), δαιμονίζομαι, and the description of demoniac as blind and mute person. Moreover, doublets are characteristic of Matthew’s composition (see notes 92 and 94 below). 92. Hawkins (ibid., pp. 81-82) divides the phenomenon of doublets into three groups: 1) combination of two sources; 2) repeated use of the same source; and 3) editorial work of the author, repeating his own phrasing. Cf. Concerning the first category, see HENGEL, Gospels (n. 19), p. 176. In relation to the third group, Matthew often uses doublets for crossreferencing, as in Matt 9,32-34 and 12,22-24. In a way, these cross-references function much like footnotes in a modern academic text. Cf. BUTLER, Originality (n. 65), pp. 138-145. 93. HENGEL, Gospels (n. 19), p. 176. E.g., Matt 8,28–9,1 par. Mark 5,1-20; Matt 9,27-31 par. Mark 10,46-52. 94. HAWKINS, Horae Synopticae (n. 77), p. 96. Cf. J. NOLLAND, The Gospel of Matthew (NIGTC), Grand Rapids, MI, Eerdmans, 2005, pp. 498-499.

COMPETITIVE TRADITIONS

257

The conflation of Markan and Lukan motifs can be seen throughout the passage. Similar to Luke, Matthew begins with the bystanders’ response to Jesus’s exorcism after the demoniac is healed. Matthew, however, underscores the uniformity of the crowd’s responses (by adding πᾶς) and verbalizes their amazement: “is this not the Son of David?” If Luke had Matthew at his disposal, it is not clear why he would have omitted Jesus’s association with ‘Israel’ (Matt 9,33) and ‘David’ in his Beelzebul episode (Matt 12,23)95. Luke’s nativity story highlights the evangelist’s interest in portraying Jesus as the son of David and the redeemer of Israel (Luke 1,32-33.53-55; 2,4.11.25-26.32)96. Moreover, retaining Matt 12,23 would have produced a more explicit triad with three different crowd responses (cf. 11,15.16)97. As it currently stands, the triad is still noticeable, but it remains unclear why Luke would have omitted Matthew’s μήτι οὗτός ἐστιν ὁ υἱὸς Δαυίδ, which would have clarified the more ambiguous ἐθαύμασαν (Luke 11,14). In contrast to the Lukan accusation that originates with a segment of the crowd, Matthew describes it as coming from the Pharisees. Matthew is likely to have changed Luke’s version here to align with his overall positive assessment of crowds in the gospel98. His inclusion of the Pharisees as the source of the accusation may have been influenced by Mark 3,22, where a group of religious leaders is presented as Jesus’ principal opposition in the story99. On the other hand, if Luke had known Mark and Matthew, the omission of the religious authorities from the episode, the Pharisees in particular, is somewhat strange. This would be another case of FH Luke omitting congenial materials found in his Matthean source100. Moreover, FH Luke would not only have rejected this element, but also replaced it with a faceless, unidentified, 95. MACEWEN, Matthean Posteriority (n. 30), p. 46. As McEwen rightly points out, Goulder does not address this question. He merely suggests that Luke wishes to condense the doublet into one account while retaining some of the divergent wording; see GOULDER, Luke (n. 58), pp. 502-503. 96. Contra SHELLARD, Light (n. 1), p. 71, who argues that Luke omits ‘Israel’ in the Beelzebul scene because he does not have a similar interest as Matthew. 97. For Lukan triads, see, e.g., Luke 3,10-14; 6,6-11//13,10-17//14,1-6; 9,57-62; 10,30-35; 15,3-32. 98. R.C. BRANDEN, Satanic Conflict and the Plot of Matthew (Studies in Biblical Literature, 89), New York, Lang, 2006, p. 107; C.L. WAHLEN, Jesus and the Impurity of Spirits in the Synoptic Gospels (WUNT, II/185), Tübingen, Mohr Siebeck, 2004, p. 138. 99. In fact, Matthew follows Mark in depicting the controversy as the climax of the animosity between Jesus and the religious authorities. Cf. GREEN, Matthew 12.22-50 and Parallels (n. 82), p. 157. 100. Especially in light of the ensuing Luke 11,37-54. But see on Luke’s overall complex view of the Pharisees, e.g., J.A. ZIESLER, Luke and the Pharisees, in NTS 25 (1979) 146-157.

258

C.M. SAULINA

and divided crowd, thus weakening the force of the accusation and minimizing the necessity of Jesus’s response. This seems illogical, especially in light of Luke 11,37-54. Matthew’s introduction of new antagonists comes with some implications for Matt 12,27. Scholars have debated the identity of οἱ υἱοὶ ὑμῶν in this verse. In Matthean context, this becomes a reference to the pharisees, instead of the more general sons of Israel in Luke. However, R.C. Branden correctly notes that “there are no New Testament passages that endorse Jewish exorcists who were members of the opposition to Jesus outside of this passage”101. Once again, Matthew seems to betray his source. Luke’s influence can also be seen in Matt 11,25a (cf. Luke 11,17a), an issue often overlooked by synoptic specialists102. The verbatim agreements in this verse are significant. Both evangelists use the participle of οἶδα at the beginning of the sentence and εἶπεν αὐτοῖς at the end. Moreover, both Matthew and Luke describe the antagonists’ mind as the object of οἶδα, although each evangelist refers to it with a different word (ἐνθύμησις for Matthew and διανόημα103 for Luke). Matthew’s reference to this in 12,25a seems at odds with the previous verses; both the crowds and the Pharisees express their opinions out loud104. Thus, the statement in v. 25a is unnecessary. The narrator’s explanation in 12,25a betrays Matthew’s source; he was influenced by Luke’s plot of the Beelzebul episode. As noted earlier, in Luke the crowd is divided into three groups. Two of them attempt to defame Jesus by 1) accusing him of using Beelzebul’s power; 2) testing him (πειράζω) with a request for a sign. Looking at this context, the narrator’s reference to Jesus knowing their thoughts (Luke 11,17a) connects very well with Luke 11,16. This knowledge aligns with the secret intention of some bystanders to entrap Jesus with their request. The intention, however, is not 101. BRANDEN, Conflict (n. 98), p. 62. 102. But see F. NEIRYNCK, Mt 12,25a / Lc 11,17a et la rédaction des évangiles, in ID., Evangelica II: 1982-1991. Collected Essays, ed. F. VAN SEGBROECK (BETL, 99), Leuven, Leuven University Press – Peeters, 1991, 481-492. Neirynck argues here for an independent redaction of Mark by Matthew and Luke. See, however, LUZ, Matthew 8–20 (n. 61), p. 200, n. 23. 103. This word is another hapax in Luke’s version of the passage. Of course, the value of this observation is limited (see n. 78 above). 104. Contra BRANDEN, Conflict (n. 98), p. 61, who argues that the Pharisees express their false charge by whispering. But if so, how can they influence the crowd? This is the main purpose of their accusation, and Jesus’ ensuing response is aimed at debunking it (in the presence of the crowd). However, Branden (ibid., p. 109) rightly notes: “The upshot of all this is that there is conflict between Jesus and the Pharisees and they are battling over the crowds”. Cf. NOLLAND, Matthew (n. 94), p. 499: “we must, rather artificially, think of Jesus as being out of earshot”.

COMPETITIVE TRADITIONS

259

so secret in Matthew’s version. As for Matthew’s decision to change the term διανόημα to ἐνθύμησις, the latter is more frequent in Greek literature105. Therefore, Matthew would have deliberately used a term with which his audience was more familiar. Having said all this, the fact that Matthew still – needlessly – included Jesus’ knowledge of the Pharisees’ mind might be considered an example of Matthew’s editorial fatigue106. Luke’s influence on Matthew can also be noticed in Matt 12,25b, the parable of the divided kingdom. The verbatim agreements are apparent (see Table 3 below). Matthew may have found Luke 11,17’s οἶκος ἐπὶ οἶκον πίπτει unclear and decided to rephrase this portion of the parable with Mark’s wording. Once again, we have an example of Matthew conflating Mark and Luke. Boring correctly observes that Markan and Lukan elements can be seen throughout the Matthean passage (see Table 3)107. In terms of Jesus’ words, Matthew attempts to follow and synchronize his two sources closely with minimal editorial changes. As Table 3 shows, the verbatim agreement is especially strong in Matt 12,27-28 and Luke 11,19-20. This could be perceived as a potential weakness for the MPH since, generally, Matthew does not have as big of a role for exorcisms in explicating Jesus’ Christological and eschatological work as the other synoptists108. However, an author is not necessarily under 105. The term διανόημα (in all cases) appears only 69 times in the TLG database, while ἐνθύμησις (in all cases) is twice as frequent, i.e., 139 times in the same database. 106. On which see, e.g., M. GOODACRE, Fatigue in the Synoptics, in NTS 44 (1998) 45-58. But see the rejoinder in T. HÄGERLAND, Editorial Fatigue and the Existence of Q, in NTS 65 (2019) 190-206. 107. BORING, Beelzebul (n. 22), p. 619 (Luke = Q). 108. Cf. WAHLEN, Impurity (n. 98), p. 114; T. LÖFSTEDT, Devil, Demons, Judas, and the Jewish Opponents of Christ in the Gospels, Eugene, OR, Pickwick, 2021, p. 134. Compare, e.g., Mark 1,21-28//Luke 4,31-37 and the scattered echoes of these materials in Matt 7,8-29; 13,54; 22,33. The exorcism of the Capernaum synagogue demoniac disappears completely, and the amazement at Jesus’ teaching – a recurring emphasis in Matthew – is scrubbed clean of the Markan/Lukan association with an exorcism. The clearest Matthean echo of the Capernaum episode (Matt 7,28-29) indicates that the evangelist was interested in the superiority of Jesus’ authority to that of the scribes and not so much in exorcistic activity. As a result of these editorial decisions, Matthew, in contrast to Mark and Luke, also deselects an exorcism as the very first miraculous deed of Jesus. Instead, this position is occupied by the story of Jesus healing a leper (Matt 8,1-4). It is important, of course, not to press the point too far. Matthew does retain the exorcistic aspect of Jesus’ activity in his summary passage Matt 4,23-25 (par. Mark 3,7-12) and in the sending of the twelve (Matt 10,8 par. Mark 6,7.13). It was therefore still an important part of Matthew’s tradition and congenial to him, even if it does not hold the same significance as in his Markan (and, on the MPH, Lukan) source. Thus, LUZ, Matthew 8–20 (n. 61), p. 204: “In Matthew the kingdom of God reaches the people without being exhausted in what occurs in the miracles and exorcisms”. On Mark 3,7-12 as the parallel for Matt 4,23-25, see F. NEIRYNCK, The First Synoptic Pericope: The Appearance of John the Baptist in Q?, in ID., Evangelica III: 1992-2000. Collected Essays

260

C.M. SAULINA

an obligation to improve on ideas borrowed from his source(s). One cannot rule out the possibility that Matthew would take a certain verse or idea from his source without further developing it. Such an idea might well have been taken ‘for granted’ and its presence in the gospel could be an indication that both the author and the intended reader are familiar and in agreement with what is being copied. I propose that this is what happened when Matthew adopted 12,27-28 from Luke 11,19-20. On both the 2DH and MPH, when it comes to Jesus’ messianic identity, Matthew does not add new material highlighting the importance of exorcisms – or the defeat of Satan and his kingdom/allies – beyond what he already finds in his sources109. It is not that Matthew somehow rejects these ideas; rather, Satan’s authority and kingdom simply do not receive further attention. The situation is different in Luke. First, although the temptation scenes in Luke and Matthew look very similar (except some obvious differences in the order), Satan only claims authority over the kingdoms of the world in Luke’s version (4,5-7). The same evangelist emphasizes the cosmic battle between God’s servants and the Satanic forces (Luke 10,17-19; 22,3.31.53; Acts 26,18). Second, unlike Matthew (on any hypothesis), Luke on at least one occasion appears to think of healings as acts of exorcizing and defeating demonic entities who are seen as the actual culprits of human afflictions110. Third, in Jesus’ response to John’s disciples, Matthew does not emphasize Jesus’ exorcisms as part of the proof of his messianic identity (Matt 11,3-4 diff. Luke 7,20-22, esp. v. 21). This passage is particularly important because in Luke it provides the closest parallel to Luke 11,19-20//Matt 12,27-28, showing how Jesus’ power in performing miraculous acts relates to God’s eschatological plan of salvation. MPH Matthew would have omitted this detail at Luke 7,21. Therefore, because MPH Matthew does not otherwise stress the motif of cosmic warfare between God’s forces and Satan’s, he is likely to have adopted Matt 12,27-28 from Luke on account of his interest in the high Christological point placed by Luke on the lips of Jesus. (BETL, 150), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2001, 209-244, pp. 235-236. 109. On the FH, Matt 12,27-28 would be one such addition (unless one turns to the ad hoc use of pre-Matthean traditions at this juncture). This addition – especially if Matt 12,27-28 is Matthean composition and not based on previous sources – runs curiously counter to the evangelist’s lack of this theme’s development elsewhere. 110. Luke 4,38-39 (diff. Mark 1,29-31//Matt 8,14-15). Note the recurring use of ἐπετιμαω in Luke 4,39.41 (diff. Mark 1,29-34//Matt 8,14-17). This verb is not used in the context of curing (non-exorcism-based) illnesses in Matthew or Mark. E.g., F. BOVON, Luke 1: A Commentary on the Gospel of Luke 1:1–9:50 (Hermeneia), trans. Christine M. Thomas, Minneapolis, MN, Fortress, 2002, p. 164.

COMPETITIVE TRADITIONS

261

A more characteristically Matthean motif, that of judgement, permeates Matt 12,15-45. To highlight this motif, Matthew concludes the Beelzebul controversy with the unforgiveable sin against the Holy Spirit (12,31-32) and defers the parable of the evil spirit’s return by inserting two other passages that heavily emphasize judgement (12,33-42). It is worth noting that the Matthean parable of the unclean spirit’s return would have been lightly edited to convey the motif of judgment (see the conclusion of v. 45: οὕτως ἔσται καὶ τῇ γενεᾷ ταύτῃ τῇ πονηρᾷ). Returning to the wording of Matt 12,31-32, the resemblance to both the Markan and Lukan parallel sayings is apparent. Matt 12,31 contains many verbatims agreements with Mark 2,28, while Matt 12,32 is closer to Luke 12,10. Once again, we see further evidence of Matthew conflating Mark’s and Luke’s accounts. The repetition of blasphemy against the Holy Spirit in Matt 12,31-32 also bolsters the view that Matthew took these sayings from two different sources and combined them here side-by-side. To summarize, MPH Matthew incorporates both Markan and Lukan ideas and occasionally modifies them to fit his own theological agenda. Drawing from Mark’s account, Matthew sharpens Jesus’ conflict with the religious leaders. Drawing from Luke, Matthew heightens the significance of Jesus’ miracles vis-à-vis his messianic identity. Thus, Matthew utilizes the previous versions of the Beelzebul controversy and their source contexts to underscore the motif of judgement against those who rejected Jesus’ ministry, especially the religious leaders. At the same time, he also stresses Jesus’ Christological identity, especially the eschatological nature of his miracles – even as the eschatological aspect of exorcisms is not necessarily developed by Matthew elsewhere. The Matthean Jesus claims that he is greater than the Temple, Solomon, and Jonah (Matt 12,6.41.42), and through his ministry the kingdom of God manifests on earth (Matt 12,28). VI. MACROANALYSIS: MATTHEW’S AND LUKE’S COMPETITIVE CONTEXTUALIZATION OF THE BEELZEBUL CONTROVERSY 1. Competitive Contextualization It is crucial to analyze the details of the Beelzebul controversies in tandem with their Matthean and Lukan contexts. In this manner, one can arrive at a more complete picture of how the passages function in each gospel vis-à-vis some major theological issues presented by each evangelist and the overall plot of each gospel. A major part of each evangelist’s objective is to present a more appropriate version of the story of Jesus for his own audience. Due to the limited space and in continuity with my adoption of the MPH, I will

262

C.M. SAULINA

focus here on Matthew as the third synoptist in relation to the other two. In this section, we will see how MPH Matthew’s utilization of his Markan and Lukan materials, especially in the context of the Beelzebul controversy, fits with his competitive (con)textualization of both sources. The latter phenomenon, I hope, will show why the MPH is more probable than the FH. The previous section showed that MPH Matthew appears to conflate the materials from Mark and Luke in his version of the Beelzebul controversy. The phenomenon of Matthean conflations can also be observed beyond this pericope. In terms of order, starting from chapter 12, Matthew follows Mark more closely than in the previous chapters. However, in the vicinity of the Matthean Beelzebul scene one can still observe that the evangelist has been influenced by both Mark and Luke. In the episode itself, MPH Matthew includes all of the sections from Luke’s version. After the episode’s conclusion, he retains one of the Lukan pericopae that follow (Luke 11,16.29-32; cf. Matt 12,38-42), while moving another to a different location in his gospel (Luke 11,33-36; cf. Matt 6,22-23). The presence of the True Relatives segment (Matt 12,46-50) also supports Luke’s influence (cf. Luke 11,27-28). In Matt 12,22-45, the evangelist therefore concentrates on working Lukan materials into his Markan framework than runs through chapter 12 and beyond. The Lukan Beelzebul controversy is surrounded by passages that have parallels in Matthew’s Sermon of the Mount (SoM), e.g., the Lord’s prayer (Luke 11,1-4//Matt 6,9-13), the teaching on perseverance in prayer (Luke 11,9-13//Matt 7,7-11), and the parable of the lamp of the body (Luke 11,33-36//Matt 6,22-23). In fact, most of the materials in Luke 10–13 can be found in Matthew’s Gospel. This shows another range of Luke’s influence on MPH Matthew’s work. While retaining some Lukan materials, Matthew follows his own compositional strategy; this necessitates relocation and omission of certain source materials. Some of the reasons for this become clear as we examine the conundrum that Matthew needs to solve when, as a conflator, he must reconcile his sources. 2. Matthew Conflator and Luke’s Travel Narrative MPH Matthew’s conflating activity shows the esteem in which he held Mark’s and Luke’s works, underscoring his role as a tradent111. He attempts to incorporate as much material as possible from both sources112. Nonetheless, 111. KIRK, Q in Matthew (n. 7), pp. 299-300 (on 2DH Matthew as a tradent). 112. Ibid., p. 300: a project of “tradition-consolidation”.

COMPETITIVE TRADITIONS

263

Matthew also has another agenda, i.e., to present the best gospel that he can and that is called for by his audience. At this point, competitive textualization comes into play. To accomplish his compositional objective, Matthew ‘reshapes’ his sources by omitting, adding, relocating, revising, and transforming them on both micro- and macrolevels. On the microlevel, we can see some variation in Matthew’s wording of the Markan and Lukan parallel passages (e.g., the Matthean editorial reworking of their versions of the Beelzebul controversy). On the macrolevel, Matthew conducts a selection process. Any materials that do not conform to his theological agenda are omitted (e.g., a large number of passages in Luke 14–17). During the macro-restructuring process, as a conflator, MPH Matthew was nevertheless faced with two major conundrums. First, he needed to reconcile his two sources when they diverged. This issue is particularly glaring in terms of how the Markan and Lukan passages parallel to Matthew appear in a different order in each respective Gospel113. Luke presents many significant events (from Matthew’s perspective) that occur in a context different from Mark. In other words, these events are assigned different roles in the overall plot of each respective gospel114. Another conundrum faced by MPH Matthew was the structure of Luke’s so-called travel narrative (TN)115, which comprises many of Jesus’ sayings

113. On the argument from order in the discussion of the synoptic problem, see B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, p. 183; C.M. TUCKETT, Q and the History of Early Christianity: Studies on Q, Edinburgh, T&T Clark, 1996, pp. 8-9; J.S. KLOPPENBORG, Excavating Q: The History and Setting of the Sayings Gospel, Minneapolis, MN, Fortress, 2000, pp. 18-33; FLEDDERMANN, Q (n. 60), pp. 46-54; GARROW, Streeter’s ‘Other’ (n. 30), pp. 219-221. 114. In Mark, Jesus’ confrontations with the religious authorities begin very early in the gospel (2,1–3,6), culminating in the Pharisees’ plan to kill Jesus (3,6). The inclusion of the law experts from Jerusalem in the Beelzebul controversy belongs to the aftermath of an already heightened animosity. The Markan Jesus’ response to the accusation signals judgment over the authorities’ rejection of his ministry. These hostility from and judgment over the religious authorities are retained by Matthew in his version of the Beelzebul controversy and the surrounding context. In contrast, Luke does not use the Beelzebul passage in the context of Jesus’ confrontations with the religious authorities. However, on his part, Luke conveys a Christological message that matches certain aspects of Matthew’s agenda. As a result, given that MPH Matthew would have found both of his source versions congenial, the differing order of those sources vis-à-vis the Beelzebul episode’s placement would likely have created some issues. 115. The extent and very existence of this section have been a subject of vigorous dispute. Cf. WOLTER, Lukasevangelium (n. 44), pp. 365-366. There are, however, some general points of agreement. 1). In Luke we can observe a middle part (9,51–18,14 [or thereabouts]), in which Luke departs from Mark’s order and mostly utilizes non-Markan materials. 2). These materials are presented as events that happen during Jesus’ journey to Jerusalem. Cf. A. DENAUX, Studies in the Gospel of Luke, Berlin, LIT Verlag, 2010, p. 39. I therefore follow

264

C.M. SAULINA

that are important for Matthew. These materials are significant due to their emphasis on the discipleship and the apologetic nature of these passages (in the context of Jesus’ confrontations with the religious leaders). In this central section of his gospel, Luke departs from Mark’s order considerably, and there appears to be no discernible organizational pattern for the section’s structure. Before continuing to MPH Matthew’s procedure in dealing with these problems, it is important to address the FH’s arguments in relation to the TN and the SoM. This issue is of obvious relevance to our discussion of the Beelzebul episode and its contexts. On the FH, Luke adopted and rearranged much of the SoM’s sayings materials116. In the process, while retaining some SoM materials in his Sermon on the Plain, FH Luke deferred a significant chunk of them to the TN. FH Luke’s motifs for this procedure presumably fall under two categories: theologically driven and ‘practical’ reasons. Mark Goodacre argues that FH Luke felt compelled to reorder Matthew, including the SoM, for a theological purpose117. In his opinion, for FH Luke, the TN signifies the importance of the Isaianic motif of the Way of The Lord, which Luke has in turn inherited from Mark118. Matthew’s structure, on the other hand, “diminishes the significance of the theme”119. Unlike Mark, Matthew spends much of Jesus’ Galilean ministry by adding materials to his Markan source120. Therefore, Matthew’s “restructuring of the Marcan narrative is unlikely to have proved appealing”121. Faced with this conundrum, FH Luke creates a new structure “that allows him to incorporate lots of the new Matthean material”122. While it may initially sound appealing, the argument seems debatable. Proposing a theological significance for the TN is one thing but claiming that this agenda leads Luke to rearrange Matthew’s structure of SoM is another matter. Even if one agrees with the importance of the ‘way of the

the conventional opinion, whereby such a section was part of Luke’s compositional plan, began at Luke 9,51, and concluded with the arrival to Jerusalem or its vicinity. 116. See now, e.g., J. ARCHER, Ancient Bioi and Luke’s Modifications of Matthew’s Longer Discourses, in NTS 68 (2022) 76-88. 117. M. GOODACRE, Re-walking the “Way of the Lord”: Luke’s Use of Mark and His Reaction to Matthew, in MÜLLER – NIELSEN (eds.), Luke’s Literary Creativity (n. 27), 26-43. Cf. J. MARCUS, The Way of the Lord: Christological Exegesis of the Old Testament in the Gospel of Mark, Louisville, KY, Westminster John Knox, 1992. 118. GOODACRE, Way (n. 117), pp. 31-34. 119. Ibid., p. 34. 120. Ibid., p. 29: “extended Galilean ministry”. 121. Ibid., p. 36. 122. Ibid.

COMPETITIVE TRADITIONS

265

Lord’ concept, it is crucial to note that, for Luke, order matters (1,3)! This observation can be supported with the manner in which Luke arranges his Markan source, whose order he mostly follows closely, with some occasional transpositions. In the TN, Luke seems to depart from his Markan source and, if one adopts the FH, certainly does not simply switch to following Matthew’s order. In this section, Luke therefore either utilizes his compositional creativity with greater force by reordering Matthew or follows the order of a different source (such as Q or L). Strikingly, the majority of the Matthean-Lukan parallel pericopae appear in the TN123. Although the reordering of the double tradition materials in Luke might be assessed as offering a better thematic unity than in Matthew, the TN’s overall arrangement does not really solve the crux of this section’s compositional structure. With a few exceptions124, scholars do not claim to have uncovered an obvious organizing principle125. First of all, even though the section is framed so as to narrate Jesus’s journey to Jerusalem, it is not a historical record of the journey. Second, there does not seem to be any apparent method in arranging the materials. The order of the episodes in the TN is neither chronological nor spatial126. Although we can see some progression in the sequence of the verbs (πορεύομαι in 9,51 and 17,11; διαπορεύομαι in 13,22; ἀναβαίνω in 18,31 and 19,28; ἐγγίζω in 19,29 vis-à-vis Jerusalem; ἐγγίζω … πρὸς τῇ καταβάσει in 19,37 vis-à-vis the Mount of Olives; with the city finally in view in 19,41), these are little more than literary devices designed simply to connect one section to another127, creating a flow128. Without them, the TN would arguably be just a collection of unrelated sayings.

123. See the list in HAWKINS, Horae Synopticae (n. 77), pp. 108-109. 124. M. GOULDER, Type and History in Acts, London, SPCK, 1964, p. 138; EVE, Relating (n. 28), pp. 167-195. 125. E.g., DENAUX, Studies (n. 115), p. 39. 126. Hence, famously, K.L. SCHMIDT, Der Rahmen der Geschichte Jesu: Literarkritische Untersuchungen zur ältesten Jesusüberlieferung, Berlin, Trowitzsch & Sohn, 1919, p. 269: “Jesus immer nach Jerusalem reist, aber auf dieser Reise gar nicht recht weiterkommt”. Cf. J.S. KLOPPENBORG, Luke’s Geography: Knowledge, Ignorance, Sources, and Spatial Conception, in J. VERHEYDEN – J.S. KLOPPENBORG (eds.), Luke on Jesus, Paul and Christianity: What Did He Really Know? (BiTS, 29), Leuven – Paris – Bristol, CT, Peeters, 2017, 101-143. 127. FITZMYER, Luke I–IX (n. 61), p. 825 (on 13,22; 17,11): “a mere convenience, since the division at these points is otherwise insignificant and somewhat arbitrary”. Cf. WEISS, Markus und Lukas7 (n. 32), p. 489: neither passage serves as “Einführung eines neuen Actes auf der Weiterreise”; WOLTER, Lukasevangelium (n. 44), p. 366. 128. Note also the accumulation of these terms toward the end of the TN (from Luke 18,31 on).

266

C.M. SAULINA

Denaux aptly notes that “there is a tension between the content (didactic saying traditions) and the frame (journey motif) of this section, a dissonance of form from content”129. I contend that this tension creates a significant obstacle for the FH’s procedure. FH Luke’s adoption, rearrangement, and relocation of a large number of sayings materials from the SoM to the TN significantly heightens this tension. The addition of Matthean materials to the TN, may, in fact, have obscured the organizational principle of this section. FH Luke’s relocation of the materials on Prayer (Luke 11,14.9-13; cf. Matt 6,9-13; 7,7-11), for instance, has no clear connection to the Beelzebul controversy (with Matthew’s ἐν πνεύματι θεοῦ removed and the Unforgivable Sin logion relocated). Moreover, it creates a complicated procedure for Luke’s source-utilization. Some FH proponents claim that Luke dislikes lengthy discourses; he generally accommodates only 12-20 verses130. Mark A. Matson has argued that this tendency gives Luke a reason to take some passages from the SoM (which consists of nearly 110 verses), leave some of the material in his Sermon on the Plain, and distribute the rest elsewhere, e.g., in the TN. By doing the latter, Luke presumably also improves the Matthean order131. Matson’s comparison of the aesthetic value of Matthew’s SoM and Luke’s rearrangement of its passages – with a focus on thematic unity – is misleading, however. It is certainly easier to identify thematic unity in smaller units, such as several isolated passages in the Lukan TN, than in a much larger collection of uninterrupted sayings like Matthew’s SoM. Nevertheless, a closer examination confirms that the latter’s structure is deliberate132. The logic behind attaching some materials from the SoM to new ones in the TN, in contrast, remains vague. Viewing the intertextuality between Luke and Matthew through the MPH’s lens provides better solutions to the problems presented by these two gospels. By placing Matthew within the author-tradent spectrum and assessing the macro-structure of his gospel, we can see his overall goal in relation to Mark’s and Luke’s work. Kari Syreeni persuasively suggests that “the evangelist’s basic plan was to produce a double yet unified gospel

129. This distinctive feature of the TN makes the structure of Luke’s gospel stands out among the other synoptics. Cf. DENAUX, Studies (n. 115), p. 39. 130. SHELLARD, Light (n. 1), p. 65; GOULDER, Luke (n. 58), p. 40. 131. M.A. MATSON, Luke’s Rewriting of the Sermon on the Mount, in GOODACRE – PERRIN (eds.), Questioning Q (n. 28), 43-70, p. 67. 132. E.g., U. LUZ, Matthew 1–7 (Hermeneia), trans. James E. Crouch, Minneapolis, MN, Fortress, 2007, pp. 172-174.

COMPETITIVE TRADITIONS

267

featuring the story and the teaching of Jesus”133. Matthew’s community seems to have valued the authority of both Mark’s and Luke’s writings. Thus, rather than making one of his predecessors superior to the other, MPH Matthew regards them equally134. Matthew’s community may also have presented certain needs that had to be met in the compositional plan (see Figure 1). This task may have at least to some degree informed Matthew’s competitive textualization vis-à-vis Mark and Luke, requiring Matthew to restructure his sources. All of these activities would have had as their aim the composition of a ‘double’ Gospel, the most complete (to date) version of the story of Jesus and his authoritative teachings135.

Figure 1. Matthew’s Basic Plan in a Competitive (Con)Textualization136

MPH Matthew solves the aforementioned twin conundrums by utilizing two basic strategies. First, he collects and categorizes his sources materials into two groups: narrative and discourse sections137. Vis-à-vis the latter, Matthew creates the five great discourses: chapters 5–7, 10, 13, 18, 24–25138. In addition to these groups of Jesus’s teachings, Matthew also collects Jesus’ miraculous deeds along with the disciples’ ministries (e.g., chapters 8–9). Matthew utilizes the formula καὶ ἐγένετο five times (7,28; 11,1; 13,53; 19,1; 26,1), using it to conclude each of the five discourses. Accordingly, these five verses also signify the transition from a particular discourse collection 133. K. SYREENI, The Making of the Sermon on the Mount: A Procedural Analysis of Matthew’s Redactoral Activity. Part I: Methodology & Compositional Analysis (AASF, 44), Helsinki, Suomalainen Tiedeakatemia, 1987, p. 83 (emphasis original). 134. Ibid., p. 84. 135. Ibid., pp. 75, 113. 136. This chart is a modification of Syreeni’s model of Matthew’s basic plan in a contribution setting. Ibid., p. 86. 137. The distinction between these two is not clear-cut. Ibid., p. 89. 138. MACEWEN, Matthean Posteriority (n. 30), p. 162.

268

C.M. SAULINA

to a narrative section. Matt 26,1 acts as the conclusion of the five pairs of narrative and teaching blocks and marks the transition to the passion narrative139. It is important to note that in incorporating and categorizing these materials, Markan and Lukan orders still play an significant role. Organizationally, Matthew chooses the path of least resistance, following the order of his principal sources140. By categorizing his materials into two categories, Matthew also manages to solve his second great conundrum, namely, utilization of Luke’s vaguely organized TN. Second, in dealing with Mark’s and Luke’s different orders, Matthew makes a number of transpositions. Here, Kirk’s observations concerning Matthew’s utilization of Q can be used by the MPH141. To accommodate the Lukan TN, Matthew follows Mark’s order and utilizes Markan pericopes as pegs to which he can attach topically related Lukan materials. In this manner the Markan narrative line allows Matthew to utilize the TN without replicating its sequence. By doing this, Matthew is able to reconcile and to align the materials from both of his sources. From the perspective of ancient media, how does MPH Matthew actually utilize his two sources to conduct these operations? Among the principal difficulties in embracing the FH are the complicated techniques that must be attributed to Luke’s adaptation of Matthew142. FH Luke appears to exhibit a considerable number of back-and-forth movements in his Matthean source as he adopts Matthew, especially in the TN143. Might the MPH be more plausible in explaining the synoptic phenomena? 139. SYREENI, Sermon (n. 133), p. 77. 140. Ibid., pp. 93-95. 141. KIRK, Q in Matthew (n. 7), pp. 278-280. 142. MCNICHOL – DUNGAN – PEABODY (eds.), Beyond the Q Impasse (n. 29), pp. 14-15; DERRENBACKER, Practices (n. 1), pp. 188-209; ID., The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, in FOSTER et al. (eds.), Studies (n. 1), 435-458, pp. 442-443; A. KIRK, Memory, Scribal Media, and the Synoptic Problem, ibid., 459-482, pp. 473-479; ID., Q in Matthew (n. 7), pp. 48-49, 149-150, 307-308. These observations problematize the FH’s envisioned procedures. The more detailed the revision plan, the greater the resistance. SYREENI, Sermon (n. 133), p. 92. 143. M.D. GOULDER, The Order of a Crank, in TUCKETT (ed.), Synoptic Studies (n. 82), 111-130, proposes that Luke moved backwards through his Matthean scroll to compose Luke 13,22–18,8. This type of a procedure seems less than probable. Cf. M. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, pp. 118-119 and n. 33 (endorsing Derrenbacker); DERRENBACKER, Practices (n. 1), pp. 192-200; KIRK, Memory (n. 142), pp. 474-475; K.-H. CHANG, Questioning the Feasibility of the Major Synoptic Hypotheses: Scribal Memory as the Key to the Oral–Written Interface, in JSNT 41 (2019) 407-432, p. 421. The backwards scrolling potentially occurs already before Luke 13,22. With regard to the Beelzebul controversy, MACEWEN, Matthean Posteriority (n. 30), p. 46, asks: “Why would Luke have bothered to scroll backwards in Matthew to use 9.32-33 when 12.22-23 was at hand?”

269

COMPETITIVE TRADITIONS

Table 3144 Matt 9,32-34

Mark 3,22-30

Matt 12,22-32

Luke 11,14-23; 12,10

14 Καὶ ἦν ἐκβάλλων

20 Καὶ ἔρχεται εἰς οἶκον· καὶ συνέρχεται πάλιν [ὁ] ὄχλος, ὥστε μὴ δύνασθαι αὐτοὺς μηδὲ ἄρτον φαγεῖν. 32 Αὐτῶν δὲ

21 καὶ ἀκούσαντες

22 Τότε προσηνέχθη

ἐξερχομένων ἰδοὺ προσήνεγκαν

οἱ παρ᾽ αὐτοῦ

αὐτῷ δαιμονιζόμενος δαιμόνιον [καὶ αὐτὸ ἦν]

ἐξῆλθον κρατῆσαι

τυφλὸς καὶ κωφός,

κωφόν· ἐγένετο δὲ

αὐτῷ ἄνθρωπον

αὐτόν· ἔλεγον γὰρ

καὶ ἐθεράπευσεν

τοῦ δαιμονίου

κωφὸν

ὅτι ἐξέστη.

αὐτόν, ὥστε

ἐξελθόντος

δαιμονιζόμενον.

τὸν κωφὸν λαλεῖν

ἐλάλησεν ὁ κωφὸς

33 καὶ ἐκβληθέντος

καὶ βλέπειν. 23 καὶ ἐξίσταντο

καὶ ἐθαύμασαν

τοῦ δαιμονίου ἐλάλησεν

πάντες οἱ ὄχλοι καὶ

οἱ ὄχλοι.

ἔλεγον· μήτι οὗτός

ὁ κωφός.

22 Καὶ

καὶ ἐθαύμασαν

οἱ γραμματεῖς

οἱ ὄχλοι λέγοντες·

οἱ ἀπὸ Ἱεροσολύμων

οὐδέποτε ἐφάνη

καταβάντες ἔλεγον

ἐστιν ὁ υἱὸς Δαυίδ;

24 οἱ δὲ Φαρισαῖοι

15 τινὲς δὲ ἐξ αὐτῶν

οὕτως ἐν τῷ Ἰσραήλ. ὅτι Βεελζεβοὺλ ἔχει ἀκούσαντες εἶπον·

εἶπον·

34 οἱ δὲ Φαρισαῖοι

καὶ ὅτι

οὗτος οὐκ ἐκβάλλει

ἐν Βεελζεβοὺλ

ἔλεγον· ἐν τῷ ἄρχοντι

ἐν τῷ ἄρχοντι

τὰ δαιμόνια εἰ μὴ ἐν τῷ ἄρχοντι

τῶν δαιμονίων ἐκβάλλει

τῶν δαιμονίων ἐκβάλλει

τῷ Βεελζεβοὺλ ἄρχοντι

τῶν δαιμονίων ἐκβάλλει

τὰ δαιμόνια.

τὰ δαιμόνια.

τῶν δαιμονίων.

τὰ δαιμόνια· 16 ἕτεροι δὲ πειράζοντες σημεῖον

144. The model for the tabulation of the synopsis in Table 3 is inspired by WAHLEN, Impurity (n. 98), pp. 180-181 (with further adjustments). This table is intended to convey only the big-picture view of the parallels between the three synoptics. As meticulously described by Boring, the agreements themselves can be further distinguished into two categories, based on the type and scope of the textual units. See the details in BORING, Beelzebul (n. 22), pp. 596-607.

270 Matt 9,32-34

C.M. SAULINA

Mark 3,22-30

Matt 12,22-32

Luke 11,14-23; 12,10 ἐξ οὐρανοῦ ἐζήτουν παρ᾽ αὐτοῦ.

23 Καὶ προσκαλεσάμενος

25 Εἰδὼς δὲ τὰς

17 αὐτὸς δὲ εἰδὼς

αὐτοὺς

ἐνθυμήσεις αὐτῶν

αὐτῶν τὰ διανοήματα

εἶπεν αὐτοῖς· πᾶσα

εἶπεν αὐτοῖς· πᾶσα

ἐν παραβολαῖς ἔλεγεν αὐτοῖς· πῶς δύναται σατανᾶς σατανᾶν ἐκβάλλειν; 24 καὶ ἐὰν βασιλεία ἐφ᾽

βασιλεία μερισθεῖσα βασιλεία ἐφ᾽ ἑαυτὴν

ἑαυτὴν μερισθῇ, οὐ

καθ᾽ ἑαυτῆς

διαμερισθεῖσα

δύναται σταθῆναι

ἐρημοῦται καὶ πᾶσα

ἐρημοῦται καὶ

πόλις ἢ οἰκία μερισθεῖσα

οἶκος

ἑαυτὴν μερισθῇ, οὐ δυνήσεται ἡ οἰκία

καθ᾽ ἑαυτῆς

ἐπὶ οἶκον

ἐκείνη σταθῆναι.

οὐ σταθήσεται.

πίπτει.

ἡ βασιλεία ἐκείνη· 25 καὶ ἐὰν οἰκία ἐφ᾽

26 καὶ εἰ ὁ σατανᾶς 26 καὶ εἰ ὁ σατανᾶς 18 εἰ δὲ καὶ ὁ σατανᾶς τὸν σατανᾶν ἀνέστη ἐφ᾽ ἑαυτὸν

ἐκβάλλει, ἐφ᾽ ἑαυτὸν ἐφ᾽ ἑαυτὸν

καὶ ἐμερίσθη, οὐ δύναται στῆναι ἀλλὰ

ἐμερίσθη· πῶς οὖν σταθήσεται

διεμερίσθη, πῶς σταθήσεται

τέλος ἔχει.

ἡ βασιλεία αὐτοῦ;

ἡ βασιλεία αὐτοῦ; ὅτι λέγετε ἐν Βεελζεβοὺλ ἐκβάλλειν με τὰ δαιμόνια.

27 καὶ εἰ ἐγὼ ἐν

19 εἰ δὲ ἐγὼ ἐν

Βεελζεβοὺλ ἐκβάλλω Βεελζεβοὺλ ἐκβάλλω τὰ δαιμόνια, οἱ υἱοὶ

τὰ δαιμόνια, οἱ υἱοὶ

ὑμῶν ἐν τίνι

ὑμῶν ἐν τίνι

ἐκβάλλουσιν; διὰ

ἐκβάλλουσιν; διὰ

τοῦτο αὐτοὶ κριταὶ

τοῦτο αὐτοὶ ὑμῶν

ἔσονται ὑμῶν.

κριταὶ ἔσονται.

271

COMPETITIVE TRADITIONS

Matt 9,32-34

Mark 3,22-30

27 ἀλλ᾽ οὐ δύναται

Matt 12,22-32

Luke 11,14-23; 12,10

28 εἰ δὲ ἐν πνεύματι

28 εἰ δὲ ἐν δακτύλῳ

θεοῦ ἐγὼ ἐκβάλλω

θεοῦ ἐγὼ ἐκβάλλω

τὰ δαιμόνια, ἄρα

τὰ δαιμόνια, ἄρα

ἔφθασεν ἐφ᾽ ὑμᾶς

ἔφθασεν ἐφ᾽ ὑμᾶς

ἡ βασιλεία τοῦ θεοῦ.

ἡ βασιλεία τοῦ θεοῦ.

οὐδεὶς εἰς

29 ἢ πῶς δύναταί τις 21 ὅταν ὁ ἰσχυρὸς καθωπλισμένος εἰσελθεῖν εἰς

τὴν οἰκίαν τοῦ ἰσχυροῦ

τὴν οἰκίαν τοῦ ἰσχυροῦ

εἰσελθὼν τὰ σκεύη

καὶ τὰ σκεύη

αὐλήν, ἐν εἰρήνῃ ἐστὶν

αὐτοῦ διαρπάσαι,

αὐτοῦ ἁρπάσαι,

τὰ ὑπάρχοντα αὐτοῦ·

ἐὰν μὴ πρῶτον τὸν

ἐὰν μὴ πρῶτον δήσῃ

22 ἐπὰν δὲ

φυλάσσῃ τὴν ἑαυτοῦ

ἰσχυρὸν δήσῃ, καὶ τότε τὸν ἰσχυρόν; καὶ τότε ἰσχυρότερος αὐτοῦ ἐπελθὼν νικήσῃ τὴν οἰκίαν αὐτοῦ

τὴν οἰκίαν αὐτοῦ

αὐτόν, τὴν πανοπλίαν

διαρπάσει.

διαρπάσει.

αὐτοῦ αἴρει ἐφ᾽ ᾗ ἐπεποίθει καὶ τὰ σκῦλα αὐτοῦ διαδίδωσιν.

30 ὁ μὴ ὢν μετ᾽ ἐμοῦ 23 ὁ μὴ ὢν μετ᾽ ἐμοῦ κατ᾽ ἐμοῦ ἐστιν,

κατ᾽ ἐμοῦ ἐστιν,

καὶ ὁ μὴ συνάγων

καὶ ὁ μὴ συνάγων

μετ᾽ ἐμοῦ σκορπίζει.

μετ᾽ ἐμοῦ σκορπίζει.

28 Ἀμὴν λέγω

31 Διὰ τοῦτο λέγω

ὑμῖν ὅτι πάντα

ὑμῖν, πᾶσα ἁμαρτία

ἀφεθήσεται

καὶ βλασφημία ἀφεθήσεται

τοῖς υἱοῖς

τοῖς

τῶν ἀνθρώπων

ἀνθρώποις,

τὰ ἁμαρτήματα ἡ δὲ τοῦ πνεύματος βλασφημία καὶ αἱ βλασφημίαι ὅσα ἐὰν βλασφημήσωσιν· οὐκ ἀφεθήσεται. (Luke 12,10)

272 Matt 9,32-34

C.M. SAULINA

Mark 3,22-30

Matt 12,22-32

Luke 11,14-23; 12,10

32 καὶ ὃς ἐὰν

Καὶ πᾶς ὃς

εἴπῃ λόγον κατὰ

ἐρεῖ λόγον εἰς

τοῦ υἱοῦ τοῦ ἀνθρώπου, τὸν υἱὸν τοῦ ἀνθρώπου, ἀφεθήσεται αὐτῷ·

ἀφεθήσεται αὐτῷ·

29 ὃς δ᾽ ἂν

ὃς δ᾽ ἂν

βλασφημήσῃ εἰς

εἴπῃ κατὰ

τῷ δὲ εἰς

τὸ πνεῦμα

τοῦ πνεύματος

τὸ ἅγιον πνεῦμα

τοῦ ἁγίου,

τοῦ ἁγίου,

οὐκ ἔχει ἄφεσιν

οὐκ ἀφεθήσεται αὐτῷ

εἰς τὸν αἰῶνα,

οὔτε ἐν τούτῳ τῷ αἰῶνι

βλασφημήσαντι οὐκ ἀφεθήσεται.

ἀλλ᾽ ἔνοχός ἐστιν αἰωνίου ἁμαρτήματος. οὔτε ἐν τῷ μέλλοντι. Key to the Synopsis in Table 3: Single Underline: agreements between Matthew and Luke Bold Fonts: agreements between Mark, Matthew, Luke Broken Underline: agreements between Mark and Matthew against Luke Grey Shade: agreements between Mark and Luke against Matthew

A number of authors have now proposed methods that might explain 2DH Matthew’s source-utilization procedure. Because both 2DH Matthew and MPH Matthew conflate their sources145, I will now continue to adapt the 2DH scholars’ explanations to the MPH’s procedure. Kirk has noted that the dichotomy between the oral and written media in how we theorize ancient authors utilizing their sources should be avoided146. The “binary media schematic”147 is “in need of serious qualification”148. As Kirk has observed, “in antiquity engagement with written artifacts occurred within a predominantly oral communications environment. Notwithstanding their distinctive properties, orality and writing stood in a highly active interfacial relationship”149. The boundary between the visual and aural aspects of a text, then, should be articulated with considerable nuance. Kirk further notes 145. GARROW, Streeter’s ‘Other’ (n. 30), pp. 209-211. 146. KIRK, Q in Matthew (n. 7), pp. 9-28 (and the literature there). 147. Ibid., p. 13. 148. Ibid., p. 15. 149. ID., Memory (n. 142), p. 461.

COMPETITIVE TRADITIONS

273

that “[s]cribes apprehended the written word ‘not as a string of written signs’ but in its reconstitution as oral utterance”150. In this environment, the role of scribal memory inevitably becomes the interfacial sphere between “oral and written tradition, oral and written media”151, and performs a vital function in the process of composing and transmitting a literary work. Scribal memory, then, appears to be one of the keys to solving the Synoptic Problem, expanding the options available to the evangelists in terms of source utilization. The commonly accepted model of an ancient author working with more than one source is to envision that author utilizing one source at a time and working in blocks152. Nonetheless, in many Synoptic Problem hypotheses it appears necessary to envision an evangelist’s conflation of two different sources153. With the scribal memory concept, this process becomes more feasible. An evangelist does not need to rely exclusively on visual contact with all of his source materials. Thus, when working with two separate sources, the author can recall one of them from memory while (potentially) maintaining visual contact with the other. According to Kirk, one possible strategy in memory-based source-utilization would be for an author to memorize and internalize his sources by dividing them into smaller topically grouped segments, whereby each section could be recalled as one unit154. For each of these units, the author may assign a certain title or theme (topos), arranging the units in a particular order and memorizing their sequence to facilitate retrieval. In this manner, by memorizing a work’s schematically organized topoi-sequence, the author acquires “an efficient search tool, moreover one that harnesses the brain’s ordinary search protocols”155. This tool also allows the author to incorporate additional 150. Ibid., p. 462. 151. ID., Q (n. 7), p. 102. 152. OLSON, Unpicking (n. 28), pp. 129-131, here 130: “Writers usually wrote with only one source – at most – in view at any one time”. Cf. DERRENBACKER, Conditions (n. 142), p. 441. 153. Any of the current major synoptic hypotheses requires at least one conflator: Matthew and Luke on the 2DH, Mark on the 2GH, Luke on the FH, and Matthew on the MPH. Cf. DERRENBACKER, Conditions (n. 142), p. 444. Derrenbacker (Practices [n. 1], p. 257; Conditions [n. 142], p. 441) argues that, in composing their writings, ancient authors almost never micro-conflated their sources. However, J.W. BARKER, Ancient Compositional Practices and the Gospels: A Reassessment, in JBL 135 (2016) 109-121, has challenged Derrenbacker’s argument by examining the phenomenon of micro-conflation in Jewish texts such as 4QDeutn (4Q41) and 8ḤevXIIgr, in the Diatessaron, and in Matthew and Luke. In his assessment, Barker notes the role of scribal memory (e.g., formulaic phrases). See the response to Barker in J.S. KLOPPENBORG, Macro-Conflation, Micro-Conflation, Harmonization and the Compositional Practices of the Synoptic Writers, in ETL 95 (2019) 629-643. 154. KIRK, Q in Matthew (n. 7), pp. 131-142; CHANG, Questioning (n. 143), pp. 416-417. 155. KIRK, Q in Matthew (n. 7), p. 141.

274

C.M. SAULINA

materials into the schematic of the memorized text. The process is especially facilitated when the new arrivals share themes with the memorized text’s topoi156. There is some discussion, however, in current synoptic criticism regarding the (potentially) memorized materials and those that are (potentially) accessed through visual contact. Some scholars suggest that a low level of verbatim agreement between a text and its source does not necessarily indicate the use of memory. In terms of the order, one might infer that when an author follows the order or his source closely, this indicates visual access. We have already seen, however, that this is not necessary on the topoisequence memory-based model. Now, Kai-Hsuan Chang has further argued that while following the order does not necessarily suggest visual contact with a source text, close verbatim agreement does157. Kirk’s and Chang’s findings may help explain MPH Matthew’s method in incorporating both Mark and Luke. In general, MPH Matthew tends to follow the Markan order, especially from Matthew 12 on. When following Mark’s order in certain segments – such as the Beelzebul controversy – Matthew may have retrieved his Markan materials from memory, adding thematically related Lukan materials through visual contact with the source (as seems suggested by the high level of double-tradition verbatim agreement in the Beezelbul episode). With this strategy, MPH Matthew was able to restructure much of the TN material in Luke 12–19. This task becomes less daunting since for attaching these materials “topical pegs in the rest of the Markan narrative are readily available”158. In chapter 12, Matthew begins to follow Mark’s order more closely starting at Mark 2,23 and until the end of the Gospel, while incorporating the parallel passages in Luke and other additional Lukan materials. This is especially illustrated by Matthew’s adaptation of Luke’s Beelzebul controversy and its surrounding materials in a Markan narrative order. Notably, MPH Matthew follows Luke in placing the accusation after the commissioning of the disciples159. In dealing with the passages that feature micro-conflation like the Beelzebul pericope, transforming Kirk’s and Chang’s Matthean source into Luke rather than Q, we can see how MPH Matthew’s frequent visual contact with Luke 10–13 assists him in writing these passages. Matthew likely scans 156. Kirk (ibid., p. 142), speaks of a “controlled random access” to a topoi-organized work. 157. CHANG, Questioning (n. 143), p. 424. 158. KIRK, Q in Matthew (n. 7), p. 277. 159. Ibid., p. 278.

COMPETITIVE TRADITIONS

275

these sections of Luke multiple times. The conflation that we see in the Beelzebul passage can be viewed as an example of Matthew’s scribal memory and source competency. The episode can be divided into several topoi, such as the political and household spheres160, and memorized in that order. In shaping this passage, Matthew recalls the Markan topoi from memory, while maintaining visual contact with Luke 11. In Table 3, we can see that the level of verbal agreements in Matt 11,27-28.30.32a is much higher than in 11,26.29.31. Therefore, in the former group, MPH Matthew is likely to have had visual contact with Luke’s text. In the latter group, he would have retrieved them from his memory of Mark’s account. VII. CONCLUSION The micro- and macro-intertextual analyses of the Beelzebul controversy and its surrounding contexts in Luke and Matthew reveal layers of similarities and contrasts (the textual and symbolic worlds). The parallels between these two versions are so stark as to strongly suggest a literary dependence. This essay has conducted a microanalysis of Matt 12,22-32, which appears to suggest Luke’s influence on Matthew (MPH) rather than the reverse (FH). The examination of the plot, the interaction between the characters in the story, and the symbolic worlds of both gospels (especially the assessment of exorcisms and their link with the advent of God’s kingdom) are key to deciphering the direction of the influence. My conclusion is that Matthew here conflates Mark’s and Luke’s versions. The macroanalysis of the surrounding contexts reveals a similar strategy. It is a part of Matthew’s goal to present the most comprehensive gospel, while honoring both of his predecessors’ texts. This objective produces the phenomena that we can observe in Matthew: generally following Mark’s order and closely reproducing Luke’s wording. The recent insights from the study of scribal memory seem to confirm this type of conflation. Seattle Pacific University United States of America

160. Ibid., p. 181.

Chakrita M. SAULINA [email protected]

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

Two proponents of the Farrer Hypothesis (FH) have recently expressed puzzlement as to why supporters of the Two-Document Hypothesis (2DH) and Matthean Posteriority Hypothesis (MPH) see Matthew’s reordering of Q/Luke as so much more plausible than Luke’s reordering of Matthew. Eric Eve describes these operations as mirror images of one another, and so fails to see why working them in one direction should be any more implausible than the other: Whatever complicated reordering FH Luke would have to have performed on material taken from Matthew, 2DH[/MPH] Matthew would necessarily have to have performed the reverse on Q[/Luke]; one set of transpositions logically must be the mirror image of the other. … It is not immediately apparent why FH Luke’s task should be any more difficult than 2DH[/MPH] Matthew’s1.

Mark Goodacre is similarly curious: The idea that Luke could not have achieved a strong reworking of Matthew’s order has always been strange given the fact that there are large-scale differences between Matthew’s and Luke’s ordering of the double tradition material. At least one of the two has been rearranging this material. If, with most twosource theorists, one broadly aligns Luke’s order with Q’s order, we give Matthew a great deal of work to do, so that the supposed unfeasibility of the large-scale rearrangement, the alleged logistical difficulty, is simply transferred from Luke to Matthew2.

Eric Eve also makes a closely related point: What is source utilization for the 2DH[/MPH] goose should be source utilization for the FH gander, so the techniques ascribed to 2DH[/MPH] Matthew for the recycling of Q[/Luke] should also be available to FH Luke for the recycling of Matthew3. 1. E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021, p. 146. 2. M. GOODACRE, Re-walking the “Way of the Lord”: Luke’s Use of Mark and His Reaction to Matthew, in M. MÜLLER – J.T. NIELSEN (eds.), Luke’s Literary Creativity (LNTS, 550), London – New York, Bloomsbury T&T Clark, 2016, 26-43, p. 42. 3. EVE, Relating (n. 1), p. 160. M. GOODACRE, Q, Memory and Matthew: A Response to Alan Kirk, in JSHJ 15 (2017) 224-233, pp. 228-229 includes a section headed “Sauce for the

278

A. GARROW

Their puzzlement, broadly speaking, is that if Matthew can perform complex operations on Luke/Q, then why could not Luke have performed the same operations (in reverse) on Matthew4? As I attempt to answer their question one issue needs clarifying from the outset. It is not the case that “one set of transpositions logically must be the mirror image of the other”5. Suppose it were possible to make an animation of the journey made by five pericopes A, B, C, D and E as they travel from one text to another. If those pericopes were arranged in a single consecutive block in the first text and are widely dispersed in the second, then playing the animation one way would not be the mirror image of playing it in reverse. The two versions tell two distinctly different stories – one of dispersal and the other of gathering. The same is true with Matthew and Luke. Playing an animation of their double tradition transpositions in one direction is not the mirror image of playing it in the opposite direction. One way is “forward” (the direction that happened in history) and the other is “reverse”. My method for determining which way is “forward” in the Luke-Matthew case is to observe what happens in other ancient examples where, importantly, there is no ambiguity about which texts are sources and which are products6. If there is a consistent pattern across all these examples, and if that pattern recurs in Matthew’s use of Luke, but not in Luke’s use of Matthew, then Matthew’s use of Luke is “forward”, and vice-versa. Before applying this method, however, a terminological issue needs addressing.

Goose” which echoes this point. Eve makes a very similar point in Relating (n. 1), p. 152: “This is not to object to the thesis that Matthew could have accessed Q[/Luke] through his scribally trained memory of text, but rather to insist that the techniques and abilities needed by 2DH[/MPH] Matthew must also be allowed to FH Luke”. The problem with this logic is that the source utilization capacity demonstrated by one author is not necessarily transferrable to every other author. The only concrete indicator of Luke’s source utilization capacity is his simple handling of Mark. This does not suggest that he had the exceptional facility required for FH Luke’s complex handling of Matthew. 4. Because I advocate for the MPH, rather than the 2DH, I will hereafter only refer to Matthew’s use of Luke rather than Matthew’s use of Q. The logic of my argument is broadly the same, however, in either case. My principal objection to the Two-Document Hypothesis, as conventionally conceived, is that it relies on the untested assumption that Matthew could not have used Luke. For an introductory presentation of this point see, A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226. 5. See full quotation from EVE, Relating (n. 1), p. 146, above. 6. Thus, there is no possibility that Kings and Chronicles might depend on Josephus’ Antiquities of the Jews, and no possibility that Matthew, Mark, Luke and John might depend on Tatian’s Diatessaron, and so on.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

279

I. A TERMINOLOGICAL ISSUE Previous discussions of the way in which recycling authors combined their sources have focussed on two distinctive, and closely related, behaviours. Such authors are said to use “one source at a time” and generally to avoid “micro-conflation”. Thus, Robert Derrenbacker Jr., speaking of Diodorus Siculus, Strabo, Arrian of Nicomedia, and Josephus, writes: These authors tend to follow one source at a time. This we see most explicitly in Josephus’ adaptation of [the] Deuteronomistic Historian and the Chronicler in his Antiquities. In addition, in the case of the account of the caste system in India, all three authors chiefly follow Megasthenes. It is only at the end of the pericope where they briefly refer to other authors. What we do not see in the above authors is a frequent and regular sort of “micro-conflation” where an author moves back and forth between sources within episodes. It is only when a pericope/episode is concluded that the author will typically move to another parallel source if he chooses7.

There are, however, problems with this observation. First, it is not fully accurate. As pointed out below, Josephus was capable of oscillating between texts to maximise their dual contributions to his recycled version8. Second, the accuracy of this statement diminishes when a wider spectrum of relevant texts is included9. Tatian’s Diatessaron, again as argued below, is particularly relevant to the current discussion since it shares with Luke and Matthew the highly specific genre “recycled gospel”. It is significant, therefore, that 7. R.A. DERRENBACKER, The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 435-457, pp. 440-441. See also ID., Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005. 8. A specific challenge to Derrenbacker’s claim is also offered in J.W. BARKER, The Use of Sources in Ancient Compositions, in S.P. AHEARNE-KROLL (ed.), The Oxford Handbook of the Synoptic Gospels, Oxford, Oxford University Press, forthcoming. 9. That various ancient authors were capable of drawing from multiple texts simultaneously is argued by J.W. BARKER, Ancient Compositional Practices and the Gospels: A Reassessment, in JBL 135 (2016) 109-121; ID., Tatian’s Diatessaron: Composition, Redaction, Recension, and Reception (OECS), Oxford, Oxford University Press, 2021, pp. 29-37; and ID., The Use of Sources in Ancient Compositions, in The Oxford Handbook of the Synoptic Gospels (n. 8). J.S. KLOPPENBORG, Macro-Conflation, Micro-Conflation, Harmonization and the Compositional Practices of the Synoptic Writers, in ETL 95 (2019) 629-643, offers a response to Barker’s 2016 article in which he contests the possibility that authors could have had simultaneous visual contact with multiple sources. Whether Kloppenborg is right about this is less significant than the fact that some ancient authors succeeded, by some means, in closely combining multiple sources.

280

A. GARROW

William Petersen observes: “the Diatessaron appears to have been a very subtle, word-by-word harmonization”10. When it comes to considering scribal practices relevant to the study of the Synoptic Problem there is also a case for including the gospels themselves. As Derrenbacker himself affirms, every literary solution to the Synoptic Problem requires at least one of the Evangelists to do something a good deal more complex than simple “one source at a time” or “block-byblock” copying11. If, therefore, Derrenbacker’s sample were enlarged to include the Synoptic Gospels, the effect of thereby increasing its relevance would be to further reduce his conclusion’s accuracy. A third problem with the language of “one source at a time” is that it is capable of being misunderstood. It would be a mistake to imagine that ancient recycling authors constructed their texts by switching from one source to another as a child might build a tower of bricks. As illustrated in greater detail below, what they actually did was choose one source as a dominant base and then fill out that base with relevant material from other sources – a technique executed by different authors with different levels of refinement. I propose, therefore, that it is time to replace binary expressions like “one source at a time” and “macro-” or “micro-conflation”, with language that embraces the full spectrum of observable activity – from ponderous switching between large blocks to the deft interleaving of minor details12. The replacement term I have in mind is “frame and fill”. 10. W.L. PETERSEN, Tatian’s Diatessaron: Its Creation, Dissemination, Significance, and History in Scholarship (SupplVC, 25), Atlanta, SBL, 1994, p. 27. Note, for contrast, S.L. MATTILA, A Question Too Often Neglected, in NTS 41 (1995) 199-217, p. 205, when she states: “It must be kept in mind that for the major part of the Diatessaron the conflation is block-by-block, only becoming more when the pressure to reconcile and combine conflicting details in the parallel gospel accounts necessitates it”. However, BARKER, Tatian’s Diatessaron (n. 9), p. 36, specifically responds: “It is highly inaccurate to characterize ‘the major part of the Diatessaron’ as ‘block by block,’ since nearly three-fourths of the time Tatian worked with three or four Gospels simultaneously”. 11. DERRENBACKER, Ancient Compositional Practices (n. 7), pp. 257-258. 12. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, p. 56, while careful to avoid the term “micro-conflation”, uses a variety of terms to describe the activity of ancient authors who combined material from more than one source. For example: supplemental materials from various sources are “grafted” into a dominant source; authors engage in “combining” elements peculiar to each source into a single version (p. 57); an author may act as a compiler and consolidator of scholarly lexical tradition (pp. 78-79, 88); authors are engaged in “assembling” the lexical interpretations of others (p. 79); the technique of “conflating” by appending a sequence of excerpted sayings (p. 87); a compiler who “effectively unites his two sources” (p. 88); a compiler who achieves the “coherent integration” of parallel accounts achieving “maximal incorporation” of their constituent elements (p. 89) – something “uncannily similar” to Matthew’s project; “pulling in” topically related

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

II. “FRAME

AND

281

FILL”

A broad range of ancient authors behaved in an essentially similar manner when creating new narratives by recycling older ones: they used a technique that might be called “frame and fill”. Before looking at some ancient examples it is worth pausing to consider how anyone at any time might be expected to go about the recycling task. Faced with two or three similar-but-different accounts of the same story the simplest approach, if plagiarism is not a concern, is to: read all the available source accounts; decide which seems the most reliable and/or complete; and then use that account as the base for the new version, supplementing that frame with additional details from the other accounts. As the following examples illustrate this appears to have been the practice of a wide range of ancient authors. None of these authors is, individually, a perfect match for Matthew and Luke in terms of their education, cultural context, or technology and tools. Nevertheless, a practice common to them all is highly likely also to have been common to both Luke and Matthew. 1. Plutarch (Early Second Century CE) Plutarch, like other Roman biographers, created his Lives by rewriting earlier biographies. His work is not an exact comparator for the Evangelists inasmuch as he belonged to a stratum of the literary elite unlikely to have included authors like Matthew and Luke. Furthermore, Plutarch was not dealing with sources he saw as preserving the words and actions of a divine material by memory (p. 119); “conflation by bringing together compatible δόγματα from different dialogues” (p. 123 – quoting J. WHITTAKER, The Value of Indirect Tradition in the Establishment of Greek Philosophical Texts, or the Art of Misquotation, in J.N. GRANT [ed.], Editing Greek and Latin Texts: Papers given at the Twenty-Third Annual Conference on Editorial Problems, University of Toronto, 6-7 November 1987, New York, AMS, 1989, 63-95, pp. 89-90); “systematic concatenation and standardization of multiple, multi-sourced, and originally distinct oracles” (p. 127); proverbs, maxims, and the like “collocated” to enable the creation of a single cognitive entity (p. 137); heterogenous items “collected” (p.136); “combining” or “laying together” (p. 144 – quoting M. CARRUTHERS, The Book of Memory: A Study of Memory in Medieval Culture (CSML), Cambridge: Cambridge University Press, 2 2008, p. 244); “slipping” material into heuristic schemes (p. 145 – cf. CARRUTHERS, Book of Memory, p. 221); and bringing together topically cohering materials from different memory locations (p. 145). Kirk also mentions: “Matthew’s … comprehensive strategy for solving the technical problem of combining two sources coherently into a new work” (p. 190); M materials are said to “supplement” the core Q sayings to make a coherent M topos (p. 196); and, “Matthew’s utilization actions in the Sermon are part of a comprehensive strategy for solving the technical problem of coherent coordination of his two sources, and more particularly, of his narrative source with his non-narrative source” (p. 223).

282

A. GARROW

figure. Nevertheless, his literary objectives are sufficiently like those of Luke and Matthew to be worthy of inclusion in this collection of comparators. A striking feature of Plutarch’s method is his practice, despite being very well read, of basing his recycled versions on a single original. Plutarch drew on a fairly wide range of material. Yet … it is still clear that the greater portion of the Lives is based on the Pollio-source alone: even on those occasions (such as Caesar’s murder) where Plutarch has other sources, it is still Pollio’s account with provides the basic narrative articulation, and Pollio’s account which provides most of the facts. The extraneous material is not more than one quarter of the whole of Plutarch’s narrative. This wide reading of sources is surprisingly unproductive; it seems to provide only a few stray supplements and additions, and occasionally to replace the Pollio-source where that account was unsuitable13.

Thus, one source provides the frame for Plutarch’s recycled version (accounting for more than three quarters of the whole) with other sources supplementing that frame where they have relevant material to add (accounting for up to one quarter of the whole)14. According to Pelling this pattern was also employed by Cassius Dio, Livy, Dionysius, and Tacitus15. 2. Josephus (Late First Century CE) A closer comparator to the activity of the Evangelists, given particular attention by F. Gerald Downing, is the Jewish author Josephus16. In his Antiquities of the Jews Josephus engages in a comparable project inasmuch as his sources had the authority of Scripture. Josephus’s technology and tools are also likely to be comparable to those of the Evangelists insofar as he wrote at a similar date17. Here again Josephus adopts the technique of frame and fill. 13. C.B.R. PELLING, Plutarch’s Method of Work in the Roman Lives, in JHS 99 (1979) 74-96, p. 91. 14. D.S. RUSSELL, Plutarch’s Life of Coriolanus, in JRS 53 (1963) 21-28, puts the ratio at 80:20. 15. PELLING, Plutarch’s Method (n. 13), p. 91, refers to these authors when noting that “This [approach] is not confined to Plutarch, nor to biography”. With specific reference to Livy (ibid., pp. 91-92): “Livy … has one principal authority for each section of his account, and uses the rest of his reading merely to supplement this principal narrative source”. 16. F.G. DOWNING, Redaction Criticism: Josephus’ Antiquities and the Synoptic Gospels (I), in JSNT 8 (1980) 46-65; and ID., Redaction Criticism: Josephus’ Antiquities and the Synoptic Gospels (II), in JSNT 9 (1980) 29-48. For details of Josephus’ conflationary practice see ID., Josephus’ Antiquities (I), the sections “Assembly” (pp. 57-60) and “Conflation” (pp. 61-64). 17. M. GOULDER, Luke’s Compositional Options, in NTS (1993) 150-152, p. 150, suggests: “We may think Josephus the closest model for Luke, since he is writing the Antiquities in the same decade (90s), and is also treating sacred texts”. Even here, however, important differences may apply. The date of Luke is debatable and his sources are unlikely to have had the status of “sacred text” in the same way as Kings, Chronicles and Samuel.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

283

[Josephus] basically adopts the order of Kings, inserting material from Chronicles at appropriate junctures. Occasionally, however, one finds Josephus reordering and re-combining the sequence of happenings proper to Kings itself18. Where the two sources generally run parallel to each other, but differ in details, Josephus typically oscillates between then, utilizing items now from one, now from the other … [in each case] Josephus evidences his intention of making maximal use of the data of both his historical sources – in so far as these do not militate against his authorial purposes19. Another factor influencing the historian’s re-arrangements seem to have been his desire to keep together related materials which he found separated in his sources20.

Thus, while there is variety in Josephus’ practice, the overall method is largely consistent. Josephus selects a frame, for example Kings, and draws in supplementary fill from, for example, Chronicles. Significantly, when Josephus is sufficiently motivated to do so, he is capable of “oscillating” between sources. Furthermore, as he works his sources together, he “brings together related topics”21. 3. Tatian (Late Second Century CE) Moving into the latter part of the Second Century, Tatian shares a striking point in common with Matthew and Luke. Like them he was a Christian seeking to recycle divergent accounts of the life of Jesus. Indeed, the resources available to the third-most Evangelist would have overlapped to a large and specific extent with the resources available to Tatian22. There should be particular profit, therefore, in observing how Tatian went about his task. William Petersen describes Tatian as adopting a strategy common to historians in general: When confronted with contradictory or inconsistent information, the historian’s task was to reconstruct “the true events.” This was done (and still is done) by carefully evaluating the reliability of each account. The one judged most reliable forms the framework of the narrative; where possible and proba18. C. BEGG, Josephus’ Account of the Early Divided Monarchy (AJ 8,212-420) (BETL, 108), Leuven, Leuven University Press – Peeters, 1993, p. 278. 19. ID., Josephus’ Story of the Later Monarchy (AJ 9,1–10,185) (BETL, 145), Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2000, p. 623. 20. Ibid., p. 629 (examples provided). Begg (ibid., p. 635) describes Josephus as “a literary juggler of no little skill”. 21. ID., Early Divided Monarchy (n. 18), p. 278. 22. Tatian combined the canonical gospels of Mark, Luke, Matthew and John. The thirdmost evangelist combined: Mark and Matthew, or Mark and Luke, or Mark and Q, along with Special Material.

284

A. GARROW

ble, what appears less reliable is fitted into that framework. … [This] is precisely what Tatian attempted to do in the Diatessaron23.

A specific example of choosing a frame, in this case Matthew, and supplementing that frame from related material in other sources is noted by Charles Hill: Another example, this one from Fuldensis, is Jesus’s saying about putting a light on a stand (Matt 5:14-16; Mark 4:21; Luke 8:16; 11:33), where all three Synoptic sources are integrated. As Zola says, “F[uldensis] manages to incorporate the objects of each instance into the Matthean form by supplying neque before each one, and by altering other small details.” The same technique is visible throughout the Arabic Diatessaron24.

As with Josephus, the integration of supplementary material into the frame appears to have been, logically enough, based on thematic grouping. As James Barker puts it: “Thematic grouping clearly emerges as one [of Tatian’s redactional tendencies], since the Diatessaron clusters similar statements and characters”25. If Tatian’s activity could be shown to be the product of conditions that only existed in the late second century, then its relevance to the current debate would be reduced. In reality, however, there are good reasons for suspecting that Tatian’s activity belonged to an extended tradition. As Nicholas Perrin puts it: “[Tatian] was self-consciously participating in a literary tradition that had already been well established for decades”26. This echoes the opinion of Petersen who concludes, after exploring the relationship between Justin Martyr’s harmony and Tatian’s Diatessaron: “These agreements admit only two explanations: either Tatian knew and used Justin’s harmony, or both relied on the same pre-existing harmonized source”27.

23. PETERSEN, Tatian’s Diatessaron (n. 10), p. 75. 24. C.E. HILL, Diatessaron, Diapente, Diapollon? Exploring the Nature and Extent of Extracanonical Influence in Tatian’s Diatessaron, in M.R. CRAWFORD – N.J. ZOLA (eds.), The Gospel of Tatian: Exploring the Nature and Text of the Diatessaron (RJT, 3), London – New York, Bloomsbury T&T Clark, 2019, 25-53, p. 30. 25. J.W. BARKER, The Narrative Chronology of Tatian’s Diatessaron, in NTS 66 (2020) 288-298, p. 298. 26. N. PERRIN, What Justin’s Gospels Can Tell us about Tatian’s: Tracing the Trajectory of the Gospel Harmony in the Second Century and Beyond, in CRAWFORD – ZOLA (eds.), The Gospel of Tatian (n. 24), 93-109, p. 100. Similarly, MATTILA, Question (n. 10), p. 205: Tatian’s technique, “while being an innovation, is not sui generis, it does not emerge out of nowhere. It stems from methods that have already been in use, but stretches these to new limits’’ (emphasis original). 27. PETERSEN, Tatian’s Diatessaron (n. 10), p. 29.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

285

Looking back to Justin’s activity Watson notes that Justin also employs the frame and fill technique with, again, Matthew providing the frame: In Justin’s two major works … Material from Matthew is cited frequently, often supplemented by or conflated with material from Luke … In many cases the primacy of Matthew is clear. The whole of the Matthean infancy narrative is recounted, at least in outline, and Luke provides only supplementary details28.

For his part, Perrin suggests that Justin does not provide the earliest example of this type of activity: There are also indications that Justin’s harmony was not the first of its kind either. A generation or two before the Apologist’s death we find evidence of a circulating composite gospel of sorts in 2 Clement, one materially reminiscent of Justin’s29.

This line of ancestry may extend further back still, given Petersen’s observation: [A]ll the canonical gospels “harmonize” earlier materials. While is true that the Diatessaron appears to have been a very subtle, word-by-word harmonization, and the canonical gospels seem to use their sources en bloc, the genre of both is, ultimately, the same30.

If the technique of frame and fill was consistently employed across the latter stages of this continuum, it makes sense to suppose that it may also have been used, albeit at a lower level of sophistication, from its start. 4. Ammonius of Alexandria and Eusebius That Tatian was not alone in his willingness to expend prodigious amounts of effort in reconciling and consolidating divergent accounts of the life of Jesus is illustrated by Eusebius’ description of the activities of Tatian’s contemporary, Ammonius of Alexandria. Thus, Matthew Crawford writes:

28. F. WATSON, Gospel Writing: A Canonical Perspective, Grand Rapids, MI, Eerdmans, 2013, pp. 473-474. 29. PERRIN, Justin’s Gospels (n. 26), p. 98. 30. PETERSEN, Tatian’s Diatessaron (n. 10), p. 27. The strength of continuity between the genre of Luke, Matthew and Tatian’s Diatessaron is enhanced by M.R. CRAWFORD, Diatessaron, a Misnomer? The Evidence from Ephrem’s Commentary, in EC 7 (2016) 253-277, who argues that Tatian understood his work to be a full “Gospel”. F. WATSON, Harmony or Gospel? On the Genre of the (So-Called) Diatessaron, in CRAWFORD – ZOLA (eds.), The Gospel of Tatian (n. 24), 69-92, p. 69, similarly refers to the Diatessaron as: “a gospel rather than a gospel harmony”.

286

A. GARROW

[Eusebius] provided a one-sentence summary of [Ammonius’] work, which is our sole surviving description of Ammonius’ composition: τὸ διὰ τεσσάρων ἡμῖν καταλέλοιπεν εὐαγγέλιον, τῷ κατὰ Ματθαῖον τὰς ὁμοφώνους τῶν λοιπῶν εὐαγγελιστῶν περικοπὰς παραθείς He has left behind for us the Diatessaron-Gospel, having placed alongside the [Gospel] according to Matthew the sections from the other evangelists that agree [with those of Matthew]. Clearly what Eusebius is describing here is something akin to a modern gospel synopsis with parallel columns. Ammonius dissected the latter three gospels in order to align the parallels he found there with corresponding passages in Matthew31.

Here again Ammonius’ activity, even though he is engaged in a different type of project, has much in common with Tatian’s compositional technique. The same is true of Eusebius’ canons, as Watson now explains: Both Ammonius of Alexandria in his synopsis and Eusebius in his canons take Matthew as their base text, partly … to reduce the considerable technical difficulties presented by their respective scholarly projects … Once established, it would be relatively easy to reintroduce elements of Lukan or Johannine sequence into a framework which remains, overall, essentially Matthean32.

Watson’s observation brings us back to where we began. The simplest method for creating a single narrative out of two or more related narratives, whether in antiquity or at any other time, is to choose one source to serve as the frame and then use the other sources to fill out that frame33. Given that this method is not only relatively simple but also universally attested amongst ancient authors who recycled related narratives, it is reasonable to expect that it would also have been used, albeit in different ways and with differing levels of sophistication, by Luke in his use of Matthew or by Matthew in his use of Luke34. However, before assessing which of these two options is more likely, there is another preliminary question to address: did Luke use Mark as a frame source or as a fill source35? 31. M.R. CRAWFORD, Ammonius of Alexandria, Eusebius of Caesarea and the Origins of Gospels Scholarship, in NTS 61 (2015) 1-29, p. 7. 32. WATSON, Harmony or Gospel? (n. 30), pp. 84-85. For a more detailed description of Eusebius’ approach, which notes that Eusebius declines to structure his whole project around Matthew, see J. COOGAN, Mapping the Fourfold Gospel: Textual Geography in the Eusebian Apparatus, in JECS 25 (2017) 337-357. 33. A further example of this same technique, with still greater levels of sophistication, is Origen’s Hexapla. 34. It is unlikely that Matthew’s and Luke’s education exceeded that of the authors considered above. It follows, therefore, that, like these authors, they would have used the simplest available method for combining multiple versions of the same narrative. 35. This question, which formed part of the ‘Proto-Luke’ debate, was a matter of heated discussion between the two World Wars. Advocates for the Proto-Luke theory (which necessarily

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

III. LUKE’S USE

OF

287

MARK

The question of whether Luke uses Mark as his frame or his fill source has a significant bearing on what follows. Specifically, it matters if FH Luke did not use Mark as his frame because, if Mark serves as one of FH Luke’s fill sources, then this provides a model for how we might expect FH Luke to use another fill source, namely Matthew. The following assessment of whether Luke used Mark as his frame begins in what might seem an odd location: the Beelzebul Controversy. This passage has long been recognised as an important one for Synoptic Problem debates. Back in 1965, Downing used this pericope to argue that, under the FH, Luke is required to “unpick”, as in “reverse conflate”, Matthew’s additions to Mark36. FH proponents sometimes concede that, if Downing’s arguments are accurately expressed, they are indeed highly damaging to the FH cause37. What they argue in response is that Luke used his sources in blocks, and here focussed on his non-Markan source alone (which from a FH perspective, is Matthew). On this basis they contend that the charge of “unpicking” no longer holds. Thus, Eric Eve states: “The pattern of agreements in the body of the Beelzebul Controversy set out in Table 5.4 [which, incidentally, shows high levels of commonality between Mark and Luke, with 25% of Mark’s text directly paralleled in Luke] is thus compatible with what one might expect if Luke were using Matthew alone”38. Similarly, Ken Olson, in his response to Downing states: “There is no need to suggest that held that Mark was not Luke’s frame) included: B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, pp. 199-222, and V. TAYLOR, Behind the Third Gospel: A Study of the Proto-Luke Hypothesis, Oxford, Clarendon, 1926; ID., Is the Proto-Luke Hypothesis Sound?, in JTS 29 (1928) 147-155. The theory’s detractors included J.W. HUNKIN, The Composition of the Third Gospel, with Special Reference to Canon Streeter’s Theory of ProtoLuke, in JTS 28 (1927) 250-262, and J.M. CREED, The Gospel according to St. Luke, London, Macmillan, 1930. Creed draws on R. Bultmann, K.L. Schmidt, J. Wellhausen, E. Klostermann, J. Weiss, A. Loisy and C. Montefiore. Despite occasional support, for example, from T.L. BRODIE, Proto-Luke: The Oldest Gospel Account, Limerick, Dominical Biblical Institute, 2006, the current consensus runs against the Proto-Luke hypothesis. 36. F.G. DOWNING, Towards the Rehabilitation of Q, in NTS 11 (1965) 169-181. Downing considers the phenomenon of FH Luke’s capacity to select Matthew’s additions to Mark, while avoiding material common to Matthew and Mark, in this and three other passages. A comprehensive tabulation of occasions when FH Luke ‘unpicks’ Matthew’s additions to Mark is now offered by O. ANDREJEVS, FH Luke’s ‘Unpicking’: Some Observations on Francis Watson’s Recent Analysis (2018) and the Extent of the Phenomenon, in JSNT 45 (2022) 3-22. 37. I. MILLS: NT Review Podcast 18. F.G. Downing, “Towards the Rehabilitation of Q” https://soundcloud.com/user-829560134/18-gerald-downing-towards-the-rehabilitation-of-q [accessed February 22, 2022]. In this podcast Ian Mills, a Farrer supporter, states his belief that, if the data were as Downing presents it, then it would be necessary to accept the existence of Q (or, alternatively, that Matthew used Luke). 38. EVE, Relating (n. 1), p. 133 (emphasis added).

288

A. GARROW

Luke would have consulted Mark’s text at all”39. Arguing that Luke uses Matthew alone in this and similar instances lessens the immediate difficulty for the FH but it has an awkward implication for the question of whether Luke used Mark as his frame. According to the FH, Luke knows both Mark’s and Matthew’s versions of the Beelzebul Controversy, but rather than choosing to integrate the two he prefers to focus on Matthew’s version while ignoring Mark’s altogether. From Luke 22,15 onwards, something similar seems to happen. Luke, we may presume, knows Mark’s version of the Passion and Resurrection but he also apparently knows another version that is consistently different from Mark’s. On the basis that Luke treats his sources one at a time, and given that Mark cannot supply even a small proportion of the vocabulary or detail found in Luke, it follows that Luke here follows his other source in preference to Mark40. The same logic applies with: The Rejection at Nazareth (Mark 6,1-6// Luke 4,16-30); The Call of the First Disciples (Mark 4,1-2; 1,16-20// Luke 5,1-11); Jesus is Anointed (Mark 14,3-9//Luke 7,36); and The Lawyer’s Question (Mark 12,28-34//Luke 10,25-28). In these cases, Luke records the incident in a way that bears comparison with Mark’s account, but the shared vocabulary is again very limited and, as with the Beelzebul Controversy, the story does not appear in its Markan sequence. If FH Luke ignores Mark in the one case, then it makes sense to suppose that Luke does not use Mark in these other cases also.

39. K. OLSON, Unpicking on the Farrer Theory, in M. GOODACRE – N. PERRIN (eds.), Questioning Q: A Multidimensional Critique, Downers Grove, IL, InterVarsity Press; London, SPCK, 2004, 127-150, p. 141 (emphasis added). 40. Luke shares very little vocabulary with Mark in the case of: The Last Supper (Mark 14,22-25//Luke 22,15-20); Jesus Foretells his Betrayal (Mark 14,22-25//Luke 22,21-23); Precedence among the Disciples and the Reward for Discipleship (Mark 10,41-45// Luke 22,24-30); Peter’s Denial Predicted (Mark 14,26-31//Luke 22,31-34); Gethsemane (Mark 14,32-42//Luke 22,39-46); Jesus Arrested (Mark 14,43-52//Luke 22,47-53); Jesus Before the Sanhedrin (Mark 14,53-65//Luke 22,54-71); Jesus’ Trial before Pilate (Mark 15,2-5// Luke 23,2-5); Jesus or Barabbas? (Mark 15,6-14//Luke 23,17-23); Pilate Delivers Jesus to be Crucified (Mark 15,5//Luke 23,24-25); The Crucifixion (Mark 15,22-26//Luke 23,33-34); Jesus Derided on the Cross and the Two Thieves (Mark 15,27-32//Luke 23,35-43); The Death of Jesus (Mark 15,33-39//Luke 23,44-48); Witnesses of the Crucifixion (Mark 15,40-41// Luke 23,49); and The Resurrection (Mark 16,1-8//Luke 24,1-12). The direct parallels between Luke and Mark in these passages never come close to 25% of Mark’s text, with the exception of the words of institution in Mark 14,22-25//Luke 22,15-20. High levels of agreement here are likely to be the product of liturgical familiarity as much as direct copying. It is also necessary to take account of the impact of the Western non-interpolation (Luke 22,19b-20) when assessing the levels of agreement in this particular passage.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

289

A similar set of phenomena occur in Luke’s coverage of John the Baptist. It seems that Luke knows at least two different streams of tradition about John the Baptist. One is extensive and the other, as represented by Mark, is relatively terse. Under the FH, of course, Luke is an author who prefers to focus on his sources in substantial blocks, rather than attempting to weave them together more finely. Mark cannot be the source for the majority of what Luke has to say about John the Baptist41, so it follows that Mark was not the source that Luke decided to follow for this block, or blocks. This logic requires, of course, that Luke’s non-Markan source happened to include the various details about the Baptist that are also featured in Mark: he was called John; he was a baptizer; he presented himself in a less-togreater relationship with Jesus; and he died at a dramatic and unjust death at the hands of Herod. This does not present a significant difficulty, however, since it is hard to imagine a tradition about John, relevant to the wider story of Jesus, that does not include these elements42. In summary, therefore, the application of the FH supporter’s view that Luke uses his sources in substantial blocks favours the conclusion that, in all the examples considered so far, Luke did not use Mark at all43. This coheres with Mark Goodacre’s view that Luke is “[t]aking Mark for a stretch (Lk. 4,31–6,19), then Matthew for a stretch (6,20–[8,3]), then returning to Mark (8,4–9,50), and so on”44. In adopting this view Goodacre picks up on Michael Goulder’s position: I maintain that Luke followed one Gospel at a time, Mark for instance from Luke 3.1 to 6.19 and 8.4 to 9.50, Matthew from Luke 6.20 to 8.3 and 9.51 to 18.14; the ‘echoes’ of the source not in use I attribute to reminiscence, arising from familiar use45.

41. The Annunciation of the Birth of John the Baptist (Luke 1,5-25); Mary’s Visit to Elizabeth (Luke 1,39-56); The Birth of John (Luke 1,57-80); John’s Preaching (Luke 3,7-14); The Baptist’s Questions (Luke 7,18-23); and Jesus’ Testimony to the Baptist (Luke 7,24-35). 42. T.M. DERICO, Oral Tradition and the Synoptic Verbal Agreement: Evaluating the Empirical Evidence for Literary Dependence, Eugene, OR, Pickwick, 2016 shows how independent recollections of a foundational narrative may overlap at this level of detail. 43. It may be of interest to note that scholars who belong to the Jerusalem Perspective see Luke as wholly independent of Mark. See, for example, R.S. NOTLEY, Non-Septuagintal Hebraisms in the Third Gospel: An Inconvenient Truth, in R. BUTH – R.S. NOTLEY (eds.) The Language Environment of First Century Judaea: Jerusalem Studies in the Synoptic Gospels (JCPS, 26), Leiden – Boston, MA, Brill, 2014, vol. 2, 320-346. 44. M. GOODACRE, On Choosing and Using Appropriate Analogies: A Response to F. Gerald Downing, in JSNT 26 (2003) 237-240, p. 239, refers to Goulder with approval. Note, however, that he does not echo Goulder’s claim that Luke is following Mark for the whole of Luke 3,1–6,19. 45. GOULDER, Luke’s Compositional Options (n. 17), p. 150.

290

A. GARROW

It is worth taking a step back at this point to observe how much of Luke, under this ‘alternating blocks’ model, was constructed without reference to Mark. Story elements in this category include: The Birth Narratives of John the Baptist and Jesus; John the Baptist’s ministry; Jesus’ Genealogy; The Temptations; The Rejection at Nazareth; The Call of the First Disciples; The Sermon on the Plain; Jesus’ Testimony to the Baptist; the long central Travel Narrative; The Lord’s Supper; The Arrest; Trial before the Sanhedrin; Trial before Pilate; Jesus before Herod; the Crucifixion; and the Resurrection. This amounts to more than three-quarters of Luke’s total narrative and includes every key juncture in the story46. If Luke did not consult Mark “at all” for these passages, then one thing is certain: Mark was not Luke’s frame. A refusal to accept this conclusion cuts against the claim that Luke used just one source at a time, which in turn removes the FH’s defence against the damaging charge of “unpicking”47. To put it bluntly, FH supporters cannot have it both ways. IV. FRAME AND FILL AND

THE

SYNOPTIC PROBLEM

With the preliminaries complete I now turn to trace the compositional processes required under the FH and the MPH. My aim here is to see which follows the patterns of frame and fill and which requires those patterns to run in reverse. 1. The Farrer Hypothesis The processes required under the Farrer Hypothesis begin, of course, with Mark. Mark gathers traditions, however received, and uses them to create his gospel narrative. 46. For what it is worth, as PELLING, Plutarch’s Method (n. 13), p. 91 observes, Plutarch’s frame source generally occupies about three-quarters of his narrative. 47. FH supporters do, nevertheless, affirm that Mark was Luke’s frame. For example, EVE, Relating (n. 1), p. 18: “Matthew and Luke are in sufficient agreement with Mark … that they can employ his gospel as the basis for their own”; p. 167: “Up to this point Luke has worked by alternating Markan and Matthean blocks, largely (though not exclusively) following the Markan narrative while mining Matthew for additional sayings material. On reaching his central section Luke once again switches back to Matthew while also bringing in material from elsewhere”; and p. 193: “[From Luke 18,15] Luke abandons Matthew to resume his use of Mark until the end of the Passion Narrative (although, as we shall see, further sequential uses of Matthew occasionally recur until the end of both gospels)”. Addressing the central section, GOODACRE, Way (n. 2), p. 39 writes: “The journey motif is a literary conceit that allows Luke to draw in his best Matthean material, while integrating it into a structure that is inspired by Mark”.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

291

The next event, according to this model, is the creation of Matthew. There can be little doubt that Matthew selected Mark as his main frame and filled out that frame using other resources. All the classic markers are present: Matthew uses almost every element of Mark; he generally adopts Mark’s order; Mark’s material usually provides the docking point onto which supplementary material is appended or into which it is inserted 48; and the removal the Markan material would leave a remainder that does not tell a coherent story (and vice versa). It is also the case, however, that Mark appears not to serve as Matthew’s frame at every juncture. This is most obvious in the case of the Sermon on the Mount. Markan material is indeed used to create the context for the Sermon, but within the Sermon Markan material is either absent or subordinate to an agenda set from elsewhere. This suggests that Matthew’s Sermon is built around a non-Markan subframe. This raises the question of where that sub-frame might have come from. The FH requires that Matthew either created it himself or drew it from an unknown source49. The former seems relatively unlikely since it would be in stark contrast to Matthew’s use of frame and fill in the remainder of the gospel. The latter, on the other hand, would require Matthew’s knowledge of a tradition remarkably like Luke’s Sermon on the Plain50. This is a subject to which I will return. Speaking more generally, Matthew is required, under the FH, to acquire or generate additional material with which to fill out Mark. On this model it is not possible to distinguish between Matthean redaction of Mark and the use of additional sources. The next event, following the FH, is the creation of Luke. Under this model Luke begins to create his gospel with knowledge of both Mark and Matthew. And, as FH proponents sometimes suggest, in a situation where 48. Thus, the Mission Discourse (Matthew 10), the Parabolic Discourse (Matthew 13), the Discourse on the Church (Matthew 18) and the Discourse on the End Times (Matthew 24–25) all open with a section of Mark that sets the agenda for the remainder of the Discourse – the contents of which are drawn from other sources as well as sometimes from elsewhere in Mark. A slightly different pattern occurs in the Sermon on the Mount, see below. 49. If the Sermon already existed as a complete unit (less perhaps a few minor additions from Mark), then there would be no difficulty since Matthew could simply have inserted this block wholesale into his text. The way Matthew operates elsewhere, however, suggest this is unlikely. In his other discourses Matthew gathers material from various sources. It is probable, therefore, that this is also how he creates his longest discourse. 50. Note that Matthew also needs a frame from elsewhere to create Matthew 11. As will be noted later, the Sermon on the Plain provides all the necessary elements of a frame for the Sermon on the Mount. It is a sermon, which the Matthean context requires, all the elements of the Sermon on the Plain are included and many of them are expanded using material from elsewhere – as Matthew does with Mark in the remainder of his text.

292

A. GARROW

Luke and his community are particularly familiar with Mark51. Under these circumstances FH Luke might be expected to accept the double attestation provided by Mark and Matthew regarding the Passion and Resurrection. Instead, however, FH Luke adopts a strategy unlikely to have been well received by audiences already well familiar with Mark (and familiar, if less so, with Matthew). He generates his own, alternative, account of the Passion and Resurrection. FH Luke also takes a surprising turn in declining to adopt Matthew as his frame. Matthew had already expanded Mark using material that Luke will similarly choose to add, albeit in a different way52. It is beyond dispute, however, that Luke does not make Matthew his frame: the order of the Matthean material in Luke is too substantially re-arranged for this to be the case. This presents the possibility that Luke chose Mark as his frame. This is the position favoured by the great majority of scholars, including FH supporters53. As noted above, however, taking this option places FH supporters in the awkward position of needing Mark to serve as Luke’s frame to avoid the problems described below, while also needing Luke extensively to ignore Mark to avoid the “unpicking” problem noted by Downing. Given the recognised severity of the latter, I continue on the basis that Luke used neither Mark nor Matthew as his frame. The difficulty for the FH, if Mark is not Luke’s frame, is that Luke is then required to adopt highly contrasting procedures when filling out his frame (sourced from elsewhere) with material from Mark and Matthew. The procedure Luke adopts with Mark is relatively simple. First, Luke looks for places in his frame where a block of Mark might be inserted. For example, the base narrative includes a sequence in which a parable about a Pharisee and a justified Tax Collector is then immediately followed by the story of Zaccheaus of Jericho – who becomes a justified Tax Collector. The humility of the Tax Collector provides a hook for the insertion of Mark’s story of the 51. EVE, Relating (n. 1), p. 23: “For as long as Mark remained the primary written narrative account of Jesus’ ministry available to them, it is likely that Matthew and Luke would have preached on it, taught from it, discussed it with friends and colleagues, deeply pondered it and generally internalized it as a central part of their tradition”. 52. F.G. DOWNING, A Paradigm Perplex: Luke, Matthew and Mark, in NTS 38 (1992) 15-36, p. 25, notes that if Luke knew both Mark and Matthew “we might well expect him only very rarely to refer to his scroll of Mark at all … As the churches quickly decided, almost everything of importance in Mark is there in Matthew”. See also the same article reproduced in ID., Doing Things with Words in the First Christian Century (SupplJSNT, 200), Sheffield, Sheffield Academic Press, 2000, pp. 174-197, esp. 185. Later authors, such as Tatian and Ammonius of Alexandria, identify Matthew as a useful frame because it has already done a lot of the heavy lifting involved in filling out Mark with additional material. 53. See note 47 above.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

293

blessing of the (humble) children – and this block of Mark also happens to include a link with Jericho, since it includes a description of how Jesus heals a blind man as he leaves Jericho. To connect this Markan block back to Luke’s frame, therefore, all that is needed is to place Mark’s blind man on the way into Jericho, rather than on the way out. Elsewhere, Luke similarly inserts whole blocks of Mark at points where such insertions expand, and do not too badly disrupt, the flow of the base narrative54. FH Luke’s pedestrian insertions of blocks of Mark stands in marked contrast to FH Luke’s adventurous treatment of Matthew. Mark Goodacre likens Luke’s handling of Matthew to radical reworkings of Matthew by Franco Zeffirelli and Martin Scorsese: One of the very things that many have claimed to be implausible about the Farrer theory’s Luke is one of the very things he shares with the Jesus films: the desire to do something radical with Matthew’s Sermon on the Mount, to abbreviate, to relocate, to redistribute, to restructure but most importantly to add some dramatic biographical plausibility to the substance of it. If we were fond of the language of trajectory and tendency, we might say that Luke is on a trajectory at the culmination of which are the Jesus films, the tendency of which is creatively and critically to rework the Sermon on the Mount55.

This contrast between FH Luke’s handling of Mark and Matthew is odd56. Odder still, however, is the fact that FH Luke’s unconventional treatment of Matthew has a very remarkable effect when viewed in reverse, as noted below. 2. The Matthean Posteriority Hypothesis The story told by the MPH begins, as does the FH, with Mark. Mark gathers traditions, however received, and uses them to create his gospel narrative. 54. The insertion of a Markan block at Luke 4,31-44 (Jesus heals Peter’s Mother-in-Law) creates a minor disruption in Luke’s narrative to the extent that Peter might be expected to receive Jesus into his house after his call to discipleship (Luke 5,1-11) rather than before. The natural continuity between the Call of the First Disciples (Luke 5,1-11), the Naming of the Twelve (Luke 6,12-16) and the Teaching of the Disciples (Sermon on the Plain) (Luke 6,17-49) is interrupted by the Markan insertion at Luke 5,12–6,11. The link between the Hospitable Women (Luke 8,1-3) and the Inhospitable Samaritans (Luke 9,51-56) is interrupted by the Markan insertion at Luke 8,1–9,50), and so on. 55. M. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, p. 130. 56. It is sometimes suggested that Luke’s treatment of Mark 4 provides a parallel to Luke’s treatment of Matthew. The minimal textual rearrangement required in the former is not, however, anything like the multiple and dramatic rearrangements required in the latter.

294

A. GARROW

The next event is the creation of Luke. As argued above, Mark does not serve as Luke’s frame and so, on this model, Luke’s frame comes from elsewhere and Markan blocks are inserted into that frame with rudimentary editorial skill. There is no evidence to suggest that this contrasts with how Luke used his other sources, if only because such are not extant57. The next event is the creation of Matthew. Under this model Matthew has a choice of two possible frames: Mark or Luke. It is beyond dispute that Matthew chooses Mark as his overall frame, which raises the question of why he rejected Luke. Matthew’s exact motivation is, of course, beyond our reach but the fact he includes almost all of Mark, and not even one quarter of Luke, shows that he had some reason for regarding Mark as the more authoritative of the two texts58. Matthew’s broad policy is then to fill out Mark’s agenda by gathering like supplementary materials – including from Luke. In the case of the Sermon on the Mount, however, Matthew does not use Mark as his frame. As noted above, it appears that Matthew drew on a separate sub-frame to support this section. An exceptionally strong candidate for that role is Luke’s Sermon on the Plain, which exhibits all the necessary features of such a frame. First, every element is reused in Matthew’s version. Second, they are re-used in the same order in Matthew’s version. Third, these elements are variously expanded upon by supplementary material from elsewhere. It would be a remarkable coincidence if Matthew happened to find a sub-frame from elsewhere that suited his purposes so perfectly. Continuing to focus on the Sermon on the Mount, an animation of its creation would show, first, the selection of the Sermon on the Plain as a frame and then the expansion of that frame by the gathering and insertion of related material from elsewhere in Luke, Mark and other sources. Such an animation would be comparable to animations of the work of Plutarch, Josephus and Tatian, each of which gathers related materials to fill out a frame. Herein lies the final obstacle to the possibility that Luke used Matthew. Luke treating Matthew like a Zeffirelli or a Scorsese makes it look like Matthew is using Luke like a Plutarch, a Josephus or a Tatian. It is simpler by far to suppose that Matthew looks like he is using Luke like one of his ancient contemporaries because that is what he was doing. 57. On a finer point of detail, I argue in A. GARROW, An Extant Instance of ‘Q’, in NTS 62 (2016) 398-417, that Didache 1,2-5a was a source for Luke. 58. The combined length of the pericopes substantially common to Luke and Matthew is less than a quarter of the total length of Luke. This does not include, for example, the Birth Narrative and those parts of the Passion Narrative where there is limited verbal similarity.

“FRAME AND FILL” AND MATTHEW’S USE OF LUKE

295

V. CONCLUSION It is sometimes claimed that FH Luke’s transpositions of Matthew are equivalent to MPH Matthew’s transpositions of Luke. These moves must have been made by someone, it is said, but there’s nothing to say that it was one way around or the other. Such statements are, however, a little like saying that action played forwards is the same as action played in reverse. A study of wider ancient compositional practices shows what normally happens when the action is played forwards. Authors who created new narratives by recycling older ones began by selecting one source as a frame and then expanded that frame by drawing in related material from other sources. The Farrer Hypothesis requires this process to run in reverse. For example, in the case of the Sermon on the Mount, “fill” material is scattered throughout Luke’s Gospel leaving the denuded “frame” in the form of the Sermon on the Plain. By contrast, the Matthean Posteriority Hypothesis follows the pattern of other ancient authors. It plays the “frame and fill” process forwards. SIIBS, University of Sheffield United Kingdom

Alan GARROW [email protected]

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

I. THE CENTURION’S SERVANT AND THE TWO-DOCUMENT HYPOTHESIS Craig A. Evans offers the Healing of the Centurion’s Servant pericope (Matt 7,28a; 8,5-13//Luke 7,1-10; 13,28-29) as an example for demonstrating the existence of Q and the usefulness of the Two-Document Hypothesis (2DH) for exegesis. Evans gives two arguments for Q based on this pericope. The first is that the presence of a very similar story in John 4,46b-54 indicates that this narrative has roots in very early tradition. The second is that Matthew’s inclusion of the logion about eschatological reversal in his version of the Centurion’s Servant, and Luke having this logion in a different context (Matt 8,11-12//Luke 13,28-29), indicates that Matthew and Luke used an earlier source, Q, independently of each other. Evans states that it is inexplicable that Luke, had he known Matthew’s Gospel, would have removed this logion from Matthew’s account of the Centurion’s Servant, since the logion seems to “fit” in Matthew’s context better than it fits with the somewhat loose collection of sayings in Luke 131. Other 2DH proponents agree with Evans’s points and offer additional arguments. One is the argument from alternating primitivity – that sometimes Matthew and sometimes Luke use wording that seems to go back to a version of the Centurion’s Servant pericope that is earlier than either of theirs. A commonly given example of Matthew’s primitivity is that he does not have Luke’s two groups of messengers sent by the centurion. There are 1. C.A. EVANS, The Two Source Hypothesis, in S.E. PORTER – B.R. DYER (eds.), The Synoptic Problem: Four Views, Grand Rapids, MI, Baker, 2016, 27-45, pp. 40-42. Note that Evans’s point about the eschatological reversal logion functions as an argument against Luke’s knowledge of Matthew, but does not work as an argument against Matthew’s knowledge of Luke. More recently, in agreement with Burkett, Evans has advocated a modified version of the 2DH, in which Matthew, Luke, and Mark all independently used as a source a “proto-Mark”, i.e., “a collection of unordered notes or memoirs that recorded Peter’s teaching”. See C.A. EVANS, Jesus and the Manuscripts: What We Can Learn from the Oldest Texts, Peabody, MA, Hendrickson Academic, 2020, pp. 24-25. Cf. D. BURKETT, The Case for Proto-Mark: A Study in the Synoptic Problem (WUNT, 399), Tübingen, Mohr Siebeck, 2018.

298

R.K. MACEWEN

aspects of Luke’s delegations that seem artificial, even odd; therefore, many scholars consider them to be a secondary feature added to the original story, one perhaps invented by Luke2. Also, the high degree of verbatim agreement between the two evangelists in this pericope is taken as evidence that a written source, and not just oral tradition, lies behind their accounts3. Before engaging with the above arguments, it is worth mentioning a problem that the Centurion’s Servant presents for the 2DH: this pericope does not look like something that belongs in Q. The Q source is theorized as a sayings Gospel, yet this pericope is a narrative. Of course, Q as commonly reconstructed has a few other narratives – the Baptism of Jesus (Q 3,21b-22)4, the Temptations of Jesus (Q 4,1-13), John’s Inquiry about the One to Come (Q 7,18-23), Confronting Potential Followers (Q 9,57-60), and Refuting the Beelzebul Accusation (Q 11,14-20)5. For an unknown reason, all these narratives about Jesus are located in the first one-third of Q; the final two-thirds consists only of sayings (including some parables). Although the Temptations of Jesus is an extended dialogue, most of the other Q narratives are quite brief – narrative introductions that provide occasions for the sayings of Jesus that follow. The Centurion’s Servant is not the only Q pericope that mentions a healing miracle. In Q 7,22, in response to a question, Jesus lists various types of healings he has performed. The pericope Refuting the Beelzebul Accusation (Q 11,14-15.17-20) contains mainly sayings of Jesus, but it begins with a brief report of a healing miracle. The Centurion’s Servant, however, stands apart; it looks like the kind of narrative one would expect to find in Mark rather than in Q. Like Mark’s account of the Syrophoenician Woman (Mark 7,24-30), it shows Jesus healing from a distance. Like several of Mark’s narratives, it emphasizes the importance of faith and features a character who requests healing for someone else (Mark 1,29-31; 2,1-12; 5,21-43; 7,24-30; 9,14-29; cf. 6,1-6). And rather than being a vehicle for presenting 2. R.A.J. GAGNON, The Shape of Matthew’s Q Text of the Centurion at Capernaum: Did It Mention Delegations?, in NTS 40 (1994) 133-142; J. NOLLAND, Luke 1–9:20 (WBC, 35A), Dallas, TX, Word, 1989, p. 314. 3. J.S. KLOPPENBORG, Excavating Q: The History and Setting of the Sayings Gospel, Minneapolis, MN, Fortress, 2000, pp. 56-59. 4. Many 2DH proponents, however, argue that Q did not include the baptism of Jesus. See the list in O. ANDREJEVS, The Double Tradition in Luke (Q) 3-7 as a Macro-Chiasm and its Significance for the Synoptic Problem, in NTS 67 (2021) 388-406, p. 398, n. 44. 5. These pericope titles are those used in J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000 (here cited: CritEd).

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

299

sayings of Jesus, in the Centurion’s Servant pericope the words spoken by the centurion outnumber those of Jesus6. Although the Centurion’s Servant seems out of place in Q, 2DH proponents generally accept it as part of Q without question or discussion7. This is probably due to two key reasons—the pericope qualifies as double tradition (DT), being found in Matthew and Luke but not in Mark, and the pericope displays strong verbatim agreement between Matthew and Luke. Nevertheless, this pericope’s lack of fit with Q raises questions, both about the assignment of the pericope to Q and about the validity of the Q theory itself. Another problem for the 2DH that is raised by the Centurion’s Servant is the way the Q version of the pericope is reconstructed. Since Q is reconstructed on the basis of Matthew and Luke’s common wording, and since Luke’s version of the Centurion’s Servant is longer than Matthew’s8, the Q version ends up looking more like Matthew’s version than Luke’s. However, in the triple tradition (TT) narratives, Matthew’s version is often the shortest while Mark’s (the earliest version) is usually the longest9. If one attempted to reconstruct a Markan pericope such as the Dying Girl and the Hemorrhaging Woman (Matt 9,18-26//Mark 5,21-43//Luke 8,40-56) on the basis of the material common to Matthew and Luke10, the resulting reconstruction would be much more like Matthew’s version than Mark’s, because the shorter version (between Matthew’s and Luke’s) would determine the results. This does not inspire confidence in reconstructions of Q. Proponents of the 2DH see Q as a collection of some of the earliest Jesus traditions. Since the Healing of the Royal Official’s Son pericope in John 4,46b-54 has some striking commonalities with the Centurion’s 6. Scholars disagree as to whether Q’s Centurion’s Servant pericope ended with Jesus’s words praising the centurion’s faith (Q 7,9) or went on to confirm that the healing occurred (v. 10). For the debate, see S.R. JOHNSON, Q 7:1-10: The Centurion’s Faith in Jesus’ Word (Documenta Q), Leuven – Paris – Sterling, VA, Peeters, 2002, pp. 355-400. This debate is somewhat surprising because, although Matt 8,13 and Luke 7,10 are completely different in wording, both narrate that the servant was healed. 7. The Documenta Q volume on this pericope (n. 6) asks and answers many questions, but “Was this pericope in Q?” is not one of them. 8. The Greek text of Matt 7,28a; 8,5-13 has a total of 176 words. Luke 7,1-10; 13,28-29 has a total of 233 words. Comparing only the units that are normally identified as the Centurion’s Servant pericope, Matt 8,5-13 has 167 words and Luke 7,1-10 has 186 words. 9. For a table comparing the lengths of many parallel pericopes in the Synoptic Gospels, see R.H. STEIN, Studying the Synoptic Gospels: Origin and Interpretation, Grand Rapids, MI, Baker Academic, 22001, pp. 53-55. 10. The idea of reconstructing Mark the way Q is reconstructed is presented by E. EVE, Reconstructing Mark: A Thought Experiment, in M. GOODACRE – N. PERRIN (eds.), Questioning Q: A Multidimensional Critique, Downers Grove, IL, InterVarsity Press, 2004, 89-114.

300

R.K. MACEWEN

Servant, this is seen as supporting both the existence of Q and the Centurion’s Servant’s placement in Q. Proponents of the 2DH often argue that these are two versions of what was originally the same traditional story, modified over time, until they became inscribed in different forms in Q and in John11. They also tend to view the Johannine pericope as an aid for the reconstruction of Q, regarding common elements between Luke and John or between Matthew and John as representing the earliest tradition12. In John’s version of the pericope, as in Matthew’s and Luke’s, a highstatus individual who lives in Capernaum seeks out Jesus, asking him to heal a member of his household who is very sick. He believes that Jesus is able to heal at a distance, and the sick person is indeed healed. In addition to a similar storyline, John’s version also has wording in common with the DT version. In common with both Matthew and Luke, John’s version of the pericope uses the words ἰάομαι (Matt 8,8.13; Luke 7,7; John 4,47), κύριε (Matt 8,6.8; Luke 7,6; John 4,49), παῖς (Matt 8,6.8.13; Luke 7,7; John 4,51), and λόγῳ in connection with the petitioner’s faith in Jesus’s speech (Matt 8,8; Luke 7,7; John 4,50). An agreement between only John and Matthew is the expression used to describe the immediacy of the healing (ἐν τῇ ὥρᾳ ἐκείνῃ [Matt 8,13]; [ἐν] ἐκείνῃ τῇ ὥρᾳ [John 4,53]). Agreements in wording between only John and Luke in the pericope include ἀκούσας (Luke 7,3; John 4,47), as well as πρὸς αὐτὸν and the verb ἐρωτάω (Luke 7,3; John 4,47), to describe the initial approach to Jesus. Particularly striking are strings of wording that are nearly identical in John and Luke: ἐπορεύετο σὺν αὐτοῖς. ἤδη δὲ αὐτοῦ οὐ μακρὰν ἀπέχοντος (Luke 7,6); ἐπορεύετο. ἤδη δὲ αὐτοῦ καταβαίνοντος (John 4,50-51). Also, both John and Luke state that the sick person was near death (ἤμελλεν τελευτᾶν [Luke 7,2]; ἤμελλεν γὰρ ἀποθνῄσκειν [John 4,47]). There are also some significant differences between the Synoptic Gospels’ pericopes and John’s version. In Matthew and Luke the petitioner is a centurion; in John he is a royal official or a member of a royal family (βασιλικός [John 4,46])13. In Matthew and Luke both Jesus and the petitioner are in Capernaum; in John the petitioner travels from Capernaum to Cana to see Jesus (v. 46). As for the dialogue between the petitioner and Jesus, in the 11. W.D. DAVIES – D.C. ALLISON, A Critical and Exegetical Commentary on the Gospel according to Saint Matthew (ICC), Edinburgh, T&T Clark, 1988-97, vol. 2, pp. 17-32; J.A. FITZMYER, The Gospel according to Luke I–IX (AB, 28), Garden City, NY, Doubleday, 1981, p. 648. 12. GAGNON, The Shape of Matthew’s Q Text (n. 2), p. 135, incl. n. 7; DAVIES – ALLISON, Matthew (n. 11), vol. 2, pp. 17-18, 22. 13. See BDAG, sub voce.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

301

synoptics the petitioner believes before Jesus assures him that the sick person is healed, while in John Jesus declares, “Your son lives”, and then the petitioner believes (v. 50). John also has a unique conclusion to the story – the petitioner’s entire household believes in Jesus (v. 53). There are several possible ways to explain the similarities and differences between John’s Healing of the Royal Official’s Son and Matthew and Luke’s Healing of the Centurion’s Servant, including: (1) They represent two divergent streams of a single, older tradition, as many 2DH proponents think. (2) They represent two distinct, though similar, events in the life of Jesus. (3) John’s version is from a recension of Q that is earlier than the recension used by Matthew and Luke. (4) John’s version is from a recension of Q that is later than the recension used by Matthew and Luke. (5) John received the story from Q and creatively rewrote it. (6) The author of Q creatively rewrote an earlier version of the story that corresponds to the Johannine tradition. (7) There was no cross-pollination between the two versions or the traditions behind them; their similarities are purely coincidental. (8) John knew Matthew and/or Luke and creatively rewrote their version. (9) John had his own version of the story, but he also knew Matthew and/or Luke; his memory of the other Gospel’s/Gospels’ version(s) led him to incorporate some of its/their language and themes when he wrote his own version. Since there are so many possibilities, and since we do not have enough evidence to determine which one(s) is/are correct, it would seem hazardous to rely on John’s version as a basis either for belief in the existence of Q or for reconstructing Q. If, as most scholars believe, John was the last of the NT Gospels to be written14, it is quite possible that John knew the Gospels of Matthew and Luke and was influenced by them. Therefore, I prefer either option 8 or 9 for explaining the relationship between the Centurion’s Servant and the Royal Official’s Son. Because of the significant differences between the two accounts, I lean toward option 9. In the quest to reconstruct the putative original wording of Q, scholars have the easiest work wherever Matthew and Luke present identical wording in the DT. Such perfect agreement allows reconstructors to feel certain that they have the original wording of Q in those places. Where Matthew and Luke have different wording, scholars pay attention to the characteristic wording of Matthew and Luke. At a given verse, if Matthew’s wording is 14. G.R. BEASLEY-MURRAY, John (WBC, 36), Nashville, TN, Nelson, 21999, p. xxv; D. GUTHRIE, New Testament Introduction, Leicester, Apollos; Downers Grove, IL, InterVarsity Press, 41990, p. 298.

302

R.K. MACEWEN

full of Mattheanisms, while Luke’s language is generic, the wording of Luke will be seen to represent Q. On the other hand, if Matthew’s language is generic, while Luke’s language has distinctively Lukan features, Matthew’s language will be preferred as closer to Q. If at a given verse, both Matthew and Luke display their own distinctive language, Q reconstructors will have less confidence in determining Q’s wording15. The International Q Project’s (IQP’s) reconstruction of the Centurion’s Servant pericope (Q 7,1-10) contains a total of one hundred words16. This includes sixty-four words common to both Matthew and Luke, twenty-nine words found only in Matthew’s version, five words found only in Luke’s version, and two words that are found neither in Matthew nor Luke in the same inflected form. The IQP views the logion on eschatological reversal (Q 13,29.28) as originally separate from the Centurion’s Servant pericope, as Luke has it. The IQP’s reconstruction of this logion contains a total of thirty-seven words, including twenty-five words found in both Matthew and Luke, seven words found only in Matthew’s version, four words found only in Luke’s version, and an additional one word found in both Matthew and Luke in an inflected form that differs from both of theirs. Thus, in the Centurion’s Servant pericope, due to the strong preference for Matthew’s wording, the IQP’s Q looks much more like Matthew’s version than Luke’s, except that it does not include the eschatological reversal logion. Contributing to the impression that Matthew, more than Luke, has preserved the original wording of Q is the IQP’s omission of Luke’s two delegations; since they are found only in Luke’s Gospel, they are considered a redactional addition by Luke. As for the eschatological reversal logion, although the IQP views Luke’s placement as corresponding to its setting in Q, its reconstruction of the logion looks much more like Matthew’s version, not only because the IQP slightly favors Matthew’s wording, but also because it prefers Matthew’s order of the phrases that make up the logion17.

15. KLOPPENBORG, Excavating Q (n. 3), pp. 101-104. 16. This word count does not include Q 7,10. The CritEd presents Q 7,10 using the sigla, “?10? ”, indicating that the editors are uncertain whether this verse stood in Q and, if it did, uncertain how to reconstruct it. CritEd (n. 5), p. 116. 17. The conclusion that Matthew better preserves Q’s wording, while Luke better preserves Q’s pericope order, runs counter to ancient compositional practices, in which a writer would normally preserve his source’s wording more closely when following that source’s order. See R.A. DERRENBACKER, The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 435-457, p. 441.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

303

I created the following synopsis (Table 1) as a tool for studying Matthew’s and Luke’s characteristic wording in the different versions of the Centurion’s Servant. Words and phrases that are clearly characteristic of Matthew are indicated with small-dash underlining. Words and phrases that are marginally characteristic of Matthew are indicated with dotted underlining. Words and phrases that are clearly characteristic of Luke are indicated with doubleline underlining. Words and phrases that are marginally characteristic of Luke are indicated with single-line underlining18. Words not underlined are not considered to be characteristic of either Matthew or Luke. One result visible in this synopsis is that Matthew uses some words and phrases that are characteristic of Luke, and Luke uses several words and phrases that are characteristic of Matthew. Table 1. Synopsis of the Centurion’s Servant Pericope19 Matt 7,28a, 8,5-13 Καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησοῦς τοὺς λόγους τούτους … 5 Εἰσελθόντος δὲ αὐτοῦ εἰς Καφαρναοὺμ προσῆλθεν αὐτῷ ἑκατόνταρχος παρακαλῶν αὐτὸν 6 καὶ λέγων· κύριε, ὁ παῖς μου βέβληται ἐν τῇ οἰκίᾳ παραλυτικός, δεινῶς βασανιζόμενος. 28a

Q 7,1-10; 13,29.28 ⟦καὶ ἐγένετο ὅτε⟧ ἐ⟦πλήρω⟧σεν .. τοὺς λόγους τούτους, εἰσῆλθεν εἰς Καφαρναούμ. 3 ἦλθεν αὐτῷ ἑκατόνταρχ⟦ο⟧ς παρακαλῶν αὐτὸν ⟦καὶ λέγων·⟧ ὁ παῖς ⟦μου κακῶς ἔχ. 1

Luke 7,1-10; 13,28-29 Ἐπειδὴ ἐπλήρωσεν πάντα τὰ ῥήματα αὐτοῦ εἰς τὰς ἀκοὰς τοῦ λαοῦ, εἰσῆλθεν εἰς Καφαρναούμ. 2 Ἑκατοντάρχου δέ 1

τινος δοῦλος κακῶς ἔχων ἤμελλεν τελευτᾶν, ὃς ἦν αὐτῷ ἔντιμος. 3 ἀκούσας δὲ περὶ τοῦ Ἰησοῦ ἀπέστειλεν πρὸς αὐτὸν πρεσβυτέρους τῶν Ἰουδαίων ἐρωτῶν αὐτὸν ὅπως ἐλθὼν διασώσῃ τὸν δοῦλον αὐτοῦ. 4 οἱ δὲ παραγενόμενοι πρὸς τὸν Ἰησοῦν παρεκάλουν αὐτὸν σπουδαίως λέγοντες ὅτι ἄξιός ἐστιν ᾧ παρέξῃ τοῦτο· 5 ἀγαπᾷ γὰρ τὸ ἔθνος ἡμῶν καὶ τὴν συναγωγὴν αὐτὸς ᾠκοδόμησεν ἡμῖν. 6 ὁ

18. See the Appendix to this essay for an explanation of how these categories of characteristic wording are defined, as well as lists of words with frequency statistics for their use in the Gospels and Acts. 19. The text of Q in this synopsis, including sigla (apart from the underlining), is from CritEd (n. 5), pp. 102-117, 414-417. Texts and quotations from the Greek NT, here and throughout this essay, are taken from NA28.

304

R.K. MACEWEN

Matt 7,28a, 8,5-13

Q 7,1-10; 13,29.28

Luke 7,1-10; 13,28-29

καὶ λέγει αὐτῷ· ἐγὼ ἐλθὼν καὶ λέγει αὐτῷ· ἐγὼ⟧ ἐλθὼν δὲ Ἰησοῦς ἐπορεύετο σὺν θεραπεύσω αὐτόν θεραπεύς⟦ω⟧ αὐτόν; αὐτοῖς. ἤδη δὲ αὐτοῦ οὐ μακρὰν ἀπέχοντος ἀπὸ τῆς 8 6b-c οἰκίας ἔπεμψεν φίλους ὁ καὶ ἀποκριθεὶς ὁ καὶ ἀποκριθεὶς ὁ ἑκατοντάρχης λέγων ἑκατόνταρχος ἔφη· ἑκατόνταρχος ἔφη· αὐτῷ· κύριε, μὴ σκύλλου, κύριε, κύριε, οὐκ εἰμὶ ἱκανὸς ἵνα μου ὑπὸ οὐκ εἰμὶ ἱκανὸς ἵνα μου ὑπὸ οὐ γὰρ ἱκανός εἰμι ἵνα ὑπὸ τὴν στέγην μου εἰσέλθῃς· τὴν στέγην εἰσέλθῃς, τὴν στέγην εἰσέλθῃς, 7 διὸ οὐδὲ ἐμαυτὸν ἠξίωσα πρὸς σὲ ἐλθεῖν· ἀλλ᾽ εἰπὲ λόγῳ, καὶ ἀλλὰ μόνον εἰπὲ λόγῳ, καὶ 7 ἀλλὰ εἰπὲ λόγῳ, καὶ ἰαθήτω ὁ παῖς μου. ἰαθή⟦τω⟧ ὁ παῖς μου. ἰαθήσεται ὁ παῖς μου. 9 8 8 καὶ γὰρ ἐγὼ ἄνθρωπός καὶ γὰρ ἐγὼ ἄνθρωπός καὶ γὰρ ἐγὼ ἄνθρωπός εἰμι εἰμι εἰμι ὑπὸ ἐξουσίαν, ὑπὸ ἐξουσίαν, ὑπὸ ἐξουσίαν τασσόμενος ἔχων ὑπ᾽ ἐμαυτὸν ἔχων ὑπʼ ἐμαυτὸν ἔχων ὑπ᾽ ἐμαυτὸν στρατιώτας, καὶ λέγω στρατιώτας, καὶ λέγω στρατιώτας, καὶ λέγω τούτῳ· πορεύθητι, τούτῳ· πορεύθητι, τούτῳ· πορεύθητι, καὶ πορεύεται, καὶ πορεύεται, καὶ πορεύεται, καὶ ἄλλῳ· ἔρχου, καὶ ἄλλῳ· ἔρχου, καὶ ἄλλῳ· ἔρχου, καὶ ἔρχεται, καὶ ἔρχεται, καὶ ἔρχεται, καὶ τῷ δούλῳ μου· καὶ τῷ δούλῳ μου· καὶ τῷ δούλῳ μου· ποίησον τοῦτο, καὶ ποιεῖ. ποίησον τοῦτο, καὶ ποιεῖ. ποίησον τοῦτο, καὶ ποιεῖ. 10 9 9 ἀκούσας δὲ ὁ Ἰησοῦς ἀκούσας δὲ ὁ Ἰησοῦς ἀκούσας δὲ ταῦτα ὁ Ἰησοῦς ἐθαύμασεν καὶ ἐθαύμασεν καὶ ἐθαύμασεν αὐτὸν καὶ εἶπεν τοῖς ἀκολουθοῦσιν· εἶπεν τοῖς ἀκολουθοῦσιν· στραφεὶς τῷ ἀκολουθοῦντι ἀμὴν λέγω ὑμῖν, λέγω ὑμῖν, αὐτῷ ὄχλῳ εἶπεν· λέγω ὑμῖν, παρ᾽ οὐδενὶ οὐδὲ ἐν τῷ Ἰσραὴλ οὐδὲ ἐν τῷ Ἰσραὴλ τοσαύτην πίστιν ἐν τῷ τοσαύτην πίστιν τοσαύτην πίστιν Ἰσραὴλ εὗρον. εὗρον. εὗρον. 11 13,28 Λέγω δὲ ὑμῖν ὅτι πολλοὶ 13,29 ⟦καὶ πολλοὶ⟧ ἐκεῖ ἔσται ὁ κλαυθμὸς ἀπὸ ἀνατολῶν καὶ δυσμῶν ἀπὸ ἀνατολῶν καὶ δυσμῶν καὶ ὁ βρυγμὸς τῶν ὀδόντων, ἥξουσιν καὶ ἀνακλιθήσονται ἥξουσιν καὶ ἀνακλιθήσονται ὅταν ὄψησθε 13,28 μετὰ Ἀβραὰμ καὶ Ἰσαὰκ μετὰ Ἀβραὰμ καὶ Ἰσαὰκ Ἀβραὰμ καὶ Ἰσαὰκ καὶ Ἰακὼβ καὶ Ἰακὼβ καὶ Ἰακὼβ καὶ πάντας τοὺς ἐν τῇ βασιλείᾳ ἐν τῇ βασιλείᾳ προφήτας ἐν τῇ βασιλείᾳ τῶν οὐρανῶν, 12 οἱ δὲ υἱοὶ τῆς τοῦ θεοῦ, ⟦ὑμς⟧ δὲ τοῦ θεοῦ, ὑμᾶς δὲ βασιλείας ἐκβληθήσονται ἐκβλ⟦ηθής ἐκβαλλομένους εἰς τὸ εἰς τὸ σκότος τὸ ἐξώτερον· ἐκεῖ σκότος τὸ⟧ ἐξώ⟦τερον⟧· ἐκεῖ ἔξω. 13,29 καὶ ἥξουσιν ἀπὸ ἔσται ὁ κλαυθμὸς καὶ ὁ ἔσται ὁ κλαυθμὸς καὶ ὁ ἀνατολῶν καὶ δυσμῶν καὶ βρυγμὸς τῶν ὀδόντων. βρυγμὸς τῶν ὀδόντων. ἀπὸ βορρᾶ καὶ νότου καὶ ἀνακλιθήσονται ἐν τῇ βασιλείᾳ τοῦ θεοῦ. 13 7,?10? 7,10 καὶ εἶπεν ὁ Ἰησοῦς τῷ

Καὶ ὑποστρέψαντες εἰς ἑκατοντάρχῃ· ὕπαγε, ὡς τὸν οἶκον οἱ πεμφθέντες ἐπίστευσας γενηθήτω σοι. καὶ εὗρον ἰάθη ὁ παῖς [αὐτοῦ] τὸν δοῦλον ὑγιαίνοντα. ἐν τῇ ὥρᾳ ἐκείνῃ. 7

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

305

The result is not completely surprising, because I do not define “characteristic wording” as that which an evangelist uses exclusively, but as wording that one evangelist uses more frequently than the other, and thus is more likely to use than the other evangelist at a given location. It is conceivable that one evangelist can occasionally use the other’s characteristic wording by coincidence, especially if the wording is from the store of commonly used vocabulary. Perhaps the most surprising instance of Matthew and Luke using each other’s characteristic vocabulary is at Matt 8,10b//Luke 7,9b, where, at the same place, Matthew uses οὐδενί and Luke uses οὐδέ. This is unexpected because, in this instance, each evangelist is using a word that is strongly characteristic of the other. On the 2DH, Q did not depend on Matthew and Luke but was a source used by them. Therefore, we might expect Q to contain mainly generic wording, rather than wording that is characteristic of either Matthew or Luke. But when Q is reconstructed, we find in it an impressive amount of both these Gospels’ characteristic vocabulary. Q 7,6-9, where there is a high degree of verbatim agreement between Matthew and Luke, contains a considerable amount of wording that is clearly characteristic of Luke: ἱκανός, ἰαθήτω, ἐγώ, ἄνθρωπός εἰμι, τούτῳ, πορεύθητι, πορεύεται, τοῦτο, ἐθαύμασεν, and εὗρον. Q 13,29.28, where there is also a high degree of verbatim agreement between Matthew and Luke, includes a great deal of wording that is clearly characteristic of Matthew: πολλοί, ἀπὸ ἀνατολῶν, ἐν τῇ βασιλείᾳ, εἰς τὸ σκότος τὸ ἐξώτερον, and ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων. To a 2DH proponent the appearance of Matthew’s and Luke’s characteristic vocabulary in Q might not seem to be a problem, but to a Q skeptic it raises suspicions. In Matt 8,8-10//Luke 7,6c-9, the two Gospels are identical in most of their wording, so Q reconstructors have no choice but to accept all that identical wording into Q. But because much of this identical wording is characteristic of Luke, the result is verses in Q that stylistically look like Luke. On the other hand, in Matt 8,11-12//Luke 13,28-29, the identical wording is more characteristic of Matthew, so the wording of the reconstruction of Q is also Matthean20. Although Q reconstructors tend to avoid Matthew’s and Luke’s characteristic vocabulary where they can (and they have an easier time of it wherever the two Gospels differ in their wording), passages in which Matthew and Luke are identical provide no option – the 20. Note that the IQP also includes some wording found only in Matthew in its reconstruction of these verses (πολλοί and εἰς τὸ σκότος τὸ ἐξώτερον), thus further enhancing its Matthean character.

306

R.K. MACEWEN

identical wording must be seen as the ipsissima verba of Q, even when the language is strongly characteristic of either Matthew or Luke. Believers in Q may find alternating primitivity in Matthew and Luke in the process of reconstructing Q, but Q skeptics see alternation between Matthew’s wording and Luke’s wording. A direct utilization hypothesis, however, obviates the need to establish which evangelist’s wording is more primitive in a given verse. On the Matthean Posteriority Hypothesis (MPH), for example, Matthew directly uses Luke as his source for his DT material21. Sometimes he closely copies Luke, resulting in strong verbal agreement between Matthew and Luke and the incorporation of characteristically Lukan wording into Matthew’s Gospel. At other times Matthew follows Luke less closely, with the result that both Matthew and Luke retain their own distinctive wording in the DT. The result is that Matthew is allowed to be Matthew and Luke is allowed to be Luke, even in the DT. Despite the problem of wording characteristic of Matthew or Luke being included in Q, however, 2DH proponents frequently argue that the wording of the DT supports their theory, especially in pericopes with high levels of verbal agreement between Matthew and Luke. It is argued that this high agreement shows that there must have been a written document (Q) on which the two evangelists both relied. John S. Kloppenborg, for example, points out that the high degree of verbatim agreement that the Gospels have with each other is quite unexpected, given the way ancient authors normally used sources. He writes, “Were Matthew and Luke using Q (or Mark) as Diodorus, Josephus, 1QapGen ar or Ps-Philo used their sources, we should expect almost no verbatim agreement, since both would sometimes paraphrase generously and, because they would have done so independently, the likelihood of coincidental agreement in not changing Q would be exceedingly low”22. To Kloppenborg, the evidence from ancient writers proves that Q must have been a written source rather than an oral source or sources, since the verbal agreement between Matthew and Luke in their DT pericopes is so surprisingly close23. However, in agreement with other Q skeptics, I would argue that this high level of agreement in the DT proves more than this – it proves that there was no Q source at all. This is because the high degree of 21. R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1992) 1-22. 22. J.S. KLOPPENBORG, Variation in the Reproduction of the Double Tradition and an Oral Q?, in ETL 83 (2007) 53-80, pp. 73-74. Emphasis original. 23. KLOPPENBORG, Excavating Q (n. 3), pp. 56-58.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

307

agreement between Matthew and Luke in the DT, higher than what they achieve with each other when they independently use Mark as a source, is much more naturally explained by one of them having used the other directly as a source than by both of them having coincidentally copied verbatim and at length from a common source24. My book on the Synoptic Problem includes a comparative study of the phenomenon of verbatim agreement25. I collected and analyzed all strings of verbatim agreement of four or more words (SVA*4) between any two of the Synoptic Gospels26. To be counted, an SVA had to include exactly the same words in the same form (thus no differences in spelling, case, tense, etc.) in the same order (thus no transpositions) with no insertions, omissions, or substitutions of words in either Gospel. I also counted the total number of words in every Synoptic Gospel pericope that had a parallel27, so that I could determine what percentage of the words in each pericope consisted of SVAs*4. A key result of this study was to show that Matthew and Luke achieve a much higher degree of verbatim agreement with each other in their DT pericopes than they do in the TT. In the Gospel of Matthew, MatthewLuke SVAs*4 make up 21.0% of the wording in the DT, but only 11.9% of the wording in the TT (see Figure 1 below). This reveals a problem for the 2DH. According to this theory, Matthew and Luke never have any direct contact with each other. In the TT they are independently using Mark; in the DT they are independently using Q. Why then do Matthew and Luke agree with each other so much more closely when their common source is Q than they do when their common source is Mark?

24. E. POWELL, The Myth of the Lost Gospel, Las Vegas, NV, Symposium, 2006, pp. 165171; M. GOODACRE, Too Good to Be Q: High Verbatim Agreement in the Double Tradition, in J.C. POIRIER – J. PETERSON (eds.), Marcan Priority without Q: Explorations in the Farrer Hypothesis (LNTS, 455), London – New York, Bloomsbury T&T Clark, 2015, 82-100; A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226, pp. 212-213. 25. R.K. MACEWEN, Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS, 501), London – New York, Bloomsbury T&T Clark, 2015, pp. 50-73. 26. I omitted shorter SVAs because they are more easily explained as reflecting coincidental agreements or as reflecting standard idioms and usages of Koine Greek. The abbreviation “SVA*4” was coined by GARROW, Streeter’s ‘Other’ (n. 24), p. 212. For my statistical data on the SVAs*4 among the Synoptic Gospels, see Tables 2.16-2.20 in MACEWEN, Matthean Posteriority (n. 25), pp. 229-241. 27. See Tables 2.4-2.14 in ibid., pp. 197-218.

308

R.K. MACEWEN

Figure 1. Verbatim Agreement between Matthew and Luke in Matthew's Gospel 25% 20% 15% 10% 5% 0%

Agreement in Markan pericopes

Agreement in non-Markan pericopes

The left column shows the total number of words in Matthew-Luke TT SVAs*4 as a percentage of the total number of words in Matthew’s TT pericopes. The right column shows the total number of words in MatthewLuke DT SVAs*4 as a percentage of the total number of words in Matthew’s DT pericopes. On the 2DH, this roughly shows the differing levels of verbatim agreement that Matthew and Luke achieve with each other through their independent use of the same sources, depending on whether the source is Mark or Q. The ability of Matthew and Luke to achieve higher levels of verbatim agreement in their DT material than in TT material is not satisfactorily explained by the fact that their DT material includes mainly sayings of Jesus while the TT material includes both sayings and narrative. It is demonstrable that the each of the evangelists (at least each one who used one or more of the others as sources) were more conservative in preserving the sayings of Jesus than in preserving narrative, since quoted material constitutes the largest number of SVAs in every category of two-Gospel combinations28. There are plenty of sayings of Jesus in the TT material, yet Matthew and Luke are unable to achieve the highest levels of sustained, word-for-word agreement in the sayings of Jesus in the TT material that they do numerous times in the sayings of Jesus in their DT material29. Matthew and Luke’s differing levels of agreement with each other in the two types of material can be most easily and naturally explained by dispensing with Q. Then the two Gospels’ lower level of agreement in the TT can be 28. See Table 2.19 in ibid., pp. 235-239. 29. Moreover, some of the longest SVAs in the DT involve sayings of people other than Jesus – John the Baptist (Matt 3,9-10//Luke 3,8-9) and the centurion of Capernaum (Matt 8,9-10//Luke 7,8-9).

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

309

accounted for by their being separated from each other by their common source Mark, while their higher level of agreement in the DT can be explained by one Gospel making direct use of the other. The preceding discussion has shown that the Centurion’s Servant pericope reveals a number of problems for the Q hypothesis. First, this pericope does not well fit the standard profile of Q as a sayings collection; it is much more at home in a narrative Gospel. Second, the standard reconstructions of the Q form of the Centurion’s Servant, which yield versions very much like Matthew’s, are problematic, since in the TT Matthew’s pericopes are typically the shortest, while the earliest (Markan) versions are longer. Third, reconstructions of the Q form of this pericope are unable to avoid incorporating the characteristic wording of Matthew and Luke; this suggests that a form earlier than that of the Synoptic Gospels has not actually been recovered. Fourth, the high level of verbatim agreement between Matthew and Luke in the DT, exemplified by the Centurion’s Servant, indicates that direct utilization of one Gospel by the other is more likely than both having separately depended on an earlier source. In light of these problems for the 2DH, I will go on to consider two alternatives – Luke’s dependence on Matthew and Matthew’s dependence on Luke. II. THE CENTURION’S SERVANT AND THE FARRER HYPOTHESIS Proponents of the Farrer Hypothesis (FH), the theory that Luke used both Mark and Matthew as sources, take the Centurion’s Servant pericope as an excellent example of the validity of their view. One reason is that this pericope contains two examples of characteristic Matthean expressions – phrases that are each used only once by Luke but appear several times in Matthew. These phrases are Καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησοῦς τοὺς λόγους τούτους (Matt 7,28a) // Ἐπειδὴ ἐπλήρωσεν πάντα τὰ ῥήματα αὐτοῦ (Luke 7,1a)30 and ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων (Matt 8,12b//Luke 13,28a). Luke uses each of these sayings only once, in parallel with Matthew, while Matthew uses each of them multiple additional times throughout his Gospel31. Other colorful expressions that Luke uses only once but Matthew 30. The phrases in Matt 7,28a and Luke 7,1a are completely different in wording but essentially have the same meaning. 31. The phrase καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησοῦς is also found in Matt 11,1; 13,53; 19,1; 26,1. The phrase ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων is also found in Matt 13,42.50; 22,13; 24,51; 25,30.

310

R.K. MACEWEN

multiplies include γεννήματα ἐχιδνῶν (Matt 3,7//Luke 3,7; Matt 12,34; 23,33) and ὀλιγόπιστοι/ὀλιγόπιστε (Matt 6,30//Luke 12,28; Matt 8,26; 14,31; 16,8). Farrerians argue that, since Luke uses each of these expressions only once while Matthew uses them multiple times, they must be seen as characteristic of Matthew, and therefore it is more likely that Luke picked them up while copying from Matthew than that Matthew found them in Q or Luke and multiplied his use of them32. A second feature of the Centurion’s Servant pericope in which FH proponents find support for their theory is an instance of what Mark Goodacre calls “editorial fatigue”: a phenomenon that will inevitably occur when a writer is heavily dependent on another’s work. In telling the same story as his predecessor, a writer makes changes in the early stages which he is unable to sustain throughout. Like continuity errors in film and television, examples of fatigue will be unconscious mistakes, small errors of detail which naturally arise in the course of constructing a narrative33.

Goodacre finds a “clear” example of this in the Centurion’s Servant34. Luke apparently prefers to identify the servant as a δοῦλος, in contrast to Matthew’s term, παῖς; he makes this change twice, at the beginning and the end of the pericope (cf. Matt 8,6//Luke 7,2; Matt 8,13//Luke 7,10). But in the middle of the pericope, at a point at which Luke is closely copying Matthew’s wording, Luke seems to slip, and takes over Matthew’s παῖς into his own text (Matt 8,8//Luke 7,7)35. According to Goodacre, though there are “several” examples like this in which either Luke or Matthew fatigues while copying Mark, and in which Luke fatigues while copying Matthew, “it seems impossible to find reverse examples, cases where Matthew has apparently become fatigued with Q” or with Luke36. Thus, Luke knew and used Matthew; Matthew did not know Luke. A third reason FH proponents find the Centurion’s Servant to be supportive of their theory is the apparent artificiality, implausibility, and awkwardness of Luke’s two delegations sent by the centurion. About these delegations Michael D. Goulder comments, “It is a great pity that Luke 32. M. GOODACRE, The Case against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, pp. 62-63. 33. ID., Fatigue in the Synoptics, in NTS 44 (1998) 45-58, p. 46, quoted in T. HÄGERLAND, Editorial Fatigue and the Existence of Q, in NTS 65 (2019) 190-206, pp. 191–192. Hägerland calls this “the most precise description of the phenomenon” (p. 191). 34. M. GOODACRE, The Synoptic Problem: A Way through the Maze (The Biblical Seminar, 80), London, Sheffield Academic, 2001, p. 155. 35. Ibid. 36. Ibid.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

311

could not resist the temptation to paint the Matthaean lily, for it has got him into difficulties”37. The delegations seem artificial because they appear designed to enhance the picture of the centurion’s humility and to make the story conform to Luke’s account of the centurion Cornelius, who sent a delegation to Peter (Acts 10)38. The delegations in Luke 7 seem implausible and awkward because (1) they contradict each other (“the first group asked Jesus to come … and the second group ask him not to come”)39 and (2) the second delegation speaks for the centurion using “the first person singular”, which is appropriate in Matthew, where the centurion stands before Jesus, but is odd as the speech of the centurion’s representatives40. When Luke’s delegations are viewed as secondary, Q reconstructors tend to end up with a version of the Centurion’s Servant pericope that looks much like Matthew’s version, except for the logion about eschatological reversal (Matt 8,11-12//Luke 13,28-29), which most 2DH proponents assign to a later part of Q, in agreement with Luke’s placement41. FH proponents view this result of Q studies as further support for their hypothesis42. As for the eschatological reversal logion, however, FH proponents find no difficulty in viewing it as originally located at Matt 8,11-12 but transferred to a new location by Luke (13,28-29). Since Luke has introduced a delegation of Jewish elders into his version of the Centurion’s Servant pericope, it would hardly be appropriate, given this new Lukan audience, for Jesus to launch into a tirade about Jews who will be cast into the outer darkness, where there will be weeping and gnashing of teeth. This is a saying that will work better elsewhere, and in Luke’s new context, in 13:22–30, it effectively joins other material that focuses on judgment, the narrow door, and those who are shut out, where a tirade about hell is narratively and thematically appropriate43. 37. M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, p. 376. 38. Ibid., pp. 376-378. 39. Ibid., p. 378. 40. Ibid. 41. JOHNSON, Q 7:1-10 (n. 6), pp. 87-131, 240-259. 42. GOULDER, Luke (n. 37), pp. 379-380. 43. M. GOODACRE, Farrer Hypothesis Response, in PORTER – DYER (eds.), Synoptic Problem (n. 1), 127-138, pp. 135-136. However, Goulder (Luke [n. 37], p. 575) thinks that, in this pericope, Luke is not threatening hell, but only “exclusion from the Messianic banquet”. He also sees clumsiness in Luke’s editing of Matt 8,11-12, e.g., the doubling of the phrase “in the kingdom of God”, which “are … signs of Lucan secondariness” that support the FH. In his discussions of Luke’s rearrangement and redaction of Matthean material, Goulder alternates between accusing Luke of clumsiness and praising his literary skill. Goodacre, except for pointing out instances of editorial fatigue, tends to emphasize Luke’s artistry as a writer. See, e.g., GOODACRE, Case (n. 32), pp. 81-132.

312

R.K. MACEWEN

III. THE MATTHEAN POSTERIORITY HYPOTHESIS VS. THE FARRER HYPOTHESIS However, from the perspective of the Matthean Posteriority Hypothesis (MPH), the theory that Matthew used both Mark and Luke as his sources, some responses can be given to FH claims about the Centurion’s Servant pericope. First, the argument that Luke apparently incorporates Matthean redaction is not conclusive for the FH; there are also many examples in which Matthew appears to be aware of Lukan redaction44. Furthermore, the fact that Matthew uses an expression such as “there will be weeping and gnashing of teeth” more often than Luke does not necessitate that Luke received it from Matthew. Otherwise, Matthew’s multiple use of Markan expressions would prove that Matthew could not have depended on Mark. Examples of Matthew multiplying expressions also found in Mark include ἐν τῷ ἄρχοντι τῶν δαιμονίων ἐκβάλλει τὰ δαιμόνια (Matt 9,34; 12,24; Mark 3,22), μετανοεῖτε· ἤγγικεν γὰρ ἡ βασιλεία τῶν οὐρανῶν (Matt 3,2; 4,17; Mark 1,15 [Mark has τοῦ θεου rather than τῶν οὐρανῶν]), υἱὸς Δαυίδ (ten times in Matthew, three times in Mark), and ἀμὴν λέγω ὑμῖν (twentyfour times in Matthew, eleven times in Mark). Clearly, on any Markan priority view, such Matthean repetitions show that Matthew had the tendency to multiply expressions from his sources that he found congenial. Helping to confirm this Matthean tendency is an interesting feature of the four expressions mentioned above that are commonly adduced by FH proponents as Matthew’s characteristic expressions picked up once by Luke. In each case, it is Matthew’s first use of the expression that is parallel to Luke’s use of it. This phenomenon suggests that, in each case, Matthew first encountered the usage in Luke, found it congenial, and chose to use it where Luke does and again in other appropriate settings45. Regarding Goodacre’s argument for editorial fatigue from Luke’s agreement with Matthew in using the word παῖς in Luke 7,7 (par. Matt 8,8), this example has some puzzling features. First, if Luke is fatigued in v. 7, it seems that by the end of the pericope, in v. 10, he regains his vigor, and once again (as he did in v. 2) replaces Matthew’s (8,13) παῖς with his preferred δοῦλος. Second, immediately after declaring his faith that Jesus is able to heal 44. See MACEWEN, Matthean Posteriority (n. 25), pp. 27-50. 45. The above two paragraphs are adapted from R.K. MACEWEN, When Mark Goodacre Asked, ‘Why Not Matthew’s Use of Luke?’, in Logos Academic Blog, 14 December 2018, http:// academic.logos.com/when–mark–goodacre–asked–why–not–matthews–use–of–luke, review of a paper read at the Annual Meeting of the Society of Biblical Literature, Denver, 18 November 2018, Synoptic Gospels section.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

313

his servant (using παῖς) with a word (v. 7), in the very next verse the centurion uses the word δοῦλος in discussing his own authority. One might think that, if it were really Luke’s intention to avoid the word παῖς throughout the pericope, his use of δοῦλος in v. 8 would have alerted him to his mistake in v. 7, leading him to go back and correct it. Perhaps Luke did fatigue in v. 7; however, these puzzling features of his usage suggest that it is not certain that he did. It is more likely that Luke was alert and intentional in all his uses of “servant” words in 7,1-10. Luke distinguishes between the speech of his characters and his own voice as narrator. Thus, δοῦλος is Luke’s narratival designation for the servant in 7,2.10, but in v. 7, where the centurion is speaking, ὁ παῖς μου (“my boy”) is the centurion’s own affectionate term for the servant, “who was precious to him” (v. 2). In v. 8, where the centurion mentions τῷ δούλῳ μου, he is probably referring to a different servant, one who is presently healthy enough to obey orders46. If Matthew knew and used Luke, he chose to use παῖς consistently throughout the pericope to refer to the sick person (editorial alertness), reserving δοῦλος for the servant mentioned as an example of the centurion’s authority, whom he, like Luke, sees as a different person (Matt 8,9//Luke 7,8). In response to Goodacre’s claim that Matthew never fatigues while using DT material, Tobias Hägerland points to the preaching of John the Baptist in Matt 3,7-12//Luke 3,7-18. In Luke’s version, Luke speaks “to the crowds who went out to be baptized by him”. He calls them “offspring of snakes” and warns them of God’s judgment unless they “produce fruits worthy of repentance” (vv. 7-9). Instead of Luke’s crowds, Matthew substitutes a group of Pharisees and Sadducees to be the recipients of John’s forceful rebuke. But when Matthew gets to his v. 11 (par. Luke 3,16), where John says, “I baptize you in water … He will baptize you in the Holy Spirit and fire”, it seems that he has forgotten to change the audience back to the crowds. Given Matthew’s consistently negative portrayal of the Pharisees and Sadducees47, it is unlikely that he thinks they would have accepted (or been granted) John’s baptism or that they would have been candidates for 46. For another example of Luke varying his wording, depending on whether the narrator or a character in the story is speaking, see Luke 5,29-30. In v. 30, in agreement with Mark 2,16, Luke quotes the Pharisees and scribes as asking, “Why do you eat with tax collectors and sinners?” But before this, in v. 29, as the narrator he refers to Jesus’s table companions as “tax collectors and others”, even though Mark 2,15 at this point has “tax collectors and sinners”. Luke, as narrator, varies his wording here, probably to show that he sides with the compassionate stance of Jesus rather than the condemning stance of Jesus’s critics. 47. See Matt 9,11.34; 12,2.14.24.38-39; 15,1-14; 16,1-12; 19,3; 21,45-46; 22,15.23-35; 23,1-36; 27,62-64.

314

R.K. MACEWEN

Jesus’s baptism with the Holy Spirit. Thus, Matthew has fatigued in his use of DT material by changing the audience in 3,7, but then failing to notice that his new audience is no longer appropriate as John’s speech continues in vv. 11-1248. Another example of Matthew’s fatigue in using DT material can be seen in the Beelzebul pericopes. In Matt 9,32-34, Jesus heals a mute man by casting a demon out of him; the Pharisees accuse him of performing this act “by the ruler of the demons”. Much of Matthew’s wording here is identical to the wording in Luke 11,14-15, which for Luke is the beginning of a longer pericope with a parallel in Mark (Luke 11,14-23//Mark 3,22-27). In Luke’s and Mark’s versions, the ruler of the demons is also identified as Beelzebul, a name not found in Matt 9,34. In a later doublet, Matthew also has a version of this longer pericope (Matt 12,22-30), which incorporates wording found in both Luke’s and Mark’s versions, this time including the name Beelzebul. In between his two “ruler of the demons” pericopes, in his Mission Discourse, Matthew (10,25b) includes this unique saying: “If they called the master of the household Beelzebul, how much more the members of his household?” Matt 10,25b appears to refer back to an insult that he expects his readers already to be familiar with. Mostly likely, Matthew is referring to the accusation in 9,34. But this is a case of editorial fatigue, because, by the time he reaches 10,25b, Matthew has forgotten that, at 9,34, he did not take over the name Beelzebul that was present in his source for the DT49. 48. HÄGERLAND, Editorial Fatigue (n. 33), pp. 203-205. See also DAVIES – ALLISON, Matthew (n. 11), vol. 1, pp. 301-304, 312. 49. MACEWEN, Matthean Posteriority (n. 25), pp. 44-45, incl. n. 61. HÄGERLAND, Editorial Fatigue (n. 33), p. 203, rejects this as an example of editorial fatigue on the grounds (1) that it is “questionable as an example of fatigue as defined by Goodacre”, probably because Goodacre only discusses editorial fatigue within single pericopes; (2) that there is a difference between being called Beelzebul and being accused of casting out demons through Beelzebul; and (3) that “[i]t is not obvious that Matt 10.25 is meant to refer specifically to any incident narrated in the Gospel of Matthew”. In response to objection 3, since Matt 10,25 refers to something that at least Jesus’s immediate audience in the narrative (i.e., his disciples) apparently knows about, and since the unusual word Beelzebul is also found in Matt 12,24, Matthew most likely did intend a connection to his other Beelzebul pericope(s). This is similar to the way Jesus’s command to teach “them to obey all the things I commanded you” (Matt 28,20) refers back to all the paraenesis found earlier in Matthew’s Gospel. In response to objection 2, this is a distinction without a difference – in both cases the insult involves an association with Beelzebul. In response to objection 1, I would suggest that Goodacre’s definition be expanded for the sake of sound exegesis. If editorial fatigue can take place within the span of a few verses, there is no reason that it could not also take place across pericopes. Here are some further examples from Luke’s use of Mark: (1) In Luke 4,23 Jesus, in Nazareth, imagines that his hearers will say to him, “Do also here the things that we heard happened in Capernaum”. However, because Luke has placed his Rejection at Nazareth pericope at the

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

315

Concerning the supposed secondary nature of the two delegations in Luke 7,3-6, some reasons can be given for regarding Luke’s account as early. First, since on a Markan priority view it is clear that Matthew frequently abbreviates Mark’s pericopes, it is quite possible that Luke’s delegations were original and Matthew omitted them50. An analogy can be seen in the writings of Plutarch. In one of his biographies, Plutarch reports that Pompey interfered in the trial of a friend by sending a representative to the trial to read an encomium he had written (Cat. Min. 48,4). But in a different biography, Plutarch depicts Pompey as attending the trial and giving the encomium himself (Pomp. 55,5). In his second version of this incident, Plutarch simplified his (probably more historically accurate) first version by omitting the messenger51. Second, the sending of delegations is historically plausible; we have many examples in ancient literature. Elsewhere in the Gospels, in fact, we find another example of Gentiles contacting Jesus “through intermediaries” (John 12,20-22)52. Third, it does not seem impossible that an emissary would deliver a message using the first person, as is depicted in Luke 7,6-8, especially if the message were short enough to be easily memorized. Luke later portrays Jesus giving Pharisees a message for Herod Antipas, as if intending for it to be delivered using the first person: “Go say to this fox, ‘Look, I am casting out demons and effecting healings…’” (13,32). The Mishnah has the statement, “The one whom a person sends [šalı̂aḥ] is like the sender” (m. Ber. 5,5)53. One implication of this would seem to be that an ancient messenger was expected to transmit the sender’s words as faithfully as possibly. Fourth, the evidence from the LXX about messengers should be noted. About thirty-eight times the LXX uses either ἀπέστειλεν or ἐξαπέστειλεν beginning of Jesus’s ministry (unlike in Mark, where it comes later, at 6,1-6), Jesus has not yet been to Capernaum in Luke’s narrative – in fact, he goes there in Luke 4,31, after the Nazareth pericope. (2) In Luke 4,38-39, Jesus goes into Simon’s house and heals his motherin-law. If we had only Luke’s Gospel and were reading it for the first time, we would not know who this Simon is, because Luke has reversed Mark’s order. Luke’s version of the call of the first group of disciples, including Simon Peter, comes later, in 5,1-11. 50. GAGNON, The Shape of Matthew’s Q Text (n. 2), pp. 139-140, incl. n. 26, argues that Matthew would not have omitted characters as significant as Luke’s delegations, yet he also gives numerous examples in which Matthew does omit characters in Markan narratives. 51. M.R. LICONA, Why Are There Differences in the Gospels? What We Can Learn from Ancient Biography, New York, Oxford University Press, 2017, pp. 60, 130. 52. I.H. MARSHALL, The Gospel of Luke: A Commentary on the Greek Text (NIGTC), Exeter, Paternoster; Grand Rapids, MI, Eerdmans, 1978, p. 278. 53. Quoted by E.J. SCHNABEL, Apostle, in J.B. GREEN – J.K. BROWN – N. PERRIN (eds.), Dictionary of Jesus and the Gospels, Downers Grove, IL, InterVarsity Press, 22013, p. 35. The word within brackets is original. Schnabel cautions that one should not assume “that the term šalı̂aḥ was fixed legally and institutionally”.

316

R.K. MACEWEN

together with ἀγγέλους in describing someone sending messengers54. In some cases, the messengers are sent to perform an action rather than to convey a message (e.g., Josh 7,22; 1 Sam 19,11-21; 2 Sam 3,26; 11,4). In other cases, a message is given, but the words of the message are not quoted (e.g., 1 Kgs 17,4; 1 Chr 19,2.16; 1 Macc 7,10). But when the words of the message are given, it is common for the message to be expressed in the first person, as if the person who sent the message were speaking directly to the recipient. There are at least sixteen clear instances of this55. Here are a few examples: “And Saoul sent messengers to Iessai, saying, ‘Send me your son Dauid who is with your flock’” (1 Sam 16,19). “And I sent messengers to them, saying, ‘I am doing a great work, and I will not be able to come down, lest the work stop. Whenever I finish it, I will come down to you’” (Neh 6,3). “And Hezekias, king of Iouda, sent messengers to the Assyrians’ king to Lachis, saying, ‘I have done wrong; turn away from me; whatever you impose on me I will bear’” (2 Kgs 18,14a)56.

Sometimes there is a back-and-forth exchange between two individuals sending messengers to each other, presented almost as if the two individuals were speaking face to face (Num 20,14-20; Judg 11,12-27; 2 Sam 3,12-13; 1 Kgs 20,2-6; 2 Kgs 14,8-10). Luke is thoroughly steeped in the LXX57. Accordingly, it should not be regarded as a case of sloppy appropriation from a source (Q or Matthew) when Luke imitates Septuagintal style by depicting a group of messengers giving their message as a first-person quotation of the sender. The influence of the LXX on Luke’s delegations can also be seen in Luke’s vocabulary (ἀπέστειλεν πρὸς αὐτὸν [Luke 7,3] and λέγων [7,6]) and by the fact that the centurion sends groups rather than individuals to transmit his words. The influence of the LXX on Luke’s description of messengers can also be clearly seen in Luke 7,27 (οὗτός ἐστιν περὶ οὗ γέγραπται· ἰδοὺ ἀποστέλλω τὸν ἄγγελόν μου πρὸ προσώπου σου); 9,52 54. This does not include verses in which God sends an angel, where the singular ἄγγελον is used. 55. Gen 32,4-6; Num 20,14-20; Judg 11,12-27; 1 Sam 16,19; 2 Sam 2,5-7; 3,12-14; 12,27-28; 1 Kgs 20,2-6; 2 Kgs 14,8-10; 16,7; 18,14; 19,9-13; 2 Chr 35,21; Neh 6,3; Jdt 3,1-4; Isa 37,9-13. Verse numbers are those found in editions of the LXX. Note that 2 Sam 3,12-14 includes three separate first-person messages. 56. The above three quotations are from A. PIETERSMA – B.G. WRIGHT (eds.), A New English Translation of the Septuagint and the Other Greek Translations Traditionally Included under That Title, New York, Oxford University Press, 2007. Because my argument in this section is based on the LXX, I have opted to quote a Bible translation that is based on the LXX rather than the MT. 57. FITZMYER, Luke I–IX (n. 11), pp. 113-125.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

317

(Καὶ ἀπέστειλεν ἀγγέλους πρὸ προσώπου αὐτοῦ); and 10,1 (καὶ ἀπέστειλεν αὐτοὺς ἀνὰ δύο [δύο] πρὸ προσώπου αὐτοῦ)58. While Luke’s presentation of his delegations, including their use of first-person speech, may appear strange to modern readers, to Luke, and probably to his original readers as well, it was entirely appropriate. Fifth, regarding the contradictory messages of the two delegations, Robert H. Gundry comments: [C]oncerning the probabilities of the centurion’s action we can hardly make negative psychological judgments with any more certainty than we can make positive psychological judgments. The unlikelihood of the centurion’s sending two successive delegations is matched by the unlikelihood of an invention of such an awkward literary device59.

Sixth, even if the delegations were invented by Luke, this would not mean that Matthew could not have known and depended on Luke’s version of the Centurion’s Servant pericope. The MPH argues that Matthew derived his DT material from Luke. It is agnostic on the question of where Luke got his Sondergut and DT material. For any non-Markan pericope in Luke, we cannot know whether Luke received it from eyewitness testimony, oral tradition, or a written source, or if he invented it60. Also, the MPH does not entail that Luke has preserved his DT material in its purest, most original form. We cannot know what Luke did with the traditions he received (although his use of Mark can give us clues) – whether he incorporated them without changes or whether he redacted them, lightly, moderately, or radically. Concerning the eschatological reversal logion (Matt 8,11-12//Luke 13,28-29), FH proponents are clearly swimming against the tide of scholarship in arguing that Matthew’s placement of it was more original than Luke’s; most 2DH proponents are convinced that Luke preserves this logion’s original setting in Q61. As cited above, Goodacre claims that Luke’s addition of the delegation of Jewish elders led him to feel the need to move the eschatological reversal logion to a setting that would be more appropriate in his Gospel. This argument would have more weight if it could be 58. Cf. esp. LXX Exod 23,20; 2 Kgs 6,32; Mal 3,1. 59. R.H. GUNDRY, Matthew: A Commentary on His Handbook for a Mixed Church under Persecution, Grand Rapids, MI, Eerdmans, 21994, p. 147. 60. I do not think that Luke invented pericopes purely from his own imagination, but I mention this here as a theoretical possibility. 61. See, e.g., JOHNSON, Q 7:1-10 (n. 6), p. 352; DAVIES – ALLISON, Matthew (n. 11), vol. 2, pp. 25-26; D.A. HAGNER, Matthew 1–13 (WBC, 33A), Dallas, TX, Word, 1993, p. 205.

318

R.K. MACEWEN

shown that Luke does the same thing in his handling of Markan narratives. However, there do not seem to be any instances of Luke removing a saying from a Markan narrative, while retaining the Markan narrative in his Gospel, and placing the saying in a new context62. On the other hand, there are a number of examples of Matthew recontextualizing Markan sayings63, which would correspond to Matthew moving Luke’s eschatological reversal logion from Luke 13 and placing it in the Centurion’s Servant pericope. It is important to note the complexity of Luke’s procedures on the FH in constructing the new home he builds for the eschatological reversal logion. This new home is a pericope that Goulder identifies as Luke 13,22-35, having the topic “The Condemnation of Israel”64. According to Goulder, Luke 13,22 is an echo of Matt 9,35 (Jesus journeying and teaching in towns and villages). Luke 13,24 combines the “narrow gate” of Matt 7,13 with the “door” of Matt 25,10. Luke 13,25 takes the shut door, the appeal to be admitted, and the reply “I do not know you” from Matt 25,10-12 (the Parable of the Ten Virgins). Luke 13,26-27 draws from Matt 7,22-23, in which sinners give pleas as to why they should be admitted but are decisively rejected. Luke 13,28-29 is the eschatological reversal logion taken from Matt 8,11-12. Luke 13,30 uses the “the last will be first” saying from Matt 20,16. The threat from Herod in Luke 13,31 is developed on the basis of Matt 14,1-14 (the death of John the Baptist and Jesus’s subsequent withdrawal)65. Luke 13,32 is based on Matt 26,1-2 (“The theme of two days and then the Passion”)66. Luke 13,33 depends on Matt 23,34-35 (the killing of the prophets). The pericope then concludes with Luke 13,34-35, which is taken directly from Matt 23,37-39 (the lament for Jerusalem)67.

62. There are, however, a few examples of Luke using (apparently) Markan sayings while not retaining in his Gospel the narratives in which they are found in Mark. Cf. Mark 8,14-21 and Luke 12,1-2; Mark 10,2-12 and Luke 16,18. 63. E.g., Matthew removes the “sheep without a shepherd” saying from Mark’s Feeding of the Five Thousand pericope and uses it in the introduction to his Mission Discourse (Matt 9,36//Mark 6,34); he transfers a statement about forgiving and being forgiven from Mark’s Withering of the Fig Tree pericope to a section on prayer in the Sermon on the Mount (Matt 6,14-15//Mark 11,25); and he removes a saying about receiving the kingdom as a child from Mark’s Jesus Blesses the Children pericope and adds it, slightly modified, to his Community Discourse (Matt 18,3//Mark 10,15). For additional examples, see MACEWEN, Matthean Posteriority (n. 25), pp. 28-29. 64. GOULDER, Luke (n. 37), p. 571. 65. Ibid., pp. 572-576. 66. Ibid., p. 576. 67. Ibid., p. 577.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

319

Thus, to construct a pericope of fourteen verses, Goulder’s Luke ranges throughout Matthew 7–26, pulling together material from at least eight different Matthean contexts. This seems like a massive amount of work for rather meager results68. The problem is compounded by the fact that Goulder views Luke as carrying out similar procedures throughout the writing of his Gospel. Goulder makes it clear that he envisions Luke gathering his material by working his way back and forth through his scroll of Matthew69. Regarding Luke’s compositional results, Goulder acknowledges, “These texts do not always dovetail very neatly into each other, and … it is not surprising if the whole passage has that over-compressed, hammered look which we noticed in 6.39-49, and which is notorious in 16.14-19”70. A serious difficulty for Goulder’s description of Luke’s procedures in 13,22-35, and for the FH as a whole, is that they run counter to what we know about the way ancient writers normally used sources. Among the procedures of ancient source users noted by Robert A. Derrenbacker in his study of the works of several Greco-Roman authors is that they do not perform “a radical reordering of source material”71. Of course, since much of the DT material appears in different orders in Matthew and Luke, someone must have performed radical reordering. Who was it? Looking at how these two evangelists used Mark as a source can help. Luke rarely changes Mark’s order. The main exceptions are a few pericopes that are so different from Mark’s versions that it is likely that Luke took them from a source other than Mark72. Matthew, however, substantially changes the order of Markan pericopes in Matthew 8–13 (cf. Mark 1,29–6,13). He also removes several sayings from Markan narratives, inserting them into new contexts, as mentioned above. Since Matthew reorders Mark much more than Luke does, it is reasonable to assume that Matthew is responsible for reordering the DT as well.

68. For other examples of the peculiarity of Luke’s use of Matthew on the FH, see MACEWEN, Matthean Posteriority (n. 25), pp. 171-172. 69. GOULDER, Luke (n. 37), pp. 572, 577. Goulder, in most cases, envisages Luke as continually having visual contact with his scroll of Matthew when incorporating Matthew’s wording. In his discussion of Luke 13,22-35, the only exception is Luke’s reference to “exorcisms and healings” in v. 32, which Goulder views as “probably” due to Luke’s “subconscious” memory of Matt 7,23 (p. 576). 70. Ibid., p. 572. 71. DERRENBACKER, Conditions (n. 17), p. 441. Emphasis original. 72. E.g., the Call of the First Disciples (Mark 1,16-20//Luke 5,1-11); the Beelzebul Controversy (Mark 3,20-30//Luke 11,14-23); the Rejection at Nazareth (Mark 6,1-6// Luke 4,16-30); the Unfruitful Fig Tree (Luke 13,6-9; cf. the Cursing of the Fig Tree in Mark 11,12-14); and a Woman Anoints Jesus (Mark 14,3-9//Luke 7,36-50).

320

R.K. MACEWEN

If Matthew has radically reordered the DT material, he certainly engaged in complex procedures. But an explanation for why Matthew would do this is at hand. The bulk of Matthew’s DT material is found in his five major discourses (Matthew 5–7; 10; 13; 18; 24–25). The first discourse, the Sermon on the Mount, is built on a core collection of sayings that are parallel to Luke’s Sermon on the Plain. To this Lukan core, Matthew appears to have added additional material – from DT, from Matthew’s Sondergut, and some from Mark. Each of Matthew’s other four major discourses is built around a Markan core with additional material added, again from DT, Matthew’s Sondergut, and from other contexts in Mark. Matthew’s primary motive for rearranging sayings material seems clear – he wanted to create five topical discourses, each of which would bring together teachings of Jesus on a particular theme. If Matthew knew and used Luke’s version of the Centurion’s Servant pericope, his changes to Luke’s version are easy to explain and understand, given what we can observe of Matthew’s procedures in using Mark, on a Markan priority view. First, Matthew often shortens Markan pericopes by eliminating narrative details, even colorful and interesting ones, in order to come more quickly to the main point of the story. For example, in the Healing of the Paralytic (Matt 9,1-8//Mark 2,1-12//Luke 5,17-26), Matthew omits mentioning Jesus being in a house, Jesus’s teaching, the crowds flocking to him so that there was no room in front of the door, that four people carried the paralytic, and that they “unroofed the roof” and lowered him into the house on his pallet (Mark 2,1-4). Second, in addition to his preference for brevity, Matthew had another motive for omitting Luke’s delegation of Jewish elders – his unwillingness to portray (non-Christian) Jewish leaders in a positive light. Matthew gives a consistently negative portrayal of Jewish leaders throughout his Gospel, whether Pharisees, Sadducees, scribes, chief priests, or elders. To focus on the elders, Matthew describes them, together with other Jewish leaders, as enforcing traditions opposed to God’s law (15,1-6), causing Jesus to suffer (16,2), challenging Jesus’s authority (21,23), plotting to kill Jesus (26,3-5), sending an armed crowd to arrest Jesus (26,47), condemning Jesus at his trial at the high priest’s house (26,57), delivering Jesus to Pilate (27,1-2), turning deaf ears to Judas’s anguished cry of remorse (27,3-4), accusing Jesus before Pilate (27,12), persuading the crowd to call for Barabbas’s release and Jesus’s execution (27,20), mocking Jesus as he dies on the cross (27,41-43), and bribing the tomb guards to lie about the resurrection (28,11-15). Consistent with Matthew’s negative portrayal of the Jewish leadership, whenever Matthew retains in his Gospel a Jewish leader portrayed positively

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

321

in Mark, he either changes him into a negative character or obscures his status as a Jewish leader. Thus, the scribe (εἷς τῶν γραμματέων) in Mark 12,28-34 who has a friendly exchange with Jesus about the great commandment is transformed by Matthew (22,34-35) into a lawyer (νομικός) sent by the Pharisees to test Jesus73. Jairus, a synagogue ruler (εἷς τῶν ἀρχισυναγώγων [Mark 5,22-23]) who approaches Jesus with faith and humility is changed by Matthew to a generic “ruler” (ἄρχων εἷς [Matt 9,18]). And Joseph of Arimathea, who piously buries Jesus after his death, Matthew changes from a member of the Sanhedrin to “a rich man” (Matt 27,57// Mark 15,43). It is also helpful to look at Matthew’s use of the word “synagogue” (συναγωγή). Matthew appears to view synagogues negatively, as the local power centers of Jewish leaders who opposed Jesus (Matt 12,9-14; 13,53-58), and probably also of Jewish leaders who opposed Matthew and his church (10,17; 23,34). Matthew’s alienation from the synagogues seems to be reflected by his frequently referring to them as “their synagogues” (4,23; 9,35; 10,17; 12,9; 13,54; 23,34 [“your synagogues”])74. Therefore, Matthew would likely not have seen it as a point in favor of the centurion of Capernaum that he built a synagogue (Luke 7,5); it is not surprising that does not mention this75. Although, as we have seen, Matthew often abbreviates his narratives in order to focus on the main point or to remove material uncongenial to him, Matthew is also known to expand the length of some pericopes. He does this in order to add material that is theologically rich or is otherwise important to him. Many examples could be given. Matthew lengthens Mark’s account of the Baptism of Jesus by adding a dialogue showing that John recognized Jesus’s superiority and had to be persuaded to baptize him (Matt 3,13-17// Mark 1,9-11). Matthew enhances the profile of Peter and teaches a lesson on faith by adding the story of Peter walking on the water with Jesus (Matt 14,22-33//Mark 6,45-52). Matthew further enhances Peter’s profile with his addition of Jesus’s lengthy blessing following Peter’s confession (Matt 16,13-20//Mark 8,27-30). In the Death of Jesus pericope, in addition to Mark’s darkness at noon and the splitting of the temple curtain, Matthew adds other miraculous portents – an earthquake, the opening of tombs, and 73. On the MPH, Matthew’s changing Mark’s friendly γραμματεύς into a hostile νομικός could have been influenced by Luke 10,25. 74. Matthew mentions synagogues only three times without using a possessive pronoun (6,2.5; 23,6). In the other Gospels the expression “their synagogues” is found only in Mark 1,23.39; Luke 4,15. 75. GUNDRY, Matthew (n. 59), p. 141.

322

R.K. MACEWEN

the resurrection of saints (Matt 27,45-54//Mark 15,33-39). To Mark’s empty tomb narrative, Matthew adds another earthquake, the angel of the Lord rolling back the stone, Jesus’s appearance to the women, and the apologetically significant account of the guard at the tomb (Matt 27,62–28,15// Mark 16,1-8). At times, Matthew expands his Markan pericopes with DT material. To the beginning of his Stilling the Storm pericope (Matt 18,18.23-27// Mark 4,35-41//Luke 8,22-25) Matthew adds Jesus’s challenges to would-be disciples, found in another context in Luke (Matt 8,19-22//Luke 9,57-62), probably in order to relate the storm pericope to the difficulties of discipleship. To the story Healing the Withered Hand (Matt 12,9-14//Mark 3,1-6// Luke 6,6-11) Matthew adds a saying about pulling an animal out of a pit, which is parallel to Luke 14,5, found in one of Luke’s Sondergut Sabbath healing stories. And to the pericope Defilement – Traditional and Real (Matt 15,1-20//Mark 7,1-23) Matthew adds a paragraph criticizing the Pharisees, which includes a saying about the blind leading the blind that is also found in Luke’s Sermon on the Plain (Matt 15,14//Luke 6,39). So, we see that, on the MPH, Matthew added Lukan sayings to a number of his Markan narrative pericopes. That Matthew added sayings to narratives seems more likely than the reverse on the FH: that Luke intentionally removed sayings that he found in Matthew’s versions of Markan narratives. Luke’s removal and recontextualization of such sayings is unlikely because there does not seem to be any reason why Luke would regularly (1) be pleased with the Markan narrative; (2) be pleased with the saying that Matthew embedded in the narrative; (3) yet be displeased with Matthew’s narrative context for the saying; and (4) feel compelled to relocate the saying to a different context. Since Matthew added Lukan sayings to Markan narrative pericopes, it would not be surprising for Matthew also to have added a Lukan saying from a different context to a Lukan narrative pericope, as he appears to have done to the story of the Centurion’s Servant by including the eschatological reversal logion. The constitutive elements of the eschatological reversal logion appear in different orders in Matthew and Luke (See Table 2 below). If one evangelist used the other’s Gospel as his source, then one version is original and the other is secondary.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

323

Table 2. Synopsis of the Eschatological Reversal Logion76 Matt 8,11-12

Luke 13,28-29

A Λέγω δὲ ὑμῖν ὅτι B πολλοὶ ἀπὸ ἀνατολῶν καὶ δυσμῶν ἥξουσιν καὶ ἀνακλιθήσονται C μετὰ Ἀβραὰμ καὶ Ἰσαὰκ καὶ Ἰακὼβ D ἐν τῇ βασιλείᾳ τῶν οὐρανῶν, E 12 οἱ δὲ υἱοὶ τῆς βασιλείας ἐκβληθήσονται εἰς τὸ σκότος τὸ ἐξώτερον· F ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων.

F 28 ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων, C ὅταν ὄψησθε Ἀβραὰμ καὶ Ἰσαὰκ καὶ Ἰακὼβ καὶ πάντας τοὺς προφήτας D1 ἐν τῇ βασιλείᾳ τοῦ θεοῦ, E ὑμᾶς δὲ ἐκβαλλομένους ἔξω. B 29 καὶ ἥξουσιν ἀπὸ ἀνατολῶν καὶ δυσμῶν καὶ ἀπὸ βορρᾶ καὶ νότου καὶ ἀνακλιθήσονται D2 ἐν τῇ βασιλείᾳ τοῦ θεοῦ.

11

Which version is more original, Matthew’s, with the ABCDEF order, or Luke’s, with the FCD1EBD2 order? We should keep in mind that Matthew’s use of Mark shows that he is not averse to changing the order of Jesus’s sayings. In Jesus’s response to the Pharisees’ challenge about hand washing, in Mark 7,6-13 he first quotes Isa 29,13 and then charges his opponents with breaking God’s commandment to honor their parents, but in Matt 15,3-9 the order of these two elements of Jesus’s reply is reversed. Similarly, in Jesus’s response to the Pharisees’ question about divorce, in Mark 10,3-9 Jesus first discusses Moses’s permission to divorce and then discusses God’s original plan at creation, but in Matt 19,4-8 the order of these two elements is reversed. If it was Matthew who used Luke as his source for the eschatological reversal logion, his procedures are understandable. Matthew introduces the logion with the phrase, Λέγω δὲ ὑμῖν ὅτι (A). This is a repetition of ἀμὴν λέγω ὑμῖν in the previous verse (Matt 8,10), indicating that, in Matthew’s context, the eschatological reversal logion will strengthen the contrast just made between the faith of the gentile centurion and the faith of many “in Israel”. Matthew’s rearrangement of Luke’s phrases simplifies the structure of the logion – instead of Luke’s negative-positive-negative-positive arrangement, Matthew puts all the positive (reward-oriented) phrases in the beginning (B-D) and all the negative (judgement-oriented) phrases at the end (E-F).

76. In this synopsis, identical wording common to the two Gospels is in bold type and similar wording is underlined. The English letters mark the two Gospels’ corresponding phrases.

324

R.K. MACEWEN

Matthew finds Luke’s “north and south” unnecessary; simply “east and west” is sufficient to make the point of people coming from everywhere (phrase B). Likewise, he finds Luke’s “and all the prophets” superfluous; just the names of the three major patriarchs are enough to indicate the honorary hosts of God’s eschatological banquet (phrase C). Matthew’s simplifying rearrangement of the logion means that he has no need to follow Luke in doubling the phrase, “in the kingdom…” (phrase D). Matthew’s change of Luke’s “the kingdom of God” to “the kingdom of heaven”, whatever the reason for it, is the same change Matthew makes many times in his redaction of Mark. Instead of Luke’s “you”, Matthew describes those who are cast out as “the sons of the kingdom” (phrase E). Elsewhere, in a Sondergut pericope (Matt 13,38), Matthew uses this phrase in a completely positive sense (and also in an eschatological context, as in 8,12). Here in the eschatological reversal logion, Matthew uses the phrase ironically – it is those who seem to be sons of the kingdom, but who really are not, who will be cast out. By using “sons of the kingdom” rather than Luke’s pointed “you”, Jesus, in Matthew’s version of the Centurion’s Servant pericope, need not be seen as condemning the whole Jewish crowd that is following him. Instead, it makes the saying a kind of “If the shoe fits, wear it” statement. By replacing Luke’s ἔξω with the phrase εἰς τὸ σκότος τὸ ἐξώτερον (phrase E), Matthew uses one of his characteristic phrases (cf. 22,13; 25,30) to enhance the eschatological force of the logion. Finally, Matthew uses Luke’s opening phrase (F) as the conclusion of the logion, seeing the picture of “weeping and gnashing of teeth” as a fitting conclusion to a stern warning to Israel. Matthew will use this phrase five more times in his Gospel. His intentionality in making the phrase the conclusion of his version of the eschatological reversal logion is confirmed by the way he uses it elsewhere. Everywhere Matthew uses the phrase it is either the last (13,42.50; 24,51; 25,30) or next to last (22,13) expression in a paragraph describing eschatological punishment. Matthew’s dependence on Luke is more likely than Luke’s dependence on Matthew for the eschatological reversal logion. Much of the wording unique to Matthew’s version of the logion consists of Matthew’s characteristic vocabulary and phrases. Matthew’s simpler structure for the logion makes Matthew’s version, rather than Luke’s, look like the more edited version77. As with Matthew’s procedure in editing Markan pericopes, here also 77. GOULDER, Luke (n. 37), p. 575, argues that “clumsy” features “and the weak expansions, are all signs of Lucan secondariness” in Luke 13,28-29. I would argue, rather, that

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

325

we can see Matthew both shortening by removing material that he sees as redundant or unnecessary and lengthening by adding theologically rich material. Because Matthew both adds to and subtracts from Luke’s version of the logion, his version ends up being only slightly shorter than Luke’s78. Finally, there does not seem to be anything in Luke’s version of the logion for which Luke would have needed to depend on Matthew. It reflects Luke’s well-known “reversal of fortunes” motif, seen throughout his Gospel79. Luke’s Sondergut parable of the Rich Man and Lazarus (16,19-31), in particular, corresponds to Luke’s eschatological reversal logion. Like the logion, this parable includes the themes of a reversal of fortunes in the afterlife, fellowship with Abraham for the one rewarded, agonizing suffering for the one punished, and even a mention of the prophets. Earlier in this essay, when discussing the 2DH, I argued that the large amount of verbatim agreement between Matthew and Luke in the DT makes a direct utilization hypothesis more likely than the two Gospels’ independent use of Q. But can we tell whether Luke used Matthew or Matthew used Luke? A statement by Goodacre suggests a testing method: “We have access to Mark, so we have an idea what Matthew and Luke look like when they are working from a shared source. We know the degree of verbatim agreement to expect”80. Figure 2 shows how Luke would have used his sources on the FH. The left column shows the total number of words in all Mark-Luke SVAs*4 as Figure 2. Luke's Use of His Sources on the Farrer Hypothesis 25% 20% 15% 10% 5% 0%

Agreement with Mark

Agreement with Matthew (in DT material)

Matthew’s smoother version supports Matthew’s secondariness, in having used and improved on Luke. 78. Matt 8,11-12 has forty-three words; Luke 13,28-29 has forty-seven words. 79. J.A. FITZMYER, The Gospel according to Luke X–XXIV (AB, 28A), Garden City, NY, Doubleday, 1985, p. 1126. 80. GOODACRE, Too Good to Be Q (n. 24), p. 90.

326

R.K. MACEWEN

a percentage of the total number of words in Markan pericopes in Luke. The right column shows the total number of words in Matthew-Luke DT SVAs*4 as a percentage of the total number of words in the Matthew-Luke DT pericopes in Luke. Clearly, on the FH, Luke is not consistent in his use of his two sources, Mark and Matthew. In his Markan pericope only 14.8% of Luke’s wording consists of SVAs*4 in agreement with Mark, while in his DT pericope, 20.8% of Luke’s wording consists of SVAs*4 in agreement with Matthew. One reason this discrepancy is surprising is that Luke generally agrees with Mark’s order, but often disagrees with Matthew’s order in the MatthewLuke DT material. Yet, in terms of ancient compositional practices, it is easier to copy the wording of a source whose order one is following than to copy the wording of a source whose order one is not following81. The assumption of Luke’s dependence on Mark supports this – Luke displays the greatest divergence from Mark’s wording and content in pericopes he places out of order relative to Mark’s outline82. So the general pattern of Luke’s strong verbatim agreement with Matthew in DT material (where he often disagrees with Matthew’s order) and weaker verbatim agreement with Mark in TT material (where he usually agrees with Mark’s order) is the opposite of what we should expect if Luke used both of the other Gospels as sources. On the MPH, Matthew is a much more consistent source user (See Figure 3). In Matthew’s Markan pericopes 22.7% of the wording consists of SVAs*4 in agreement with Mark, while in Matthew’s DT pericopes 21.0% of the wording consists of SVAs*4 in agreement with Luke. In Figure 3, the left column shows the total number of words in all Matthew-Mark SVAs*4 as a percentage of the total number of words in Markan pericopes in Matthew. The right column shows the total number of words in Matthew-Luke DT SVAs*4 as a percentage of the total number of words in the Matthew-Luke DT pericopes in Matthew. To summarize the results of my statistical study, in general, in Markan pericopes, we see high agreement between Matthew and Mark and considerably lower agreement between Mark and Luke and between Matthew and

81. DERRENBACKER, Conditions (n. 17), p. 441. 82. Cf., e.g., the Call of the First Disciples (Mark 1,16-20//Luke 5,1-11); the Beelzebul Controversy (Mark 3,20-30//Luke 11,14-23); the Rejection at Nazareth (Mark 6,1-6// Luke 4,16-30); the Unfruitful Fig Tree (Luke 13,6-9; cf. the Cursing of the Fig Tree in Mark 11,12-14); and a Woman Anoints Jesus (Mark 14,3-9//Luke 7,36-50).

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

327

Figure 3. Matthew's Use of His Sources on the Matthean Posteriority Hypothesis 25% 20% 15% 10% 5% 0%

Agreement with Mark

Agreement with Luke (in DT material)

Luke83. In the Matthew-Luke DT, we normally see high agreement between Matthew and Luke. Note that Matthew is a factor wherever the agreement is highest. The simplest explanation for this is to credit one writer – the evangelist Matthew – as having caused all of this high agreement. As Kloppenborg tells us, ancient writers normally did not copy from their sources verbatim at length84. So, on the evidence of the Gospels, at least one of the evangelists has behaved abnormally. On the FH, Matthew behaved abnormally by closely copying Mark, and Luke behaved abnormally by closely copying Matthew (but not so closely copying Mark). On the 2DH, Luke behaved abnormally by closely copying Q (but not so closely copying Mark), and Matthew behaved abnormally by closely copying both Q and Mark. On the MPH, there is only one evangelist, Matthew, who must be seen as behaving abnormally, though self-consistently, in using sources85. Matthew would have copied Mark closely in Markan pericopes and have copied Luke closely in non-Markan pericopes that he took over from Luke. Also on the MPH, Luke can be seen as a more normal type of ancient writer in his use of Mark – mostly paraphrasing rather than copying word-for-word at length. As for Luke’s non-Markan pericopes (both his DT and L material), 83. Exceptions to the rule of low Matthew-Luke agreement in Markan pericopes are seen in the Minor Agreements and in the so-called Mark-Q overlaps. These cases are rightly seen as problems for the theory of Matthew and Luke’s independence from each other. See GOODACRE, Synoptic Problem (n. 34), pp. 144-151; MACEWEN, Matthean Posteriority (n. 25), pp. 61-62. 84. KLOPPENBORG, Variation (n. 22), pp. 73-74. 85. As Kirk observes, “Matthew’s patterns of utilization are the same for both the double and triple tradition”. See A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, p. 309.

328

R.K. MACEWEN

Luke’s sources for them are unknown; but on the MPH we are free to assume that Luke used all his sources similarly to the way he used Mark. On the 2DH, however, our understanding of Luke’s compositional procedures would have to be quite different. The construction of Luke’s Gospel alternates between blocks consisting mainly of Markan material and blocks consisting mainly of non-Markan material86. In the Markan blocks, Luke appears to behave as a normal ancient source user – he employs one source at a time and follows that source’s order. But in the non-Markan blocks, Luke makes numerous quick jumps back and forth between Q and the source of his Sondergut, which scholars commonly refer to as “L”. For example, in chapters 13–14 (containing 70 verses) Luke makes thirteen shifts between Q and L (or Luke’s redaction), with some of the individual blocks of Q or L being only one, two, or three verses long87. Such rapid shifting between sources would have been both very challenging and abnormal for an ancient author88. On the MPH, Matthew’s compositional procedures must certainly be seen as complex, and thus unusual for an ancient author, particularly in Matthew’s major discourses, where he brings together material from at least three sources – Mark, Luke, and the source of his Sondergut. However, the 2DH faces exactly the same problem of Matthew’s compositional complexity, except that it substitutes Q for Luke89. Nevertheless, the MPH is a stronger theory than the 2DH. This is because the MPH requires only one 86. This can be seen graphically in A. BARR, A Diagram of Synoptic Relationships, Edinburgh, T&T Clark, 21995. 87. Luke 13,1-17 (L); 13,18-21 (Q); 13,22-23 (L or redaction); 13,24-30 (Q); 13,31-33 (L); 13,34-35 (Q); 14,1-10 (L); 14,11 (Q); 14,12-15 (L); 14,16-23 (Q); 14,24-25 (L or redaction); 14,26-27 (Q); 14,28-33 (L); 14,34-35 (Q). Note that this situation would be similar on the FH – Luke would be rapidly moving back and forth between Matthew and L. For the assignment of verses to Q, see CritEd (n. 5), pp. 400-460. For discussions of whether certain verses are Lukan redaction, see FITZMYER, Luke X–XXIV (n. 79), pp. 1021, 1052, 1060. For another example of Luke’s quick transitions between Q and L on the 2DH, see GARROW, Streeter’s ‘Other’ (n. 24), p. 217. 88. MACEWEN, Matthean Posteriority (n. 25), pp. 174-175. 89. Derrenbacker explains how Matthew could have achieved his conflation of sources by proposing that Matthew possessed Q, and perhaps also Mark, in codex form. If Matthew had codices of his sources, he would have had “random access” to their contents, rather than the mere “sequential access” provided by scrolls. This would have enabled Matthew to achieve both his frequent reordering of his sources’ pericopes and sayings and his occasional conflation of his sources’ wording. See R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005, pp. 253-254; ID., Conditions (n. 17), p. 444. It should be noted that a similar explanation would work equally well for Matthew’s compositional procedures on the MPH – i.e., that Matthew possessed copies of Luke and Mark in codex form. See GARROW, Streeter’s ‘Other’ (n. 24), p. 218.

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

329

evangelist to engage in complex compositional procedures, requires only one evangelist to be responsible for all the highest verbatim agreement among the Gospels, and provides the best explanation for the discrepancy in Matthew-Luke verbatim agreement between the TT and the DT90. IV. CONCLUSION The arguments of 2DH proponents that the phenomena of the Centurion’s Servant pericope support their theory have been found wanting. The claim that the pericope of the Healing of the Royal Official’s Son (John 4,46b-54) supports the presence of the Centurion’s Servant in Q is unconvincing because multiple explanations for the similarity of these two pericopes are possible. The argument from the implausibility of Luke removing the eschatological reversal logion from Matthew’s context and recontextualizing it works against the FH but not against the MPH. The argument from alternating primitivity between Matthew’s and Luke’s versions of the Centurion’s Servant is vitiated by the alternating use of Matthew’s and Luke’s characteristic vocabulary in the reconstructed Q version of this pericope. Furthermore, since the Centurion’s Servant is a healing narrative, it does not well fit the standard description of Q as a sayings collection. The arguments of FH proponents that the phenomena of the Centurion’s Servant pericope support their theory have also been found wanting. Luke’s one-time use of some of Matthew’s characteristic expressions does not prove that Luke used Matthew any more than Mark’s less frequent use of expressions used more often by Matthew prove that Mark depended on Matthew. The argument from editorial fatigue fails because there are reasons to think that Luke’s use of παῖς in 7,7 was intentional and because there are some good examples of Matthew fatiguing elsewhere in the DT. The argument that Luke’s delegations look like artificial additions to Matthew’s version of the pericope does not succeed because they cohere with Luke’s literary style (and thus could have been part of an original, Lukan version of the Centurion’s Servant) and because Matthew’s omission of them would cohere with his known redactional behavior. 90. The material in this essay on verbatim agreement in the Gospels is taken from R.K. MACEWEN, An Evaluation of Common Arguments for the Independence of Matthew and Luke: Do They Hold Up against the Matthean Posteriority Hypothesis?, Paper read at the Annual Meeting of the Society of Biblical Literature, San Antonio, TX, 21 November 2021, Interrelations of the Gospels section, which is based on ID., Matthean Posteriority (n. 25), pp. 50-73.

330

R.K. MACEWEN

The MPH is the best source-critical explanation for the phenomena of the Centurion’s Servant pericope, both because of the weaknesses of the arguments for the alternative theories and because Matthew’s editing of Luke’s Centurion’s Servant conforms to his procedures in editing Mark. In his use of Mark, Matthew abbreviates, deleting unnecessary and uncongenial material, and he also expands, adding theologically rich material, which may include sayings material taken from other contexts in Luke. Finally, the patterns of verbatim agreement seen in the Centurion’s Servant pericope, and throughout the Synoptic Gospels, are most easily and economically explained by the MPH, which sees Matthew’s source utilization procedures as responsible for all the closest agreement. Appendix: Matthean and Lukan Wording in the Centurion’s Servant Pericope Words are phrases in the following list are divided into four categories: clearly Matthean, marginally Matthean, clearly Lukan, and marginally Lukan. I have attempted to include all words and phrases from Matt 7,28a; 8,1-5; Luke 7,1-10; 13,28-29 that fit one of these categories. For the purposes of this study, words are considered Matthean or Lukan simply in terms of comparison of usage between these two Gospels, regardless of frequency elsewhere in the NT, except that Acts is sometimes taken into consideration in evaluating whether a word or phrase is to be considered Lukan. The point is to be able to judge whether Matthew or Luke is more likely to use a given word or phrase. A word or phrase is categorized as clearly Matthean if Matthew uses it one-third or greater times more than Luke uses it, or if only Matthew uses it. A word or phrase is categorized as marginally Matthean if Matthew uses it more frequently than Luke uses it, but less than one-third more frequently. A word or phrase is categorized as clearly Lukan if Luke uses it one-third or greater times more than Matthew uses it, or if only Luke uses it. A word or phrase is categorized as marginally Lukan if Luke uses it more frequently than Matthew uses it, but less than one-third more frequently, or if the word is more frequent in Acts than it is in Matthew. I did not consider it necessary or helpful to categorize words based on usage in triple tradition, double tradition, or Sondergut; the evangelists chose to use certain words, whether or not they found them in their sources. I have not included words for which usage frequencies between Matthew and Luke do not seem significantly different. Admittedly, I have had to make some subjective judgments in including or not including words and in categorizing words. The statistics following each word or phrase give frequency numbers for all the NT narrative books in the order Matthew/Mark/Luke/John+Acts. Although word usage statistics for Mark and John are not necessary for this study, they are included to allow comparison with other NT narrative literature. When the lexical form of a word is given, the following statistics cover the occurrences of all inflected forms of that word. When an inflected form of a word is given, the following statistics

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

331

cover only that specific form of the word. In each section, words and phrases are given in alphabetical order, excluding articles. Clearly Matthean ἀκολουθέω 25/18/17/19+4 ἄλλος 29/22/11/33+8 ἀμήν 31/14/6/50+0 ἀμὴν λέγω ὑμῖν (incl. γὰρ and δὲ) 29/12/5/20+0 ἀνατολή 5/1/2/0+0 ἀπὸ ἀνατολῶν 3/0/1/0+0 βάλλω 34/18/18/17+5 βασανίζω 3/2/1/0+0 ἡ βασιλεία τῶν οὐρανῶν 32/0/1/0+0 γενηθήτω σοι 2/0/0/0+0 εἰς τὸ σκότος τὸ ἐξώτερον 3/0/0/0+0 ἑκατόνταρχος (second declension) 3/0/0/0+1 ἐκεῖ ἔσται ὁ κλαυθμὸς καὶ ὁ βρυγμὸς τῶν ὀδόντων 6/0/1/0+0 ἐν ἐκείνῃ τῇ ὥρᾳ or ἐν τῇ ὥρᾳ ἐκείνῃ 4/1/1/1/+1 ἐν τῇ βασιλείᾳ 10/1/6/0+0 Ἰησοῦς 152/82/88/244+69 ὁ δὲ Ἰησοῦς 17/17/6/4+0 καὶ ἐγένετο ὅτε ἐτέλεσεν ὁ Ἰησοῦς 5/0/0/0+0 καὶ λέγει αὐτῷ 7/8/0/6+1 κακῶς 7/4/2/1+1 λέγει αὐτῷ 19/12/4/44+1 μόνος 14/6/10/15+8 οὐδέ 26/10/19/16+11 οὐρανός 82/18/35/18+26 ὁ παῖς μου 3/0/1/0+0 παραλυτικός 5/5/0/0+0 πολλοί 16/13/8/15+8 προσέρχομαι 51/5/10/1+10 σκότος 7/1/4/1+3 τελευτάω 4/2/1/1+2 οἱ υἱοὶ τῆς βασιλείας 2/0/0/0+0 ὑπάγω 19/15/5/32+0 ὕπαγε 11/8/0/4+0 Marginally Matthean βασιλεία 55/20/46/5+8 ἐκβάλλω 28/18/20/6+5 κύριε 33/1/26/33+15 λέγοντες 47/15/36/10+23 ὄχλος 50/38/41/20+22

332

R.K. MACEWEN

Clearly Lukan ἀγαπάω 8/5/13/37+0 ἀξιόω 0/0/1/0+2 ἡ βασιλεία τοῦ θεοῦ 5/14/32/2/+6 διό 1/0/2/0+8 ἐγένετο 13/18/69/17+54 ἐγὼ ἄνθρωπός εἰμι 1/0/2/0+2 ἔθνος for the Jewish people or Israel 0/0/2/3+4 εἰς τὰς ἀκοὰς 0/0/1/0+1 εἰς τὸν οἶκον 4/5/15/1+5 (cf. εἰς τὴν οἰκίαν 8/3/5/0+1) εἰσῆλθεν εἰς 2/3/5/3+5 ἑκατοντάρχης (first declension) 1/0/3/0+12 ἐκβάλλω + ἔξω 1/1/3/4+2 ἐμαυτοῦ 1/0/2/16+4 ἐν τῷ Ἰσραὴλ 2/0/4/0+0 ἔντιμος 0/0/2/0+0 ἐπειδή 0/0/2/0+3 ἐπλήρωσεν 0/0/1/0+3 ἐρωτάω 4/3/15/28+7 εὑρίσκω 27/11/45/19+35 εὗρον 4/3/14/0+12 ἤμελλεν 0/0/4/4+3 θαυμάζω 7/4/13/6+5 ἰάομαι 4/1/11/3+4 ἱκανός 3/3/9/0+18 καὶ ἀποκριθεὶς 6/7/13/0+0 καὶ γάρ 3/2/6/1+1 καὶ ἐγένετο 7/7/29/0+6 καὶ εἶπεν 9/15/49/39+7 λαός 14/2/36/3+48 μακράν 1/1/4/1+3 μέλλω 9/2/12/12+34 νότος 1/0/3/0+2 οἰκοδομέω 8/4/12/1+4 οἶκος 10/13/33/5+25 ὃς ἦν 0/0/2/3+3 ὅταν 19/20/29/17+2 οὐδείς 19/26/35/53+28 οὗτος 147/79/229/239+236 ταῦτα 22/17/47/61+28 τούτῳ 7/4/12/13+17 πάντα τὰ ῥήματα 0/0/3/0+1 παραγίνομαι 3/1/8/2+20 παραγενόμενοι 0/0/2/0+4 παρέχω 1/1/4/0+5 πέμπω 4/1/10/32+11

THE CENTURION’S SERVANT AND THE SYNOPTIC PROBLEM

333

περὶ τοῦ Ἰησοῦ 0/1/2/0+3 πορεύομαι 29/3/51/16+37 πρεσβύτεροι τῶν Ἰουδαίων 0/0/1/0+1 πρός 42/65/166/102+133 πρὸς τὸν Ἰησοῦν 1/4/5/2+0 ῥῆμα 5/2/19/12+14 στραφεὶς 2/0/7/1+0 σύν 4/6/23/3+51 σὺν αὐτοῖς 0/0/3/0+7 συναγωγή 9/8/15/2+19 τίς 112/106/195/136+171 τινος 3/4/10/2+8 ὑγιαίνω 0/0/3/0+0 ὑποστρέφω 0/0/21/0+11 φίλος 1/0/15/6+3 Marginally Lukan διασῴζω 1/0/1/0+5 εἰσέρχομαι 36/30/50/15+34 λέγω ὑμῖν or ὑμῖν λέγω without ἀμὴν 28/4/37/7+1 πίστις 8/5/11/0+15 τάσσω 1/0/1/0+4 ὡς 40/22/51/31+63

Tyndale Theological Seminary Badhoevedorp, The Netherlands

Robert K. MACEWEN [email protected]

LUKE AND THE WAX TABLET REVISITED: AN ASSESSMENT OF SUPPOSED DIFFICULTIES

In 2012, I published an article with two objectives: (1) In response to those who had ridiculed Michael Goulder’s suggestion that Luke moved backward through a scroll of Matthew, I visited the physicality of scroll reading to show the reasonableness of Goulder’s belief that an author might move backward through a source scroll, and I invoked Annewies van den Hoek’s demonstration that Clement of Alexandria appears to have done exactly that1. I intended my comparison between Goulder’s Luke moving backward through Matthew and Clement’s moving backward through Philo’s De Posteritate (in Strom. 2,5,3–6,4) strictly to show that the oftrepeated response of Two-Document hypothesis (2DH) proponents to Goulder’s scenario is not as well grounded in ancient compositional practices as they allege. (2) In the end, I rejected Goulder’s scenario on the grounds that Luke’s movement doesn’t appear as linear as Goulder imagines. I therefore also advanced some positive arguments, one of which has to do with Luke’s likely use of wax tablets as a compositional aid. We know that wax tablets were the preferred medium for composing texts in Luke’s day2, so it makes sense a priori to consider whether the use of tablets can help explain 1. J.C. POIRIER, The Roll, the Codex, the Wax Tablet and the Synoptic Problem, in JSNT 35 (2012) 3-30. Cf. M. GOULDER, The Order of a Crank, in C.M. TUCKETT (ed.), Synoptic Studies: The Ampleforth Conferences of 1982 and 1983 (SupplJSNT, 7), Sheffield, JSOT, 1984, 111-130, pp. 121-130; M. GOULDER, Michael Goulder Responds, in C.A. ROLLSTON (ed.), The Gospels According to Michael Goulder: A North American Response, Harrisburg, PA, Trinity Press International, 2002, 137-152, p. 146. See J.L. VAN DEN HOEK, Clement of Alexandria and his Use of Philo in the Stromateis: An Early Christian Reshaping of a Jewish Model (SupplVC, 3), Leiden – New York –København – Köln, Brill, 1988, p. 152. 2. A. BÜLOW-JACOBSEN, Writing Materials in the Ancient World, in R.S. BAGNALL (ed.), The Oxford Handbook of Papyrology, Oxford, Oxford University Press, 2009, 3-29, pp. 11-12, notes: “Wax tablets were clearly the everyday notebook for bookkeeping, business correspondence, and literary drafts”. Cf. Ovid, Metam. 9,522-570; R.H. ROUSE – M.A. ROUSE, Wax Tablets, in Language & Communication 9 (1989) 175-191. See the discussion of wax tablets in R. CRIBIORE, Gymnastics of the Mind: Greek Education in Hellenistic and Roman Egypt, Princeton, NJ, Princeton University Press, 2001, pp. 153-157. Note too that F. WATSON, Q as Hypothesis: A Study in Methodology, in NTS 55 (2009) 397-415, p. 406, had suggested Luke’s use of wax tablets.

336

J.C. POIRIER

the shape of Luke’s gospel relative to his sources. Moreover, the ease with which wax tablets could be moved about presents a ready explanation for how Luke could so easily alter the sequence of Matthew’s pericopes. Robert Derrenbacker and F. Gerald Downing both responded to my article, taking exception to remarks I made about their work, and to the arguments I advanced in support of the Farrer Hypothesis (FH). Unfortunately, they mishandle several issues, and they both put forth a skewed assessment of my contribution – heavily skewed, in Downing’s case. Three years later, Alan Kirk also challenged my understanding of how a Farrerian Luke might have used Matthew, but he too mishandles some matters. In what follows, I revisit the scenario I put forward in 2012, defending it in the face of these three critics3. My 2012 article has an unmistakable zigzag aspect to it, in that I devoted a lot of space to defending Goulder, only to abandon his scenario in favor of a different explanation. This was bound to cause some confusion among would-be detractors. This can be seen, for example, in Downing’s attempt to problematize the idea of a table holding two open scrolls and a wax tablet simultaneously (see below), in that it was Goulder (and not I) who imagined two scrolls opened side by side, whereas it was I (and not Goulder) who imagined Luke using a wax tablet. (Downing erred in supposing that wax tablets require table space.) The reader is advised that the zigzag aspect of my original article might haunt his/her reading of the present essay. Before addressing Derrenbacker’s and Downing’s responses individually, I wish to mention an amusing aspect of their back-to-back publication within the same periodical, having to do with their respective malignings of my proposed scenario. On the one hand, Derrenbacker objects to my suggestion that Luke could have run out of wax tablets after making it only a third of the way through his writing project, as he thinks Luke’s social status was higher than I allowed: “it strains credulity”, he writes, “to suggest that the ‘great omission’ was simply caused by Luke running out of appropriate writing supplies, particularly given Luke’s apparent social status, seen most readily in his likely employment of an amanuensis to assist in his

3. See E. EVE, Writing the Gospels: Composition and Memory, London, SPCK, 2016, pp. 41-42, 143-144. Alan Garrow appears not to have read my 2012 article, as he recently claimed that “scholars have shown that Luke’s use of Matthew, as proposed under the Farrer Hypothesis (FH), requires Luke to treat Matthew in ways that are physically and mechanically improbable, and without contemporary precedent”. A. GARROW, Plausibility, Probability, and Synoptic Hypotheses: A Response to F. Gerald Downing, in ETL 96 (2020) 131-137, p. 131.

LUKE AND THE WAX TABLET REVISITED

337

rather complicated writing project”4. On the other hand, Downing scoffs at the supposedly great number of wax tablets my scenario entails, and imagines that Luke would require some sort of “specialised furniture” to pull off the task imagined by Goulder5. Thus, Derrenbacker objects that my Luke is too low in social status, while Downing objects that he is too high6! I. DERRENBACKER’S CRITIQUE Of my three respondents, Derrenbacker voices the most reasonable concerns. I am especially pleased that he likes (and even extends) my suggestion that Luke used wax tablets, as he knows more about scribal realia than I do. But I do not regard his criticisms of my scenario as especially strong. He complains that, although my appeal to wax tablets might explain how Luke could rearrange his source material so freely, I had not “offered an appropriate literary analogy from antiquity that would mirror what Luke is apparently doing with Matthew on FH”, nor had I “offered a motivation for this re-sequencing”7. If it falls to me to explain why Luke should want to resequence the material in the travel narrative, I would simply repeat what I have noted elsewhere: a Luke who is aware of the shape of Matthew’s 4. R.A. DERRENBACKER, Texts, Tables and Tablets: A Response to John C. Poirier, in JSNT 35 (2013) 380-387, p. 385. 5. F.G. DOWNING, Waxing Careless: Poirier, Derrenbacker and Downing, in JSNT 35 (2013) 388-393, p. 391. 6. G.H.R. HORSLEY, Classical Manuscripts in Australia and New Zealand, and the Early History of the Codex, in Antichthon 27 (1993) 60-85, p. 81, minimizes Luke’s social status even more than Downing; he considers the gospel writers too low in status to have much familiarity at all with scrolls (!), which he characterizes as the “traditional vehicle for the preservation of high literature”. Thus, he considers it “implausible … that the autograph gospels were written on rolls”. V.K. ROBBINS, The Social Location of the Implied Author of Luke-Acts, in J. NEYREY (ed.), The Social World of Luke-Acts: Models for Interpretation, Peabody, MA, Hendrickson, 1991, 305-332, p. 320, has argued that “the thought of the implied author [of Luke-Acts] is located near the artisan class”. I tend to agree with Landry’s view of Luke: “Luke was not only a competent writer, but also a professional – one with enough skill, education and past success as a writer to have had, in effect, a publisher”. D.T. LANDRY, Luke’s Revision of Matthew’s Infancy Narrative, in T. BENNY LIEW (ed.), Reading Ideologies: Essays on the Bible and Interpretation in Honor of Mary Ann Tolbert (Bible in the Modern World, 40), Sheffield, Sheffield Phoenix, 2011, 45-75, p. 54. Cf. Loveday Alexander’s well-known comparison of Luke’s writings to “middlebrow” technical writings. L. ALEXANDER, Luke’s Preface in the Context of Greek Preface-Writing, in NT 28 (1986) 48-74. Downing had used “middlebrow” to describe Luke’s style a year earlier. F.G. DOWNING, Ears to Hear, in A.E. HARVEY (ed.), Alternative Approaches to New Testament Study, London, SPCK, 1985, 97-121, p. 108. 7. DERRENBACKER, Texts, Tables and Tablets (n. 4), pp. 384-385.

338

J.C. POIRIER

gospel should be expected to adopt a noticeably different shape, simply for the sake of his own gospel’s marketability8. (Downing earlier accepted this explanation as having been “cogently argued” by Goulder)9. The first thing Derrenbacker challenges is my invoking of Clement’s use of Philo as an example of reverse scrolling through a source – a corroboration of Goulder’s implicit claim that an ancient writer might do such a thing10. Derrenbacker questions van den Hoek’s assumption that Clement knew Philo in scroll form11. He points out that the codex had become more widely used by Clement’s day, and this allows him to suggest that Clement knew De Posteritate in codex form. I agree that we should not casually assume that a given source had been accessed in scroll form12, but, in this case, I find a good reason to uphold that assumption: a codex would facilitate a reverse-linear accessing of Philo’s writings, but it does not explain why Clement should proceed that way, whereas the scroll form does. Scrolls required rerolling to the beginning before being read again (in a forward direction), and they often were not rerolled after being read. (That is the most probable explanation for why the titles of books were usually written at the end of a scroll)13. The form of the scroll therefore explains why Clement would move backward through Philo’s text. It should also be noted that Derrenbacker’s codex-reading solution raises questions of its own – most obviously: if he will allow that Clement might have moved backward through a codex of Philo for no discernible reason (other than convenience), should he not also allow Luke to have moved backward through a codex of Matthew for no discernible reason (other than convenience)? 8. The word “marketability” is not intended to imply an economic motivation. See J.C. POIRIER, Introduction: Why the Farrer Hypothesis? Why Now?, in J.C. POIRIER – J. PETERSON (eds.), Marcan Priority without Q: Explorations in the Farrer Hypothesis (LNTS, 455), London – New York, Bloomsbury T&T Clark, 2015, 1-15, pp. 2-6, 10-11. 9. See F.G. DOWNING, Disagreements of Each Evangelist with the Minor Close Agreements of the Other Two, in ETL 80 (2004) 445-469, p. 465. 10. See GOULDER, The Order of a Crank (n. 1). 11. DERRENBACKER, Texts, Tables and Tablets (n. 4), p. 381. 12. But cf. Tertullian’s references to “volumen” when discussing the OT writings in Adversus Judaeos 8,13; 13,5. Tertullian also uses “volumen” to refer to his own writings, as well as those of Justin Martyr and Soranus of Ephesus. J.E.L. VAN DER GEEST, Le Christ et l’Ancien Testament chez Tertullien: recherche terminologique (LCP, 22), Nijmegen, Dekker & van de Vegt, 1972, pp. 56-57. Christian preference for the codex was tied to writings of scriptural status. See L.W. HURTADO, The Earliest Christian Artifacts: Manuscripts and Christian Origins, Grand Rapids, MI, Eerdmans, 2006, p. 57. 13. F. SCHIRONI, ΤΟ ΜΕΓΑ ΒΙΒΛΙΟΝ: Book-ends, End-titles, and Coronides in Papyri with Hexametric Poetry (ASP, 48), Durham, NC, American Society of Papyrologists, 2010, p. 84, n. 229, cf. pp. 83-84, admits the advantage of an end title in the case of a roll that had not been rewound, but, for some reason, she treats this as exceptional circumstance. She notes (ibid., p. 29) that “all rolls of the Roman era without exception have an end-title”.

LUKE AND THE WAX TABLET REVISITED

339

Derrenbacker also questions whether one can compare a Farrerian Luke’s use of Matthew with Clement’s use of Philo, since Luke uses so much of Matthew, whereas Clement uses Philo only sporadically: “Philo is being consulted, at best, infrequently and occasionally”, and “[t]his is very different from what Luke is doing with Matthew on the FH”14. I do not see how this matters on a purely mechanical level, but it may be that Derrenbacker assumes that Luke’s use of Matthew should be more respectful of his order. If so, this simply takes us to the widely discussed issue of the Farrerian Luke’s rearrangement of Matthew, which mirrors the issue of a Streeterian Matthew’s rearrangement of Q. Derrenbacker also drives a wedge between Clement and Goulder’s Luke by citing “differences in length” between Clement and Luke, and between Clement’s source and Luke’s source. I cannot see how the difference in length between Clement’s and Luke’s texts matters. As for the Philonic material that Clement employs, Derrenbacker writes that it “could conceivably have been all on the same ‘page’ open in front of Clement”15, obviating the need for any reverse scrolling. (It is worth pointing out that this collides with Downing’s contention that a scroll would have been “unrolled to three columns at most”)16. In this, Derrenbacker implies that things are significantly different with the Matthean parallels in front of Goulder’s Luke17, but if the totality of Philo’s De Posteritate used by Clement could be displayed in a single view, then the totality of Matthew used in Luke 13,22–18,18 could be displayed in only two views. (De Posteritate 5-20 contains 3269 words, while Matthew 16–25 contains 6528 words.) Derrenbacker again emphasizes the extent and intricacy of Luke’s use of Matthew (on Goulder’s view), and opposes this to Clement’s use of Philo, but, once again, I cannot see how this makes a difference on a mechanical level. And I do not understand how Derrenbacker can say that I fail to provide an “analogous example” to “Luke’s reverse re-contextualization of Matt 25–16 … from the literary world of antiquity”18, especially in responding to an article in which I discuss Clement’s use of Philo as just such an example. Derrenbacker also questions my reference to the use of tabletops for displaying source texts, by noting the lack of consensus over whether the

14. DERRENBACKER, Texts, Tables and Tablets (n. 4), p. 382. 15. Ibid., p. 382. 16. DOWNING, Disagreements (n. 9), p. 446. Downing assumes a column to contain about 500 characters. 17. “Goulder’s Luke” is a narrow subset of “Farrer’s Luke” – it stands for the particular ways in which Goulder tried to explain Luke’s activity on the terms of the Farrer Hypothesis. 18. DERRENBACKER, Texts, Tables and Tablets (n. 4), p. 382.

340

J.C. POIRIER

table-like structure in room L30 at Qumran was used for copying texts19. I wrote my 2012 article fully aware, of course, of that lack of consensus, and it is partly because of it that I drew attention to other tables in the Greco-Roman world. Apparently, I should have said more. There is no question that tables could be found in many homes and workshops in the early second century20. I took care to distinguish between using tables as writing surfaces and using them to display one’s source text[s] – a distinction not always made by Goulder’s detractors – as only the latter matters for any appeal to micro-conflation21. At this point, we are confronted with John Kloppenborg’s insistence (following Jocelyn Penny Small) that tables were not used for such a purpose22. There is a good deal of evidence that some tables were indeed sufficient for the task, and it is hard to believe that a writer wishing to conflate two scrolls would scruple to use an available tabletop for such an end. I specifically noted in my 2012 article that m. Kel. 24,6 refers to the sort of table “which lies … before scriveners”, so presumably there was such a thing23. (I realize that some ancients might scruple to unroll a scroll on a dedicated dining surface – such as the low rectangular tripods that were sometimes hidden away beneath beds)24. 19. Ibid., p. 383. Cf. POIRIER, Roll (n. 1), p. 14. K.G. PEDLEY, The Library at Qumran, in RQum 2 (1959) 21-42, p. 33, writes, “[T]he ledge [in room L30 at Qumran] … resembles one at Nippur, of which a photograph has been published; ledges have been found at Pergamum, Lagash, and Ephesus; and the existence of such a ledge is commonly considered to be one of the proofs for identification of ancient libraries”. 20. Glaucon objected to Socrates’ omission of tables (among other things) as a common need for individuals, claiming that a city without couches, tables, prepared dishes, and sweetmeats would be a “city of pigs” (Plato, Resp. 372d). Admittedly, he had in mind dining surfaces, which might not be socially or religiously amenable for reading and writing, but some of my point still remains. See D. ANDRIANOU, Chairs, Beds, and Tables: Evidence for Furnished Interiors in Hellenistic Greece, in Hesperia 2 (2006) 219-266, pp. 220-221. 21. According to DERRENBACKER, Texts, Tables and Tablets (n. 4), p. 83, “over-whelming evidence suggests that writing desks … did not exist in antiquity”. 22. J.S. KLOPPENBORG, Macro-Conflation, Micro-Conflation, Harmonization and the Compositional Practices of the Synoptic Writers, in ETL 95 (2019) 629-643, p. 636, writes, “Although tables existed, they were not used, as far as can be determined, as writing or copying surfaces but instead for placing or counting objects”. See J.P. SMALL, Wax Tablets of the Mind: Cognitive Studies of Memory and Literacy in Classical Antiquity, London, Routledge, 1997, pp. 150-151. 23. POIRIER, Roll (n. 1), p. 14. 24. ANDRIANOU, Chairs, Beds, and Tables (n. 20), p. 258, writes, “Tables are likely to be found only in the setting of a symposium, where they are regularly depicted in front of beds as part of the ‘ritual.’ There they serve as portable trays to carry the dishes, and are placed next to the beds at a height convenient for the occupant. It might not be far-fetched to assume that tables came with beds as a ‘set’ in houses where such entertainments were customarily held. Supporting evidence is provided by a unique reference in the Delian inventories to attached tables that could be pulled out from under the beds”.

LUKE AND THE WAX TABLET REVISITED

341

While it might be true that the table in the ancient Greek household “had much fewer uses than with us”25, it would be a mistake to think of all tables as either low-top dining surfaces or small end-tables26. Tables could be used as working surfaces, and often were designed as such. One sort of table featured a large round top made of marble, like the fourth-/thirdcentury BCE table found at Pella measuring 1.30 m. in diameter27. Should we presume that Luke would not have made use of such surfaces if he had them to hand? Can we presume he did not have them to hand? And if his available surfaces were too small for multiple texts, why not posit the use of multiple surfaces28? (Given that benches sometimes doubled as tables29, we need only to envision two such benches being placed together to allow parallel reading of two open scrolls.) I also noted in 2012 that Juvenal referred to Cicero’s Second Philippic as a volume one “unroll[ed] next to the first” (volveris a prima quae proxima [Satires 10,126])30, so there is some sense in which Juvenal imagines two scrolls lying side by side, with at least one of them open. II. DOWNING’S CRITIQUE I take no pleasure in responding to someone who writes as stridently as Downing. He characterizes me as “hav[ing] fallen for the quite common argumentative device of devoting most of a study to showing others in a bad light, paraphrasing their contentions to sound as vulnerable as possible, so as to leave the stage clear for the new writer’s own rather scanty main suggestion to shine more brightly, against the gloom induced”31. Although 25. Apud G.M.A. RICHTER, Furniture of the Greeks, Etruscans, and Romans, London, Phaidon, 1966, p. 63. 26. Ibid. Cf. ANDRIANOU, Chairs, Beds, and Tables (n. 20), pp. 251-257. 27. Ibid., p. 253. 28. Andrianou (ibid., pp. 224, 258) emphasizes the portability of Greek furniture, a feature complementing the “multifunctional rooms of ancient houses”. She specifically mentions that tables “were possibly quite light in weight so that they could be moved around as needed” (p. 251), citing Xenophon Symposium 2,1; Plato Lakonians fr. 71; Menander Kekryphalos fr. 209. 29. Andrianou (ibid., p. 255) writes, “[I]n a number of Hellenistic houses [benches] apparently functioned as tables as well. In house IB in Eretria, part of a wall in front of the west wall of the kitchen (room u) possibly functioned as a bench”. See K. REBER, Die klassischen und hellenistischen Wohnhäuser im Westquartier (ERETRIA Ausgrabungen und Forschungen, 10), Lausanne, Editions Payot, 1998, p. 55 fig. 75. 30. POIRIER, Roll (n. 1), p. 15. 31. DOWNING, Waxing Careless (n. 5), p. 390.

342

J.C. POIRIER

I was already familiar with the way Downing writes, this charge surprised me a great deal. He also expresses a wish that I might someday apply my insights “in greater and clearer evidenced detail than the scant amount vouchsafed” in my 2012 article. I regard these sentiments as unfair characterizations of both the negative and positive goals of my argument. As I hope to show, the negative claims of my article do not overreach, as they were all intended as substantive table-setting moves, and the positive claims are scarcely “scant” or unclear in detail. My ultimate goal in 2012 was to defend “Farrer’s Luke”, but I also devoted considerable space to showing that “Goulder’s Luke” had been dismissed on inadequate grounds. My point was to note that “nothing about the FH – or, indeed, Goulder’s version of it – is technically prohibitive or even uncharacteristic of how ancient writers worked”32. I was driven to this argument by Downing’s repeated claims that Farrer’s Luke had “stepped intellectually, technically, and even technologically right out of his contemporary culture”33. I aimed to show that, contrary to the whole bent of Downing’s case, Goulder’s Luke was completely – “intellectually”, “technically”, and “technologically” – at home in his culture. It has long been Downing’s regular practice to exaggerate the difficulty of all tasks related to writing and information retrieval in the ancient world. One passage in one of his articles has become somewhat emblematic of his whole approach: “If Luke is doing his own writing, rather than dictating, he is balancing a springy new scroll on a board on his knees, together with a pen, and somewhere to hand are his scroll of Mark, and a sharpener, a pricker or other line-marker, ruler, dividers, sponge, bowl of water, pumice and ink (perhaps some sand?)”34. In recent years, Downing has made a commotion about the supposed difficulty of finding one’s place in a column of words written in scriptio continua, inviting readers to attempt to find matching strings of letters in parallel columns and comparing the difficulty 32. POIRIER, Roll (n. 1), p. 4. 33. F.G. DOWNING, Compositional Conventions and the Synoptic Problem, in JBL 107 (1988) 69-85, p. 82. 34. ID., A Paradigm Perplex: Luke, Matthew and Mark, in NTS 38 (1992) 15-36, p. 21. M. GOODACRE, Q, Memory and Matthew: A Response to Alan Kirk, in JSHJ 15 (2017) 224-233, pp. 226-227, notes, “It is of course true that codices are easier to handle than scrolls, but it is a mistake for modern authors to be stressed about how ancient authors coped with their media realities, just as future authors might be amazed at how people managed to navigate their way through print books without being able to run electronic searches or to click on, mouse over or manipulate hypertext. … The expression of bafflement about the difficulties involved with handling scrolls comes from a kind of anachronistic anxiety about how ancient authors could have managed something apparently so unwieldy”.

LUKE AND THE WAX TABLET REVISITED

343

of that task with what (he thinks) certain source theories imagine the evangelists to be doing35. (He even assizes the parallel material in terms of counted letters rather than of counted words!) He is apparently unaware of recent experiments showing that those who have been taught to read scriptio continua are generally able to differentiate words within that format just as quickly as modern westerners are able to differentiate words in scriptio discontinua. (Jan Heilmann traces the view that words written in scriptio continua only became recognizable through being read aloud to nineteenthcentury “cultural pessimism”: “[T]he fact that the Romans replaced their habit of writing scriptio discontinua with scriptio continua in the second century CE shows that scriptio continua is not a characteristic of deficient writing systems”)36. Another respect in which Downing tries to tie Farrer’s Luke to overspecialized circumstances has to do with his reference to the “specialised furniture” of “Alexandrian scholarly procedure” – furniture that would have been unavailable to Luke. He agrees with me that “[r]eading desks that could support two scrolls are evidenced”, but he objects that these are “from social contexts where Luke is not obviously at home”37. In point of fact, of course, Luke does not require anything out of the ordinary in the way of furniture to accomplish what the FH envisions38: there were many general35. The argument is used in DOWNING, Disagreements (n. 9), pp. 445-469; ID., Writers’ Use or Abuse of Written Sources, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 523-548; ID., Plausibility, Probability, and Synoptic Hypotheses, in ETL 93 (2017) 313-337. He considers it probative that “the ‘shapes’ of words are much less distinct [in scriptio continua] than we are used to” (Disagreements [n. 9], p. 450), but he gives scant attention to the fact that they are more distinct to those trained to read that sort of text. He mentions that an ancient reader “would have been used to scanning texts in Greek majuscules in scriptio continua”, but he still imagines that reading aloud was a part of what enabled such readings. His understanding of ancient media, in this respect, is out of date. 36. J. HEILMANN, Reading Early New Testament Manuscripts: Scriptio continua, ‘Reading Aids’, and Other Characteristic Features, in A. KRAUSS – J. LEIPZIGER – F.S. JUNGBLUT (eds.), Material Aspects of Reading in Ancient and Medieval Cultures (MTK, 26), Berlin – New York, De Gruyter, 2020, pp. 178, 182, appeared too late for Downing, but Heilmann points to earlier experiments by J. Epelboim and by A. Vatri that show that “spaces between words are not the primary factor controlling eye movement”, and quotes the latter’s judgment that “[T]here is no reason to assume that reading unspaced text is a particularly demanding cognitive task in itself” (pp. 180-181). Heilmann warns, “We should be careful not to project our difficulties with reading scriptio continua onto ancient readers” (p. 183). 37. Waxing Careless (n. 5), p. 390 (emphasis original). 38. See ibid., p. 391. ANDRIANOU, Chairs, Beds, and Tables (n. 20), p. 259, writes, “As a result of the incomplete, fragmentary, and often poorly published evidence, we cannot label certain houses as ‘houses of the elite’ on the basis of furniture alone”.

344

J.C. POIRIER

purpose tables in the ancient world that could have accommodated two scrolls (see above). And in defense of my own scenario, it bears mentioning that the rigidity of the wax tablet allowed a measure of freedom in choosing where one should write. (According to Élisabeth Lalou, “Le format des tablettes plus long que large convenait probablement fort bien aux comptables et la rigidité du bois des tablettes facilitait l’écriture dans toutes les circonstances”)39. It was not the case, as Downing supposes, that Luke’s working surface would have needed room for two open scrolls and the wax tablet that I posit40 – that is not how wax tablets were used, not to mention that I had offered my wax-tablet scenario as an alternative to the two-scroll scheme. It is ironic that Downing should characterize my argument as an attempt to show others “in a bad light” when his own modus operandi entails wildly exaggerated claims about the 2DH being the only source theory that respects ancient compositional practices41. As he rehearses it anew for my sake, one of Downing’s arguments is that it is remarkable … that a Luke or any evangelist writing third should refuse to include even in paraphrase some of the most closely similar matter appearing in both the others, when ancient writers in practice and expressly were content to encounter, meet with, come across … general agreement in their sources for themselves in turn to paraphrase without more ado42. 39. Les tablettes de cire médiévales, in BEC 147 (1989) 123-140, p. 131. 40. DOWNING, Waxing Careless (n. 5), pp. 390-391, writes: “What seems to be demanded is a desk wide enough for at least one open and one in double-roll, as well as, perhaps, room to support ‘Luke’s’ wax tablet, if he is making his own notes (and thus at least one scroll has props of some sort either side of the column lying open to be perused). Applications of common technologies are culturally formed, are they not? Are we now asking ‘Luke’ for cultural innovation – not in the technology of wax tablets, but in its application? Or is Alexandrian scholarly procedure common rather than in-group knowledge (…), and is its specialised furniture available?” 41. Part of what lies behind Downing’s “bad light” charge may be a misunderstanding of why I mentioned a certain detail. He mischaracterizes my correction of “D.S. Russell” to “D.A. Russell” as an accusation (of slovenliness), complaining that I “accused” him of “getting a modern author’s second initial wrong” (ibid., p. 390, n. 4; see POIRIER, Roll [n. 1], p. 4, n. 2). My correction was not meant to be accusatory (or picayune), but was offered as a correction of a long-lived and confusing misreference: D.S. Russell is indeed a real person, whose books many readers will have on their shelves, but he has nothing to do with Downing’s argument. I intended it as a prod for Downing to correct a misreference that has appeared in numerous articles he has written, spanning several decades. (I tried to stop the parade of misreferences years ago, when I read a prepublication version of Downing’s essay for the Oxford volume – an essay Downing accuses me of ignoring – and I encountered there yet another reference to “D.S. Russell”. I brought the error to the attention of the volume’s editors, but they allowed the volume to be published with the wrong reference). 42. DOWNING, Waxing Careless (n. 5), p. 389. Downing accuses me of “dramatiz[ing]” my argument “with serious misrepresentations” of others, of which the “most egregious” is my “insistence that Downing bases an important argument on ancient writers’ supposed

LUKE AND THE WAX TABLET REVISITED

345

He seems to imply that whatever Luke’s contemporaries did when they use multiple sources, we should expect Luke himself to do – end of story. But should we not first ask if Luke’s motivation and circumstance were similar to those of the writers with whom Downing compares him? Downing treats the tendencies of ancient writers as universally applicable43, without taking account of a writer’s knowledge of the mutual relationship between his/her sources. I would explain the behavior of the classical writers he discusses as an attempt to expose historical bedrock by seeking out what is multiply attested among supposedly independent witnesses: if A and B wrote independently of each other, then it naturally follows that their common witness will represent their most valuable tradition. (Indeed, this logic is made explicit in two sources that he quotes: in Tacitus – “Where the authorities are unanimous, I shall follow them” – and in Arrian – “Whenever Ptolemy son of Lagus and Aristobulus son of Aristobulus have both given the same accounts … it is my practice to record what they say as completely true”)44. But does Farrer’s Luke stand in a position where such a canon of logic applies? Hardly! In the Farrer universe, Luke’s two sources are not independent – Matthew used Mark, and Luke knows that Matthew used Mark45. It therefore would not make sense for Luke to count Mark’s and Matthew’s common witness as possessing more worth than the other material in those gospels, and there should be no expectation for Luke to ‘search’ ‘apparently’ for ‘verbatim agreements’ in their sources”. He calls this “simply false”, and speaks instead of a search for “a common ‘story line’, a shared ‘narrative logic’”. But that doesn’t seem to represent Downing’s actual procedure – as T.J. Mosbø points out, Downing sometimes speaks of Luke excluding material from Matthew, when in fact he has not excluded it but rather reworked it: “What Downing appears to be objecting to is that Luke has not reproduced the exact wording where Matthew and Mark agree exactly. For Luke to have done so, he would have had to compare Matthew to Mark word for word in order deliberately to include those phrases where they matched exactly. This is basically the same absurd procedure that Downing has suggested Luke would not have followed if he had used Matthew as a source”. T.J. MOSBØ, Luke the Composer: Exploring the Evangelist’s Use of Matthew, Minneapolis, MN, Fortress, 2017, p. 181. 43. DOWNING, Disagreements (n. 9), p. 445, refers to “the widely evidenced contemporary predilection for common witness among ‘historical’ sources”. 44. Both quoted in ibid., p. 447. See Tacitus, Annals 13,20, in J. JACKSON, Tacitus, Annals (LCL, 322), Cambridge, MA, Harvard University Press, 1937, p. 35 (incorrectly listed by Downing as Tacitus’ Histories); Arrian, Anabasis of Alexander 1.praef.1, in P.A. BRUNT, Arrian, History of Alexander, and Indica (LCL, 236), Cambridge, MA, Harvard University Press, 1976, p. 3. 45. Luke’s knowledge of Matthew’s use of Mark is implicit in E. Franklin’s and B. Shellard’s understandings of Luke as a response to Matthew’s designs on Mark – see E. FRANKLIN, Luke: Interpreter of Paul, Critic of Matthew (SupplJSNT, 92), Sheffield, Sheffield Academic, 1994; B. SHELLARD, New Light on Luke: Its Purpose, Sources and Literary Context (SupplJSNT, 215), London, Sheffield Academic, 2002, pp. 261-288.

346

J.C. POIRIER

imitate Downing’s control group46. Downing ridicules the “fertile and adaptable imaginations” of those who can envision Luke doing what Farrer’s Luke must have done47, but the evidence suggests that those who cannot see a reason for Luke’s behavior have no imagination at all. Downing’s representations of ancient compositional practices are usually written with an air of confidence, but they tend to lack any real grasp of how things were done. Take, for example, his complaints about Goulder’s references to Luke “marking” his place in his source texts. My 2012 article defended Goulder’s assumption that Luke could easily have “marked” his spot within a scroll, in a sense, by explaining that the reader’s place could be preserved merely by moving the spindles of the open scroll together. Downing rightly notes that this fails to satisfy Goulder’s language, which specifically refers to placing a “marker” within the scroll – as if to mark a passage more specifically48. He notes that Goulder’s language “must denote [marking one’s place] more than just somewhere in an open column”49. It falls to me, therefore, to explain more exactly what Goulder meant by a “marker”. Goulder himself explains: Our earliest papyrus texts of the Gospels, 𝔓66, … are divided by ektheses; that is, a unit like the good Samaritan is marked in the text by a line with a letter sticking out as the unit ends, probably to signal to a lector in church that the unit is finishing. Such markers would enable Luke to pick out one Matthean pericope from another in a moment50.

46. DOWNING, Paradigm Perplex (n. 34), p. 27, had earlier described Mark and Matthew as “agreed witnesses” in Farrer’s Luke’s purview. My point is that they are nothing of the sort – they agree by dint of collusion. See J.C. POIRIER, Delbert Burkett’s Defense of Q, in POIRIER – PETERSON (eds.), Marcan Priority without Q (n. 8), 191-225, p. 221, and the response in J.S. KLOPPENBORG, The Farrer/Mark without Q Hypothesis: A Response, ibid., 226-244, p. 236. Kloppenborg says he “wonders” how I know that Luke had knowledge of Matthew’s dependence on Mark. My response is to enlist the common back story for why Farrer’s Luke wrote his gospel: he was unhappy with what Matthew did to Mark. Cf. GARROW, Plausibility (n. 3), p. 134: “There is nothing to suggest that the authorities Downing cites saw particular virtue in dual verbatim testimony. Indeed, common sense suggests the opposite. The more extensively two witnesses agree verbatim the greater the suspicion that, far from offering independent corroboration, the second witness is merely parroting the first”. 47. See DOWNING, Disagreements (n. 9), p. 468. 48. E.g., M.D. GOULDER, Luke: A New Paradigm (SupplJSNT, 20), Sheffield, JSOT, 1989, p. 346, writes, “Luke rolls up his scroll of Mark, with a marker at 3.19, and opens his Matthew again with an eye for 5.1 …”. Cf. pp. 197, 291. See DOWNING, Paradigm Perplex (n. 34), p. 18. 49. Waxing Careless (n. 5), p. 389. 50. GOULDER, Michael Goulder Responds (n. 1), p. 146.

LUKE AND THE WAX TABLET REVISITED

347

Thus, when Goulder envisions Luke placing a “marker” in a text, he evidently has in mind the drawing of some sort of line or symbol at the point he wants to mark. As William Johnson notes repeatedly, the introduction of sigla by the reader is a regular feature in scrolls – particularly in those texts that he (following earlier studies) calls “scholars’ texts”51. In a survey of a group of Oxyrhynchite texts, Johnson notes that most of the texts are … marked up with chi or other sigla in the left margin (…). Such sigla have sometimes been taken as keys to commentaries, but, to follow McNamee (1992, 19-22), these are best understood as a variety of ways of signaling nota bene; that is, they are marks by readers that signal passages of interest or those that need further attention52.

In an earlier article, Johnson had argued that the paragraphus mark served to anchor one’s place within the text, to assist the one reading the text aloud to find his/her place after having looked away from the text53. This anchorage for reading seems almost certainly to represent what Goulder had in mind by a “marker”. Luke’s use of such devices, therefore, would have been wholly within character for scholarly readers of his day. In fact, Jaap Mansfeld and David T. Runia describe a proceeding similar to that envisioned by Goulder when they refer to “marked passages … copied onto wax tablets (πύκτια, pugillares), from where they were transferred to collections of notes (ὑπομνήματα, commentarii), thus allowing a first sorting out of material”54. This description presents a convenient bridge between Goulder’s suggestion that Luke used “markers” and my own suggestion that Luke sorted the material during a waxen stage of composition, although my own scenario no longer requires markers of this sort.

51. W.A. JOHNSON, The Ancient Book, in R.S. BAGNALL (ed.), The Oxford Handbook of Papyrology, Oxford, Oxford University Press, 2009, 256-281, pp. 262, 272-273, 276. Johnson describes these texts as “annotated texts that contain textual remarks and variants” (p. 272). 52. Johnson (ibid., p. 273) notes that “[m]ost of the texts (ten of the sixteen) show signs of substantial additions of readers’ marks (such as added punctuation, word division, or elision markers), and almost all have at least some readers’ marks (as is generally common in bookrolls, especially for difficult texts). Often multiple readers are involved”. See the discussion of the paragraphus in SCHIRONI, ΤΟ ΜΕΓΑ ΒΙΒΛΙΟΝ (n. 13), p. 10 (cf. the studies listed there at p. 16, n. 31). 53. W.A. JOHNSON, The Function of the Paragraphus in Greek Literary Prose Texts, in ZPE 100 (1994) 65-68. But see the response to Johnson in HEILMANN, Reading Early New Testament Manuscripts (n. 36), pp. 186, 188. 54. J. MANSFELD – D.T. RUNIA, Aëtiana: The Method and Intellectual Context of a Doxographer, vol. 1: The Sources (Philosophia Antiqua, 73), Leiden – Boston, MA – Köln, Brill, 1997, pp. 204-205.

348

J.C. POIRIER

Downing allows that wax tablets “might seem to allow for much greater flexibility for a third evangelist using the other two as sources than many commentators, including Derrenbacker and Downing, have supposed and proposed”55. He asks for more from me, however, than this sort of allowance: “How in detail wax tablets could have been used to the end indicated, and whether anyone (other than an imagined ‘Farrerian Luke’) ever did so use them, is left not only unevidenced, but unquestioned”56. He does not hint as to what sort of “detail” I failed to provide. His objection seems to be that I have shown only that wax tablets were widely used, but not that anyone ever used them to combine sources in the way Farrer’s Luke did57. I sit comfortably with that objection – at some point, one has to wonder if the supply of objections is really endless58. Anyone wanting to combine two sources, and who had an adequate supply of wax tablets, could probably put two and two together. I suggested, in 2012, that Luke’s supply of tablets may have been limited, and that this can explain the three-panel appearance of his gospel (with the so-called travel narrative intercalated between 3,1–9,50 and 18,15–24,53), while the ease of rearranging tablets can explain the resequencing of Matthean material that we find in the travel narrative59. Downing estimates that, “[i]f one were to imagine a close correlation between a fairly common papyrus column size and similar sized writing on wax, say, about 500 characters for each”, it would have taken Luke some 180-200 tablets to write his gospel at a single pass, or some 60-70 tablets to write it in three parts60, and he balks at the idea that someone in Luke’s position could have had access to that many tablets. This strikes me as odd, and I will explain why. 55. DOWNING, Waxing Careless (n. 5), p. 390 (emphasis original). 56. Ibid. 57. Downing (ibid., p. 391) writes: “[I]s there any evidence for anyone in any contemporary socio-cultural context using so many tablets, for anything, let alone for collating two sources – better, two narrative sources – to produce a third?” 58. Downing (ibid., p. 391, n. 10) notes that “[n]either Small nor Winsbury suggests shuffling tablets”. So is that the issue? Does Downing concede that Luke was mechanically capable of “shuffling tablets”, but insists that we need evidence in the form of cultural practices before we can reasonably suggest that he did so? Where then, we might ask, is the evidence of cultural practices in support of what Matthew does to Q on the grounds of the Two-Document Hypothesis? (Some twenty years earlier, Downing apparently thought it sufficient for Matthew’s designs on Q to be facilitated on a strictly mechanical level; cf. his Paradigm Perplex [n. 34], p. 20). 59. A similar instance of one’s writing being limited by one’s supply of wax tablets can be found in Jerome, Epist. 18,16 (J.-P. MIGNE [ed.], Patrologia Latina vol. 22, Paris, Migne, 1845, 361-376, col. 372) (mentioned in ROUSE – ROUSE, Wax Tablets [n. 2], pp. 186-187). 60. DOWNING, Waxing Careless (n. 5), p. 391 (emphasis original).

LUKE AND THE WAX TABLET REVISITED

349

Given the widespread use of wax tablets as compositional aids, one wonders why a number like 60-70 should be considered excessive. It might repay us to ask how that number would have struck writers during the centuries when wax tablets were common. Downing reasonably dismisses the account of Plato’s secretary (Philip of Opus) copying The Laws from wax tablets onto papyrus61 – a task requiring 1064 wax surfaces, if we use Downing’s estimate of 500 characters per surface62 – but he ignores later accounts of works comparable in size to Luke being stored in a wax medium. I list four: (1) We have the twelfth-century account of Anselm composing his Proslogion entirely on wax tablets, and of storing it in that medium, only to have the collection of tablets disappear, at which time he wrote out the entire work a second time on wax tablets – before a second disaster forced him to have the work copied onto parchment63. By my count, the Proslogion contains 30,636 characters (sans punctuation, spaces, main title, and outline)64. This would have required 62 wax surfaces, on Downing’s terms. Given that Anselm wrote the work out twice, he must have had ready access to some 124 wax surfaces. (And we should bear in mind that Anselm intended those tablets to be dedicated to long-term storage of the text of the Proslogion, whereas Luke intended his tablet use as a strictly temporary arrangement.) (2) There is a relevant story behind Willibald’s eighth-century Life of St. Boniface. According to an appendix to this work, Willibald wrote his account of Boniface on wax tablets, which he then presented to Bishops Lull and Megingoz for their corrections and suggestions65. Once the account met their approval, it was transferred to a more permanent medium. Unfortunately, we do not know how often Willibald brought his material to Lull 61. Ibid. The account comes from Diogenes Laertius (Lives of Eminent Philosophers 3,37), who does not own it as necessarily true. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, p. 49, n. 97, follows Downing at this point. 62. The estimate of 1064 tablets is based on my count of 531,760 characters in the Burnet edition of The Laws (sans punctuation and word spacing) – see J. BURNET (ed.), Platonis Opera, Oxford, Oxford University Press, 1903. Downing’s own estimate for a waxen copy of The Laws is between 800 and 1000 tablets. See J.L. SHARPE III, The Dakhleh Tablets and Some Codicological Considerations, in É. LALOU (ed.), Les tablettes à écrire, de l’antiquité à l’époque moderne: actes du colloque international du Centre National de la Recherche Scientifique, Paris, Institut de France, 10-11 octobre 1990 (BIB, 12), Brepols, Turnhout, 1993, 127-148, pp. 128129 (cited in KIRK, Q in Matthew [n. 61], p. 49, n. 97). DERRENBACKER, Texts, Tables and Tablets (n. 4), p. 384, on the other hand, accepts Plato’s account. 63. The account was given by Anselm’s biographer Eadmer. See ROUSE – ROUSE, Wax Tablets (n. 2), p. 179; J. RIDER, God’s Scribe: The Historiographical Art of Galbert of Bruges, Washington, DC, Catholic University of America Press, 2001, pp. 257-258, n. 10. 64. Alcuin restored spacing between (Latin) words in the eighth century. 65. G.H. PERTZ (ed.), Monumenta Germaniae Historica, Scriptores, Hannover, Hahn, 1829, vol. 2, p. 357.

350

J.C. POIRIER

and Megingoz. If he brought the entire work at once, he presented a text consisting (by my count) of some 58,024 characters (66,646 if we include spaces), requiring some 117 wax surfaces (or some 134 wax surfaces, if written with word-spacing). And if Willibald presented the bishops with whole chapters (of which there are nine), we are forced to imagine the longest of the chapters (viz. chap. 8) composed in wax66. That chapter alone is 16,227 characters long, without word-spacing (= 33 wax surfaces), or 18,673 characters long, with word-spacing (= 38 wax surfaces)67. (3) In the prologue to his Vita Sancti Cuthberti, Bede speaks similarly of circulating that work in rough-draft form among those who knew Cuthbert personally, “I made it my business to put down on parchment the results of my rigorous investigation”68. Bede’s writing presumably existed only in wax before this point. (By my count, this work is 42,045 characters long [= 85 wax surfaces] – 48,314 if we include spaces [= 97 wax surfaces])69. (4) Mention should also be made of the arrangement among Hugh of Saint Victor’s students, in which one of them was commissioned by his fellow students to approach Hugh once a week with the wax account of the master’s lectures, so the master could correct and add to them70. One can only imagine how many wax tablets were involved in these weekly meetings. Although I have used Downing’s estimated wax-tablet capacity of 500 words in compiling my counts, none of the foregoing is meant to concede that estimate. The real figure is probably much higher. At the end of her article on “Les tablettes de cire médiévales”, Lalou presents a list of wax tablets that she cites, which she divides into full-sized tablets and “carnets” (= polypteuchs smaller than 10 cm on all sides)71. The following recounts the sizes of those she lists in the former category – in the order in which she lists them – along with the number of tablets and wax surfaces, and the dimensions of the wax surfaces: 66. The chapter divisions appear not to be arbitrary, in that they are marked in every case by discursive closure (e.g., “Amen”). 67. My counts are based on the text in W. LEVISON, Vitae Sancti Bonifatii, Archiepiscopi Moguntini, Hannover, Hahn, 1905, pp. 4-57. 68. Trans. B. COLGRAVE, Two Lives of Saint Cuthbert: A Life by an Anonymous Monk of Lindisfarne and Bede’s Prose Life, Cambridge, Cambridge University Press, 1940, p. 145. See RIDER, God’s Scribe (n. 63), p. 257, n. 10. Rider calls attention to Guibert of Nogent, who tells us (in Vita 1,17) that he did not use wax tablets in writing his commentary on Genesis, but rather “committed [the words] to the written page in their final form [scribenda … immutabiliter paginis] as I thought them out”. Rider points out, following an observation by Lalou, that Guibert’s bothering to mention this shows it was an unusual way of writing. See LALOU (ed.), Les tablettes de cire médiévales (n. 31), pp. 133-134. 69. Text from I IntraText Edition CT, 2007 (http://www.intratext.com/X/LAT0423.htm). 70. See RIDER, God’s Scribe (n. 63), p. 259, n. 16. 71. Les tablettes de cire médiévales (n. 39), pp. 135-138.

LUKE AND THE WAX TABLET REVISITED

351

14 tablets (= 26 faces) 475 mm × 205 mm 16 tablets (= 30 faces) 130 mm × 300 mm 11 tablets (= 21 faces [“manquent 4 tablettes au début”]) 135 mm × 272 mm 14 tablets (= 26 faces) 169 mm × 330 mm 14 tablets (= 26 faces) 170 mm × 343 mm 1 tablet (= 2 faces) 130 mm × 338 mm 8 tablets (= 16 faces) 170 mm × 343 mm 4 tablets (= 6 faces) combined with 9 tablets (= 16 faces) 132 mm × 340 mm 6 tablets (= 11 faces) 140 mm × 348 mm 8 tablets (= 16 faces) 168 mm × 330 mm 1 tablet (= 2 faces) size not listed 8 tablets (= 15 faces) 384 mm × 168 mm 1 tablet 326 mm × 160 mm 1 tablet (= 2 faces) 122 mm × 242 mm 15 tablets (= 29 faces) 155 mm × 360 mm 1 tablet (= 2 faces) 351 mm × 155 mm 8 tablets (= 14 faces) 275 mm × 153 mm 6 tablets (= 10 faces) 75 mm × 210 mm

By taking the dimensions of these tablets, and by weighting them according to the number of wax surfaces found within each (reconstructed) tablet, we arrive at an average wax surface size of 336.3 mm × 155.6 mm (= 13.24 in. × 6.12 in.). (I weight these according to the number of surfaces, in case that number was somehow a determining factor in the size of the surfaces.) Downing equates the capacity of a wax tablet with a typical column of text in a papyrus, but given the average size of the tablets in Lalou’s list, and the lack of reason for those writing on tablets to create significant margins, it might be nearer the mark to equate the capacity of a tablet with two columns of writing. If this represents a better estimate, then the preceding estimates of required wax tablets (including those required by Luke) should all be reduced by half. III. KIRK’S CRITIQUE Alan Kirk approaches issues like wax tablet use with forced skepticism, to make room for a more purely mnemonic approach to the order of the double tradition. According to him, “Synoptic scholarship has recognized the significance of … material media realities for source criticism, though less care has been taken to ensure that Synoptic phenomena actually fit the media practice in question”72. He counts me among those who take “less 72. KIRK, Q in Matthew (n. 61), p. 42. As Kirk writes of the “disappoint[ment]” of anyone wishing to find a discussion of “ancient media realities” in POIRIER – PETERSON (eds.), Marcan Priority Without Q (n. 8), I should perhaps mention that the original draft of my

352

J.C. POIRIER

care” than matters require – as though (perhaps) echoing Downing’s charge that I somehow had “wax[ed] careless”73. Kirk criticizes my suggestion that Luke could have composed his gospel with the help of wax tablets, partly on the grounds that “[i]t was quite a feat for the twelfth-century writer Baudry of Bourgueil to compose a poem of 112 verses on a 14-leaved wax tablet”74. Unfortunately, this comparison between Luke and Baudry is misguided, as the latter’s tablet appears to have been a sort of mini-tablet (a pugillaris): Baudry’s poem celebrating his tablet refers to its writing surfaces as paginulae (“little pages”)75. (Kirk should have referred to Baudry as having a “14-paged tablet” [or “8-leaved tablet”] rather than a “14-leaved tablet”, as “leaf” [folium] refers to a physical board which in turn has one or two writing surfaces.) Baudry’s love for his polyptych was based not merely on its capacity to write a 112-line poem, but also on its portability76. (He remarks that its length could “hardly take a hexameter,” and he could write only eight lines per page!) We can safely assume that Luke’s wax tablets were not the sort of satchel-sized miniatures that appealed to traveling poets, and which were held in the hand (rather than balanced on the knee) while being written in77. And if Lalou’s list of wax tablets is any indication of typical size, then Kirk is misguided in making J.B. Poynton’s much smaller specimens (9 in. × 7 in.; 7 in. × 4 in.) exemplary78. The oddest aspect of Kirk’s critique lies in his attempt to spoil the simplicity of my scenario, by ascribing to it a sort of duality: Poirier first imagines Luke’s usage along the lines of the use of tablets in Greco-Roman authorial projects: to facilitate collection and gestation of materials, subsequently worked up into a preliminary draft (in Poirier’s view also on wax tablets), which is then refined and transferred in clean copy to a papyrus roll. … But then Poirier tacitly switches scenarios to having Luke use “Introduction” to the volume included a discussion of such “realities” as one among several things that made the FH especially timely. When it became necessary to cut several hundred words from the volume (in view of contract stipulations), I scuttled that discussion, supposing that it might sound self-congratulatory. 73. See Eve, Writing the Gospels (n. 3), pp. 133-138. KIRK, Q in Matthew (n. 61), p. 168, refers to my wax tablet scenario as “simply a variant of [Ulrich] Luz’s proposal”. 74. Ibid., p. 49. 75. See ROUSE – ROUSE, Wax Tablets (n. 2), p. 187; ROUSE – ROUSE, The Vocabulary of Wax Tablets, in Harvard Library Bulletin, 1.3 (1990) 12-19, p. 14. 76. On Baudry and his poem, see P. ABRAHAMS, Les Œuvres Poétiques de Baudri de Bourgueil (1046-1130), Paris, Librairie Ancienne Honoré Champion, 1926, pp. 53-55 (no. 47: “Ludendo de tabulis suis”); J.-Y. TILLIETTE, Baudri de Bourgueil, Poèmes (Auteurs Latins du Moyen Âge), Paris, Les Belles Lettres, 1998, pp. 36-37 (no. 12: “Ludendo de tabulis suis”). 77. On the distinction, see esp. ROUSE – ROUSE, Vocabulary (n. 74), p. 13. 78. KIRK, Q in Matthew (n. 61), p. 42. See J.B. POYNTON, Books and Authors, in Greece & Rome 3 (1934) 94-104, p. 94.

LUKE AND THE WAX TABLET REVISITED

353

wax tablets in more of a direct source-utilization mode: to excerpt passages from Matthew and transfer them into their new arrangement in the Travel Narrative. … Poirier therefore awkwardly combines two quite different compositional procedures79.

Kirk appears to intend this accusation quite seriously, as he alludes to it later in his book when he states that my “wax tablets proposal … do[es] not get off the ground”80, and he claims that my overall scenario (which he thinks is actually a case of “switch[ing] scenarios”) “suffers from serious difficulties”81. I am frankly dismayed by this criticism: where Kirk sees a mountain, I don’t recognize so much as a molehill. Do I really “switch scenarios” by suggesting that Luke both wrote a “preliminary draft” on his tablets and “excerpt[ed] passages from Matthew and transfer[red] them into their new arrangement in the Travel Narrative”? Shuffling a deck of playing cards and then dealing them out is technically a “switching” of “scenarios”, but that doesn’t obviate the fact that one of those activities naturally follows the other. Is Luke’s decision to do what I described really materially different from normal “Greco-Roman authorial projects”? Are my suggestions dualistic – much less flight-grounding? According to Kirk, Excerption to wax tablets, if posited as an intermediate step, does not dispense with the problems scroll utilization raises for Matthew’s rearrangement of Q materials (or on the FGH, Luke’s rearrangement of Matthean materials). It just creates a collection of excerpted materials, in the sequence of the source, still to be rearranged and coordinated with other source materials82.

This is plainly false: the point in positing excerption to wax tablets is that it explains how an evangelist can rearrange the material – easily, quickly, and efficiently. If “problems” remain for that evangelist’s “rearrangement” of materials, they are not mechanical but literary. It seems to me that Kirk is trying to conjure a problem out of thin air. In point of fact, Kirk elsewhere defends exactly the sort of fell-swoop obtainment of a two-step process that he considers problematic in my work. Consider, for example, his response to Derrenbacker, who “balk[s] at [Kirk’s] rejection of the need for intermediary media – provisional drafts (ὑπομνήματα); wax tablets; open-leaf notebooks; proto-codices – to account on sourceutilization grounds for Matthew’s rearrangement of the Q materials”83. Apparently, fell-swoop obtainments are acceptable, so long as they promote 79. KIRK, Q in Matthew (n. 61), pp. 48-49. 80. Ibid., p. 307. 81. Ibid., p. 149. 82. Ibid., p. 54. 83. ID., The Synoptic Problem, Ancient Media, and the Historical Jesus, in JSHJ 15 (2017) 234-259, p. 234.

354

J.C. POIRIER

the correct source theory. As Derrenbacker points out, Kirk’s thinking on this matter – and, I would add, his criticism of my proposal – revolves around a stark separation between “excerpting sources” and “drafting a work”: according to Kirk, “Difficulties in Matthew’s or Luke’s use of their sources cannot be resolved by appeal to excerption practices. Excerpting was mostly a copying operation”84. The sort of separation that Kirk defends here is frankly hard to believe, especially since a rapprochement between excerpting and drafting seems to be involved in the very idea of writing a gospel. Kirk also lays out the standard line about multiple scrolls not being opened simultaneously, given the mechanics of scroll use and the size and supposedly restricted use of tables, etc. Small’s claim that there are “no … ancient representations of multiple texts open simultaneously” finds in Kirk another mouthpiece, in spite of my pointing out in 2012 that Juvenal referred to Cicero’s Second Philippic as a volume that one “unroll[ed] next to the first” (Satires 10,126)85. Writers typically did not spread their scrolls out on tables or desks. A desk would have been of no greater use than the floor for management of two or more scrolls. Tables could support objects with flat, stable surfaces such as vases, statuary, tablets, and coins; in scenes depicting financial transactions they are used ‘not for recording the transaction, but for counting the money’. Iconographic evidence of writing on tables features tablets, not scrolls, and the scribes are taking shorthand, not copying. A table or chest if present held writing accessories such as inkwells. Since the average scroll was between ten and thirty feet long, tables would hardly have advantageously accommodated even a single scroll, and ‘the likelihood of a number of rolls being open simultaneously … while actually writing was, therefore, extremely low’86.

I already dealt with these issues in my response to Derrenbacker, but it is at least worth noting Kirk’s strange assumption that a scroll would have been completely unrolled when placed on a table. IV. CONCLUSION At the beginning of the twentieth century, Lewis A. Muirhead commended Paul Wernle’s book on the Synoptic Problem for “its dislike of merely possible documents”: “A critic must have space – the larger the better – for possible documents. But the space is the lumber-room to which the 84. Q in Matthew (n. 60), p. 53. See R.A. DERRENBACKER, Jr., Matthew as Scribal Tradent: An Assessment of Alan Kirk’s Q in Matthew, in JSHJ 15 (2017) 213-223, p. 222 and n. 48. 85. KIRK, Q in Matthew (n. 61), p. 54. See SMALL, Wax Tablets of the Mind (n. 22), p. 167. 86. KIRK, Q in Matthew (n. 61), pp. 54-55.

LUKE AND THE WAX TABLET REVISITED

355

mind consigns its unsolved problems. In the workshop all the space is needed for what is actual and what is probable”87. This division of an argument’s playing pieces into “actual”, “probable”, and “merely possible” documents is helpful for a metacriticism of synoptic source criticism. Supporters of the 2DH naturally think of Q as a “probable” document, but they have made an effort recently to characterize Farrerian Luke as scarcely even possible88. I have tried to show that a proper understanding of ancient compositional practices retains and even improves the probability of a Farrerian Luke89. Independent scholar United States of America

John C. POIRIER [email protected]

87. L.A. MUIRHEAD, review of P. Wernle, Die Synoptische Frage, in Critical Review of Theological and Philosophical Literature 12 (1902) 205-208, p. 205. 88. The realm of the possible may have opened wider with Idan Dershowitz’s recent argument for literal razor-and-paste redaction in the Hebrew Bible. I. DERSHOWITZ, The Dismembered Bible: Cutting and Pasting Scripture in Antiquity (FAT, 14), Tübingen, Mohr Siebeck, 2021. 89. I wish to thank the editors for their careful review and for correcting a well-hidden citational error.

NEW AVENUES

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

The challenge of explaining the literary relationship of the Synoptic Gospels stands at the core of the Synoptic Problem. Since my own picture of the Gospel tradition largely differs from all of the theories under discussion so far, I consider it appropriate to provide here a guide to some of its cornerstones. Therefore, the following gives a brief outline of (1) my theory’s foundational assumptions and (2) the resulting conclusions concerning the tradition history of the Gospels1. Only afterwards can (3) the literary character of the Gospel tradition and its methodological implications be addressed, indicating the proposed model’s impact on some of the questions in the area of Gospel production, such as orality or scribality. The conclusion (4) attempts to shed some light on the principal assumption shared by all of the theories under discussion. I. THE BASIC PREMISE

FOR

RECONSTRUCTING THE GOSPEL TRADITION

Instead of exclusively concentrating on the interrelations of the Synoptic Gospels, my proposal seeks to explain the tradition history of all four Gospels. This includes, of course, an explanation of the synoptic interrelations. It differs, however, in one basic assumption from the prominent synoptic theories dominating the recent scholarly debate, i.e., the TwoSource Theory (2ST) and the Mark-Without-Q Hypothesis (MwQH)2. 1. The following is a very brief summary of M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021 (here cited: OG). For a similar summary in German cf. ID., Überlieferungsgeschichte der kanonischen Evangelien, in ZNT 22 (2019) 39-60. 2. Three short remarks on terminology: (1) With respect to the 2ST, I prefer to speak of a ‘theory’ rather than of a “hypothesis”, since the 2ST combines diverse observations and prerequisites to a complex system which claims to explain more than simply the literary relations between the synoptics. In terms of the philosophy of science, the criteria required to validate a “theory” are well defined (extension, consistency, falsifiability, etc.). In spite of a widespread prejudice, a “theory” does not carry a higher degree of certainty than a “hypo-

360

M. KLINGHARDT

The principal difference between the 2ST and the MwQH is the assumption of an additional source (‘Q’) by the 2ST, which the MwQH counters by relying on the simple interdependency of the three Synoptic Gospels. The basic question at the heart of all the synoptic debates since the end of the 19th century is whether it is possible to explain the complex synoptic interrelations without postulating an additional source. As a methodological challenge, this problem is interesting insofar as it reveals the complexity of the materials under discussion. As an historical problem, however, it is insignificant. For there does exist an additional text that precedes the canonical Gospels and serves as their source. It is the Gospel that belonged to the collection of eleven writings (one Gospel and ten Pauline letters) used by Marcion and the Marcionites. Based on the heresiological attestation, this Gospel can be reconstructed fairly well. Its title is simply “Gospel” (εὐαγγέλιον, or evangelium, respectively); below, I refer to it as *Evangelium (or *Ev, in short)3. In contrast to the widespread opinion ranging from Justin Martyr, Irenaeus, and Tertullian to Harnack and beyond, my theory postulates that *Evangelium is not an edited version of Luke but its most important source: the editorial direction clearly runs from *Evangelium to Luke, not the other way around. *Evangelium is therefore a pre-Lukan text. This changes the basic configuration of the synoptic interrelations considerably: the resulting model is in many ways different from all the theories that have been discussed since the beginning of the 19th century. An analysis of the synoptic tradition must take into consideration four texts: *Evangelium, Mark, Matthew, and Luke. To place them in a diachronic development, it is necessary to compare these four texts to each other and to determine the relative literary dependence between them. Methodologically, this requires only well-established instruments: a simple source-critical analysis which relies on determining the editorial direction, thesis”. (2) Also, in “2ST” I prefer the more general term “source” to “document”, for the latter implies unproven premises which are better avoided (by definition, a document is a text that exists only once; a copy is not document). (3) Finally, the designation “Mark-WithoutQ Hypothesis” (MwQH) follows a suggestion of J.S. KLOPPENBORG, On Dispensing with Q? Goodacre on the Relation of Luke to Matthew, in NTS 49 (2003) 210-236, pp. 212f. I agree with it, for it refers to the model’s factual logic rather than to one representative. The wellknown alternative “Farrer Theory” is not very accurate, since the basic model is both older than Austin Farrer’s 1955 article and has been promoted on better grounds by other scholars. 3. The asterisk indicates that the text is not directly attested in manuscripts but must be reconstructed based on secondary witnesses. In accordance with the long-standing practice of Classical studies to use Latin titles for ancient writings irrespective of their language (e.g., “Plato, resp.”), I use the Latin evangelium (rather than the Greek εὐαγγέλιον). J.D. BEDUHN, The First New Testament. Marcion’s Scriptural Canon, Salem, OR, Polebridge, 2013, chose a mix of Greek and Latin (evangelion) in order to give this text a distinct name.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

361

based on the redaction-critical plausibility. Surprisingly, this is not a challenging task and can be done without generating an unsolvable problem. The main result is that *Evangelium is not only older than Luke but also precedes Mark and Matthew; therefore, it is the oldest Gospel known so far. The tradition history resulting from the source-critical comparison does not only apply to *Evangelium and the Synoptic Gospels but also to John. The tradition history develops in four stages from *Evangelium to Mark to Matthew to John and then to Luke: *Evangelium represents the oldest stage of the Gospel tradition, Luke the latest. Many observations indicate that every single stage of the tradition process presupposes all respective previous texts. This is a rough outline of the model. Since the whole model depends on the basic insight of *Evangelium-priority over Luke, this element should be substantiated in more detail; even more so, because it opposes a very long tradition, which already started in the 2nd century and is still prominent today. According to this tradition, Marcion, the ‘arch-heretic’, supported his own theological proclivities by altering (basically shortening) canonical Luke: The “Pontic rat gnawed the Gospel to pieces” (Tertullian) which now has holes like “a shirt eaten away by moths” (Epiphanius)4. This portrait of Marcion who “cleansed” and edited Luke according to his theological ideas has been conveyed to 20th century scholarship by Harnack’s influential book on Marcion5. Yet, this portrait is untenable. Doubts should have been raised because of Marcion’s reverse accusation. The Marcionites claimed that the “defenders of Judaism” falsified and interpolated *Evangelium and combined it with the Law and the Prophets. Tertullian who reports this accusation states: “I say that my (Gospel) is true, Marcion makes that claim for his; I say that Marcion’s is falsified, Marcion says the same of mine”6. Since both assertions contest each other mutually, the tug-of-war over the question of the editorial direction remains unresolved, although Tertullian struggles to win. Like the other heresiologists, he refutes the “heretical” theology using Marcion’s own Gospel (i.e. *Evangelium) as the textual basis and insists: “I am sorry for you, Marcion: your labor has been in vain. The Christ Jesus in your gospel is mine”7. Despite this glorious declaration of victory, one must ask: How is it possible that the heresiologist could refute 4. Quotations from Tertullian, Marc. 1,1,5, and from Epiphanius, Haer. 42,11,3. 5. A. VON HARNACK, Marcion: Das Evangelium vom fremden Gott. Eine Monographie zur Geschichte der Grundlegung der katholischen Kirche (TU, 45), Leipzig, Hinrich, 21924. 6. Tertullian, Marc. 4,4,1. 7. Tertullian, Marc. 4,43,9.

362

M. KLINGHARDT

Marcion’s theology based on Marcion’s own Gospel, if Marcion’s altering, editing, and cleansing of Luke aimed at a consistent conformity with this very theology? The logical contradiction is irresolvable and proves the claim of Marcion’s alleged edit wrong. Tertullian realized this contradiction but was unable to rebut it: his attempt at an explanation is embarrassingly helpless8. Whether Tertullian believed his own argument or not, the editorial direction between *Evangelium and Luke must be determined based on a comparison of the texts themselves. This can be done without difficulty, and with surprising unambiguousness. Since the text of *Evangelium can be reconstructed based on the heresiological attestation (next to Tertullian, Epiphanius and the Adamantiusdialogues are the most important sources) it is wise to concentrate first on unambiguously attested passages to avoid a circular argument. The most prominent example is the well-attested beginning of *Evangelium: Compared to Luke, it lacked the prologue (Luke 1,1-4) and the childhood stories (1,5–2,25) as well as the complete tradition about John, the baptism, the genealogy, and the temptation (3,1aβ-38). Instead, *Evangelium began with the dating (*3,1aα) and the exorcism in Capharnaum (*4,31-37), followed by the pericope of Jesus’ rejection in Nazareth (with a considerably different profile: it contained only *4,16.23f.28-30), and then *4,42f.9. A sourcecritical comparison of both texts indicates the priority of *Evangelium: it is much more plausible that Luke edited and supplemented *Evangelium, than that *Evangelium altered and abbreviated Luke. The most important single observation to consolidate this judgment is the position of the Nazareth pericope in both texts. In Luke 4,16-30 it has a fundamental, programmatic function. For that reason, Luke positioned this narrative at the very beginning of Jesus’ public service, even condoning the fact that this operation seriously flawed the narrative’s consistency. In Luke, the anaphoric reference to “the things that we have heard you did at Capernaum” (4,23) is aimless, since Luke reports the exorcism in Capernaum only afterwards (4,31-37). As has long been recognized, this unfortunate position is the result of Luke’s editorial intention to create the programmatic scene and,

8. Tertullian made Marcion responsible for the weakness of his argument and accused him of a perfidious cover-up: “Now here Marcion, on purpose I believe, has abstained from crossing out of his gospel certain matters opposed to him, hoping that in view of these which he might have crossed out and has not, he may be thought not to have crossed out those which he has crossed out, or even to have crossed them out with good reason” (Marc. 4,43,7). Cf. OG (n. 1), pp. 127-133. 9. Cf. ibid., pp. 143-165, 512-535. An asterisk before chapter and verse always refers to the reconstructed text of *Evangelium (English translation available, ibid., pp. 1281-1322).

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

363

therefore, secondary. Yet, Luke did not editorially change Mark 6,1-6 but the sequence of *Evangelium. The reverse explanation (that Marcion altered the Lukan sequence) is highly improbable. Not only is it that none of Marcion’s alleged theological motives would explain this correction, but this solution still must also assume a text prior to Luke, which then was disimproved by the Lukan redaction. Furthermore, Luke not only re-positions this pericope, but also complements it. It comes as no surprise that the key verses expressing Luke’s editorial concept (Luke 4,17-22.25f.) are lacking in *Evangelium: they are redactional additions by Luke. Only the mention of Naaman the Syrian (Luke 4,27) is found in *Evangelium, yet not in the Nazareth pericope but in the different context of the Healing of the Ten Lepers (most probably after *17,18). Concerning the transposition of this verse, too, one could ask which editorial direction is more plausible: did *Evangelium find it in Luke 4 and transferred it into the story of the Ten Lepers, or did Luke find it in this context and integrated it in his newly created programmatic Nazareth scene? The answer to this question again is unambiguous: Luke edited *Evangelium, not the other way round. This is but one of several examples establishing the basic editorial direction between *Evangelium and Luke, regardless of an exact reconstruction of *Evangelium. A detailed verse-to-verse comparison between both texts makes the following conclusion clear: In a great number of instances, *Evangelium-priority over Luke is cogent or at least highly probable. Then there is a rather small group of instances for which the editorial direction cannot be established unambiguously. Lacking throughout, however, are observations leading to the assumption of Luke-priority10. Based on this observation, critical scholarship from early on had objected against the traditional view of Luke-priority: Marcion’s alleged redaction of Luke cannot be established; it is a sheer claim without source-critical substantiation or, as Semler already put it with regard to Tertullian’s argument, it is “declamation, not honest history”11. This source-critical result confirms the Marcionites’ claim that Luke is an interpolated redaction of *Evangelium. Historically, this is probable. For a charge similar to that directed against the Marcionites is attested also against other “heretics”: they are

10. This is the general result of the detailed reconstruction of *Evangelium. Cf. ibid., pp. 511-1279. 11. J.S. SEMLER, Thomas Townsons Abhandlungen über die vier Evangelien: Mit vielen Zusätzen und einer Vorrede über Marcions Evangelium, Leipzig, Weygand, 1783, p. 26 (the quote is found in the Vorrede, although the pages are not numbered).

364

M. KLINGHARDT

said to have used Luke only but altered it according to their discretion12. From the end of the 2nd century onwards, there were Christians who used only a short version of Luke but not the canonical four-Gospel-book. In the perspective of the emerging orthodox church and its heresiologists, the sheer use of a “deficient” bible defined the “heretics”. To sum up: Tertullian’s report of the Marcionites’ claim that “their” Gospel (i.e. *Evangelium) is older than Luke is not only historically plausible, it can also be substantiated with a redaction-critical comparison. All the evidence points in the same direction: Luke is an edited version of *Evangelium; *Evangelium is older than Luke13. II. RECONSTRUCTING THE GOSPELS’ TRADITION HISTORY The consequence of the preceding discussion is as follows: the position of *Evangelium in the history of the Gospel tradition is prior to Luke, in a place analogous to Mark and “Q” in the 2ST, or Mark and Matthew according to the MwQH, respectively. Thus, the premise for reconstructing the Gospel tradition proposed here differs significantly from all models discussed in the last 200 years. Establishing a diachronic model of the emerging Gospel tradition on this premise would require a source-critical comparison between each two Gospels. Since some relations are already well established (most importantly, Mark-priority over Matthew), it may suffice to concentrate on the critical examples. 1. *Evangelium and Mark Since *Evangelium is similar to Luke, many observations establishing the editorial relation between Luke and Mark are similar to those between 12. Cf. Origen, Hom. Luc. 16,5 (“there are innumerable heresies receiving the Gospel of Luke”); 20,2 (“all heretics who receive Luke condemn what is written in it”). Next to the Marcionites, Origen reckons also the Valentinians and the followers of Lucanus among those heretics receiving Luke only (Cels. 2,22); cf. Irenaeus, Haer. 3,15,1f. (the Valentinians use only Luke but not Acts). 13. In spite of many differences in details and consequences, this basic insight is shared, e.g., by: BEDUHN, The First New Testament (n. 3); P.A. GRAMAGLIA, Marcione e il Vangelo (di Luca), Torino, Accademia University Press, 2017; D.A. SMITH, Marcion’s Gospel and the Synoptics. Proposals and Problems, in J. SCHRÖTER – T. NICKLAS – J. VERHEYDEN (eds.), Gospels and Gospel Traditions in the Second Century: Experiments in Reception (BZNW, 235), Berlin – New York, De Gruyter, 2019, 129-173; M. VINZENT, Methodological Assumptions in the Reconstruction of Marcion’s Gospel (Mcn). The Example of the Lord’s Prayer, in J. HEILMANN – M. KLINGHARDT (eds.), Das Neue Testament und sein Text im 2. Jahrhundert (TANZ, 61), Tübingen, Francke, 2018, 183-222.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

365

*Evangelium and Mark. To determine the editorial direction, it seems advisable to begin by analyzing the great differences. Thus, the most prominent features in need of an explanation are the lack of an equivalent of Mark 6,45–8,26 in Luke (in the traditional terminology of the 2ST: the “great omission”) and that Luke 9,51–18,14 has only marginal analogies in Mark (i.e., the “great insertion”). In these aspects, *Evangelium and Luke are very similar. Almost all the materials from Luke 9,51–18,14 are attested for *Evangelium which, on the other hand, does not betray any knowledge of Mark 6,45–8,26. Of course, the terms ‘omission’ and ‘insertion’ imply Mark-priority as the key element of the 2ST; using this terminology to describe the relation between Mark and *Evangelium would beg the question of the editorial direction. In order to determine the relation between Mark and *Evangelium, the following questions must be answered: Did *Evangelium insert *9,51–18,14 into a Markan Vorlage or did Mark omit this material when editing *Evangelium? Did *Evangelium leave out Mark 6,45–8,26 or did Mark add this material to his source? An answer to these questions provides surprising insights. First, Markpriority over Luke, being the basis of both the 2ST and the MwQH, appears to be so firmly established that these great differences are seldom substantiated source-critically. That these features are, in fact, an “omission” or an “insertion”, respectively, is usually taken for granted: there remains a serious gap of substantiation. Discussions of the Synoptic Problem in the past concentrated so much on the details or aspects of the respective theories that these great differences attracted only marginal attention. Second, an attempt to close this gap reveals a methodological concern that is not always respected. The lack of material cannot be reliably explained by a sourcecritical comparison of only two texts; it always requires further information. Usually, such additional information is tacitly derived from the greater (diachronic) model that is supposed to be the result of the comparison in the first place. Keeping this problem in mind helps one understand the first example, the non-existence of *9,51–18,14 in Mark. The reasons for its absence cannot be substantiated; they can only be guessed14. The reverse solution (*Evangelium ‘inserted’ this material into its source, Mark) is of little help, either. It remains as inconclusive as the many attempts to find a well-reasoned literary structure in the respective Lukan section, the bulk of the so-called travel narrative (though not its entirety). Since it is almost 14. The most recent objection against Mark-priority over *Evangelium (SMITH, Marcion’s Gospel and the Synoptics [n. 13], p. 145) seems to put more confidence in this methodologically weak premise than I find advisable.

366

M. KLINGHARDT

impossible to detect a meaningful literary structure for the whole section, little can be said about a possible editorial intention that would explain its insertion in *Evangelium. Thus, this example cannot contribute substantially to determining the literary direction between *Evangelium and Mark. The other example, however, is more significant. On the one hand, the reasons why *Evangelium does not contain Mark 6,45–8,26 can only be guessed. On the other hand, the editorial plausibility of a Markan addition to *Evangelium is obvious: Mark’s supplementing 6,45–8,26 created an ambitious and elaborate narrative unit in which every single element has a meaningful and well-reasoned narrative function at its exact position, advancing the narrative and connecting different story lines. The section fits neatly into the literary and theological concept and presents itself as an editorial improvement15. When the alternative editorial directions between Mark and *Evangelium are weighed against each other, these examples of great differences clearly point to *Evangelium-priority over Mark. There are further observations substantiating this result, of course, not only with regard to the omission or addition of material, but also with regard to the change of phrases and single words16. 2. *Evangelium, Mark, and Matthew In the next step, Matthew becomes part of the emerging picture. Assuming Mark’s priority over Matthew, which needs no further substantiation17, this does not prove difficult. Structurally, Matthew’s integration into the tradition is similar to other models reckoning with Mark-priority: Matthew integrated additional material into his main source, Mark. According to the 2ST, part of this additional material came from the hypothetical source “Q”; according to the present model, the source is the well-attested *Evangelium.

15. M. KLINGHARDT, Boot und Brot. Zur Komposition von Mk 3,7–8,21, in BTZ 19 (2002) 183-202. 16. Cf. OG (n. 1), pp. 199-237. 17. Matthew’s dependence on Mark is so well-founded that it is hardly in question. The Griesbach- or Two-Gospel hypothesis (William R. Farmer and others under his inspiration) is the only alternative discussed more recently. The merits for which it attracted some attention for a while lay in its being a resort from the aporias of the 2ST rather than in its convincing source-critical observations.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

367

For a demonstration, the composition of the Sermon on the Mount can serve as the prime example: it is fiercely debated between the representatives of the 2ST and MwQH. The materials of the Sermon on the Mount mostly belong to the double tradition and, according to the 2ST, came from “Q” and were used by Luke and Matthew according to their respective editorial concepts. In Luke, these materials are distributed over about a dozen different places, some of which appear somewhat infelicitous. An explanation assumes that Luke preserved the original order closer than Matthew did; the bumpy distribution is therefore due to “Q” (and not to Luke’s editorial intention) and thus poses no further problems. In Matthew, however, the composition makes a lot of sense and is hardly debatable. The 2ST seems to provide a satisfying explanation. The MwQH, however, faces a serious problem. It must assume that Luke dissolved the Matthean composition and then scattered its material across his gospel. This unlikely assumption has been criticized for a long time and provoked Streeter’s famous judgment that “such a proceeding would only be tenable if, on other grounds, we had reason to believe he was a crank”18. Assuming *Evangelium-priority over Matthew avoids this problem. Luke was no crank but followed the sequence of his source, *Evangelium. Besides integrating the materials from *Evangelium into his Gospel, Matthew made other additions as well. Among them, the infancy stories (Matthew 1f.) provide an instructive example. While the MwQH has no problems understanding Luke 1f. as an extrapolation from Matthew 1f., the 2ST cannot satisfyingly explain the similarities and differences between both accounts; usually, they are simply ignored. On the one hand, both accounts appear to be too different to be derived from the same source; on the other hand, the analogies are much too striking to be accidental. The birth-stories are original in Matthew who uses them to demonstrate Jesus’ Davidic parentage. In doing so, Matthew introduces the topic of Jesus’ legitimate claim of power into his narrative, which creates a bridge to the very end of his Gospel. Luke depends on the Matthean birth-story, adopts some important elements (e.g., virginal conception; arranging Jesus’ birth in history; naming Jesus etc.), corrects some aspects, and changes the narrative perspective from Joseph to Mary in the central passages (announcement; consenting to God’s plan); a logical technique if the goal was to integrate additional aspects into a known story without creating discrepancies.

18. B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, p. 183.

368

M. KLINGHARDT

3. *Evangelium, John, and Luke Perhaps surprisingly, the next stage in the proposed tradition history is not Luke but John. The narrow focus on the Synoptic Problem in scholarship of the last two centuries is understandable in light of the complexities: it was difficult enough to reconcile the literary relations among the Synoptic Gospels within a plausible model; to many, incorporating John into the picture seemed too great a task. However, both methodologically and factually, excluding John from the reconstruction of a coherent Gospel tradition proved to be an obstacle. More and more, research bifurcated into separate discourses on the Synoptic Problem and on the “Johannine question”, which rarely interacted with each other. In retrospect, this proved to be a disadvantage for both camps. The dominance of the 2ST (at least in German scholarship) obscured the fact that there are many important similarities between Luke and John against Mark and Matthew, predominantly in the Passion-narrative. The special relation between Luke and John is incompatible with the 2ST19. Attempts to force a solution and to harmonize both aspects must postulate up to five additional sources; this is not a solution but the confession of an aporia20. These and other observations have led some scholars to assume John-priority over Luke21. In spite of very convincing observations pointing to such priority, these solutions were not widely received. This may be due not only to the dominance of the 2ST but also due to other features that point into the opposite direction of Lukepriority. Here, the reference of Luke 24,12 to John 20,3-10 is but one prominent example. Since the evidence pointing to Luke-priority (mostly the layout and many details in the Passion narrative) is already present in *Evangelium, the ambiguity of the editorial direction is easily explained: John used *Evangelium and John was used by Luke. Taking into consideration that John also knew and used Matthew22, the development of the Gospel tradition can be summarized in a diagram: 19. Cf. M. WOLTER, The Gospel According to Luke, Volume 1 (Luke 9:51–24) (BMSEC), trans. Wayne Coppins and Christoph Heilig, Waco, TX, Baylor University Press, 2017, p. 443: “No model of the tradition history exists, which could explain this outcome without leaving questions open”. 20. Cf. H. KLEIN, Die lukanisch-johanneische Passionstradition, in ZNW 67 (1976) 155186; F. SCHLERITT, Der vorjohanneische Passionsbericht. Eine historisch-kritische und theologische Untersuchung zu Joh 2,13-22; 11,47–14,31 und 18,1–20,29 (BZNW, 154), Berlin – New York, De Gruyter, 2007; OG (n. 1), pp. 290ff. 21. E.g., B. SHELLARD, The Relationship of Luke and John: A Fresh Look at an Old Problem, in JBL 46 (1995) 71-98; M.A. MATSON, In Dialogue with Another Gospel? The Influence of the Fourth Gospel on the Passion Narrative of the Gospel of Luke (SBL DS, 178), Atlanta, GA, Society of Biblical Literature Press, 2001. 22. Cf. J.W. BARKER, John’s Use of Matthew (Emerging Scholars), Minneapolis, MN, Fortress, 2015.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

369

*Evangelium Mark Matthew John

Luke Fig. 1: The editorial relations between the Gospels

This diagram shows all the editorial relations23. It is a simple development from *Evangelium over Mark, Matthew, and John to Luke, where each stage used all the texts available from the previous stage(s). Thus, the tradition appears as a tightly-knit net, or rather as a tree which arises from a single root and around which the branches of the subsequent tradition entwine. 4. The Canonical Redaction of the Gospels A last addition is necessary, though. The Gospels representing the stages of the tradition are not identical to the canonical Gospels. A final (“canonical”) redaction went over all the previous texts. This redaction unified the Gospels into one coherent collection: it added the uniformly formulated titles with the ascriptions to the respective evangelists. It also increased the coherency between the four parts of this (sub-)collection, smoothing out a few differences and inserting narrative conjunctions between the Gospels. One example of such coherency signals was mentioned above: Peter’s race to the empty tomb (Luke 24,12). If read exclusively within the narrative of Luke, this verse appears enigmatic and serves no recognizable narrative function. If, however, it is read within the collection of the four-Gospel book, this verse constitutes a link to John 20,3-10, thus indicating that John’s narrative about the footrace to the empty tomb is correct although no other Gospel mentions it. Another, much more prominent example is 23. The dotted arrow indicates that John’s use of Mark is the only relation that cannot be substantiated positively, mostly because the evidence is inconclusive since the material is also contained in Matthew whom John certainly used, or in *Ev.

370

M. KLINGHARDT

the longer ending of Mark (16,9-20): its coherency signals establish the narrative unity between all four Gospels as well as Acts. Typically, the evidence for this element can be obtained only by means of textual criticism: that Mark 16,9-20 is a later addition to an older version of Mark is clear from the manuscript evidence alone. Luke 24,12 is similarly text-critically disputed. It is not present in D 05 and a few Old Latin manuscripts; also, and significantly so, it is unattested for *Evangelium. This brings us back to the beginning, the priority of *Evangelium, for this aspect makes it clear that Luke’s redaction of *Evangelium is part of the fourth and final step of the canonical redaction. The picture of the complete Gospel tradition should therefore also indicate that the previous versions are only hypothetically concluded (hence the asterisks in the diagram).

*Evangelium *Mark

I

*Matthew

II *John

Matthew Mark

Luke

John

III IV

Fig. 2: The canonical redaction of the Gospels

The bold arrows indicate the work of the final redaction that was responsible for the four-Gospel book as we know it from the NT manuscripts. This implies that the Gospels did not gain their canonical textual form until the time when the four-Gospel collection was completed. As the connection between Luke and Acts or the cross-references in Mark’s longer ending demonstrate, this coincided with the completion of the canonical New Testament24. Although this reconstruction of the Gospel tradition and the idea of the canonical redaction of the NT are independent of each other, these two theories are compatible and therefore strengthen the historical probabilities of the reconstruction. 24. Cf. D. TROBISCH, The First Edition of the New Testament, Oxford, Oxford University Press, 2000.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

371

III. THE GOSPEL TRADITION AND ITS LITERARY CHARACTER: SOME METHODOLOGICAL CONSIDERATIONS Based on this model of the tradition history, I would like to contribute to a discussion of some aspects of the editorial work that must be assumed for the Gospel tradition. I will concentrate on a few basic, methodological aspects only. As indicated above with respect to the inconsequential application of source-criticism, it is reasonable to check from time to time the methodological foundations of our scholarship and to ascertain that they still apply to the questions we ask. The biggest problem is awareness of one’s own presuppositions. The more evident they appear, the less one feels the need to explain them. In some ways, we seem to leave the presuppositions behind us, where we cannot see them. Yet, as the discourse between the 2ST and the MwQH proves, others can see them clearly: the critical arguments raised against the respective other position are much more convincing than the arguments supporting one’s own view25. Therefore, I will use the model sketched out above to focus on the shortcomings of the other theories, first in general, than with respect to “scribality”. I must leave it to others to be critical of my presuppositions in hopes that they will uncover the insufficiencies of my argumentation. 1. Synoptic Theories or Tradition History of the Gospels: Assessing Alternate Theories A close look at the presuppositions of the 2ST and the MwQH reveals categorically diverse results. To begin with the 2ST, its weakness is not that it is just a mere hypothesis (as Austin Farrer bluntly declared)26 or that it is the “corollary of the hypotheses of Markan priority and the independence of Matthew and Luke” (as John Kloppenborg precisely corrected)27. Rather, it is the “Minor Agreements” that prove to be fatal to the 2ST. If they are not “explained away” in one way or another (none of which is even rudimentarily satisfying), they prove a literary link between Matthew and Luke that, for methodological reasons, must not exist. For if, conversely, such a link existed, the postulate of “Q” would be inadmissible: it contradicts the 25. Cf. M. KLINGHARDT, The Marcionite Gospel and the Synoptic Problem: A New Suggestion, in NT 50 (2008) 1-27, pp. 1-4. 26. A.M. FARRER, On Dispensing with Q, in D.E. NINEHAM (ed.), Studies in the Gospels: Essays in Memory of R.H. Lightfoot, Oxford, Blackwell, 1955, 55-88, p. 66: “The Q hypothesis is a hypothesis, that is its weakness”. 27. KLOPPENBORG, Dispensing (n. 2), p. 211.

372

M. KLINGHARDT

independence of Matthew and Luke, the very foundation on which the postulate of “Q”, the central element of the 2ST, is based in the first place. The 2ST is flawed by an internal logical inconsistency and, therefore, is principally untenable on the grounds of methodology. I contend that it is indeed “built on sand”28. This insight, of course, is neither new nor original. I should nevertheless emphasize said insight to indicate two consequences. First, together with the 2ST, its terminology no longer makes sense. The use of designations like “Q”, “Q-material”, “Q-community” or “Minor Agreements” in the context of explaining synoptic interrelations muddies the water, to say the least. This terminology still perpetuates the logic of the 2ST. Particularly irritating is the use of this terminology when assuming some form of *Evangelium-priority over Luke. Because *Evangelium explains the Matthean-Lukan agreements against Mark in the material shared with Mark, there is simply no legitimate place for “Q”29. The 2ST and the model based on *Evangelium-priority (hereafter, *EPM) are mutually exclusive. The second insight relates to the critical assessment of the 2ST in comparison to other models such as the MwQH or the Two-Gospel Hypothesis (2GH)30. Whereas the MwQH and the 2GH can, and must, be weighed against each other in order to determine their respective greater or lesser plausibility or (as F. Gerald Downing corrected)31 probability, this is not the case with the 2ST: for the MwQH, the 2GH, and other models, their explanatory power or ‘plausibility’ can be assessed only in comparison. The 2ST, on the other hand, fails in itself. For the MwQT and the 2GT, such a comparison is possible for they both share the same methodological premises: they provide a source-critical explanation of the synoptic interrelations based on the three Synoptic Gospels alone, without considering additional sources. Obviously, such a comparison is not possible with the *EPM, for the basic premises are far too different. Including *Evangelium adds a completely new element to the 28. Cf. C. HEIL – G. HARB – D.A. SMITH (eds.), Built on Rock or Sand? Q Studies: Retrospects, Introspects and Prospects (BiTS, 34), Leuven – Paris – Bristol, CT, Peeters, 2018; despite the intriguing alternative indicated by the title, the possibility that “Q” is, indeed, “built on sand” is not even discussed in this volume. 29. Against BEDUHN, The First New Testament (n. 3), pp. 92-97; also, e.g., S. MATTHEWS, Does Dating Luke-Acts into the Second Century Affect the Q Hypothesis?, in M. MÜLLER – H. OMERZU (eds.), Gospel Interpretation and the Q-Hypothesis (LNTS, 573), London – New York, Bloomsbury T&T Clark, 2018, 245-265, pp. 260ff. 30. W.R. FARMER, The Synoptic Problem, New York, Macmillan, 1964, and others under his inspiration. 31. F.G. DOWNING, Plausibility, Probability, and Synoptic Hypotheses, in ETL 93 (2017) 313-337, pp. 314f.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

373

tradition of the Gospels which must be considered in reconstructing its history. In light of this categorical difference, how can this model be assessed and how can its explanatory value be determined? Two answers come to mind. The first answer is simple and states the obvious: Of course, one must use all available data, and this includes *Evangelium. The claim that the *EPM is (so far) the only theory to consider the complete evidence justifies its superiority over the MwQH or the 2GH. Consequently, it is insufficient to criticize the *EPM on the grounds of its allegedly smaller explanatory power in comparison to other theories. Such an attempt has been made by Hildegard Scherer who claims that, in a number of instances, the 2ST offers better explanations for the synoptic phenomena than the *EPM32. Apart from her examples being debatable and the 2ST being refuted by its internal inconsistency, this critique is insufficient for it simply neglects a portion of the available data, i.e. *Evangelium. Instead, Scherer uses the relative plausibility of the 2ST in order to refute *Evangeliumpriority. This is not acceptable. A comparison of two theories based on their respective explanatory value would be possible only if they were of the same methodological status, which, in this case, they are not. This leaves two methodologically valid possibilities to criticize the *EPM. One of them is to refute its methodological basis by disproving *Evangelium-priority over Luke. In order to avoid circularity, this requires refraining from the use of the 2ST and relying exclusively on a source-critical comparison between Luke and *Evangelium. Such critique is theoretically possible33 but may prove difficult. Alternatively, one could accept *Evangelium-priority over Luke but come to different conclusions with regard to the subsequent course of the Gospel tradition. I would not exclude the possibility that there are better ways to reconstruct the Gospel tradition (based on *Evangeliumpriority!) but no attempt has been made so far. For assessing the explanatory power of the *EPM, there are other aspects to be considered as well. One of them, as mentioned above, is the inclusion of John into the picture of the emerging Gospel tradition, which allows resolving the contradicting observations with regard to the debate on John and the synoptics. Without 32. H. SCHERER, Königsvolk und Kinder: Der Entwurf der sozialen Welt im Material der Traditio duplex (BBB, 180), Göttingen, Bonn University Press – V&R Unipress, 2016, pp. 61ff. 33. The need to emphasize this possibility is explained by the critique of C.M. HAYS, Marcion vs. Luke: A Response to the Plädoyer of Matthias Klinghardt, in ZNW 99 (2008) 213-232, p. 228, who raised the objection of “internal unfalsifiability” against the *EPM. Were this objection correct, it would deny the *EPM any scientific value. However, since *Evangelium-priority over Luke is possible but not necessary, the objection is unwarranted.

374

M. KLINGHARDT

reverting to categorically different approaches (such as, e.g., secondary orality), the *EPM explains John’s place in the Gospel tradition with the same methodological means of source-criticism as have been used for all other stages of the tradition. The *EPM can explain more data with less assumptions than any other theory. This theoretical economy could be an important element when assessing the plausibility of Gospel theories. 2. The Literary Character of the Gospel Tradition: Some Remarks on Scribality In the last two decades, the discourse concerning the understanding of the concrete processes of writing and composing the Gospels has gained momentum: How did ancient authors compose their texts? How must we envision the actual process of writing or dictating in which they were engaged? How did they use, consult, or cite other works? These questions (and many more in this area) are important, for they broaden our understanding of early Christianity and the social conditions under which the New Testament was first produced. What stimulated this debate, however, was the use of some of its results to assess the greater or lesser probability of certain synoptic hypotheses. Robert Derrenbacker’s Ancient Compositional Practices inaugurated this line of thinking and may serve as an example34. Based on insights from ancient authors whose editorial practices with regard to their use of sources can be determined, he concluded that ancient authors usually followed one source at a time, adapted its wording but did not rearrange its material in a large-scale way35. Although the synoptic interrelations with their micro-conflations and large-scale rearrangements of material are not “consistent with the known practices of writers in antiquity”, some are even less so than others36. This argument has been thoroughly debated – criticized, defended, diversified, corroborated, questioned, improved – ever since. 34. R.A. DERRENBACKER, Ancient Compositional Practices and the Synoptic Problem (BETL, 186), Leuven – Paris – Dudley, MA, Leuven University Press – Peeters, 2005; cf. ID., The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 435-458; ID., Ancient Literacy, Ancient Literary Dependence, Ancient Media, and the Triple Tradition, in W.E. ARNAL – R.S. ASCOUGH – R.A. DERRENBACKER – P.A. HARLAND (eds.), Scribal Practices and Social Structures Among Jesus Adherents. Essays in Honour of John S. Kloppenborg (BETL, 285), Leuven – Paris – Bristol, CT, Peeters, 2016, 81-95. 35. DERRENBACKER, Practices (n. 34), pp. 116f. 36. Ibid., p. 257.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

375

Since I contend that all known synoptic theories are unconvincing on other grounds, I cannot contribute to this debate. Instead, it is more useful to look at the line of arguing with which it is concerned. Although the main objective of this debate was to evaluate the relative probability of the known synoptic theories, some of the results became independent from this context and began to be used as an objection against the *EPM: presumably, this model does “not reckon with the conditions of ancient text-production and -reception”37. Although this very brief remark does not do justice to the sophistication of the larger debate, it reveals a fundamental problem. The procedure employed in evaluating evidence must be turned around, of course. The primary evidence is the Gospels and their literary data which reveal the intricate interrelations: this is the undebatable foundation. What ancient sources reveal about the writing practices and the procedures of composing and compiling texts constitutes secondary evidence at best; this evidence can only show how certain authors solved their particular tasks but hardly allows for a robust generalization. Applying this secondary evidence to the assumed procedure of Gospel writing is methodologically even farther removed. This would require a general comparability between both areas regarding the genre and functionality of the respective literature or the general working conditions; for some contributions to the debate, such comparability is debatable. With respect to the whole area of “scribality” and its application to Gospel writing, the hierarchy of categories unanimously refers to the Gospels and the Gospel tradition as the primary evidence from which all further investigation must proceed. It is not advisable first to separate the Gospels from ancient literature and then use the findings from this literature to evaluate whether the Gospel writing was possible in the first place. An answer to any of the following questions must include the Gospels: What were ancient writers capable of with respect to composition or compilation of sources? How closely were they able to observe several source-texts at a time, and for what length of passages? What kinds of (micro- or macro-) conflations or large-scale rearrangements were feasible under their particular writing conditions? Answering these questions must not only include the Gospels but also must begin with them, because they are the indisputable primary evidence. They provide some surprising insights into the abilities of ancient authors. These insights change considerably when taking the *EPM as the basis instead of the known synoptic theories. I will concentrate on the two most important aspects. 37. A. STANDHARTINGER, Review of M. Klinghardt, Das älteste Evangelium etc., in TLZ 112 (2016) 387.

376

M. KLINGHARDT

(1) Gospel Production and Gospel Tradition One characteristic difference between the *EPM and other synoptic theories is the close relation of all Gospels to *Evangelium: there is only one main source for the Gospel tradition. In this respect, the *EPM is closer to the MwQH than any other theory. Next to some consequences for the historical questions (see below), this “single root” model begs an inevitable question that otherwise was raised only rarely: If a Gospel was edited and, presumably, improved by the later stages of the tradition, why did it survive? Under the 2ST, this question applies to Mark only: if Matthew edited Mark and “Q”, incorporating Mark almost entirely and, thus, replacing it, why did Mark not disappear as it is assumed for “Q”? While this question could largely be ignored under the 2ST38, it is inevitable for the *EPM which reckons with four editorial stages of the tradition. The answer to this question is obvious. In contrast to a generally held assumption39, Matthew did not intend to replace Mark but to complement it. Complementation implies that a post-text is added to the pre-text and handed down together with it: “Gospel tradition” is the tradition of Gospel collection. It is possible that this process started with Matthew (or more precisely, *Matthew); it was most likely in place by the stage of the supplementation of *John. This formation process is both an extrapolation and a supplementation: a text with the same plot but a different, though similar, narrative is added next to the pre-text, thus creating a completely new work. This procedure allowed for amendments and small corrections without directly contradicting the other narratives. Mostly, other narrative perspectives were added. The Gospel tradition is a process of growth. The only verifiable replacement of a Gospel in this process is *Evangelium being superseded by Luke as part of the four-Gospel book. As is well known, this replacement did not go down quietly, the arguments between the emerging catholic church and the Marcionites being audible echoes in its wake. Clearly, the authors’ working method that must be assumed for this process is not adequately captured by “composition” based on “compilation of sources”, although both aspects are part of the picture. What the *EPM could add to the recent debate on “scribality” is an increased awareness of the literary character of the Gospel tradition. Production of a

38. I am aware of only one explicit attempt to answer this question: J. DEWEY, The Survival of Mark’s Gospel: A Good Story?, in JBL 123 (2004) 495-507. 39. E.g., G.N. STANTON, The Fourfold Gospel, in NTS 43 (1997) 317-346, pp. 341f.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

377

sophisticated multilayer text held together by a complex net of coherency signals requires a different level of literary ability than a mere compilation of sources. One example would be the phenomenon of similarity (or even identity) and variation in the Gospel tradition. Beyond the question of actual feasibility, all the aspects serve a literary function. Verbatim agreements underscore the reliability; after all, the same story is being told. Variations increase the credibility of an account from a different perspective, particularly if an “eyewitness” references events known from other “sources”. A prime example of this is “Peter’s” report of the heavenly voice in 2 Peter which is clearly based on different synoptic accounts yet it is not identical to any of them40. Large-scale rearrangements of material clearly express the author’s editorial intention, for example, the Matthean composition of the Sermon on the Mount based on *Evangelium. This, at least, is the perspective of source-critical analysis (regardless of whether the source is ‘Q’ or *Evangelium). The readers’ perspective is different: they must understand that the material of the Sermon on the Mount in Luke is correctly reported, as the prologue promises (Luke 1,1-4). Yet, the readers would probably prefer the account of the eyewitness Matthew, because from their perspective “Matthew” and “John” are presented as eyewitnesses in the same way as “Peter” in 2 Peter. Certainly, the authors who mastered intellectually challenging tasks like these were anything but simple “scribes”. With regard to the “feasibility”, the phenomena discussed in the scribality discourse appear as finger exercises when compared to the literary achievements of the Gospel tradition. Admittedly, the picture that the Gospel tradition developed from *Evangelium to the four-Gospel book in the complex manner described here cannot be proven. It is an interpretation of the available data: a sensible, yet unfamiliar, conclusion. Some might call it speculation. Yet, what is the alternative? The 2ST must assume that two single (independent) texts (Mark; “Q”) were in existence, which then were editorially worked together in two independent editorial acts (Matthew; Luke); later, all three texts found their way (again: independently?) into the New Testament Gospel collection. Although this alternative is theoretically possible, its plausibility is limited. By separating the formation of the individual texts from the formation of their collection, it must assume several independent acts. In comparison, and with regard to the theoretical design, this traditional picture of the emergence of the Gospels is certainly the more complex one.

40. Cf. 2 Pet 1,17; Matt 17,5; Mark 9,7; Luke 9,35.

378

M. KLINGHARDT

(2) Canonical Redaction of the Gospels The *EPM allows for another important insight in the literary character of the Gospel tradition, namely, as indicated above, the Gospels’ canonical redaction. Since in this final redaction the creation of canonical Luke coincides with the editorial harmonization of the previous Gospels, this element further corroborates the assumed unity of producing the Gospels and the collection. In order to validate the canonical redaction, it is necessary to turn to the more elementary level of manuscripts, for they provide the only traces for earlier stages of the pre-canonical versions of *Matthew, *Mark, and *John. For Luke, the primary witness of its earlier stage is *Evangelium, of course; yet, traces of this pre-canonical version can be found in the manuscript tradition (of Luke), too41. A brief, yet highly instructive example is the so-called Passion Predictions with the date of the Son of Man’s resurrection. In several instances, the Synoptic Gospels report that Jesus predicted that the Son of Man would suffer and be killed but resurrected “after three days” (μετὰ τρεῖς ἡμέρας), or “on the third day” (τῇ τρίτῃ ἡμέρᾳ), respectively. In the critical editions (NA28; GNT5), the phrase “after three days” appears in Mark 8,31; 9,31; 10,34, the phrase “on the third day” in Matt 16,21; 17,23; 20,19; Luke 9,22; 18,33. At first glance, this “Minor Agreement” (Matthew and Luke agreeing against Mark) is easily explained as an understandable redactional change: since the period from Good Friday to Easter Sunday is two days at best42, the change can be interpreted as a Matthean-Lukan editorial alignment of the prediction with the events actually told. Yet, the situation is more complex. In the manuscripts of all three Synoptic Gospels, both formulations are attested. In five of eight instances both attestations are found, in varying manuscripts. Only in three of eight instances do all of the manuscripts present the same formulation: in Mark 8,31 μετὰ τρεῖς ἡμέρας is unanimously attested, in Matt 20,19 and Luke 18,33 all manuscripts have τῇ τρίτῃ ἡμέρᾳ43. Thus, even in these instances with a uniform attestation the same ambiguity shows up as in all the other instances. In order to establish the Gospel tradition according to the *EPM, *Evangelium must be included in this survey: for *9,22, μετὰ τρεῖς ἡμέρας is attested, *18,33 is attested as lacking44. 41. For the text-critical implications of the *EPM, cf. M. KLINGHARDT, Die Schrift und die hellen Gründe der textkritischen Vernunft, in ZNT 39/40 (2017) 85-104; OG (n. 1), pp. 73-114. 42. Actually, it is only 39 hours from the “ninth hour” of Friday (Matt 27,46; Mark 15,38; Luke 23,45) until sunrise on Easter Sunday (Mark 16,2). 43. For the full presentation of the mss attestation, cf. OG (n. 1), p. 334. 44. Testimony for *9,22: Tertullian, Marc. 4,21,7 (post tertium diem); Epiphanius, Haer. 42,11,6(16) (μετὰ τρεῖς ἡμέρας); Adamantius, Dial. 5,12 (μεθ’ ἡμέρας τρεῖς). – Attestation

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

379

This irregular attestation has not been noticed very often. If the discrepancy between “after three days” and “on the third day” is commented on at all, then it is usually with a remark along the lines that “the tension between these two formulations puzzling modern scholars would not have troubled first-century Jewish exegetes”45. Had this discrepancy not puzzled the editor(s) responsible for the canonical redaction, the manuscript tradition would hardly have reflected it at all. The solution, in any case, is easily comprehensible. At the beginning of the tradition, *Evangelium used the formulation “after three days” (*Ev 9,22), reflecting the Danielic pattern of the eschatological week which is divided in two equal halves each of which lasts “one time, and (two) times, and a half time” (Dan 7,25; 12,7). In the widespread tradition based on Daniel, at this point in the middle of the week (that is, after three and a half days), God’s eschatological prophets are killed. At the end of the week, after another three and a half days, they are resurrected. “After three days” (*Ev 9,22) is a simplification of Daniel’s complicated formulation which is attested elsewhere in the same tradition, next to similar simplifications46. (2) *Mark 8,31 adopted this formulation from *Evangelium, supplemented it in the second instance where it was not present in his Vorlage (*Mark 9,31; *Ev 9,44), and created a third prediction anew (*Mark 10,34), using the identical formulation ‘after three days’. (3) Based on *Mark as his main source, *Matthew adopted the formulation in *Matt 16,21, 17,23, and probably even in 20,19. Apart from this last instance, the formulation μετὰ τρεῖς ἡμέρας is still visible in some manuscripts. (4) At the end of the tradition process, Luke aligned the prediction with the events of the narrative. Instead of the older formulation (μετὰ τρεῖς ἡμέρας) with its recognizable reference to the Danielic week pattern, he cautiously put τῇ τρίτῃ ἡμέρᾳ in Luke 9,22 and also in Luke 18,33 which he created anew, based on Mark 10,32-34 par. Matt 20,29-34. Since Luke 18,33 does not for *18,33 as missing: Epiphanius, Haer. 42,11,6(52). – Luke 9,44 and *9,44 agree with each other in not presenting a prediction of the resurrection. 45. H.K. MCARTHUR, ‘On the Third Day’, in NTS 18 (1971) 81-86, p. 85. 46. The simplifications include “they were dead three days” and “on the fourth day” (for full references cf. OG [n. 1], pp. 720f., n. 9). The formulation “after three days” is attested by: Latin Tiburtina (E. SACKUR, Sibyllinische Texte und Forschungen. Pseudo-Methodius, Adso und die Tiburtinische Sibylle, Halle, Max Niemeyer, 1898, p. 186: post tres dies a Domino suscitabuntur); Adso, Antichr. (J.-P. MIGNE [ed.], Patrologia Latina, 101, Paris, Migne, 1851, 1289-1298, col. 1297a: post tres dies a Domino suscitabuntur); cf. Lactantius, Inst. Div. 7,14,3 (S. BRANDT – G. LAUBMANN [eds.], L. Caeli Firmiani Lactanti Opera omnia [CSEL, 19], Prague – Vienna, F. Tempsky, 1890, p. 638: post diem tertium reviviscet). For the complete background, cf. K. BERGER, Die Auferstehung des Propheten und die Erhöhung des Menschensohns. Traditionsgeschichtliche Untersuchungen zur Deutung des Geschickes Jesu in frühchristlichen Texte (SUNT, 13), Göttingen, Vandenhoeck & Ruprecht, 1976, pp. 107ff.

380

M. KLINGHARDT

have a pre-text in *Evangelium, it is no surprise that in this instance there are no variants in the manuscripts. Likewise, this formulation appears where Luke formulates on his own, without relying on any of his sources (e.g., Luke 24,7). (5) Luke – or more precisely: the same editorial hand that created “Luke” – also revised the other pre-canonical Gospels and replaced the old formulation “after three days” with “on the third day”. Only one single instance might have escaped his attention: there are no variants in the manuscripts for Mark 8,31 with the pre-canonical formulation. This example is instructive in several respects. Leaving aside the theological insights, it demonstrates that, at the end of the Gospel tradition, a thorough redaction revised the pre-canonical texts and harmonized them. Since the completion of the canonical four-Gospel book most probably happened at the same time, and was part of the same enterprise as the production of the canonical edition of the New Testament, it is likely that this revision also included all other NT writings. It is therefore not possible to say whether “on the third day” in 1 Cor 15,4 stems from a pre-canonical version or from the canonical redaction. The claim that the editorial changes from “after three days” to “on the third day” are part of one uniform editorial act is easily substantiated. If the editor intended a consistent alignment of the prediction with the events, there would have been several ways to do so, “after two days” or “after one and a half day” would have worked as well, if not even better. The variety of simplifications of the complicated Danielic formulation that is attested in the widespread tradition indicates what range of formulations was readily available. Yet the redaction chose only one alternative and inserted it in all instances: this would seem to be the result of one encompassing redaction. The unifying redaction at the end of the Gospel tradition as part of the *EPM adds a completely new dimension to the “scribality” debate that goes far beyond the adaptations, variations, and conflations that must be assumed for the earlier stages of the tradition. The emergence of the Gospel tradition was a “scribal” act from beginning to end, involving up to four written sources. The editors of the pre-canonical Gospels (according to the four assumed editorial stages of the tradition) were highly capable authors whose literary skills combined a high degree of diligence in their handling of the sources with creativity exhibited in inventing new story-lines in order to pursue their well-reflected upon theological aims47. They were anything but mere “compilers”. The canonical edition was a large-scale literary enterprise of text production. Even limiting the consideration to the completion of the 47. For an outstanding example, cf. KLINGHARDT, Boot und Brot (n. 15).

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

381

four-Gospel book (and leaving aside the editorial work on the other NT writings), this is an enormous task for which the work of an experienced team is easier to imagine than a one-man show. What does not come to mind at all is a single “scribe”. 3. Gospels as Literature What comes to mind even less is orality. In light of these results, there is obviously no room for an “oral tradition” as a means of explaining the complex literary relations between the Gospels. Under the *EPM, all the intricacies that constitute the Synoptic Problem – the similarities, verbatim agreements, dissimilarities, rearrangements, etc. – can be explained as purely textual phenomena in the process of editing and re-editing of texts. There is simply no need to postulate oral tradition. To propose this neither means that people did not talk about the Gospels nor implies that something that could be labeled “oral tradition” never existed. It does imply, however, that orality was not a factor that shaped the Gospel tradition in a recognizable way; this process was purely literary and “scribal”. Thus, it comes as no surprise that recent scholarship only rarely relies on orality to explain these literary phenomena, either for the Synoptic Problem48 or for the relation between John and the synoptics for which the artificial category of secondary orality has been discussed49. The reluctance to rely on orality in order to understand literary phenomena is well-justified methodologically: it would imply a metabasis eis allo genos. This, as we learn from Aristotle, is principally inadmissible. Consequently, the important strands of recent orality research on the Gospels are not concerned with the written Gospel tradition at all but rather engage its oral pre-history. Oral tradition, of course, is not simply “there”; it is assumed to be “behind the Gospels”50 and therefore must be postulated. But what for and on what grounds? Since blind spots by definition are not obvious, it is not easy to identify the blind spot of this postulate and to determine its 48. A recent exception with regard to the Synoptic Problem is A.D. BAUM, Der mündliche Faktor und seine Bedeutung für die synoptische Frage. Analogien aus der antiken Literatur, der Experimentalpsychologie, der Oral-Poetry-Forschung und dem rabbinischen Schulwesen (TANZ, 49), Tübingen, Francke, 2008. 49. E.g., M. LABAHN, “Secondary Orality” in the Gospel of John: A “Post-Gutenberg” Paradigm for Understanding the Relationship between Written Gospel Texts, in S.E. PORTER – H.T. ONG (eds.), The Origins of John’s Gospel (Johannine Studies, 2), Leiden – Boston, MA, Brill, 2015, 53-80. 50. Cf. E. EVE, Behind the Gospels: Understanding the Oral Tradition, London, SPCK, 2013.

382

M. KLINGHARDT

justification. Most probably, the reason for this postulate lies in the Gospels’ contents, in particular the passages about Jesus’ commissioning the disciples, installing them as apostles, and sending them out to preach the Gospel. These passages seem to create a bridge between the narratives and the evangelists’ reality. This bridge is robust only if the tradition itself is historically reliable. However, the principal load-bearing pillars of this bridge have collapsed together with the 2ST. Without “Q”, the two “independent” origins of the tradition that could be checked against each other in order to establish the respective historical credibility have ceased to exist. Without “Q”, there is no “Q-community” that might represent the postEaster apostles as the bearers of the oral tradition. Without “Q”, there is no interface between “oral tradition” and written Gospels. What is left is simply the wish that the events narrated in the Gospels (and in Acts) might be identical to the events in history. This wish is supported by theologically noble motives but not by methodologically valid arguments. Yet, even without “Q”, one might object that the Gospels’ genre is not fictional but factual and claim that, because the Gospels are historiography, they essentially constitute reliable sources51. This claim implies a certain literary concept that is the way in which the implied readers are related to the events of narrative. The readers must consider the events to be historically true and, in order to do so, they must be given the means to establish this truth by themselves. It is debatable, however, whether historiography is the basic literary concept of the Gospels. A literary analysis based on the *EPM can establish different ways in which the Gospels intended for the readers to relate themselves to the respective narrative. *Mark’s literary concept is elaborate and ambitious. The “short ending” (16,8) deliberately leaves out the reports about the appearances of the risen Jesus as found in *Evangelium52. Instead, *Mark re-directs the readers (whom else?) back to where the narrative started: “to Galilee he will be going before you, there you will see him” (*Mark 16,7). If Jesus is going before them (προάγω), they will follow him and, therefore, will be in the same position as the disciples of the narrative who follow (ἀκολουθέω) Jesus. Instead of having the risen Jesus appear, *Mark sublimely urges the readers to start over and read the Gospel again. *Matthew, too, makes the implied readers equi-originary with the disciples of his narrative, albeit 51. Cf., e.g., E.-M. BECKER, Das Markus-Evangelium im Rahmen antiker Historiographie (WUNT, 194), Tübingen, Mohr Siebeck, 2006. Cf. EAD. (ed.), Die antike Historiographie und die Anfänge der christlichen Geschichtsschreibung (BZNW, 129), Berlin – New York, De Gruyter, 2005. 52. For the profile of the ‘Easter narratives’ in *Evangelium, cf. OG (n. 1), pp. 1224ff.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

383

through different literary means. His literary concept can be gleamed from the closing scene of his gospel: the risen Lord commissions the disciples to “make disciples” by instructing them “all I have commanded you”; consequently, the implied readers will hear the same voice of Jesus as the original disciples. *Matthew provides these disciples “of the second order” with the means to find out what exactly the “Lord had commanded” the disciples of the narrative (Matt 28,19f.) – by reading Matthew53. John’s concept is much simpler. He has Jesus promise to send the “Counselor” (παράκλητος) who “will guide you in all truth” (John 16,13). This is even more than Jesus himself could tell the disciples of the narrative right away because they “cannot bear it now” (John 16,12). The knowledge of the implied future readers’ is therefore superior to that of the disciples of the narrative: the former group will participate in a fuller amount of truth than the latter group does. Therefore, the pre-canonical Gospels from *Mark over *Matthew to *John rely on different concepts to bridge the gap between literature and reality. Their claim for truth is not justified by means of historiography but by means of literature. Their narratives are not true because the events can be corroborated historically through a critical comparison of sources, but because they are aesthetically compelling by virtue of imaginative evidence54. Among the evangelists, Luke is the only one to implement elements of historiography to foster credibility. He references events from the larger and well-known political history and borrows stylistic elements from well-known historiographical works. He assures the readers of his own diligent research. He claims to rely on eyewitnesses55 and ministers of the word (Luke 1,2). In addition, he presents himself as a reliable link in the chain of witnesses, as the We-passages of Acts demonstrate. The readers of the complete New Testament will find out that he is an eyewitness of, and a player in, the later events. Interestingly, these “historiographical” elements only occur on the last stage of the tradition according to the *EPM, in Luke and in the elements of the canonical edition. In view of the prior stages of the tradition, these assurances are simply one of most basic features of literature as such: the fiction of authenticity.

53. In (canonical) Mark, however, this elaborate literary concept is destroyed by the supplementation of the longer ending. For the literary concepts of Mark and Matthew, cf. M. KLINGHARDT, Erlesenes Verstehen. Leserlenkung und implizites Lesen in den Evangelien, in ZNT 21 (2008) 27-37. 54. For mechanisms of literary persuasiveness in the NT, cf. ID., Inspiration und Fälschung. Die Transzendenzkonstitution der christlichen Bibel, in H. VORLÄNDER (ed.), Transzendenz und die Konstitution von Ordnungen, Berlin – New York, De Gruyter, 2013, 331-355. 55. If Luke is seen as part of the canonical edition, the implementation of the Gospels’ titles with the authors’ names makes both Matthew and John eyewitnesses.

384

M. KLINGHARDT

Admittedly, this claim of authenticity is more subtle than the bold assurance “Written by Himself!” in the first edition of Robinson Crusoe56. Yet, it uses some of the same basic methods, such as resorting to the first-person narrative perspective, e.g., in the We-passages or in Peter’s reference to his witnessing the heavenly voice (2 Pet 1,17). Feigning authenticity does not mean that all the content is purely fictitious, though. While Defoe’s Robinson is fictional literature, the same must not be true for his main source, Alexander Selkirk’s report about being a castaway for four years on an island in the Pacific57. The same may apply to canonical Luke with respect to his sources: the assurance of authenticity is a fictitious literary feature; the credibility of his sources and even the earliest of all the Gospels cannot be ascertained. IV. TAKING A LOOK BEHIND THE *EPM: WAS THERE A GOSPEL TRADITION? In the end, one should attempt the impossible and examine one’s own presuppositions in search for the blind spots of the *EPM. The only promising way to do this would be to identify the premises that are taken for granted and therefore go unchallenged. One of these premises stands out in a rather visible manner: It is the assumption that there, indeed, was a Gospel tradition. All synoptic theories assume a process in which the Gospels were written one after another, among them, of course, also the *EPM with the four assumed editorial stages of the Gospel tradition as outlined above. That the Gospels were written sequentially, one at a time, is a common premise that has never been challenged. It is not self-evident at all, though: There is no hard evidence supporting this assumption. The evidence we have confirms the existence of two collections of writings. There is, on the 56. The title of the 1719 first edition of Daniel Defoe’s Robinson claims: “The Life and Strange Surprizing Adventures of Robinson Crusoe of York, Mariner: Who lived Eight and Twenty Years, all alone in an un-inhabited Island on the Coast of America, near the Mouth of the Great River of Oroonoque; Having been cast on Shore by Shipwreck, wherein all the Men perished but himself. With An Account how he was at last as strangely deliver’d by Pirates. Written by Himself”. Readers understand this last claim was made by the editor, W. Taylor, who claims in the preface: “The Editor believes the thing to be a just History of Fact; neither is there any Appearance of Fiction in it”. The readers must believe the editor’s belief, of course. 57. E. WOODE, A Voyage to the South Sea, and Round the World […] Wherein an Account is given of Mr. Alexander Selkirk, his Manner of living and taming some wild Beasts during the four Years and four Months he liv’d upon the uninhabitated Island of Juan Fernandez, London, 1712.

THE LITERARY CHARACTER OF THE GOSPEL TRADITION

385

one hand, the collection containing *Evangelium and ten Pauline letters as attested for Marcion and the Marcionites by the heresiologists. On the other hand, there is the canonical edition of the New Testament with the four-Gospel book as attested by the broad and uniform manuscript tradition. We do not have the slightest evidence that any of the Gospels ever existed outside these two collections. What, then, is the justification for assuming that there was a “tradition”? The initial answers that come to mind are not satisfying. Usually, the idea of a written Gospel tradition is seen as a prolongation of the multi-strand oral tradition which at some point was put into writing; the process of transformation and adaptation typical for oral tradition then continued within the written Gospels, creating the written “Gospel tradition”. Thus, the postulate of a written Gospel tradition depends on the postulate of an oral tradition. Since the only methodologically valid reason to assume the existence of an oral tradition in the first place is the literary character of the written tradition, this argument is circular and therefore untenable. Another answer is based on the assumed process of “canonization”, that is, the process of collecting, rejecting, and accepting single writings as part of a longer anonymous process that finally feeds into the canonical NT. Apart from this picture’s discordance with the manuscript tradition58, it also contradicts the model of the *EPM. The final redaction of the Gospels, for which the unifying redaction of the passion predictions can serve as an example, implies an altogether different picture: the four-Gospel book is the result of one editorial act rather than a longer coincidental emergence. Thus, neither the fantastic ideas about oral tradition and its “scribalization” nor the vague picture of a multi-coincidental process from which the four-Gospel book presumably emerged are of any use as the means of corroborating that there ever was a written Gospel tradition. Despite my firm conviction that the Gospel tradition according to the *EPM best explains the literary relations between the Gospels, I cannot sufficiently answer the basic question on which the whole theory rests: Was there ever a process in which one Gospel was written after the other? This is what asking methodological questions usually does: instead of providing answers, it unearths inconvenient questions. If such questions help shatter long-held matters, of course, they have achieved their aim. Technische Universität Dresden Dresden, Germany

Matthias KLINGHARDT [email protected]

58. Cf. TROBISCH, First Edition (n. 24), pp. 8-44.

DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT

In his contribution to this volume, Matthias Klinghardt provides an intriguing alternative solution to the Synoptic Problem according to which all four canonical Gospels are ultimately rooted in Marcion’s Gospel (MG). He gives as his sequence of synoptic development (1) MG; (2) Mark; (3) Matthew; (4) John; and (5) Luke. Like Augustine, Klinghardt insists that “each stage used all the texts available from the previous stages”1, with the result that “the tradition appears as a tightly-knit net, or rather as a tree which arises from a single root and around which the branches of the subsequent tradition entwine”2. In addition, Klinghardt also proposes a final canonical redaction, temporally coinciding with the creation of canonical Luke, that went back over all four previous Gospels in order to harmonize them3. Klinghardt’s contribution represents a “very brief summary” of a much larger work. This provides him with an opportunity to feature in a short space what he regards as the highlights of his evidence and his most compelling arguments. I. FH + PROTO-LUKE Klinghardt’s solution combines two pieces. If one leaves MG out of consideration what is left looks very much like the Farrer Hypothesis (FH), with a dash of John’s gospel thrown in. But when one adds MG, it becomes a sort of hybrid of the FH and Proto-Luke Hypothesis causing things to change in ways even Klinghardt may not realize. This is partly because, despite Klinghardt’s frequent use of the word “all” in reference to other synoptic solutions, he is apparently unaware of the existence of the Matthean Posteriority Hypothesis (MPH). The only MwQH (Mark-without-Q Hypothesis) he mentions is 1. Augustine, Cons. 1,2,3 and 1,2,4. 2. Klinghardt’s essay in this volume, p. 369. 3. Klinghardt’s essay in this volume, pp. 369, 378.

388

R.V. HUGGINS

the FH. This is interesting because at odd moments he argues very much like an MPH defender. So, for example, in relation to the direction of dependence of the Sermon on the Mount/Plain, Klinghardt states that the MwQH (by which he means the FH) faces the serious problem of having to “assume that Luke dissolved the Matthean composition and then scattered its material across his gospel”4. The MPH, however, which is also a MwQH, agrees with Klinghardt completely and sees this as an important point in support of its position as over against the FH. II. EXPLANATORY POWER At one point, Klinghardt curiously insists that “it is insufficient to criticize [his solution] on the grounds of its allegedly smaller explanatory power in comparison to other theories”5. Actually, the first question one must consider when assessing any solution to the Synoptic Problem is its relative explanatory power as over against other solutions. Indeed, it is a question Klinghardt does not hesitate to ask of solutions other than his own. But it is actually the explanatory power of one aspect of Klinghardt’s theory that I find intriguing. Instead of running into the difficulty that the FH faces with regard to the Sermon on the Mount, Klinghardt can simply say that Matthew reworked the material of MG and that Luke instead accurately reproduced what he found in MG, “as the prologue promises (Luke 1,1-4)”6. The same can be done in many other places where canonical Luke appears to be more primitive than Matthew. But this seeming strength is accompanied by a more general weakness. In the case of the Sermon on the Mount, Klinghardt’s immediate advantage over the FH is attended by a significant diminution of his justification for claiming that Matthew came before canonical Luke and was known and used by him. Only those places count where Luke can be shown to agree with Matthew against MG. And then more problematic still are those places where Klinghardt’s reconstruction of MG is said to include Matthean material not found in canonical Luke (e.g., Luke 6,20//Matt 5,3; Luke 6,27-29// Matt 5,38; Luke 11,43//Matt 23,6-7)7.

4. Klinghardt’s essay in this volume, p. 367. 5. Klinghardt’s essay in this volume, p. 373. 6. Klinghardt’s essay in this volume, p. 377. 7. M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021, pp. 597, 606, 848.

DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT

389

III. THE INFANCY NARRATIVES Klinghardt can of course appeal more broadly to the infancy narratives and genealogies, present in Matthew and Luke but absent from MG. But doing so puts him in the same position as defenders of the FH who have to explain how it was that, assuming Luke used Matthew, their two infancy narratives and genealogies are so different. Then second, it renders problematic his claim that a final redactor (Luke?) went back over all four canonical Gospels to harmonize them. If such a final redaction occurred, why was so little effort at harmonization made? Why, for example, was Luke’s account not modified to make space for Matthew’s flight into Egypt? Or again, why the striking differences between Luke’s and Matthew’s genealogies, with Jesus being said to descend from two different sons of David and Joseph from two different fathers? Klinghardt seems to answer that such things did not matter to the Evangelists. Why? The Evangelist’s narratives are “true [not] because the events can be corroborated historically … but because they are aesthetically compelling by virtue of imaginative evidence”8. The exception to this, according to Klinghardt, was Luke, who “is the only one to implement elements of historiography to foster credibility”9. But if that is so then why did Luke not harmonize these pieces? In any case, the claim that the differences did not matter is contradicted by the puzzlement over the differences from the time of Julius Africanus (2nd/3rd cent.), who offers a harmonizing explanation of the genealogies credited to the desposyni (the family of Jesus), right down to the present time10. IV. LUKAN BOOKENDS Positing, as Klinghardt does, what basically represents a single Gospel (MG/Luke) bookending his synoptic solution, which can be said to both influence and be influenced by all the other Gospels, makes it doubly difficult to pin down the direction of inter-literary dependencies. For one thing, the approach delegitimizes Klinghardt’s appeals to basic consensuses of synoptic studies in support of particular parts of his solution. So, when Klinghardt claims that “Matthew’s dependence on Mark is so well-founded that it is 8. Klinghardt’s essay in this volume, p. 383. 9. Klinghardt’s essay in this volume, p. 383. 10. Eusebius, Hist. eccl. I,7. For examples, see my The Generation of the Deportation to Babylon: Matthew 1,17 as the Key to Counting Matthew’s Genealogy, in ETL 97 (2021) 293-312.

390

R.V. HUGGINS

hardly in question”11, he seems to overlook the fact that in the very act of incorporating Mark into his MG Priority solution, Markan Priority ceases to be Markan Priority, and gives way instead to a kind of Lukan Priority. But scholars argue that Luke follows Mark for the same kinds of reasons that they think Matthew does. One cannot just assert Markan priority over Matthew but not Luke (= MG) on the grounds that it is a widely held consensus among those who do not embrace your solution. Consequently, Klinghardt’s theory suffers the same kind of weaknesses as the Augustinian and Griesbach hypotheses in terms of directionality of dependence. V. FLIPPING MARK AND LUKE Because Klinghardt starts with a form of Lukan priority, it behooves him to argue that Luke’s so-called great omission (Mark 6,45–8,26) really represents Mark’s great insertion and Luke’s great insertion (Luke 9,15–18,14), Mark’s great omission. And he claims the reasons for doing so are “obvious”, because “every single element [in Mark’s great insertion] has a meaningful and well-reasoned narrative function at its exact position”12. Is that so? Could it not be, viewing it as Luke’s omission, that MG/Luke felt no need to include the material? Was it really necessary, for example, for him to have two feedings of the multitudes like Mark does? Or, on the flip side of that same redactional coin, was it really so “obvious” why Mark, already having one feeding of the multitude, needed to add another? VI. WHY NOT RETAIN ACTUAL MARKAN PRIORITY? What inspired Klinghardt to place Mark where he does in the synoptic order in the first place? Where is the downside to arguing that Mark was prior not only to Luke but also to MG? From Klinghardt’s perspective, there are two obvious reasons for having Mark second. First, placing Mark before MG would leave Klinghardt with a much less attractive synoptic solution. The underlying dramatic story of Marcion speaking truth to ecclesiastical power, boldly but authentically proclaiming against all odds that his was the single real Gospel, certainly has an appeal, while viewing MG as just another Synoptic Gospel based on Mark is less exciting. 11. Klinghardt’s essay in this volume, p. 366, n. 17. 12. Klinghardt’s essay in this volume, p. 366.

DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT

391

Second, placing Mark second removes a large body of synoptic material from view. When Mark comes second, everything in MG that parallels Q and L material is effectively placed beyond scrutiny on the grounds that it is anybody’s guess where MG got it. But once Mark is moved into first place, where it would seem to belong, all that material suddenly becomes extra-Markan material. This would force Klinghardt to account for places where MG and Matthew share material not found in Mark (as opposed to unique MG material paralleling or not paralleling L), making it necessary, in other words, for him to propose something very like Q. Putting Mark second removes the material from view without having to account for it. VII. THE HOLE AT ONE END OF KLINGHARDT’S THEORY Proposing, as Klinghardt does, a solution in which the earliest and latest Gospels basically represent two different editions of the same work creates a rather large hole at its earlier end that needs to be plugged before confidence can be achieved about which of the other gospels follow which. Klinghardt can only insist that John’s gospel influenced canonical Luke if he can first establish that it did not influence MG, which then influenced canonical Luke. The kind of difficulties he faces is illustrated in the “prominent example” he chooses to make that very case, namely John’s alleged influence on Luke provided by John 20,3-10 (= Luke 24,12). There are a couple of issues here. The first is that there is a difference between unattested material and material attested as absent. Luke 24,12 is simply unattested in MG13. Naturally that raises questions as to why Klinghardt featured it as his “one prominent example” rather than, say, a passage that was actually attested as being either present or absent from MG. Having said that, it is quite possible that Luke 24,12 may have been absent from MG, as it was from Codex Bezae (D). But, since that passage represents one of Westcott and Hort’s so-called Western non-interpolations, it may also have been absent from the original canonical Luke as well. Or, it might have been present in our current canonical Luke and absent from MG because the text of the latter was more pristine at that point. In either case it tells us more about the textual tradition Marcion was using than directions of dependency between MG and canonical Luke. It certainly does nothing to prove that John influenced Luke directly rather than indirectly through MG. 13. D.T. ROTH, The Text of Marcion’s Gospel (NTTSD, 49), Leiden – Boston, MA, Brill, 2015, p. 435.

392

R.V. HUGGINS

Failure to establish that John influenced Luke directly rather than indirectly through MG potentially creates a kind of domino effect for Klinghardt’s solution in which, as John repositions itself temporally prior to MG, Matthew, which Klinghardt counts as a source for John, repositions itself prior to that, and Mark, as a source for Matthew, jumps back earlier still. In his article, Klinghardt states regarding the direction of dependence between MG and canonical Luke that “[a]ll the evidence points in the same direction: Luke is an edited version of [MG]; [MG] is older than Luke”14. Indeed, Klinghardt insists that this conclusion can be established “without difficulty, and with surprising unambiguousness”15. I disagree. For Klinghardt, the “most important single observation” establishing this is Luke’s seemingly out of place remark in Nazareth: “Do here in your hometown what we have heard that you did in Capernaum” (4,23). The issue is that canonical Luke does not mention any activity by Jesus in Capernaum prior to this passage. Only later (4,31) does Luke have Jesus go down to Capernaum where he casts out a demon in the synagogue. By way of contrast, in Mark the casting out of the demon in Capernaum comes first (Mark 1,21-28//Luke 4,31-37), and Jesus’s visit to Nazareth later (Mark 6,1-6//Luke 4,16-30). Klinghardt writes: “As has long been recognized, this unfortunate position [of the reference to Capernaum] is the result of Luke’s editorial intention to create the programmatic scene and, therefore, secondary”16. It is curious that in addition to reversing the order of these two events from Mark, Luke also made his arrangement problematic by including the reference to Capernaum, which Mark’s version does not have. For consistency’s sake he might have left that out, which would have removed the problem completely, but for some reason he included it. When Klinghardt refers to Luke’s editorial intention in this passage and how it implies that his arrangement is secondary, he is certainly correct, but in the interest of clarity we must ask, secondary to what? Certainly, secondary to Mark’s arrangement, but secondary to MG? Unfortunately, establishing the latter is not so easy. The odd placement of Luke’s Capernaum statement is so striking that it is hard to miss, and so naturally invites attempted explanation and/or correction. Klinghardt insists that MG’s order is more original than Luke’s. But it can be argued to the contrary that MG was actually secondary to Luke and was simply attempting a correction there. The entire question is complicated by the fact that the 14. Klinghardt’s essay in this volume, p. 364. 15. Klinghardt’s essay in this volume, p. 362. 16. Klinghardt’s essay in this volume, pp. 362-363.

DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT

393

problematic Capernaum reference is unattested. We don’t actually know whether it was present or absent from MG. Klinghardt insists that “none of Marcion’s alleged theological motives would explain this correction”. But is he correct? Could Marcion’s placing Capernaum first have been motivated by something other than a mere interest in correcting Luke’s order? That such might be the case is suggested by MG’s interpretation of “going down” (κατῆλθεν) to Capernaum not in the geographical sense as in Luke 4,31 (i.e., down from Nazareth), but rather in the sense of going down from higher realms17. Did Marcion assign some special significance to Capernaum? Others certainly did. When Origen noticed the mention of Capernaum in Luke 4,31, he interpreted the unexpected placement as a signal of deeper meaning in the passage, which he thought he found in the idea that “Capernaum” represented a type of the Gentiles and “Nazareth” of the Jews18. The Valentinian Heracleon reacted in a similar way to the mention of Jesus going to Capernaum in John 2,12. In his case, he thought it odd that Jesus did nothing there. So he too sought a deeper meaning and concluded that Capernaum “means these farthest-out parts of the world, these districts of matter, into which He [Jesus] descended, and because the place was not suitable, he says, He is not reported either to have done anything or said anything in it”19. In MG, to be sure, Jesus does do and say things in Capernaum prior to going to Nazareth, but the question remains open whether whoever produced MG took the strange placement as a signal of Capernaum’s special significance. More significant, however, is Tertullian’s point about MG’s removing references connecting Jesus to Nazareth20. Regrettably, Marcion’s early opponents were more interested in refuting him on the basis of what he left in MG rather than providing a comprehensive description of what he left out. The reason for this was that although Marcion’s revision, if that is what it was, was substantial, it was not thoroughgoing21. This is not to say that 17. Tertullian, Marc. 4,7,1. 18. Origen, Hom. Luc. 33,1. 19. Quoted in Origen, Comm. Jo. 10,9 (ET: ANF 9.386) 20. Tertullian, Marc. 4,8,2. Tertullian attributed this to Marcion’s desire to avoid overlaps between his Jesus and the Christ prophesied by the creator God. See also Adamantius, Dial. 1,24 (818b-c), 2,14 (829e) and Celsus in Origen, Cels. 4,74. 21. Contra Klinghardt who seems to assume that if Marcion had undertaken editing Luke at all he must have pursued the project in a thoroughgoing manner. “How it is possible”, he asks, “that the heresiologist could refute Marcion’s theology based on Marcion’s own Gospel, if Marcion’s altering, editing, and cleansing of Luke aimed at a consistent conformity with this very theology? The logical contradiction is irresolvable and proves the claim of Marcion’s alleged edit wrong” (Klinghardt’s essay in this volume, pp. 361-362).

394

R.V. HUGGINS

Marcion and his followers did not feel they could answer their critics point for point by applying Marcion’s dominating Good God vs. creator god hermeneutic. Still, the limited extent of Marcion’s own editing of the text, again if that is what it was, does appear to have resulted in an ongoing process of “correcting” of alleged corruptions that eventually extended to involve the other Gospels as well22. Hence, Tertullian’s reference to the Marcionite’s “daily retouching their work” and Adamantius’ claim that “right up to the present” the Marcionites “remove anything that does not agree with their opinion”23. Because the heresiologists focused more heavily on passages that Marcion left in to refute him, it sometimes becomes necessary to approach the question of whether something was left out on the supposition that had it been included it would have been sized upon by Marcion’s opponents for comment. The suppression of references to Jesus’ Nazareth connections is a case in point. Beyond the opening chapters, the reference to Nazareth being Jesus’s hometown, the mention of his being the son of Joseph, and the reference to prophets not being accepted in their hometown are also missing from MG Luke 4,16.22.24. Then there is the attested reduction of “Jesus of Nazareth” to just “Jesus” in 4,34, 18,37, and 24,19. Sometimes, as noted earlier, what the heresiologists assumed were corruptions were actually evidence of MG’s reflecting a text form evidenced in Codex Bezae (D) and a few Latin mss. But in all the cases just mentioned regarding Jesus’s Nazareth connections D agrees with the standard text against the readings of MG24. VIII. WAS MARCION THE AUTHOR? If one fails to establish MG’s priority over canonical Luke the only real way to plug the hole mentioned earlier is to demonstrate instead that Marcion

22. As in the strange case of Matt 5,17, which Tertullian repeatedly faults Marcion for removing from his version of Luke (sic!) (Marc. 4,7,4; 4,9,15; 4,12,14). This despite Tertullian’s insistence that Marcion had not touched the other Gospels (Marc. 4,5,5). But then Adamantius’ Marcionite interlocutor insists that Judaizers had corrupted Matt 5,17, which, he insisted, had originally read: “I did not come to fulfill the Law but to destroy it” (Dial. 2,15 [830d]) (ET: Robert A. Pretty). 23. Tertullian, Marc. 4,5,7 (ET: ANF 3); Adamantius, Dial. 2,18 (867a). Cf. Celsus’ possible reference to the Marcionites in Origen, Cels. 2,27. 24. The possible exception being the reference to Jesus’s hometown in Luke 4,16, which is present in the Latin text of D but written in the line by a corrector in the Greek. It is difficult to say, however, whether the corrector had added the word τεθραμμένος at that point or merely reinforced it.

DID MARCION’S GOSPEL COME FIRST? A RESPONSE TO M. KLINGHARDT

395

himself was its author. But that is really not an option since Marcion and his followers viewed the process of the Judaizing corruption of the gospels as already occurring by the time Paul wrote Galatians (1,7, 2,4.14; 2 Cor 11,13)25, and, as Tertullian aptly notes, “Emendatio culpam non antecedit”26. This way is especially barred for Klinghardt’s solution because it would make Marcion himself both the terminus ad quo and the terminus ad quem for the composition of all four canonical gospels. Very likely the first time anyone saw MG would have been after c. 138 when Marcion arrived in Rome. So, if we assume for example with Markus Vinzent that Marcion “produced [his gospel] in his Roman classroom”, say, between his arrival c. 138 and excommunication in 14427, then what are we to do with the fact that Marcion was already condemning canonical Luke, the end point of Klinghardt’s canonical fourfold development, in his Antitheses28? Can we really even be sure that the Antitheses were not written prior to MG? IX. TRADING ACCUSATIONS Klinghardt does not help himself with the emphasis he places on Tertullian’s statement: “I affirm that Marcion’s Gospel is adulterated; Marcion, that mine is”29. From his perspective Tertullian’s comment is made untenable by the mutuality of the accusation, so that “doubts should have been raised because of Marcion’s reverse accusation”. But, as John Knox pointed out eighty years ago, “These claims are precisely what we would expect from the two rival camps”30. One needs to inquire into the actual substance of Marcion’s counter accusation. MG cannot be shown to have enjoyed any geographical diversity prior to Marcion, and that which came after depended on the dispersion of Marcion’s followers. On the other hand, canonical Luke can be shown to have been preferred to MG both within and beyond the boundaries of the great church, where we find the opening chapters missing in MG being attested, for example, by Basilides31, Valentinus32, the 25. Tertullian, Marc. 4,3,1-4. 26. Ibid., 4,4,5. 27. M. VINZENT, Christ’s Resurrection in Early Christianity and the Making of the New Testament, Surrey – Burlington, VT, Ashgate, 2011, p. 86. 28. Tertullian, Marc. 4,4,4. 29. Tertullian, Marc. 4,4,1. 30. J. KNOX, Marcion and the New Testament: An Essay in the Early History of the Canon, Chicago, IL, University of Chicago Press, 1942, p. 78. 31. Hippolytus, Haer. 7,26,9 (ANF 7.14) = Luke 1,35. 32. Hippolytus, Haer. 6,35,3 (ANF 6.30) = Luke 1,35.

396

R.V. HUGGINS

Valentinian Ptolemy33, Marcus34, the Testimony of Truth35, Tatian’s Diatessaron36, and the Protevangelium of James37. Origen noted that there were “countless heresies that accept the Gospel According to Luke”38, and in many of those cases it was canonical Luke, not MG, that was being accepted. How seriously, then, should we take the word of a wealthy ship builder (Marcion) who shows up in Rome in the mid-second century with a version of the Gospel of Luke that nobody had ever seen39, claiming it was the true version and that the one currently in use was corrupt40? On what grounds should Marcion be considered as in any way equipped even to know one way or the other? Even supposing MG was the version of Luke that Marcion was familiar with in Sinope, as the people in Rhossus were with the Gospel of Peter41, why should we regard his assessment of canonical Luke as having more substance to it than, say, a Protestant fundamentalist’s claim who, ignorant of the history of the text, condemns the Roman Catholics for “adding books” to the Bible? Independent scholar United States of America

Ronald V. HUGGINS [email protected]

33. Irenaeus, Haer. 1,3,2 = Luke 2,42; Haer. 1,8,4 = Luke 2,28.36. See also the interpretation of Luke 2,13-14 by unspecified Gnostics mentioned in Haer. 3,10,3. 34. Irenaeus, Haer. 1,15,1 = Luke 1,35. 35. NHC IX,3,45,6-22 (Julius Cassianus? See Clement of Alexandria, Strom. 3,86-95). 36. See the evidence conveniently displayed in D. PASTORELLI, The Genealogies of Jesus in Tatian’s Diatessaron, in C. CLIVAZ – A. DETTWILER – L. DEVILLERS – E. NORELLI (eds.), Infancy Gospels: Stories and Identities (WUNT, 281), Tübingen, Mohr Siebeck, 2011, 216-230, pp. 218-221. 37. Prot. Jas. 11–17. The reason I include this here is because of the virginity in partu reference in 20,1-3 which, in that period, usually reflected the docetic/gnostic dislike of the idea of Jesus being actually born of a virgin rather than his simply passing through her, as was often said, as water through a tube (Adamantius, Dial. 5,9 [855e]; Irenaeus, Haer. 1,7,2; 3,11,3; Tertullian, Carn. Chr. 23,2-6; cf. Ignatius of Antioch, Smyrn. 1,1; Trall. 9,1, Eph. 18,2). Origen (Comm. Matt. 10,17; cf. Hom. Luc. 7,4) and Clement of Alexandria (Strom. 7,16,93-94) both held a kind of perpetual virginity position, seemingly under the influence of the Protevangelium, although the latter author does not mention the work by name in that connection (cf., Ascension of Isaiah 11,7-10). 38. Origen, Hom. Luc. 16,5 (ET: Joseph T. Lienhard). 39. Tertullian, Marc. 4,5,2. 40. Especially if, as Tertullian surmises, he had previously for a time accepted the common version: “that portion of [Luke] which we alone receive is so much older than Marcion, that Marcion himself once believed it, when in the first warmth of faith he contributed money to the Catholic church, which along with himself was afterwards rejected when he fell away from our truth into his own heresy” (Marc. 4,4,3). 41. Eusebius, Hist. eccl. VI,12,2-6.

MARCION’S GOSPEL DID COME FIRST! A REJOINDER TO R.V. HUGGINS

I am grateful to Ron Huggins for his response and for his critical remarks on my paper and the book it summarizes. Huggins’ main question is stated clearly in the title of his paper: “Did Marcion’s Gospel Come First?” The problem of the editorial direction between *Evangelium and Luke is indeed fundamental. All further considerations depend on the answer to this question. For this reason, I maintain that, with regard to evaluating the whole model, the ‘explanatory power’ of any Synoptic theory is, and must be, secondary to this basic problem (contra Huggins section II). Clearly, any model of the Gospels’ tradition history that does not include the most basic and critical evidence has no explanatory power to speak of. For this reason, I regard the evaluation of the strengths or weaknesses of any of the ‘Synoptic theories’ (or some elements thereof) over each other or over the *EPM1 in order to assess the validity of the respective theories as irrelevant – unless, that is, the basic question is sufficiently answered first. Consequently, characterizing the *EPM as a combination of the ‘FH and ProtoLuke’ or stating similarities between the *EPM and the MPH (Huggins section I) amounts, at best, to an illustration of some aspects of the *EPM in the terminology of other theories. Because this terminology is alien to the framework of the *EPM, it is of little help. It mostly muddies the methodological waters and is best avoided. Even though the ‘explanatory power’ has no legitimate function in the comparison of theories whose methodological status differs, it may nevertheless be instructive to consider the examples mentioned by Huggins (section II): the agreements between Luke and Matthew against *Evangelium (*Ev & Matth = Luke) and, “more problematic[ally] still” (why?), agreements between *Evangelium and Matthew against Luke (*Ev = Matth & Luke). Why should this be “a significant diminution of [the] claim that Matthew came before canonical Luke”? Huggins seems to imply that the Evangelists 1. Against Huggins’ choice of abbreviations, I maintain the designations for both the text in question (*Ev) and the Synoptic theories, as explained in my essay above. Therefore, the designation I continue to use here is *Ev instead of (Huggins’) MG etc.

398

M. KLINGHARDT

were mere redactors who integrated all the elements of the tradition in their writings and did not leave out anything. Since any ‘Synoptic theory’ must accept the fact that materials were left out in the editorial process, this implication is not based on evidence. One wonders where exactly its foundation lies. The evidence, in any case, points in a different direction. The Evangelists did make choices about what they deemed useful for their purposes and what they preferred to pass over. Examples include the ‘great omission’ according to the 2ST and the ‘great insertion’ of Luke 9,51–18,14 according to the MPH. Since any given ‘Synoptic theory’ relies on assuming authorial choices in the editorial process, this is not a peculiar feature of the *EPM and certainly does not constitute a valid argument against it. The same applies, of course, to the differences between the Matthean and Lukan infancy narratives (Huggins section III). These discrepancies are simply part of the evidence and pose the identical problem to any given solution. (Yet, I will concede that scholarship has largely ignored this problem.) Therefore, they cannot discredit a particular explanation. Apparently, someone chose to create these discrepancies. Even more importantly, someone chose to put these narratives next to one another in the four-Gospel book. We must assume that this was neither accidental nor inevitable, but a deliberate choice, since the only form in which we have the Gospels is in the manuscripts containing all four of them. Acknowledging this fact, then, sheds some light on the principles of Gospel writing. Meticulous preservation of the sources and faithfulness towards the tradition probably did not rank as high among those principles as some assume. Instead, variance and multi-perspectivity were probably more important. A major part of Huggins’ critique deals with the priority of *Evangelium over Mark in the *EPM: “What inspired Klinghardt to place Mark where he does in the synoptic order in the first place?” (section VI). Since I changed my mind about the sequence of Mark and *Evangelium2, it is clear that this sequence is not at all self-evident: It deserves full attention and foremost methodological accuracy. The basic argument is that of the respective greater editorial probability: is it easier to assume that A edited B or the other way around? As I point out elsewhere, it is much easier and, methodologically, safer to determine positive editorial motives (why does an editor include certain material?) rather than to muse about why something does

2. In an earlier stage of my thinking, I claimed to leave the relation between *Ev and Mark open but, inconsequently, included a diagram indicating Mark-priority over *Ev. For the retractio, cf. M. KLINGHARDT, The Oldest Gospel and the Formation of the Canonical Gospels (BiTS, 41), Leuven – Paris – Bristol, CT, Peeters, 2021, p. 195 and n. 16 (here cited: OG).

MARCION’S GOSPEL DID COME FIRST! REJOINDER TO R.V. HUGGINS

399

not exist. Although far from being exhaustive, I examined the relation between *Evangelium and Mark at some length3. In my opinion, the result (*Ev-priority over Mark) holds up against Huggins’ questions. Mostly, he raises the possibility of the alternative editorial dependence, be it with respect to the so-called ‘great omission’ or with respect to the second feeding story (section V): Could it not be the other way round? Yes, of course, it could – there is more than only one possible answer, which is why we seek an explanation in the first place. For that reason, it is necessary to determine which alternative carries a greater probability and to provide arguments for it. So far, with respect to the editorial direction between *Evangelium and Mark, I have not seen an argument that would convince me to reverse the direction. To be fair, the lack of convincing arguments cannot be held against Huggins’ very short response which cannot be expected to provide sufficient counter-arguments. Yet, so far not even more extensive critiques could convince me4. A particular problem is raised by Huggins invoking ‘historical’ arguments that are external to the textual world, such as Julius Africanus’ interpretation of the genealogies (section III) or what Marcion did or did not do (sections VII, VIII, IX). These historical questions are interesting and have their legitimate place. Yet, they do not bear on the literary questions under discussion. Methodological precision requires that we keep these two areas as distinct as possible: historical judgment relies on sources, not the other way round. Therefore, securing the sources (including, if necessary, their reconstruction) must precede the historical judgment. The history of the Gospel tradition is principally a literary problem that must be solved by literary means, i.e. by source criticism. For this reason, I am reluctant to spend too much time speculating about what Marcion did or did not do. Instead, I find it advisable to concentrate on the literary phenomena. Among the ‘historical’ data is also the mutuality with which Tertullian and Marcion accuse each other of having altered and falsified the ‘original’ Gospel. Huggins is correct to note that pointing to these opposite claims does not really help my case (section IX). However, I still find it important to point out that there is, indeed, such a counter-argument to the seemingly unanimous claim – from Tertullian to Harnack and beyond – that Marcion mutilated canonical Luke. Huggins correctly invokes John Knox as his witness that such mutuality is to be expected; yet, Knox is the only one who assumed the priority of the Marcionite Gospel over Luke within the last 170 years of scholarship – against an abundance of voices who repeated 3. Ibid., pp. 199-237. 4. Cf. the examples mentioned in ibid., pp. 428ff.

400

M. KLINGHARDT

rather than argued Tertullian’s position. Deciding between these mutual accusations cannot rely on speculations about the dispersion of Marcion’s Gospel (about which we know next to nothing, as is the case, by the way, with any other Gospel and, in fact, with all early Christian literature); nor is the later success of the Canonical edition an admissible argument supporting the claims for its originality. There is only one solution to this problem: “One needs to inquire into the actual substance of Marcion’s counter accusation” (Huggins section IX). Apparently, I fully subscribe to this solution: This is exactly what I did (or tried to do) in the most extensive, and even tiring, way. The overall result of my source-critical (i.e., literary!) inquiry unambiguously confirms Marcion’s claim, not Tertullian’s: Marcion’s Gospel did come first! Aside from Ron Huggins surprisingly confirming my approach (if not its results), the most important insight I gained from his response is the gravitational pull that many decades of scholarship on the Synoptic problem exert on both (a) the perception of the problems in question and (b) their solution. The impact of this familiarization is apparent in different areas and on different levels. I will mention only three of them. Most obvious is the use of the traditional terminology in comparing the *EPM to any ‘Synoptic theory’. It ignores the fact that the *EPM is not, and does not intend to be, a solution to the Synoptic problem. On the contrary, perceiving the literary relations between the Gospels as a mere ‘Synoptic’ problem is a reduction that conceals the fact that a proper solution must include all four Gospels (and *Evangelium, of course). On a second level, attempts at the critique of the *EPM by offering alternative interpretations of limited aspects remain unconvincing as long as they do not acknowledge the overall picture in their argument. The complexity of the *EPM requires a critique that at least takes notice of its claim to include many single aspects and to combine different research areas in one model. An even more fundamental level concerns the principles of the Gospels’ tradition history. The editorial process that we assume for the formation of the four-Gospel book was not dominated by ideas such as unambiguousness or consistency. Had they been the principles, a single Gospel would have been the preferred solution. Yet, there are four of them with greater and lesser discrepancies, and even clear contradictions. To understand why this is so means to understand the principles guiding the production of the four-Gospel book. This is the most complex task. It should be the result of all the scholarship on the Gospels, rather than its premise. Technische Universität Dresden Dresden, Germany

Matthias KLINGHARDT [email protected]

THE WIDOW AT NAIN, THE SINFUL WOMAN WHO LOVED MUCH, AND THE EXTENT OF Q

I. LUKE’S COMPOSITIONAL PRACTICE AND

THE

EXTENT

OF

Q

In his study of ancient compositional practices, Robert A. Derrenbacker notes that ancient authors tended to follow one source at a time. For example, in Antiquities, Josephus alternates between blocks of material where he follows the Deuteronomistic Historian and blocks where he follows the Chronicler. Very rarely does an author move “back and forth between sources within episodes. It is only when a pericope/episode is concluded that the author will typically move to another parallel source if he chooses”1. This is not surprising. It is difficult enough to keep one scroll open and accessible while writing on another; it is nearly impossible to keep two scrolls within view while writing on a third, especially if the author needs to move to different parts of one of the scrolls to find the parallels that fit the current passage in the other source2. These observations present problems for every solution to the Synoptic Problem, for the type of microconflation that is so rare elsewhere can be seen in Matthew, and we also find Matthew accessing Q out of order, which is also extremely rare in the com1. R.A. DERRENBACKER Jr., The “External and Psychological Conditions under Which the Synoptic Gospels Were Written”: Ancient Compositional Practices and the Synoptic Problem, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 435-458, pp. 440-441. 2. A. KIRK, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS, 564), London – New York, Bloomsbury T&T Clark, 2016, pp. 54-55 writes: “reading a scroll monopolized both hands. Some complicated maneuvering lay between reading a source and copying it; likewise, consultation of more than one scroll at a time was impracticable”. He quotes Jocelyn Small, a historian of Greco-Roman art, who finds “no … ancient representations of multiple texts open simultaneously” and argues that with the size of scrolls and the disuse of tables and desks, “the likelihood of a number of rolls being open simultaneously … while actually writing was … extremely low”. J.P. SMALL, Wax Tablets of the Mind: Cognitive Studies of Memory and Literacy in Classical Antiquity, London, Routledge, 1997, p. 167; EAD., Artificial Memory and the Writing Habits of the Literate, in Helios 22 (1995) 159-166, p. 161.

402

D.B. SLOAN

positions Derrenbacker considered. Derrenbacker solves this problem by suggesting “that Matthew relies on Q by memory, particularly at those places when he deviates from the order of Q”3. More recently Alan Kirk devoted a monograph to explaining Matthew’s practice here4. What neither Derrenbacker nor Kirk addresses is the problem of Luke’s source utilization. Derrenbacker is satisfied with the fact that Luke keeps his Markan and Q material in separate blocks and that Luke follows the order of his source in each of those blocks, practices “both feasible and consistent with the known practices of writers in antiquity”5. But what Derrenbacker does not note is that Luke presents not just double tradition material in his non-Markan blocks but also special material, and he repeatedly moves back and forth between the two. It is true that this could be accomplished if Luke pulls the special material from his memory (or if he is making it up as he goes), but what this does not explain is why Luke would do this only in his non-Markan blocks. Consider how rarely Luke adds new material in his six Markan blocks6: 1. Luke 4,31–6,19 is a 71-verse block. Every verse here was taken from Mark except for one eight-verse passage about a miraculous catch of fish (Luke 5,1-8)7. 2. Luke 8,4–9,50 is a 103-verse block. Only two of these verses are not found in Mark (Luke 9,31-32), but these verses seem to be an explanation of what Luke takes to be implicit in Mark’s text, not an independent tradition8. 3. Luke 18,15-43 is a 29-verse block. Every single verse here comes from Mark. 4. Luke 19,28–24,11 is a longer block (241 verses) and Luke incorporates non-Markan traditions in the entry narrative (19,39-44), the Last Supper 3. DERRENBACKER, Conditions (n. 1), p. 444. 4. KIRK, Q in Matthew (n. 2). 5. DERRENBACKER, Conditions (n. 1), 443. 6. Luke’s block approach to composing his gospel has often been noted. See, for example, H.J. CADBURY, The Style and Literary Method of Luke (HTS, 6), Cambridge, MA, Harvard University Press, 1920, pp. 76-77; B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, p. 167; R. BULTMANN, The History of the Synoptic Tradition, trans. John Marsh, New York, Harper & Row, 21968, pp. 362-363; J. JEREMIAS, New Testament Theology. Part One: The Proclamation of Jesus (The New Testament Library), trans. John Bowden, London, SCM; New York, Scribner’s, 1971, pp. 39-40; etc. 7. Luke 5,39 is a possible exception, but this verse is probably a non-Western interpolation. 8. Luke notes that Moses and Elijah were speaking about Jesus’s coming exodus and that Peter and John and James were drowsy but then became fully awake. There is also a shorter insertion at 9,42b-43a, which notes that Jesus gave the previously demonized boy back to his father and that the people were astonished.

THE WIDOW AT NAIN

403

(22,15-16.28-38), the trial scene (23,2-17), and the crucifixion (23,2732.39-43), but otherwise this lengthy block is simply Luke’s version of Mark 11–169. We see here that Luke will occasionally devote space to explaining something in Mark and occasionally but rarely insert a non-Markan tradition (Luke 5,1-8; 19,39-44; 23,27-32.39-43) in his Markan blocks. But, in general, the Markan blocks are simply Luke’s rewriting of Mark’s material, almost always in fewer words than Mark devoted to this material. Now consider the non-Markan blocks: 1. Luke 3,1–4,30 starts with a mix of Markan, Q, and other material in 3,1-6 and then gives a speech from Q in 3,7-9, special material in 3,10-15, Q material in 3,16-1710, a mix of Markan and special material in 3,18-20, Markan material in 3,21-22, special material in 3,23-38, Q material in 4,1-13, and special material with Markan material mixed in in 4,14-3011. 2. Luke 6,20–8,3 starts with Q’s beatitudes (6,20-23), to which Luke adds woes (6,24-26), and then gives the rest of the Sermon on the Plain from Q, but with additions at 6,33-35a.37b-38a. Luke then gives Q’s passage about the centurion’s servant but adds a delegation of elders in 7,3-6a. Then he includes an L passage (7,11-17), followed by a Q passage (7,18-35) with redactional additions at 7,20-21 and possibly 7,29-30, another L passage (7,36-50), and some closing verses that are either redactional or from L (8,1-3)12. 3. Even worse is Luke 9,51–18,14, where Luke supposedly switches sources 71 times, while drawing 153 verses from Q and 198 verses from L13. 4. In the short non-Markan block that is Luke 19,1-27, Luke supposedly includes an L passage (19,1-10) followed by a Q passage (19,12-14.16-26) where Luke again adds details about a delegation (19,15.27), along with his own introduction (19,11)14. 9. Luke also adds explanatory details in the entry narrative (19,37), the parable of the Tenants (20,18), the Apocalyptic Discourse (21,18.22.24.34-36), the arrest of Jesus (22,48.51), the crucifixion (23,48), and the empty tomb (24,7-8). Luke 22,43-44 is likely an interpolation. 10. For a close analysis of Luke 3,1-17 in the context of the Synoptic Problem, see O. ANDREJEVS – D.B. SLOAN, Matthean Conflations and Luke’s Utilization Procedure under the Farrer Hypothesis: In Defense of F. Gerald Downing, in JBL 141 (2022) 737-759. 11. J.M. ROBINSON – P. HOFFMANN – J.S. KLOPPENBORG (eds.), The Critical Edition of Q: Synopsis Including the Gospels of Matthew and Luke, Mark and Thomas with English, German, and French Translations of Q and Thomas (Hermeneia), Minneapolis, MN, Fortress; Leuven, Peeters, 2000 (hereafter, CritEd), pp. 2-43. 12. Ibid., pp. 44-149. 13. Ibid., pp. 150-523. 14. Ibid., pp. 524-557.

404

D.B. SLOAN

In other words, Luke operated entirely differently in the non-Markan blocks than he did in the Markan blocks. But what if our picture of Q is distorted and Luke actually works the same way in the non-Markan blocks as he does in the Markan blocks? We generally identify Q by looking for non-Markan verses that are shared by Matthew and Luke. But if Matthew is not working consecutively through Q as he is through Mark, then it could be that Matthew misses a lot of the Q material. Luke works consecutively through both his sources, and just as he rarely inserts a non-Markan tradition into the Markan blocks, the non-Markan blocks may simply be Q material, much of which Matthew omits. In my forthcoming book, I work through each of the non-Markan blocks and argue this hypothesis verse-byverse. For the purpose of this essay, I will limit myself to two passages in Luke’s second non-Markan block. As far as I am aware, no one in the last hundred years has suggested that either of these passages comes from Q15, so these passages should be a good test of whether Luke’s special material can be adequately explained as Q material that Matthew omitted. These passages are the only two in the second non-Markan block that are entirely unparalleled in Matthew: the Widow at Nain (Luke 7,11-17) and the Sinful Woman Who Loved Much (Luke 7,36-50). II. THE WIDOW AT NAIN (LUKE 7,11-17) In the Gospel of Luke, the Centurion’s Servant (from Q) and the Raising of the Widow’s Son (unparalleled in Matthew) form a pair. Both passages exhibit an Elijah-Elisha typology16, and specifically correspond to the two 15. H.H. WENDT, Die Lehre Jesu, Göttingen, Vandenhoeck & Ruprecht, 1886, vol. 1, pp. 79-81, argues that Luke 7,36-50 came from Q. Nineteenth-century German scholarship tended to attribute much more of Luke’s special material to Q than twentieth- and twentyfirst century scholarship does. 16. J.D. OTTEN, The Bad Samaritans: The Elijah Motif in Luke 9.51-56, in JSNT 42 (2020) 375-389, p. 379, notes: “Connections between the centurion narrative and the account of Naaman in 2 Kgs 5 include: a well-respected Gentile officer as the primary character (vv. 2, 4-5; cf. 2 Kgs 5.1), the recommendation of high-ranking officials on his behalf (vv. 3-5; cf. 2 Kgs 5.5), the fact that neither Elisha nor Jesus interacts with the subject directly, but communicates through emissaries (vv. 6-9; cf. 2 Kgs 5.5-10), and a healing from a distance (v. 10; cf. 2 Kgs 5.14)”. See also J. SHELTON, The Healing of Naaman (2 Kgs 5:1-19) as a Central Component for the Healing of the Centurion’s Slave (Luke 7:1-10), in J.S. KLOPPENBORG – J. VERHEYDEN (eds.), The Elijah-Elisha Narrative in the Composition of Luke (LNTS, 493), London – New York, Bloomsbury T&T Clark, 2014, 65-87. J.B. GREEN, The Gospel of Luke (NICNT), Grand Rapids, MI, Eerdmans, 1997, p. 290, notes “resounding echoes” between Jesus’ raising of the widow’s son and Elijah’s raising of the widow’s son in 1 Kgs 17,17-24, most notably “the identification of the dead man as the only son of a

THE WIDOW AT NAIN

405

stories Jesus relates of Elijah and Elisha in Luke 4,25-27 (which I would also argue is from Q). Both passages focus on the plight not of the person who will get healed but of the person who most relies on the one in need of healing17. The two are also a gender pair18. Gender pairs are a regular phenomenon in Q (7,29-30; 11,31-32; 13,18-21; 15,4-9; 17,34-35; cf. 7,32; 12,53; 14,26)19. The two passages are also designed to contrast one another: “The centurion was wealthy and had influence and power. … [The woman] is a poor and powerless widow … [with] no one to fend for her and no hope for her future”20. And finally, these passages show that Jesus can heal either a person “at the point of death” (7,2) or a person “who had [already] died” (7,12). Rather than attributing this pairing to the design of Q, it is standard to see this as an example of Luke doing something that we never see him do with Markan material: mix two sources to create a paired set of stories. François Bovon summarizes well the general position on this passage: “Although it is transmitted in one of Luke’s Q passages (6,20–7,10; 7,18-35), it did not originate in Q. If Matthew had known this account, he would surely have not omitted it”21. Bovon gives no reason why Matthew “would surely have not omitted it”. Matthew does not exhibit the same interest in widows as Luke, omitting Mark’s reference to widows in Mark 12,40 and the story of the Widow’s Offering in Mark 12,41-4422. Matthew also has a Markan story about a child being raised from the dead (Matt 9,18-26), widow, the meeting of the prophet and the widow at the gate of the city, and the return of the resuscitated son to his mother”. See also T.L. BRODIE, Towards Unravelling Luke’s Use of the Old Testament: Luke 7:11-17 as an Imitatio of 1 Kgs 17:17-24, in NTS 32 (1983) 247-267. 17. J.R. EDWARDS, The Gospel according to Luke (PNTC), Grand Rapids, MI, Eerdmans, 2015, p. 214. 18. Ibid., p. 213. 19. S. PARKS, Gender in the Rhetoric of Jesus: Women in Q, Lanham, MD, Lexington Books – Fortress Academic, 2019, pp. 77-109. Parks also includes two more gender pairs in Q (rightly as we will argue below): Luke 4,25-27 and Luke 11,5-8//Luke 18,2-5 (ibid., pp. 84-89). Of course, Luke also creates gender pairs of his own, though not as frequently (Luke 1,5-23; 2,25-38; Acts 9,32-43), so this Q-like feature does not necessitate that this passage is from Q. 20. D.E. GARLAND, Luke (ZECNT), Grand Rapids, MI, Zondervan, 2012, p. 299. 21. F. BOVON, Luke 1: A Commentary on the Gospel of Luke 1:1–9:50 (Hermeneia), trans. Christine M. Thomas, Minneapolis, MN, Fortress, 2002, p. 266. 22. H.P. WEST Jr, A Primitive Version of Luke in the Composition of Matthew, in NTS 14 (1967-68) 75-95, pp. 80-82. West also notes that in all the Matthean miracle stories, the person in need of healing or someone related to that person comes humbly to Jesus to request healing, but here “Jesus approaches the woman involved and performs the cure without any prior display of devotion or humility on her part” (ibid., p. 82). This too may have contributed to Matthew’s disuse of this passage.

406

D.B. SLOAN

so Matthew may have seen this second resuscitation account as unnecessary. The circularity of the typical reasoning about Matthew’s use of Q is problematic: we tend to assume that Matthew rarely omitted material from his sources and based on this assumption limit Q to the double tradition, which then reinforces the idea that Matthew wouldn’t have omitted material like this if he knew it. The fact is that Matthew had two sources that often overlapped, and he had a preference for his Markan source. If both related a story of Jesus raising a child from the dead, he may very well have omitted the Q version. Bovon recognizes that the presence of non-Lukan elements in this passage necessitates Luke’s use of a source here, but he assumes that this source is not Q: “Although, as a rule, Luke alternates long passages from Mark and Q, he generally takes smaller individual accounts from his special material and inserts them into whatever passage from Mark or Q he is currently following”23. But we have seen that Luke does not generally do this in the Markan material! If we have a passage from an unknown source sandwiched between two passages where the source is known, the default assumption should be that the middle passage also comes from the same source. Scholars regularly note how well this passage fits the sequence of material in Luke 6,20–7,3524. To quote John Carroll: “The beatitudes pronounced blessed those who weep (6:21). When Jesus comes upon a funeral procession, he presses beyond words to actions that give them substance. Restoring to life a widow’s only son, he turns her weeping to joy and revives her hopes for the future”25. And of course, Q 7,22 assumes that John’s disciples have seen Jesus raise the dead. Luke 7,11-17 not only provides the climactic miracle that Q 7,22 assumes26, but the wording particularly anticipates Q 7,22. Luke 7,14 has Jesus give the command ἐγέρθητι, anticipating Jesus’ 23. BOVON, Luke 1 (n. 21), p. 266. Bovon gives Luke 5,1-11 as an example of Luke doing this with Mark. We will discuss this below. 24. I.H. MARSHALL, The Gospel of Luke: A Commentary on the Greek Text (NIGTC), Exeter, Paternoster; Grand Rapids, MI, Eerdmans, 1978, p. 283. 25. J.T. CARROLL, Luke: A Commentary (New Testament Library), Louisville, KY, Westminster John Knox, 2012, p. 163; cf. EDWARDS, Luke (n. 17), p. 215. 26. This is typically seen as the reason for Luke’s “insertion” of this story at this point: MARSHALL, Luke (n. 24), p. 283; J.A. FITZMYER, The Gospel according to Luke I–IX (AB, 28), New York, NY, Doubleday, 1981, p. 655; H. SCHÜRMANN, Das Lukasevangelium. Erster Teil: Kommentar zu Kap. 1,1–9,50 (HTKNT, 3/1), Freiburg i.B. – Basel – Wien, Herder, 1969, p. 398; J. NOLLAND, Luke (WBC, 35), Dallas, TX, Word, 1989-93, vol. 1, p. 320; BOVON, Luke 1 (n. 21), p. 266; etc. Of course, if Luke simply needed a resurrection story to prepare for Q 7,22, he could have either move the story of Jairus’ daughter here or placed Q 7,18-35 with Q 9,57-60, since the story of Jairus’ daughter is recounted in the next chapter of Luke’s Gospel (8,40-56).

THE WIDOW AT NAIN

407

statement in 7,22 that “the dead are raised [ἐγείρονται]”27. Q 7,18-19 also makes more sense if more than one miracle has just been narrated: “And John, ⟦on hearing about all these things⟧, sending through his disciples, ⟦said⟧ to him: Are you the one to come, or are we to expect someone else?”28. Even if the CritEd is wrong in following Luke’s “all these things” rather than Matthew’s “the works of Christ”, it is clear that Q 7,18 assumes that Jesus has already performed multiple miracles, not just the healing of the centurion’s servant. Why not assume that the Raising of the Widow’s Son is one of those miracles29? In fact, there are several indications that Q 7,18-35 was constructed with this passage in mind. Q 7,18 assumes that a report has reached John. The going out of a report is narrated in Luke 7,17, which uses language (“spread through the whole of Judea and all the surrounding country”) that is reminiscent of Matt 9,26//Luke 4,1430, which is likely also from Q, and of Matt 3,5//Luke 3,3, which is often attributed to Q31. Q 7,26-28 may have been designed as a response to an acclamation like the one in Luke 7,16. The people assume in 7,16 that Jesus is “a great prophet”, and so the following passage (shared with Matthew) points out that even John was “more than a prophet” and yet the least in the kingdom is greater than he. Jesus is therefore much more than a great prophet32. It seems that Luke has preserved Q’s sequence of Centurion’s Servant, Widow at Nain, and John’s Question. III. THE SINFUL WOMAN WHO LOVED MUCH (LUKE 7,36-50) The next passage in Luke builds on the preceding passages in several ways. After Jesus performed miracles reminiscent of Elijah and Elisha, the people naturally concluded, “A great prophet has arisen among us” (7,16). This is probably what earned Jesus an invitation to the home of the Pharisee, 27. CARROLL, Luke (n. 25), p. 166. 28. CritEd (n. 11), pp. 118-119. 29. We could also note that Q 7,28 continues the interest in Elijah-Elisha typology, identifying John as the prophesied Elijah (cf. Mal 3,1; 4,5), and “the least in the kingdom” as greater than John (EDWARDS, Luke [n. 17], p. 217). The Elijah-Elisha typology continues in the next double tradition passage as well (Q 9,57-62). 30. S. HULTGREN, Narrative Elements in the Double Tradition: A Study of Their Place within the Framework of the Gospel Narrative (BZNW, 113), Berlin – New York, De Gruyter, 2002, p. 314. 31. So CritEd (n. 11), pp. 4-7. See the discussion in ANDREJEVS – SLOAN, Matthean Conflations (n. 10). 32. GARLAND, Luke (n. 20), p. 299.

408

D.B. SLOAN

for the first words out of the Pharisee’s mouth when Jesus does something unexpected are “If this man were a prophet…”33. And in the immediately preceding passage, Jesus has just claimed that tax collectors and prostitutes accepted John’s message but Pharisees and lawyers did not (Q 7,29-30)34. These tax collectors and prostitutes have thereby entered the kingdom and become greater than the greatest of those born of women, including the Pharisees and lawyers (Q 7,28). Here we have a sinful woman, probably a prostitute, demonstrate this in her act of greater love than the Pharisee who hosts Jesus (7,44-47)35. John Nolland suggests that the “woman’s readiness for Jesus and evident prior acquaintance with the forgiveness that comes through him [are] an indication that her life had already been touched by the ministry of John”, illustrating the principle of 7,29-3036. And verse 34 has just claimed that the accusation the Pharisees would have against Jesus is that he is “a glutton and a drunkard, a friend of tax collectors and sinners”. Sure enough, in verse 39 the Pharisee’s complaint is that Jesus is allowing himself to be touched by “a sinner”37. As Joel Green notes, our passage demonstrates that Jesus’ companionship with sinners is not only characteristic of his ministry but is itself “an outgrowth and material manifestation of the ministry of salvation”38. That Jesus could experience such intimacy with a sinner demonstrates the fact that he has caused sinners to justify God. But it is not just that this passage builds on the previous ones. It is also designed in much the same way as Q 7,18-35. James Edwards notes: “In both accounts sinners appear as protagonists of the gospel (7:29, 47-50) and Pharisees as antagonists (7:30, 40-47). In both Jesus makes his decisive point by means of a brief parable in the midst of the story (7:32, 41-42). And finally, in both the fate of Jesus’ counterpart – John the Baptizer in the

33. K.E. BAILEY, Poet & Peasant and Through Peasant Eyes: A Literary-Cultural Approach to the Parables in Luke, Combined Edition, Grand Rapids, MI, Eerdmans, 1983, vol. 2, pp. 10-11; cf. GREEN, Luke (n. 16), pp. 307-308. 34. In my book I argue that these words are to be seen as a quotation of Jesus, not as a parenthetical comment (as in most translations). 35. SCHÜRMANN, Lukasevangelium (n. 26), pp. 429-430; MARSHALL, Luke (n. 24), p. 304; NOLLAND, Luke (n. 26), vol. 1, p. 251. Cf. FITZMYER, Luke I–IX (n. 26), p. 684; BOVON, Luke 1 (n. 21), p. 291; EDWARDS, Luke (n. 17), p. 225. The extravagance of the woman’s expressions of love here – in comparison to what we might have expected from the host – is often noted. 36. NOLLAND, Luke (n. 26), vol. 1, p. 351. 37. SCHÜRMANN, Lukasevangelium (n. 26), p. 430; GREEN, Luke (n. 16), p. 306. Cf. FITZMYER, Luke I–IX (n. 26), p. 684. 38. GREEN, Luke (n. 16), p. 306.

THE WIDOW AT NAIN

409

first, and Simon the Pharisee in the second – is left open-ended”39. The passage also centers on the connection between cancelling debts and forgiving sins, a Q theme most evident in the Lord’s Prayer. According to either the Farrer Hypothesis or the standard Two-Document Hypothesis, these connections are due to Luke’s “literary creativity”40, but this is clearly not the case. Luke has replaced “prostitutes”, which Matthew preserves in his version of Q 7,29, with “all the people”, losing a key connection between the preceding passage and this one41. And we have seen already Luke’s habit of working with a source in blocks. It is likely that something like Luke 7,36-47 already existed after Q 7,18-35 in Luke’s source, something further suggested by the lack of influence of Mark 14,3-9 on Luke 7,36-50 and by Luke’s placement of this passage here rather than in its Markan location42. As is now widely recognized, both the scene and the parable in its center have come to Luke from tradition43. Why conclude that this tradition is not Q44? Simply because Matthew omits the passage? There are several reasons why Matthew might have omitted this passage. First, Matthew already has an anointing story from his favorite source, Mark. If he knew another from Q, he may well be likely to omit it just as Luke omitted Mark’s anointing story in favor of this one45. Second, Matthew does not share Luke’s (and Q’s) interest in women. We already saw that he omitted the raising of a widow’s son (Luke 7,11-17). According to The Critical Edition of Q, he also omits the woman-centered parable of the Lost Coin but keeps the man-centered parable of the Lost 39. EDWARDS, Luke (n. 17), p. 225. 40. E.g., see the enclosed essays in M. MÜLLER – J.T. NIELSEN (eds.), Luke’s Literary Creativity (LNTS, 550), London – New York, Bloomsbury T&T Clark, 2016 (with the exception of J.S. Kloppenborg’s Response). 41. On the woman of Luke 7,36-50 as a prostitute, see n. 35 above. 42. It is possible (though not certain) that the alabaster jar, the name Simon, and the act of anointing Jesus’ feet with myrrh are all intrusions into the story based on Mark 14,3-9, but if these points of contact are removed, the story still stands on its own (B.W. BACON, Studies in Matthew, New York, Henry Holt, 1930, p. 111; SCHÜRMANN, Lukasevangelium [n. 26], pp. 441-442). Bacon and Schürmann both argue that these are intrusions, but both think that they must have been inserted before Luke took up this story, or Luke would not have thought this story similar enough to Mark 14,3-9 for him to decide to omit the latter. 43. See especially NOLLAND, Luke (n. 26), vol. 1, p. 351. Cf. SCHÜRMANN, Lukasevangelium (n. 26), p. 441; MARSHALL, Luke (n. 24), p. 307; FITZMYER, Luke I–IX (n. 26), pp. 684685; BOVON, Luke 1 (n. 21), pp. 291-292; etc. 44. WENDT, Lehre (n. 15), pp. 79-81, previously argued that this passage came from Q and that Luke preferred the Q version to the Markan version. 45. R.K. MACEWEN, Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS, 501), London – New York, Bloomsbury T&T Clark, 2015, p. 104.

410

D.B. SLOAN

Sheep (Q 15,4-10). And as we have noted, in working with Mark he omits the Widow’s Offering (Mark 12,41-44), which is the only pericope in the last eight chapters of Mark that Matthew omits46. Third, this passage is rather scandalous. If a woman, seemingly a prostitute, letting down her hair and kissing Jesus’ feet is not scandalous enough47, Jesus’ willingness to publicly shame his host may be even more scandalous48. Bailey remarks: “To attack the quality of the hospitality offered, regardless of the circumstances, is unknown in fact or fiction, in personal experience or in traditional story”49. From beginning to end, this passage is embarrassing and ripe for omission had Matthew known it from his source. Even the parable would have been unnecessary for Matthew who has his own parable of two debtors (Matt 18,23-35)50. Robert H. Gundry argues that Matthew composed Matt 18,23-35 “by adapting the parable of the two debtors told to Simon the Pharisee”51. There may be too many differences between Matt 18,23-35 and Luke 7,41-43 to suggest direct dependence, but the fact that the two are parallel may further explain Matthew’s omission of this passage. MacEwen notes: “If Matthew omitted a short parable that he found in Luke in favor of a longer Sondergut parable containing similar elements, this is exactly what he did when he omitted Mark’s parable of the Seed Growing Secretly (4,26-29) and replaced it with the parable of the Weeds among the Wheat (Matt 13,24-30)”52. Considering that Luke tends to work with one source at a time, that this passage is intricately connected to the preceding Q passages, that Matthew would have had several reasons to omit this, that the passage is traditional, and that it runs with the Q themes of gender pairing (Simon vs. the prostitute), debt forgiveness (cf. Q 11,4), and contrasting Pharisees and sinners (cf. Q 7,29-30); the natural conclusion to draw is that this is a Q passage that Matthew has omitted. 46. BACON, Studies in Matthew (n. 42), p. 110, also notes Matthew’s omission of Mark 12,41-44 as parallel to his potential omission of the Lukan anointing story. 47. On the inappropriateness of the woman’s actions, see BAILEY, Poet (n. 33), vol. 2, pp. 9-10, and GREEN, Luke (n. 16), p. 310. But see also the qualifications in M.C. PARSONS, Luke (Paideia), Grand Rapids, MI, Baker Academic, 2015, pp. 129-130; EDWARDS, Luke (n. 17), p. 225. See also MACEWEN, Matthean Posteriority (n. 45), p. 104, n. 122: “Possibly, Matthew was as uncomfortable as Simon the Pharisee with Jesus’ feet being wiped by the woman’s hair and kissed by her lips”. 48. SCHÜRMANN, Lukasevangelium (n. 26), p. 431. 49. BAILEY, Poet (n. 33), vol. 2, p. 15. 50. MACEWEN, Matthean Posteriority (n. 45), p. 104. 51. R.H. GUNDRY, Matthew: A Commentary on His Handbook for a Mixed Church under Persecution, Grand Rapids, MI, Eerdmans, 21994, p. 371. 52. MACEWEN, Matthean Posteriority (n. 45), p. 104.

411

THE WIDOW AT NAIN

But that does not mean that all fifteen verses were taken from Q. The first twelve verses are intricately connected and concentrically designed (Simon and the woman’s actions – Simon’s conclusion – Jesus’ illustration – Simon’s conclusion – Simon and the woman’s actions), and verses 48-50 are somewhat out of place53. Not only this, but the pronouncement of forgiveness and the claim that her faith (expressed through her actions?) has saved her are at odds with the passage itself, in which the woman was forgiven before she expressed these faithful actions54. Verses 48-49 also closely parallel Luke 5,20-21, which is taken from Mark 2,5b-7, and verse 50 parallels Luke 8,48, which is taken from Mark 5,3455: Mark 2,5-7

Luke 5,20-21

Luke 7,48-49

καὶ ἰδὼν ὁ Ἰησοῦς τὴν πίστιν αὐτῶν λέγει τῷ παραλυτικῷ· Τέκνον, ἀφίενταί σου αἱ ἁμαρτίαι. ἦσαν δέ τινες τῶν γραμματέων ἐκεῖ καθήμενοι καὶ διαλογιζόμενοι ἐν ταῖς καρδίαις αὐτῶν· Τί οὗτος οὕτως λαλεῖ; βλασφημεῖ· τίς δύναται ἀφιέναι ἁμαρτίας εἰ μὴ εἷς ὁ θεός;

καὶ ἰδὼν τὴν πίστιν αὐτῶν εἶπεν· Ἄνθρωπε, ἀφέωνταί σοι αἱ ἁμαρτίαι σου. καὶ ἤρξαντο διαλογίζεσθαι οἱ γραμματεῖς καὶ οἱ Φαρισαῖοι λέγοντες· Τίς ἐστιν οὗτος ὃς λαλεῖ βλασφημίας; τίς δύναται ἁμαρτίας ἀφεῖναι εἰ μὴ μόνος ὁ θεός;

εἶπεν δὲ αὐτῇ· Ἀφέωνταί σου αἱ ἁμαρτίαι. καὶ ἤρξαντο οἱ συνανακείμενοι λέγειν ἐν ἑαυτοῖς· Τίς οὗτός ἐστιν ὃς καὶ ἁμαρτίας ἀφίησιν;

Mark 5,34

Luke 8,48

Luke 7,50

ὁ δὲ εἶπεν αὐτῇ· Θυγάτηρ, ἡ πίστις σου σέσωκέν σε· ὕπαγε εἰς εἰρήνην καὶ ἴσθι ὑγιὴς ἀπὸ τῆς μάστιγός σου.

ὁ δὲ εἶπεν αὐτῇ· Θυγάτηρ, ἡ πίστις σου σέσωκέν σε· πορεύου εἰς εἰρήνην.

εἶπεν δὲ πρὸς τὴν γυναῖκα· Ἡ πίστις σου σέσωκέν σε· πορεύου εἰς εἰρήνην.

53. BOVON, Luke 1 (n. 21), p. 293, notes: “The repeated introductory formulas (vv. 48 and 50 after v. 44) already betray the redactional expansion of the conclusion”. They also draw from Mark “and are narrative elements that allow a dignified though slightly formulaic and banal conclusion to the episode” (ibid., p. 298). 54. SCHÜRMANN, Lukasevangelium (n. 26), pp. 438, 440. While some have seen a similar idea in the use of ὅτι in verse 47, this is not the correct way to read that verse. See FITZMYER, Luke I–IX (n. 26), p. 687. 55. SCHÜRMANN, Lukasevangelium (n. 26), p. 440; MARSHALL, Luke (n. 24), p. 307; BOVON, Luke 1 (n. 21), p. 298.

412

D.B. SLOAN

It seems, then, that Luke found the original ending (7,47) insufficient. Marshall writes: “Luke could not leave the woman as simply the object of a discussion between Jesus and the Pharisee, and so his story had to contain some saying addressed to her”56. It is very telling, though, that Luke uses material from stories in Mark to close out this “incomplete” passage. The words of Jesus in 7,48.50 are identical to the words of Jesus in other episodes, and even the words of the other guests are almost identical to the words of others present when Jesus forgave sins at another time. In this way, Luke can add to his source (as he adds a new ending to the Entry Narrative or to the Apocalyptic Discourse in Mark), but when he does so he will take material from elsewhere if possible. And even still, this particular addition follows the rule we have observed in the Markan blocks that Luke does not really add new material but tries to interpret what he sees before him. He has read in his source that the woman is “forgiven much”, and he has found a pronouncement of forgiveness in his Markan source to use to have Jesus actually address the woman (Luke 7,48//Mark 2,5). Luke has also assumed that the guests in the home of the Pharisee would be as incensed as the Pharisees and scribes were when something similar happened in Mark. Q leaves the reactions unstated to invite the audience to offer their own reactions and draw their own conclusions about how Simon and the guests should have responded57. Luke reuses Mark to imagine what the guests’ reaction must have been in this case (Luke 7,49//Mark 2,6-7). Finally, Luke imagines how Jesus must have addressed the woman (Luke 7,50//Mark 5,34). To Luke, Mark gives him the keys for imagining what the text leaves implicit. These actions correspond with what Derrenbacker notes, that “[i]t is only when a pericope/ episode is concluded that the author will typically move to another parallel source if he choose”58. And Luke probably does not open his copy of Mark to find those verses but recalls them as best he can from memory. IV. CONCLUSION Not everything in Luke 7 is from Q. The echoes of Mark 2,5-7 and 5,34 at the end of the final passage, along with the likelihood that the story originally ended at 7,47, suggest that Luke has provided his own conclusion 56. MARSHALL, Luke (n. 24), p. 307. 57. BAILEY, Poet (n. 33), vol. 2, p. 20. 58. DERRENBACKER, Conditions (n. 1), p. 441.

THE WIDOW AT NAIN

413

here using Markan material. But the traditional view that our two passages do not come from Q is problematic. We have seen numerous points of contact between the double tradition and the special material here, including points at which the double tradition assumes something that we find in the special material and a point at which Luke obscures a connection through his redaction. We have also seen numerous reasons why Matthew might have omitted this material. Some of our observations, such as Matthew’s tendency to omit stories that have a parallel in Mark or that highlight women or that are scandalous for some other reason, have implications for other material in Luke’s non-Markan blocks. If Luke’s non-Markan versions of the Nazareth pericope (Luke 4,16-30) and the two greatest commandments (Luke 10,25-37) came from Q, might Matthew have omitted those versions in favor of Mark’s simpler versions? If Matthew sometimes omits women-centered passages, might Matthew have simply omitted the story of Mary and Martha (Luke 10,38-42), or the story of the woman who pronounces Jesus’ mother blessed (Luke 11,27-28), or the story of the woman who was healed on the Sabbath (Luke 13,10-17)59? And if Matthew avoids controversial or even scandalous details, might Matthew have chosen not to include the parable of the Prodigal Son (Luke 15,11-32), who squanders his wealth with prostitutes and is celebrated by his father without punishment? Perhaps Matthew’s own parable of the Two Sons (Matt 20,28-32) is a way of telling the story without including its scandalous details60. What about the parable of the Shrewd Manager (Luke 16,1-13) who is commended for dishonestly using his master’s money to save his own skin61? Or the unjust judge (Luke 18,1-8), who is surprisingly used as an analogy for God62? The so-called L parables are the most scandalous parables63. Might this be simply because Matthew omits the more scandalous stories in his second source? I contend that trusting the double tradition to give us a good idea of the extent of Q is problematic. Matthew’s ad hoc approach to utilizing Q makes him far more likely to omit Q material than Luke’s sequential approach, and Q’s material may have been less favorable to Matthew than Mark’s material. Therefore, I think we would get a better idea of the contents of Q 59. So WEST, Primitive Version of Luke (n. 22), pp. 80-82. 60. M. HENGEL, The Four Gospels and the One Gospel of Jesus Christ: An Investigation of the Collection and Origin of the Canonical Gospels, trans. John Bowden, Harrisburg, PA, Trinity, 2000, p. 183. 61. Notably Matthew does include the less scandalous conclusion of this parable (Luke 16,13) at Matt 6,24. 62. PARKS, Gender (n. 19), pp. 86-89, argues that the core of this story is from Q. 63. See G.P. ANDERSON, Seeking and Saving What Might Have Been Lost: Luke’s Restoration of an Enigmatic Parable Tradition, in CBQ 70 (2008) 729-749.

414

D.B. SLOAN

by looking at Luke’s non-Markan blocks than by looking at the double tradition. Perhaps our fascination with the Gospel of Matthew has caused us to miss what would have been clear if we knew nothing of the Gospel of Matthew, namely that Luke simply alternates between two sources in block in Luke 3,1–24,1164. In this case, almost all of the material in Luke 3,1–4,30; 6,20–8,3; 9,51–18,14; and 19,1-27 (and probably also 22,15-16.28-38) would be from Q, with Luke very rarely adding an explanatory comment or a new tradition. I argue this in my forthcoming volume on Luke’s second source. One might object that we do not expect to find narrative episodes like the Widow at Nain (Luke 7,11-17) and the Sinful Woman Who Loved Much (Luke 7,36-50) or the story of Zacchaeus (Luke 19,1-10) or the Healing of a Disabled Woman (Luke 13,10-17) in a collection of sayings of Jesus, but passages like the Temptation narrative (Q 4,1-13), the Centurion’s Servant (Q 7,1-10), John’s Question (Q 7,18-35), the Potential Followers (Q 9,57-60), and the Beelzebul Accusation (Q 11,14-26) suggest that Q was never just a collection of sayings but a series of episodes that highlight what Jesus would say on each occasion65. Matthew’s use of Q for material he could add to his five great discourses would have caused him to omit narrative introductions to other passages in his second source66. If this is the case, Q 12–19 may have looked much more like Luke 12–19 than like its Matthean counterparts67. 64. Luke 24,12 is a non-Western interpolation taken from the Gospel of John. 65. M. GOODACRE, The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, pp. 170-185, surveys key evidence in the CritEd that Luke’s source had narrative properties. Narrative features of Q have also been noted in J.G. WILLIAMS, Parable and Chreia: From Q to Narrative Gospel, in Semeia 43 (1988) 85-114; F.G. DOWNING, Quite Like Q: A Genre for ‘Q’: The ‘Lives’ of Cynic Philosophers, in Bib 69 (1988) 196-225; HULTGREN, Narrative Elements (n. 30); M. LABAHN, Der Gekommene als Wiederkommender: Die Logienquelle als erzählte Geschichte (ABG, 32), Leipzig, Evangelische Verlagsanstalt, 2010; H.T. FLEDDERMANN, The Plot of Q, in ETL 88 (2012) 43-69; D. DORMEYER, Narrativität und Theologie der Wunder in Q, in C. HEIL – G. HARB – D.A. SMITH (eds.), Built on Rock or Sand? Q Studies: Retrospects, Introspects and Prospects (BiTS, 34), Leuven – Paris – Bristol, CT, Peeters, 2018, pp. 213-230; etc. 66. R.V. HUGGINS, Looking for the Wrong Q in the Wrong Place, Paper read at the Annual Meeting of the Society of Biblical Literature, Baltimore, MD, 24 November 2013, Q section; cf. J.C. HAWKINS, Probabilities as to the So-Called Double Tradition of St. Matthew and St. Luke, in W. SANDAY (ed.), Studies in the Synoptic Problem, Oxford, Clarendon, 1911, 95-138, esp. p. 124. 67. A distinction between Q 3–11 and Q 12–22 is often assumed, simply because we find narrative introductions throughout the former and none in the latter, but notably such a distinction is found only in Matthew’s version of these passages, where he places most of Q 12–22 in his five great discourses. The passages in Luke 11,37–18,14; 19,1-27 are structured much like those in Q 3–11, and the episodic nature of Luke’s central section (which still

THE WIDOW AT NAIN

415

In Derrenbacker’s conclusions for the Synoptic Problem, he argues that “[t]he theory that has [the] subsidiary gospel(s) displaying the least amount of conflation at the pericopal level works best”68. The assumption that Q contained little more than the double tradition requires Luke to have engaged in micro-conflation throughout his non-Markan blocks, despite the fact that he almost never does this in his Markan blocks. A modified Two-Document Hypothesis, one in which Q includes most of Luke’s special material, minimizes the amount of conflation Luke does at the pericopal level, making Luke’s use of Q and his use of Mark consistent. In the two examples we provided, there is much evidence that this is exactly what has happened – Luke has preserved in his non-Markan blocks Q passages that Matthew has omitted. Therefore, I contend that this modified TwoDocument Hypothesis is the synoptic theory that “works best”69. Ashland Theological Seminary United States of America

David B. SLOAN [email protected]

devotes 83% of its words to sayings of Jesus) likely better reflects the structure of Q than CritEd does. 68. DERRENBACKER, Conditions (n. 1), p. 444. 69. I have not addressed here the theory that Matthew used Luke as well as Mark, as argued by R.V. HUGGINS, Matthean Posteriority: A Preliminary Proposal, in NT 34 (1992) 1-22; HENGEL, Four Gospels (n. 60); E. POWELL, The Myth of the Lost Gospel, Las Vegas, NV, Symposium, 2006; MACEWEN, Matthean Posteriority (n. 45); A. GARROW, Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis, in NTS 62 (2016) 207-226. I agree with many of the arguments made by these scholars, who have demonstrated that Matthew’s knowledge of much of the special material in Luke is quite possible. For several of these authors, however, the source for Luke’s non-Markan blocks is left unaddressed. A simple comparison of Mark with Luke suggests that Luke knows two sources, Mark, and what I think we can call Q. The Matthean Posteriority Hypothesis does not obviate the need for a hypothetical source; some written source must explain Luke’s non-Markan blocks. The question, then, becomes whether Matthew obtained the double tradition material from Luke or from Luke’s source. For reasons beyond the scope of this paper, it seems to me that the evidence favors the latter.

THE (MINOR) MINOR AGREEMENTS: A STATISTICAL ANALYSIS OF MATTHEW’S AND LUKE’S PARAPHRASTIC TENDENCIES

It is easier to find qualitative studies in biblical scholarship than statistically informed quantitative analysis, and the Synoptic Problem is no exception. Shorter studies focus on specific passages to highlight the strengths of this or that hypothesis. Even larger, more holistic studies in monographs and commentaries often effectively pile short studies of individual passages one on top of the other to amass evidence and note patterns, but without much in the way of quantitative assessment overall. Approaching the Synoptic Problem in this piecemeal fashion has persuasive force, but it carries some problems. First, the passages are so small and the redactional behavior so particular that the quantitative differences between Matthean and non-Matthean, Lukan and non-Lukan are often insignificant1. Second, there is always the danger that a convincing argument reflects the interpretative cleverness of the scholar, not the strength of the data. After all, the paradigm shift within synoptic studies from models of Matthean priority such as Augustine’s or Griesbach’s (still humming along in some circles) to models of Markan priority sees a reversal in the explanations for passages shared by Matthew and Mark: the reasoning seamlessly switches from Mark’s redaction of Matthew to Matthew’s redaction of Mark without any actual new data2. Such a paradigm reversal serves 1. See also below on Mark 8,31; 12,22. 2. The change is rather on which data to focus. Take a simple case in Mark 5,22//Luke 8,41//Matt 9,18: Griesbachians David B. Peabody, Lamar Cope, and Allan J. McNicol see Mark’s εἷς τῶν ἀρχισυναγώγων ὀνόματι Ἰάϊρος as borrowing εἷς from Matthew and the rest from Luke’s ὄνομα Ἰάϊρος … ἄρχων τῆς συναγωγῆς (One Gospel from Two: Mark’s Use of Matthew and Luke, Harrisburg, PA, Trinity Press International, 2002, p. 146). Michael D. Goulder, on the Farrer model, instead has Matthew simplify εἷς τῶν ἀρχισυναγώγων to ἄρχων εἷς since it never uses ἀρχισυνάγωγος, which Luke then conflates as ἀνήρ … ἄρχων τῆς συναγωγῆς (Luke: A New Paradigm [SupplJSNT, 20], Sheffield, JSOT, 1989, p. 424). A Two-Document Hypothesis explanation might largely agree with Goulder on Matthew’s use of Mark while explaining Luke’s formulation on its own terms (e.g., ἄρχων τῆς συναγωγῆς adapted from LXX). M. WOLTER, The Gospel According to Luke, Volume 1 (Luke 1–9:50)

418

J.M. TRIPP

as a warning that our models influence how we perceive the data as much as the data inspire models to explain them. Truly holistic quantitative analyses of the synoptic data are less common, especially ones that go beyond tallying numbers. Although quantitative claims are (explicitly or implicitly) made in synoptic studies, sifting data from three texts whose word-lengths number in the tens of thousands is, well, tedious, and it is difficult to publish quantitative studies that present all the data for scrutiny. I don’t provide mine here, after all, only summaries. As a result, methods for measuring and assessing the data remain unrefined, while the symbols and vocabulary of statistical analysis can appear unfamiliar and “abstruse”3. Yet studies of the Synoptic Problem are replete with (often implicit) quantitative claims about coincidence, likelihood, and authorial tendencies. Good research design balances quantitative and qualitative analysis. An anecdotal case inspires an idea and the idea is tested more widely. Or conversely, patterns are noted in the broader data that provoke us to understand them at the level of the particular. This essay reflects a small effort to contribute to quantitative discussions of the Synoptic Problem. My goal is to analyze the Minor Agreements (MAs) in the context of Matthew’s and Luke’s wider redactional habits, understood as a collection of paraphrases of the sayings and stories recorded in the Gospel of Mark. If Matthew and Luke adapted Mark independently, we still expect them to agree to alter Mark on a great many words, perhaps thousands. Even allowing for so many coincidental agreements, as we will see the actual number of MAs is significantly higher. Furthermore, conditional probability consistently demonstrates sufficient evidence that Luke is an influential variable on Matthew (as defined below), but not enough evidence that Matthew is an influential variable on Luke. While not without its caveats, the evidence from these quantitative assessments coheres well with the Matthean Posteriority Hypothesis (MPH), but is less coherent with the Two-Document Hypothesis (2DH) and the Farrer Hypothesis (FH).

(BMSEC), trans. Wayne Coppins and Christoph Heilig, Waco, TX, Baylor University Press, 2016, p. 359. Any interpretive lens can explain the phrasing since there is so little data to work with. 3. See the review of A. ABAKUKS, The Synoptic Problem and Statistics, Boca Raton, CRC, 2015, in D.L. MEALAND, The Synoptic Problem and Statistics: A Review, in JSNT 40 (2017) 236-241, p. 241.

THE (MINOR) MINOR AGREEMENTS

419

I. THE MINOR AGREEMENTS Broadly speaking, Minor Agreements are points in the triple tradition where the Gospels of Matthew and Luke share textual features that differ from those in their presumed source text, the Gospel of Mark4. An example highlighted by Mark Goodacre appears in Mark 12,225. Some Sadducees describe a woman married successively to seven brothers who each die when: Mark 12,22: Last of all also (ἔσχατον πάντων καί) the woman died. Luke 20,32: Later also (ὕστερον καί) the woman died. Matt 22,27: And later than all (ὕστερον δὲ πάντων) she died, the woman.

Both Matthew and Luke agree in changing Mark’s ἔσχατον to ὕστερον6. The question is why. Did one influence the other? Or is it a coincidence? Occasionally quantitative reasoning is employed to answer these questions. As Goodacre points out, Matthew uses ὕστερον this way (×7) more often than Luke (once), suggesting to him that Matthew influenced Luke’s substitution7. Frans Neirynck claims coincidental redaction cannot be ruled out, but Goodacre counters that “while all things are possible, not all things are equally probable”8. Granted, the probability that a random gospel passage comes from Matthew given that it uses ὕστερον is higher (7/9) than from Luke (1/9), the probability that Matthew or Luke uses ὕστερον given a sequential use of ἔσχατον in Mark is considerably more balanced (2/2 and 1/2 respectively [see Mark 12,6 and parallels])9. The MA is easy to spot, but its individual significance is unclear. 4. Not all MAs are clear-cut, leading to disagreement over what constitutes triple tradition, the level and nature of bias in synopses, and how to count agreements; see M.E. BORING, The “Minor Agreements” and Their Bearing on the Synoptic Problem, in P. FOSTER – A. GREGORY – J.S. KLOPPENBORG – J. VERHEYDEN (eds.), New Studies in the Synoptic Problem: Oxford Conference, April 2008. Essays in Honour of Christopher M. Tuckett (BETL, 239), Leuven – Paris – Walpole, MA, Peeters, 2011, 227-251. 5. The Case Against Q: Studies in Markan Priority and the Synoptic Problem, Harrisburg, PA, Trinity Press International, 2002, pp. 154-155; earlier ID., Goulder and the Gospels: An Examination of a New Paradigm (SupplJSNT, 133), Sheffield, Sheffield Academic Press, 1996, pp. 115-116. 6. 1 Cor 15,8 also deals with resurrection and pairs ἔσχατον with πάντων; F. NEIRYNCK, Goulder and the Minor Agreements, in ETL 73 (1997) 84-93, p. 89 (= ID., Evangelica III [BETL, 150], Leuven – Paris – Sterling, VA, Leuven University Press – Peeters, 2000, 307-318, p. 313). 7. Case (n. 5), p. 155. John 13,36 uses ὕστερον in the context of death and resurrection. 8. Ibid. 9. See below on conditional probability. Luke 20,13 eliminates any sequential modifier from Mark 12,6. Although not sequential, see also πάντων ἔσχατος in Mark 9,35 and ἔσχατοι (twice) in Mark 10,31. Matthew and Luke opt to change or eliminate Mark’s uses of ἔσχατος

420

J.M. TRIPP

Beyond stoking some mild intellectual curiosity, MAs gather scholarly attention due to the role they play in arguments against the 2DH, which poses Matthew and Luke as independent adaptations of Mark10. Advocates of the 2DH highlight the many different choices the authors of Matthew and Luke made in adapting Mark to support the claim that they worked independently. It seems reasonable then that opponents of the 2DH counter with the many similar choices they made in adapting Mark, especially when they both altered Mark – the Minor Agreements. If the later authors worked independently, so the thinking goes, they would rarely alter Mark in the same way without one text influencing the other. Michael Goulder takes the argument to an extreme: “Logically, it would take only one really striking Minor Agreement to destroy [the 2DH]”11. No such case has been found. Instead, advocates of direct dependence models such as FH may focus on quantity over (insurmountable) quality. As Michael McLoughlin remarks while commenting on a MA in Mark 5,27 par., “A few cases of this sort might be explained by coincidence, but there are several hundred”12. Over the last few decades, 2DH scholarship has built up defenses against the qualitatively stronger MAs that include appeals to shared oral tradition13, Mark-Q overlaps or multiple recensions of Mark14, or textual more often than not. On the FH, the probability that Luke adapts Matthew’s ὕστερον is only 1/7. Quantitatively in Mark 12,22, we should also take account of changes on which they do not agree (Luke subtracting πάντων; Matthew substituting δέ, transposing ἀπέθανεν). Qualitatively, a case could be made for Matthew’s phrasing as a conflation of Mark’s and Luke’s (ἔσχατον πάντων + ὕστερον = ὕστερον πάντων). 10. M. MCLOUGHLIN, Listing the Minor Agreements, in ETL 87 (2011) 201-228, p. 202 claims, “the MA may be regarded as the principal literary data which separate the [2DH] from its many competitors”, while Goodacre calls them “the Achilles heel of the two-source theory” (Case, p. 152). 11. Luke’s Knowledge of Matthew, in G. STRECKER (ed.), Minor Agreements: Symposium Göttingen 1991 (GTA, 50), Göttingen, Vandenhoeck & Ruprecht, 1993, 143-160, p. 143. S.D. BLACK, One Really Striking Minor Agreement: TIΣ EΣTIN O ΠAIΣAΣ ΣE in Matthew 26:68 and Luke 22:64, in NT 52 (2010) 313-333, attempts to position an addition to Mark 14,65 as one such case. 12. Listing the Minor Agreements (n. 10), p. 201 (emphasis added). Luke and Matthew add τοῦ κρασπέδου in the same place to Mark’s ἥψατο τοῦ ἱματίου αὐτοῦ, “she touched the hem of his clothes”. One may speculate why they coincidentally add this element, of course. The position McLoughlin notes here is that this happens too often for coincidence to provide the best explanation for all the data. 13. Compare already I.H. MARSHALL, The Gospel of Luke: A Commentary on the Greek Text (NIGTC), Exeter, Paternoster; Grand Rapids, MI, Eerdmans, 1978, p. 358 on Luke 9,10-17: “there are one or two agreements with Mt. against Mk. which may go beyond mere coincidence, but they hardly testify to anything more than the existence of continuing oral traditions” (emphasis added). 14. In other words, largely untestable literary solutions employing hypothetical texts (either Q was not what we think, or Mark was not). For the former, see, e.g., H.T. FLEDDERMANN,

THE (MINOR) MINOR AGREEMENTS

421

corruption15. But the sheer number of MAs continues to pose a problem on its own. 2DH scholars defend their position by presenting anecdotal MAs as coincidental similarities in redaction. For example, neither Luke 9,22 nor Matt 16,21 repeat the articles from a list of Jesus’s adversaries, Mark 8,31: τῶν πρεσβυτέρων καὶ τῶν ἀρχιερέων καὶ τῶν γραμματέων. Goulder argues for Matthew’s influence on Luke since this and other MAs in the verse are “un-Lucan”. He tallies 19 other lists in Luke, of which only three drop the article after the first element, giving a probability of 3/19 (16%) that Luke would coincidentally drop the articles along with Matthew16. Goulder’s statistical method is oversimplified, but situating Luke’s alterations in the context of its broader behavior is a step in the right direction17. Soon after, Robert H. Gundry and Timothy A. Friedrichsen (along with Frans Neirynck) picked up the argument over a series of articles that unfortunately dropped focus on Goulder’s statistical argument, however limited it was18. These studies argue over things like whether to count Acts, how many lists Luke includes or whether case matters (nominative or otherwise), and why either author might drop articles. Whether any of their reformulations shift the statistical significance of Goulder’s findings is mostly left behind. Mark and Q: A Study of the Overlap Texts (BETL, 122), Leuven, Leuven University Press – Peeters, 1995. For the latter, see, e.g., A. ENNULAT, Die “Minor Agreements”: Untersuchungen zu einer offenen Frage des synoptischen Problems (WUNT, II/62), Tübingen, Mohr Siebeck, 1994. 15. See C.M. TUCKETT, The Minor Agreements and Textual Criticism, in STRECKER (ed.), Minor Agreements (n. 11), 119-142. It is unclear what percentage of MAs result from textual corruption. Since we actually have access to text-critical data, unlike the other solutions, quantitative analysis of the impact of text-critical issues would be immensely helpful. 16. GOULDER, Luke (n. 2), pp. 49-50. 17. Goulder notes four MAs here (the substitution of ἀπό for ὑπό, p = 5/66; of “on the third day” for “after three days,” p = 1/1; and of ἐγείρω for ἀνίστημι, p = 2/8). Treating them as independent, he calculates the likelihood of Luke making all four changes by multiplying the probabilities to get p = 1/334 (0.3%). Goulder’s point is only that these specific changes are “un-Lucan”, not that Luke’s treatment of the verse as a whole is uncharacteristic. For that, he would have to account for all agreements and disagreements. Although I would not use this as a method to determine (in)dependence, comparing the probability of those choices against the MAs would contextualize the latter. 18. R.H. GUNDRY, Matthean Foreign Bodies in Agreements of Luke with Matthew against Mark, in F. VAN SEGBROECK – C.M. TUCKETT – G. VAN BELLE – J. VERHEYDEN (eds.), The Four Gospels 1992: FS Frans Neirynck (BETL, 100), Leuven, Leuven University Press – Peeters, 1992, 1467-1495; see also ID., A Rejoinder on Matthean Foreign Bodies in Luke 10,25-28, in ETL 71 (1995) 139-150; ID., The Refusal of Matthean Foreign Bodies to Be Exorcised from Luke 9,22; 10,25-28, in ETL 75 (1999) 104-122. Cf. F. NEIRYNCK – T.A. FRIEDRICHSEN, Note on Luke 9,22. A Response to M.D. Goulder, in ETL 65 (1989) 390-394 (= Evangelica II, ed. F. VAN SEGBROECK [BETL, 99], Leuven, Leuven University Press – Peeters, 1991, 43-48); T.A. FRIEDRICHSEN, Luke 9,22 – A Matthean Foreign Body?, in ETL 72 (1996) 398-407.

422

J.M. TRIPP

Gundry agrees with Goulder that Luke was unlikely to make this change, so Luke must have gotten the idea from Matthew – despite the fact that Luke drops articles in lists elsewhere. Friedrichsen notes that Luke may have had similar reasons as Matthew to drop the articles, and that it is not impossible that Luke would alter Mark like this given other lists without articles, so the MA could be only a coincidence. Most MAs are unremarkable on their own, and even the more notable cases can perhaps be explained – again, individually – by coincidental redaction. So, the present state of the argument regards a situation where: 1) no single case carries enough weight to decide the issue, and 2) one side claims there are few enough cases that agreements could be coincidental, while the other side claims there are too many to be coincidence. Situations like this are precisely what statistical hypothesis testing is designed to address. With almost any quantitative claim, there is bound to be evidence supporting it and evidence against it. For example, if I flip a fair coin ten times, it is reasonable to expect something like five heads, five tails – but this is only an average. Sometimes I will get six heads, less often three, and very rarely (probability less than 0.1%) I may flip ten heads in a row. Now, if I have a thousand people flip a coin ten times independently, I expect the average tally to be very close to five heads with maybe one tally getting ten heads. If after collecting the data I find that 50 people got ten heads, I may suspect my original claim is not true, e.g., the coin is not fair or the tallies are not independent, say, if many people copied tallies off each other. Instinctively we suspect something is amiss because getting all heads once is one thing, but getting it so many times is just such an unlikely outcome. In fact, we can calculate the probability that at least 50 trials out of a thousand will turn out this way, and it is astronomically small – so low my calculator gave up and just said 0. In other words, the probability (or p-value) that our original claim is true and we just happened to get so many all-heads trials is infinitesimally close to impossible. If instead we got only two trials with all heads, however suspicious we may be, this outcome has a roughly 13% chance of happening. The extra round of all heads is a tiny red flag, but too weak to falsify our claim that one trial with all heads was a reasonable expectation using a fair coin. Deviations this small from our expectations will happen from time to time.

THE (MINOR) MINOR AGREEMENTS

423

Hypothesis testing assigns a probability that counterevidence is only coincidental. This is the p-value. As counterevidence mounts, in our example as one participant after another turns in a tally with all heads, the likelihood that the hypothesis is true and we just happened upon so much counterevidence starts to fall. At a certain point, an extremely low probability serves as a sign that something else is going on. That is, the low probability suggests the counterevidence is a significant (but not definitive) indication either that our original hypothesis does not hold, or that it is incomplete and additional variables are influencing the data. Below I apply methods of hypothesis testing to the 2DH claim regarding MAs that they are the result of coincidences in redactional choices made by Matthew and Luke writing independently. As previous statistical analyses of the MAs have shown, there is a highly significant overlap in the editorial choices made by Matthew and Luke. It is not that there are many MAs, as we might expect with substitutions of δέ for καί, but that there are too many to be consistent with coincidental redaction. Such a result cannot disprove the independence of Matthew and Luke, but it does issue a serious challenge to the 2DH. Furthermore, attending to conditional probabilities within the data may point to a particular direction of dependence: Luke is in some ways an influential variable on Matthew, making space to explore the MPH. II. STATISTICS

AND THE

SYNOPTIC PROBLEM

I am certainly far from the first to apply statistics to some question raised by this or that aspect of the Synoptic Problem19. Statistical analyses of just the MAs can be found, and they are fairly consistent in finding significant evidence against claims for Matthew’s and Luke’s independence20. So why another quantitative study of the MAs? One reason is that a question this large with such complicated issues of methodology is worth returning to, 19. A very incomplete list includes A.M. HONORÉ, A Statistical Study of the Synoptic Problem, in NT 10 (1968) 95-147; R. MORGENTHALER, Statistische Synopse, Zürich, GotthelfVerlag, 1971; the works surveyed by J.C. POIRIER, Statistical Studies of the Verbal Agreements and Their Impact on the Synoptic Problem, in CBR 7 (2008) 68-123; and, more recently, Z. GARSKÝ, Das Wirken Jesu in Galiläa bei Johannes (WUNT, II/325), Tübingen, Mohr Siebeck, 2012, pp. 7-27, with some similarities to Andris Abakuks’s methods (see below). This is not even my first attempt to apply statistics to the Synoptic Problem; see J.M. TRIPP, Measuring Arguments from Order for Q: Regression Analysis and a New Metric for Assessing Dependence, in Neotestamentica 47 (2013) 123-148. 20. As ENNULAT, Minor Agreements (n. 14), Vinson, and Abakuks (see below) all do, with differing interpretations of the significance.

424

J.M. TRIPP

worth testing through different lenses to see whether our results consistently agree. A second is that, despite challenging one of the theoretical pillars of the 2DH, earlier studies have not persuaded many 2DH scholars. It is worthwhile taking a moment to ask why. Here, two studies illustrate some of the strengths and weaknesses of prior studies, and the barriers they face in gaining broader influence in synoptic studies. Richard Vinson’s dissertation provides an early attempt to apply proper statistical analysis to the question of Matthew-Luke agreements beyond merely tabulating data21. Vinson counts MAs under five categories: changes in grammatical form, word used, word order, omissions, and additions22. He then compares the relative frequency of MAs in Matthew and Luke with two sets of data presented as analogous: quotations of Greek writers from early patristic apologists, and experimental data taken from independent redactions performed by contemporary graduate students. Vinson is not misguided in seeking comparative data, but the analogy must be very strong if it is going to tell us anything. As Friedrichsen points out, Vinson’s analogues do not meet this standard23. On the other hand, Friedrichsen critiques Vinson for counting MAs by word since, for example, a change in preposition may also force a change in case for the object that follows24. Yet the choice to alter Mark’s prepositions is not always taken, does not always alter the case of the object, and when it does it causes more lexical changes than it would otherwise, outcomes that counting by word can differentiate to some extent. Friedrichsen fears “inflation of the number of Minor Agreements” but fails to demonstrate that Vinson’s counting method inflates them disproportionately in one text or the other25, and misses the point that their significance does not lie in their absolute but relative number. More recently, Andris Abakuks published a mathematically rigorous but so far largely overlooked study, The Synoptic Problem and Statistics26. 21. R.B. VINSON, The Significance of the Minor Agreements as an Argument against the Two-Document Hypothesis, Ph.D. dissertation, Duke University, 1984. 22. His categories are similar to mine, with changes in grammatical form and word collapsed into substitutions. 23. T.A. FRIEDRICHSEN, The Minor Agreements of Matthew and Luke against Mark: Critical Observations of R.B. Vinson’s Statistical Analysis, in ETL 65 (1989) 395-408, here pp. 404-408. 24. Ibid., p. 401. 25. Ibid., p. 403, an objection VINSON, Significance of the Minor Agreements (n. 21), p. 424, anticipates and answers appropriately. 26. See also A. ABAKUKS, A Statistical Time Series Approach to the Use of Mark by Matthew and Luke, in J.C. POIRIER – J. PETERSON (eds.), Marcan Priority without Q: Explorations in the Farrer Hypothesis (LNTS, 455), London – New York, Bloomsbury T&T Clark, 2015, 119-139.

THE (MINOR) MINOR AGREEMENTS

425

Abakuks seeks to improve upon Honoré’s triple-linked study of agreements between the Synoptic Gospels and to use more refined statistical methods to determine the likely sequence of writing27. He finds the strongest evidence in favor of the FH while admitting there is also evidence for the MPH28. It is refreshing to see the proper use of statistical notation and vocabulary, but they may have contributed to the somewhat frosty reception of Abakuks’s work. The methodological discussions may have proven too “abstruse”, as Mealand puts it, for readers to follow much less appreciate or challenge29. Although we reach somewhat different conclusions, Abakuks’s methods and rigor merit consideration in synoptic studies. He limits his studies to Markan words that are retained exactly in Matthew or Luke, while I wish to broaden my scope to all Markan words as they are adopted, adapted, or omitted in later gospels. In what follows, my discussion and methods are notably less sophisticated than Abakuks’s, relying only on hypothesis testing and conditional probability using χ2 and two-proportion z-tests, calculations that are readily accessible and easy to perform. Furthermore, I attempt to ground the categorization of agreements and disagreements in a mode of understanding textual alterations widely available in the first century: παράφρασις or paraphrase exercises. III. PARAPHRASIS IN

THE

FIRST CENTURY

Matthew and Luke occasionally follow Mark’s wording quite closely. At other times, one or both varies wildly from the source text. A lively debate has emerged in synoptic studies over how to explain these variations in how they treat their source, with different camps crediting variations to literary concerns, oral performance, or (failures in) memory. Fortunately, a single widespread first-century practice incorporates all three modes of interacting with a text, perhaps allowing us to draw insights from each approach: paraphrase30. 27. See HONORÉ, Statistical Study (n. 19). 28. ABAKUKS, Synoptic Problem (n. 3), pp. 36-37, 41 (“the cases of the triple-link model that identify Luke as the last of the Synoptic Gospels to have been written provide the best fit … the cases Lk-Mk-Mt and Mk-Lk-Mt, where Matthew is the last gospel to be written, come into contention as providing as good a fit as the cases for which Luke is the last”), 181 (“inconclusive as to whether it was more likely that Luke had Matthew as a source or Matthew had Luke as a source”). 29. MEALAND, Review (n. 3) p. 241. 30. For earlier applications of paraphrase to the Synoptic Problem, see F.G. DOWNING, Compositional Conventions and the Synoptic Problem, in JBL 107 (1988) 69-85; T.A. BROOKINS,

426

J.M. TRIPP

Paraphrase exercises were widespread in grammatical and early rhetorical education31. Although perhaps only the author of Luke went beyond elementary education, learning was not so rigidly segregated as it is today. Elementary students learned side-by-side with more advanced students, in the same room or in the open air32, so they would most likely be exposed to paraphrase exercises. The practice of paraphrase is evident among the schoolhand papyri33, discussions of pedagogy (e.g., Suetonius, Gramm. 4; Quintilian, Inst. Or. 1,9,2; 10,5,4-11), and explained in detail in the firstcentury Progymnasmata of Aelius Theon. Paraphrase exercises begin and end with written texts, but develop through the use of memory and oral performance. A teacher assigned texts to memorize and recall verbatim, but this was only a first step (Theon, Prog. 141; Quintilian, Inst. Or. 2,7,2-4). The student would take a newly memorized text and begin to rework it in various ways, “deliberately taking up some sentences and turning them in as many ways as possible, just as one shape after another are usually made out of the same piece of wax” (Quintilian, Inst. Or. 10,5,9)34. (S)he might change it to sound like a different author, apply a saying to a new situation, or improve the clarity or style. At this stage, the student may perform the story multiple times to the teacher or to classmates for feedback. Only when all the kinks had been worked out would they commit their paraphrase to writing. Students worked their way up from paraphrasing famous sayings to short chreiai and eventually to whole narratives or speeches. Deviations from the original text Luke’s Use of Mark as παράφρασις: Its Effects on Characterization in the ‘Healing of Blind Bartimaeus’ Pericope (Mark 10.46-52/Luke 18.35-43), in JSNT 34 (2011) 70-89; E. EVE, Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS, 592), London – New York, Bloomsbury T&T Clark, 2021. 31. For detailed discussion, see J.M. TRIPP, Direct Internal Quotation in the Gospel of John (WUNT, II/493), Tübingen, Mohr Siebeck, 2019, pp. 25-32. Theon places paraphrase exercises among the early lessons of the rhetor (see M. PATILLON [ed.], Aelius Theon: Progymnasmata, Paris, Les Belles Lettres, 1997, pp. xxiv-xxvii), while Quintilian (Inst. Or. 1,9,1-3) insists a grammarian teach it. Suetonius (Gramm. 4,5) leaves it up to the teacher, but admits that paraphrase was often taught earlier in the past. 32. R. CRIBIORE, Gymnastics of the Mind: Greek Education in Hellenistic and Roman Egypt, Princeton, NJ, Princeton University Press, 2001, pp. 16, 26-28, 37-43. 33. T. MORGAN, Literate Education in the Hellenistic and Roman Worlds (CCS), Cambridge – New York, Cambridge University Press, 1998, pp. 200-209. 34. While teachers expected students to recall sayings of great teachers verbatim, these respected sayings were far from off limits. Students would “vary the grammatical constructions of notable sayings in all possible ways” (Suetonius, Gramm. 24,4); translation from R.A. KASTER (ed.), Suetonius: De Grammaticis et Rhetoribus, Oxford, Clarendon, 1995. Teachers may have gravitated toward well-known sayings so that students could understand them more deeply by playing with them.

THE (MINOR) MINOR AGREEMENTS

427

were prized as long as its sense could be maintained while hopefully improving its presentation – especially (and not despite) if the original text were still available. One wanted to put his or her own mark on the text because the original was well-known35. Mark’s sayings and short anecdotes lend themselves perfectly to revision through paraphrasis. Using this basic model, one can imagine the author of Matthew or Luke reading a miracle story, putting it to memory, then performing it multiple times in front of an audience to determine how best to revise it, clean up Mark’s grammar and syntax, improve the vocabulary and style, clarify what a pronoun referred to by changing the syntax or adding a referent36, or making connections to other passages they had recently been telling in the same manner37. The incorporation of this performative element helps to explain not only substitutions with the same sense in a different style, say replacing “kingdom of God” with “kingdom of heaven”, but also phonetic substitutions where the later text carries a different meaning using similar sounds. For example, Matthew substitutes Mark’s, “Now the mother-in-law (ἡ δὲ πενθερά) of Simon” (1,30) with, “He saw the motherin-law (εἶδεν τὴν πενθερὰν)” (8,14). Seeing her is a novel addition to the scene, but one can imagine a slip of the tongue or a noisy crowd shifting ἡ δέ into εἶδεν. As long as the author or the audience thought the change worked in the scene, it could feasibly work itself into the final written text38. Students practiced four types of paraphrase, first individually, then in combination: transposition, addition, subtraction, and substitution. Putting addition to the side due to its inherent methodological difficulties, I will examine Matthew’s and Luke’s changes to Mark under these categories, beginning with subtraction.

35. In pedagogical contexts, this is evident in that Homer by far provides the most paraphrased texts. 36. For example, ὁ δὲ εἶπεν (Mark 5,34), ὁ δὲ Ἰησοῦς … εἶπεν (Matt 9,22). 37. J.S. KLOPPENBORG, Variation in the Double Tradition and an Oral Q?, in ETL 83 (2007) 53-80, p. 61, argues for a similar model regarding the author of James’s use of Q and also in the pre-gospel transcription history of Q. It is unclear whether he allows the authors of Matthew or Luke to interact with Q in this manner, but if they could paraphrase Q, they could just as easily paraphrase Mark. 38. See also Mark 2,26 (πῶς) // Luke 6,4 (ὡς) (cf. Mark 12,26 // Luke 20,37); Mark 4,20 (ἐν) // Matt 13,23 (μέν); Mark 6,10 (ἐάν) // Luke 9,4 (ἣν ἄν); Mark 6,20 (εἰδὼς) // Matt 14,5 (ὅτι ὡς); Mark 8,4 (ὧδε) // Matt 15,33 (ὥστε); Mark 8,21 (oὔπω συνίετε) // Matt 16,11 (πῶς οὐ νοεῖτε); Mark 14,70 (γὰρ Γαλιλαῖος) // Matt 26,73 (γὰρ ἡ λαλιά σου). In Mark 14,32, “he said to his (αὐτοῦ) disciples, ‘Sit here (ὧδε)’”. In Matt 26,36, “he told the disciples, ‘Sit here (αὐτοῦ)’”. Since Matthew never uses αὐτοῦ in a locative sense elsewhere, it is most likely only due to Mark’s use of the phonetically/lexically identical possessive αὐτοῦ just before his verb that Matthew replaces ὧδε with it.

428

J.M. TRIPP

1. Subtraction Theon presents subtraction simply as the opposite of addition (Prog. 140). Subtraction involves the removal of elements from a sentence or story to quicken the pace or to eliminate extraneous detail, to remove objectionable material or to focus on what is important in this particular retelling39. As early as Plato’s Protagoras (342a-343b), both Socrates and his sophistic opponent advocate for laconic brevity in wise sayings, which may lead a paraphrast to subtract words if only to combat a perceived wordiness. Although they add much material in and around Mark, Matthew and Luke tend to shorten Mark’s stories, perhaps to fit more of their other material into Mark’s framework. This tendency to streamline Mark may have contributed to the subtraction of extraneous or repetitious details, or of long-winded descriptions and rambling conversations. Sometimes Mark’s adapters eliminate whole stories, perhaps because they objected to their content40 or because the stories were simply unnecessary to the larger narrative. When Matthew and Luke subtract the same word, phrase, or sentence, they create a negative Minor Agreement41. These shared subtractions are methodologically the least problematic: in most cases, they either remove a word or they do not. There is occasionally some ambiguity whether a shorter phrase in place of a longer one should be considered a substitution or a mixture of substitution and subtraction, but most of the time it is fairly straightforward to decide whether the later texts exclude an element from Mark. Matthew and Luke often subtract the same material, but again the counterclaim is that they subtract too often. In order to address that claim, we need to get an expected value for how often they should both subtract the same word, and there is no reason to assume that value will be small. To explain what I mean, let us take a small digression on expected values. (1) Sidebar: Expected Values The question is not whether there are a lot of Minor Agreements, but whether there are significantly more (or indeed significantly fewer) than we would expect on the assumptions we have made about how Matthew and Luke were composed. But how do we assign a value for how many agreements 39. See TRIPP, Direct Internal Quotation (n. 31), pp. 55-62. 40. Already B.H. STREETER, The Four Gospels: A Study of Origins, London, Macmillan, 1924, pp. 151-152, notes Mark’s “use of phrases likely to cause offense, which are omitted or toned down in the other Gospels”. 41. See the literature discussed in MCLOUGHLIN, Listing the Minor Agreements (n. 10), p. 204.

THE (MINOR) MINOR AGREEMENTS

429

we might expect in order to judge whether there are too many, and more to the point significantly too many? Returning to our coins, let’s imagine a preposterous situation: the authors of Matthew and Luke are each adapting Mark in separate cities, completely ignorant of each other. They are independent – so far, so good. Pushing the situation further, they adapt Mark in a completely ridiculous manner: for every word they come across in Mark, they flip a coin to decide whether to keep it (say, if the coin lands on heads) or omit it (if tails). We expect that sometimes both authors would flip heads, so both keep the word. Occasionally, however, both coins land on tails, so both omit against Mark, creating an MA. But taken individually shared omissions would not be signs of one author’s dependence on the other’s text. In this scenario, they are adapting Mark not only independently but randomly. Still, at least some of the time both authors will flip tails, so the number of negative MAs will not be 0. Quite the contrary: if they never agreed against Mark, we would be just as suspicious that they were not truly independent as if they agreed too often. Since they act independently, we can calculate how often we expect them to make certain choices. Let us take, for example, the case where both authors keep Mark’s words. The probability that Matthew keeps a word is 50%, as it is for Luke. So the probability that both authors independently keep the word is: P(Matt keeps AND Luke keeps) = (0.5)(0.5) = 0.25 = 25%42

Now, according to the NA28 edition, the Gospel of Mark has 11,082 words43. If we take this as a given, then we can calculate the expected number of words for each scenario: Table 1: Expected Values Using Independent Coin Flips Luke Keeps

Luke Omits

Matt Keeps

2770.5

2770.5

Matt Omits

2770.5

2770.5

42. We multiply the probabilities because we assume they are independent (although not strictly so in the technical sense). If they were not independent, calculating this probability becomes more complicated. 43. I include words bracketed in the critical text. I did not run a statistical analysis on the impact of NA28’s bracketed words, but since they tend toward harmonization, cases that bring Matthew and Luke into agreement against Mark may be outweighed by cases that harmonize Matthew or Luke with Mark, or that harmonize all three. For passage pairing (but not for the text), I use K. ALAND, Synopsis Quattuor Evangeliorum, Stuttgart, Deutsche Bibelgesellschaft, 151996.

430

J.M. TRIPP

In other words, in a situation that is completely (and artificially) independent and even random, we would still expect both Matthew and Luke to agree on omitting almost 3,000 words against Mark. So just finding ‘a lot’ of negative MAs is not enough to rule out even this artificially independent scenario44. Furthermore, it highlights why we cannot focus only on MAs when evaluating results. A 95% confidence interval for just the case where both omit ranges from 2,682 to 2,861 words. If we focus only on the MAs, e.g., as Goulder does with Mark 8,31, the actual number (2,839) sits comfortably within the confidence interval45. That is, paying attention only to MAs cannot rule out even this ridiculous compositional method. Yet none of the other results lie within their respective confidence intervals, and we can calculate the probability that a scenario in which two authors edit Mark like this would yield all the actual results in Matthew and Luke using a χ2 analysis46. The actual results of subtractions in Matthew and Luke’s adaptation of Mark are well outside the margins of error for flipping a coin (p < 0.00001%). At least we can rule out flipping fair coins as a compositional method. That is all right though – no one thinks the authors of Matthew and Luke flipped coins to decide what to keep from Mark. Instead, I am interested in how we can use a portrait very much like Table 1 based not on theoretical probabilities but on the behavior we observe in the later texts. To do that, let us look at how Matthew and Luke omit material from Mark in general. Since it is unrealistic to think that all words have an equal chance of being subtracted (or transposed or substituted) as Matthew and Luke adapt Mark, the tests we use act as a sort of low bar: significant results cannot prove dependence or influence (although they support such a hypothesis)47, but failing to show a significant difference from what effectively amounts to flipping two differently weighted coins suggests the evidence for widespread dependence or influence is extremely weak. 44. As I have discovered through personal correspondence, David B. Sloan is developing a similar line of thought from a 2DH perspective. 45. Similarly, M. MCLOUGHLIN, Using the Minor Agreements, in ETL 94 (2018) 93-114, p. 95. McLoughlin, seeking (qualitatively) significant MAs, claims 13 MAs (λέγει → εἶπεν) are coincidental. While the shift toward agreement on this particular point is not (statistically) significant, using his numbers, 13 MAs is still two more than we would expect taking all treatments of λέγει into account. 46. Chi-square analysis compares groups by categories (see Table 2) to determine the probability that results with the variation demonstrated in one group (say, expected results) would result in a number profile like the second group (i.e., actual results) without the influence of some other variable. Explanations for using χ2 are readily available online, as are calculators for inputting numbers to test them. 47. As ABAKUKS, Statistical Time Series Approach (n. 26), p. 121, recognizes.

431

THE (MINOR) MINOR AGREEMENTS

(2) Subtraction in Matthew and Luke Of Mark’s 11,082 words, by my count the Gospel of Matthew shows no trace of 4,186 (37.8%) of them, either exactly or in substituted form, while the Gospel of Luke shows no sign of 5,942 (53.6%) words48. With these proportions, we can produce a table with the expected values much as we did with our hypothetical scenario, and compare it with the actual results. Table 2: Expected and Actual Values of Subtraction

 Expected Results

 Actual Results

Luke Keeps

Luke Subtracts

Matt Keeps

3236 (30%)

3660 (33%)

Matt Subtracts

1905 (17%)

2153 (20%)

Luke Keeps

Luke Subtracts

Matt Keeps

3794 (35%)

3102 (28%)

Matt Subtracts

1347 (12%)

2839 (26%)

Based on their general patterns of subtraction from Mark, even while ignoring the specific content of the words, we expect Matthew and Luke to mutually subtract more than 2,000 words if they subtract independently. So, we already expect a great number of negative Minor Agreements, but as we can see, in the actual results there is a shift toward agreement in what they both keep (35% of words compared to 30% expected, a 5% shift or over 500 words) and mutually subtract (6% shift). This is a highly significant result (p < 0.001%). Since adapters who subtract independently at the rates the authors of Matthew and Luke do generally are incredibly unlikely to agree so often, there is considerable evidence to suspect that something has pushed them toward agreement. We should be wary of overstating the results, however. Saying they are significantly “not independent” does not necessarily mean they are “dependent” on each other. Mutually influential variables like shared ideologies, similar educations or basic stylistic concerns could shift the numbers toward agreement. If Matthew found a story offensive or disagreeable, it is possible that Luke would feel the same way. If Matthew finds this or that sentence overly wordy because of how he was taught to write in school, it is entirely possible that Luke would have learned similar stylistic values. Grammatical changes are preferable because they are measurable and testable. For example, Mark uses ὅτι to mark direct speech more than any other NT text (×34)49. Matthew and Luke are less likely to do so, a tendency 48. For tallies from HONORÉ, Statistical Study (n. 19), and J.B. TYSON – T.R.W. LONGSynoptic Abstract (The Computer Bible, 15), Wooster, OH, Biblical Research Associates, 1978, see ABAKUKS, Synoptic Problem (n. 3), pp. 17-21. 49. M.G. SIM, Marking Thought in New Testament Greek: New Light from Linguistics on the Particles ἵνα and ὅτι, Eugene, OR, Pickwick, 2010, p. 154.

STAFF,

432

J.M. TRIPP

shared with more literary writings. Removing this grammatical change from the tally as an outlier may be justified if it met the following quantitative criteria in addition to precedents set in literature and pedagogical discussions: • Both Matthew and Luke are significantly more likely to subtract ὅτι before direct speech than they are to subtract in general; • both Matthew and Luke are significantly more likely to agree on subtracting ὅτι before direct speech than they are to agree on subtracting words in general; • there is a significant shift toward Matthew’s and Luke’s typical usage of ὅτι before direct speech in non-Markan material. This is not without its methodological difficulties, however, and any analysis would have to establish clear criteria for problematic cases50. However, if such criteria could be established, 2DH scholars may begin to pick away at MAs due to grammatical concerns and to reassess the negative MAs to see whether their number remains significant. The following statistical observation makes it somewhat dubious that grammatical considerations will eliminate the significant overlap in negative MAs. We want to separate out whether Matthew and Luke subtract together so often because one influences the other or a prior variable influences both of them. One way to test that is to see how each behaves when material from the other is present versus when the other is absent. Afraid of overcounting negative MAs, Vinson does not count shared subtractions when either Matthew or Luke omits the whole passage51. I have counted them, in part because I cannot say for sure that the choice of the earlier adapter (whoever it was) to omit the whole passage had no influence whatsoever on the later adapter. However, Vinson makes a fair point that when Matthew omits the 93 words of Mark 1,23-28 altogether, while Luke 4,33-37 keeps the story but subtracts 16 words, it is difficult to credit Luke’s subtraction of a εὐθύς here or there specifically to its absence in Matthew since everything is missing in Matthew. 50. Having said that, ὅτι is likely to meet the first two criteria. In the 13 clear-cut cases, Matthew and Luke never agree to keep ὅτι. Luke Keeps

Luke Subtracts

Matt Keeps

0 (0%)

2 (15%)

Matt Subtracts

1 (8%)

10 (77%)

Ignoring subtractions of ὅτι with direct speech may be justifiable. 51. VINSON, Significance of the Minor Agreements (n. 21), pp. 17-18.

THE (MINOR) MINOR AGREEMENTS

433

One way to safeguard against this is to check whether the adapters behave the same way in the presence of the other or not. Is Luke’s subtraction of 16 words from Mark 1,23-28 (or from any passage that Matthew omits) otherwise normal? To do this, we would use a statistical concept called conditional probability, which is beginning to influence biblical studies through the application of Bayesian analysis52. Specifically, we could use conditional probability to test whether a variable is influential on our texts or not. Let me give an example: the empirical probability that Matthew subtracts a word from Mark is 37.8% by my count. Suppose, however, that instead of using NA28, I used a particular manuscript, say Codex Alexandrinus which includes the longer ending of Mark (16,9-20) and many variations from the reconstructed text. We might then expect the probability that Matthew subtracts a word from Mark given the condition that I use a different text to shift one way or the other. If the shift were significant, the choice of manuscript would be an influential variable. On the other hand, we expect the probability that Matthew subtracts a word from Mark given that it is Tuesday not to move at all. It would still be 37.8% since the day of the week is a completely independent, indeed unrelated variable from whether Matthew subtracts words from Mark, not an influential one. In short, truly unrelated variables will not affect the probability at all, while some relevant variables will not shift the probability much. If a variable does shift the probability significantly, we consider it an influential variable and we might try to ascertain the nature of the influence. Now, let us apply this reasoning to Matthew and Luke. First, we divide passages by whether Matthew includes them or not. According to Aland, Matthew shares no material in common with Mark in seven listed passages. If we look at how Luke behaves in these passages versus passages where Matthew includes material that may have influenced Luke’s adaptation of Mark, we get the following results:

52. See ABAKUKS, Synoptic Problem (n. 3), pp. 23-40. Bayes’ theorem focuses on the direction of conditionality. For example, the probability that Luke subtracts a word given that Matthew also does is not necessarily the same as the probability that Matthew subtracts a word given that Luke does. Bayesian analysis draws attention to what is known and what we conclude from it, information that is surprisingly easy to confuse. For an introduction to Bayesian analysis, see C. HEILIG, Hidden Criticism? Methodology and Plausibility of the Search for a Counter-Imperial Subtext in Paul, Minneapolis, MN, Fortress, 22017, pp. 27-35.

434

J.M. TRIPP

Table 3: Luke’s Subtractions in Matthean Material vs. Non-Matthean Material

Passages Matthew Includes Passages Matthew Omits Total

Markan Words

Luke Subtracts

Percentage

10666

5708

54.0%

416

234

43.5%

11082

5942

53.6%

Luke is less likely to subtract words from passages that appear elsewhere only in Mark, which at first glance could be interpreted as a move toward preservation. But is Luke significantly less likely to subtract in these passages? As it turns out: no. Using a two-proportion z-test, p = 27.3%, an insignificant result53. There is insufficient evidence in this data to support the claim that Matthew is an influential variable on Luke’s tendency to subtract material from Mark. If the results continue this way, it would be a positive sign that grammatical and stylistic considerations might have a shot at explaining the MAs. To be thorough, however, we should also test Matthew’s behavior when Luke is present versus when Luke is absent. Table 4: Matthew’s Subtractions in Lukan Material vs. Non-Lukan Material Markan Words

Matt Subtracts

Passages Luke Includes

10049

3660

36.4%

Passages Luke Omits

1033

526

50.9%

11082

4186

37.8%

Total

Percentage

Here we see the opposite tendency on Matthew’s part: it seems to subtract at a higher rate when Luke is absent. When we run the same twoproportion z-test on Matthew, however, we get a surprising result: the probability that Matthew subtracts at such a higher rate coincidentally when Luke is absent is less than 0.0001%, highly significant54. In other words, there is significant evidence that Luke is an influential variable on Matthew. This does not easily cohere with the 2DH, which posits them as independent (so we would not expect the variance to be significant), or with the FH where we would expect the results to run in the opposite direction, with Matthew an influential variable on Luke. 53. A two-proportion z-test is used to test whether two proportions are significantly different provided the sample sizes are big enough (they are). 54. The p-value is approximately 5.66 × 10-20, or 19 orders of magnitude smaller than Luke’s p-value.

THE (MINOR) MINOR AGREEMENTS

435

These results are consistent, however, with the MPH. We might even offer a tentative interpretation of Matthew’s behavior: if Luke omits an entire passage, it sets a precedent and the author of Matthew more freely subtracts material from Mark – either partially or entirely – than it does when Luke includes the passage. 2. Transposition Theon presents paraphrase according to order as its most basic form, when the paraphrase keeps “the same words and we get diverse formulations by a transposition of elements” (Prog. 139-140)55. Since Greek word order is more fluid, a variety of sentences can be arranged from the same lexical elements without changes in grammar. Still, shifting word order can also shift emphasis and introduce or alleviate ambiguity56. I note here two choices I made in counting transpositions. First, I counted only transpositions within a sentence or between two sentences as marked in the NA28 critical text. Luke is rather more likely to move whole sentences around within a passage, while Matthew is more likely to move whole passages around within Mark’s outline. The latter are not covered by paraphrase exercises, and there does not seem to be enough evidence to warrant counting the former under this paradigm, so I did not count transpositions of whole sentences. Second, shuffling lexical elements can get complicated quickly. Just because both adapters move an element later does not mean they move it to the same place. Therefore, I only count them as equivalent transpositions when the words are adjacent to the same Markan element or in the same absolute position within a sentence, such as its beginning or end. This choice renders the counts consistent, but deeper methodological discussion is perhaps warranted in light of ancient examples of paraphrase. To begin with, I consider whether each adapter transposes the word or not. Of the 5,141 words that Luke keeps, 692 (13.5%) are transposed. Of the 6,896 words that Matthew keeps, only 699 (10%) are transposed. Predicting an expected value of shared transpositions introduces a new wrinkle, though: initially, I consider only words they both keep so that each has a chance of transposing a word or phrase and we consider the same sample text from Mark57. 55. PATILLON (ed.), Progymnasmata (n. 31), p. 108; Quintilian agrees with Theon (Inst. Or. 1,9,2). 56. See TRIPP, Direct Internal Quotation (n. 31), pp. 42-51. 57. We could consider a transposition in one hypertext and a subtraction in the other as a disagreement in how to paraphrase Mark. For methodological ease, however, I have chosen

436

J.M. TRIPP

Table 5: Expected and Actual Values of Transposition

 Expected Results

 Actual Results

Luke Keeps

Luke Transposes

Matt Keeps

2951 (77.8%)

459 (12.1%)

Matt Transposes

333 (8.8%)

52 (1.4%)

Luke Keeps

Luke Transposes

Matt Keeps

3052 (80.4%)

367 (9.7%)

Matt Transposes

244 (6.4%)

131 (3.5%)

Again, there is a shift toward agreement, although in this case more weight shifts toward agreement with Mark (80.4% compared to an expected 77.8%, or a 2.6% shift) than against Mark (2.1% shift). Still, they agree to alter Mark’s word order more than twice as often as expected, and the total results are again highly significant (p < 0.001%). We may also be curious whether they not only agree to transpose a word but also move it to the same position. Of the 131 words they both transpose, 81 (61%) are equivalent transpositions58. There is strictly no way to measure the significance of this observation, although I may note that 81 words moved to the same position is substantially more than the 52 words we expected them to move together at all. Now let us check the conditional probability. Here we are not concerned just with words they both adapt but with the total words in passages of each type since the given information is only whether or not the other text adapts the passage. Table 6: Luke’s Transpositions in Matthean Material vs. Non-Matthean Material Markan Words (Kept)

Luke Transposes

Passages Matthew Includes

4959

673

13.6%

Passages Matthew Omits

182

19

10.4%

5141

692

13.5%

Total

Percentage

to compare only words they both include which should balance out between the expected and actual tallies. 58. Matt 3,11 and Luke 3,16 move elements from Mark 1,8 (ἐγὼ ἐβάπτισα ὑμᾶς ὕδατι) before all material adapted from Mark 1,7-8. Matthew’s ἐγὼ μὲν ὑμᾶς βαπτίζω ἐν ὕδατι does not match Luke’s syntax (ἐγὼ μὲν ὕδατι βαπτίζω ὑμᾶς), so only three words are counted as equivalent transpositions.

THE (MINOR) MINOR AGREEMENTS

437

Again, Luke seems more conservative in passages omitted by Matthew than in passages Matthew includes. Yet there is no significant difference between the groups (p = 22.4%), which is to say there is insufficient evidence to claim that Matthew was an influential variable on Luke. Table 7: Matthew’s Transpositions in Lukan Material vs. Non-Lukan Material Markan Words (Kept)

Matthew Transposes

Percentage

Passages Luke Includes

6389

633

9.9%

Passages Luke Omits

507

66

13.0%

6896

699

10.1%

Total

Yet again, Matthew transposes more often in passages that Luke omits than in passages included by Luke, but here the difference is significant (p = 2.6%), although not highly significant. With transposition, one suspects Matthew’s p-value is lower than Luke’s more so because the sample is larger (507 words versus 182) than because Matthew’s behavior shifts so dramatically. However, it is worth noting that these results hold even if we consider just the 3,794 words they both adapt from Mark with the words in passages only one adapts. Luke transposes 498 or 13.1% of words shared with Matthew, and 10.4% in passages Matthew omits – but this difference is insignificant (p = 29.3%). Matthew meanwhile transposes 375 or 9.4% of words shared with Luke, but 13.0% in passages Luke omits. Although the samples are closer in size, Matthew’s differential is now larger than Luke’s and remains significant (p = 2.9%). Still, the results for transposition are rather weaker than for subtraction. 3. Substitution According to Theon, substitution is the most advanced form of paraphrase, one to be avoided unless the student is confident (s)he can retain or improve the sense of the original text (Prog. 140-142)59. The category is broad for Theon, including grammatical changes, using synonyms or metaphorical words in place of more literal ones. I have followed Theon’s broad definition of substitution. Of the 5,141 words that Luke keeps, it substitutes 2,006 (39.0%). Of the 6,896 words that Matthew keeps, it substitutes 2,222 (32.2%). Again, 59. See TRIPP, Direct Internal Quotation (n. 31), pp. 62-73.

438

J.M. TRIPP

based on these general patterns in how they treat the text of Mark overall, we can calculate expected values for the 3,794 words they both keep from Mark: Table 8: Expected and Actual Values of Substitution in Matthew and Luke

 Expected Results

 Actual Results

Luke Keeps

Luke Substitutes

Matt Keeps

1568 (41.3%)

1004 (26.5%)

Matt Substitutes

745 (19.6%)

477 (12.6%)

Luke Keeps

Luke Substitutes

Matt Keeps

1954 (51.5%)

632 (16.7%)

Matt Substitutes

454 (12.0%)

754 (19.9%)

As we can see, the actual results shift substantially toward agreement. A χ2 analysis gives p < 0.001%, a highly significant result. However, when they agree to alter Mark’s wording, Matthew and Luke are not terribly likely to substitute with the same word or phrase. Whereas we expect Matthew and Luke to agree to substitute more than 12% of their shared material, they alter less than half of that (225 words or 5.9%) using the same wording. One could argue for example that one text changing Mark’s καί to δέ may spur the other to consider a τότε more frequently, but the discrepancy lowers the impact of these results. Finally, if we turn to conditional probabilities, we see familiar outcomes. Table 9: Luke’s Substitutions in Matthean Material vs. Non-Matthean Material

Passages Matthew Includes

Markan Words (Kept)

Luke Substitutes

Percentage

4959

1929

38.9%

Passages Matthew Omits

182

77

42.3%

Total

5141

2006

39.0%

Unlike in previous cases, Luke is less conservative with substitution when Matthew is absent but not by a wide enough margin to be significant (p = 35%). Table 10: Matthew’s Substitutions in Lukan Material vs. Non-Lukan Material

Passages Luke Includes

Markan Words (Kept)

Matthew Substitutes

Percentage

6389

2033

31.8%

Passages Luke Omits

507

189

37.3%

Total

6896

2222

32.2%

THE (MINOR) MINOR AGREEMENTS

439

Matthew is also less conservative with material not found in Luke (as usual), but by a wider margin (5.5% compared with Luke’s 3.4%) that marks a significant difference (p = 1.1%). Yet again, Luke is an influential variable on Matthew while there is not enough evidence to support the claim in reverse. IV. CONCLUSION In paraphrasis, we have a popular first-century model for editing sayings and short narratives like those in the Gospel of Mark. The model was available to the authors of Matthew and Luke through their Greek education, evident in the production of medium-length, coherent Greek narratives that reflect Greek literary conventions. Whether they reached the level of the grammarian themselves, they very likely witnessed more advanced students practice and present paraphrase exercises. They were undoubtedly exposed to the storytelling and literary cultures of the first century, which prized creative modifications to sayings and stories to clarify them or draw out the depths of their meaning. This includes the scriptures Matthew and Luke cite throughout their lives of Jesus, and is evident in the modifications they make to the same sacred writings when they cite them60. My focus, however, is not on the role of paraphrase in the first century, not directly at least. If we adopt one of the many other models for understanding how the authors of Matthew or Luke may have gone about adapting Mark, focusing on their subtractions, transpositions, and substitutions still seems to be an appropriate way of subdividing the data while assessing the quantitative character of the Minor Agreements61. In each case, we get consistent results: • The number of MAs is significantly higher than we would expect from purely independent treatments of Mark. In other words, the quantitative data supports a case for direct influence, or at least fails to rule it out.

60. For what could be interpreted as paraphrase in Hebrew Bible narratives, see G. SAVRAN, Telling and Retelling: Quotation in Biblical Narrative (Indiana Studies in Biblical Literature), Bloomington, Indiana University Press, 1988. For the gospels’ paraphrastic citations of scripture, see R.J. MILLER, Helping Jesus Fulfill Prophecy, Cambridge, Lutterworth, 2016. 61. It is no surprise that previous statistical analyses have employed similar breakdowns. However, the model one employs affects how the data is categorized, which then affects the summary statistics one reaches.

440

J.M. TRIPP

• There is insufficient evidence to support the claim that Matthew is an influential variable on Luke. • There is sufficient evidence to support the claim that Luke is an influential variable on Matthew, strongest of all with subtraction but not ruled out with transposition or substitution. While Luke’s direct influence on Matthew is not the only possible explanation, it remains an hypothesis worth pursuing. At a glance, these findings are inconsistent with the 2DH, which minimizes mutual influence between Matthew and Luke, at least without ad hoc appeals to oral tradition, hypothetical textual reconstructions, or textual corruption. The data are also inconsistent with the FH, where Matthew influences Luke, not vice versa. But the data are consistent with the MPH. Under this model, one might expect Matthew to behave differently when adapting a saying or story knowing (or consulting) Luke’s version than when only Mark’s version is available. As has hopefully become clear, ‘influence’ does not imply direct borrowing from Luke, i.e. making changes only when Luke does or especially the same changes. We are not testing individual choices. ‘Influence’ means that knowing how Luke behaves in a passage makes it easier to predict how Matthew will behave. I do not want to fall prey to overstating how decisive these findings are but rather to encourage further statistical analysis. So I would like to end by offering paths for critiquing the conclusions I have drawn about the data. One may object that counting MAs by words inflates their number, but doing so proportionally inflates the expected value as well. Again, it is not that there are many MAs by this counting method, but that there are too many. Some previous scholars have preferred to measure by editorial choices62, but this is nebulous and subjective, and it fails to reflect the different weight that omitting an article has when compared to subtracting a whole story. Most of the people reading this essay are also writers; I’m sure they can appreciate the difference between the many minor revisions they make as they refine their work, and pressing delete on a whole page. Measuring by words is a consistent, well-defined metric that reflects this difference63. Still, if another scholar were to choose a different metric that is also 62. See MCLOUGHLIN, Listing the Minor Agreements (n. 10); BORING, Minor Agreements (n. 4), and effectively FRIEDRICHSEN, Minor Agreements (n. 23). 63. A somewhat obvious self-critique: under the influence of paraphrase as practiced in the first century, I have failed to count transpositions of whole sentences, changes in order that are surely more disruptive than simply shifting a word here and there. My choice greatly lowers the number of words counted as transpositions to avoid the methodological headache

THE (MINOR) MINOR AGREEMENTS

441

well-defined – that is, any analyst applying the same method would reach the same tally – then the conclusions they reached would be worthy of consideration. Another line of questioning might focus on the imbalance of influence, with decisions that Luke made apparently shifting the behavior in Matthew but not the other way around. One critique here may be that Matthew deletes passages wholesale less often than Luke does, so the sample of Lukan material unfound in Matthew is quite a bit smaller. Smaller samples raise the risk of Type II errors, more colloquially “false negatives”64, because the smaller sample size contributes to the lower significance (higher p-values) in Luke’s behavior. Some factors mitigate this concern. First, the samples are not tiny (at the smallest, 182 and 507 words), although Luke’s constitutes less than 4% of the total words adapted from Mark. Second, in two cases Luke simply doesn’t vary as much between Matthean passages and nonMatthean ones as Matthew does in the reverse circumstances, which also lowers the significance of the Lukan data65. A final issue is the influence of the particular synopsis used to determine which material was omitted. A different synopsis would almost certainly alter the counts, shifting probabilities. In some cases, Luke’s may slide into significance or (more likely) Matthew’s into insignificance. This is worth testing. If it were the case, it would be noteworthy if only to better understand how large an influence synopses have over our (perception of the) data66. A way to mitigate the influence of the synopsis and to refine the data may be to divide by sentences rather than whole passages, at least for subtraction (although this choice comes with its own methodological difficulties)67. of deciding which passages to count as transposed when there is no logical relationship between them (on this difficulty, see TRIPP, Arguments from Order [n. 19], pp. 127-134). 64. In more formal language, one commits a Type II Error when one fails to reject a false hypothesis, e.g., that Luke was not influenced by Matthew. Margins of error shrink as sample sizes grow, so choosing samples that are too small (even if it is all the material that fits the criteria) makes it more difficult to show a significant difference between them. This is why I have used the phrasing of having insufficient evidence to support a claim for difference rather than pushing the evidence too far and claiming that there is no difference. 65. Matthew’s subtractions in Lukan vs. non-Lukan passages have a difference of 14.5%, Luke’s difference between Matthean and non-Matthean passages is only 9.5%; Matthew’s difference is 5.5% in substitution, Luke’s 3.4%. However, since Matthew’s difference is significant at 3.1% for transposition, while Luke’s is slightly larger (3.2%) yet insignificant, the size of the sample clearly had an impact (see above). 66. Although ABAKUKS, Statistical Time Series Approach (n. 26), p. 122 finds the results of his statistical analysis not seriously affected by the choice of Honoré or Tyson and Longstaff, the matter is still worth exploring. 67. See also GARSKÝ, Das Wirken Jesu (n. 19), pp. 44-80 for quantitative methods for determining best-fit parallels using exact word-matches, which may be applied to build a data set rather than a synopsis.

442

J.M. TRIPP

A more basic critique that will undoubtedly form in some readers’ minds is the use of statistics at all. After all, the use of p-values in the tests I have run here assume that one word is equally subject to subtraction, transposition, and substitution as any other. This is not strictly the case68. It is for this reason that I am cautious about overstating the implications of the evidence I have put forward. There are significantly more Minor Agreements than we would expect on the assumption of independence, and Matthew behaves significantly differently when Luke adapts a passage than when it does not. All that is claimed is that the shifts are signs of something, that direct dependence or influence cannot be ruled out. Perhaps a 2DH scholar will demonstrate that the probability of paraphrasing in certain ways is significantly higher than others. In that case, the data will still stand, but the interpretation of its significance will shift. My goal is not to lay the matter to rest, but to spur challenges in whatever small way I can so that statistical methods for testing quantitative claims in synoptic studies may be refined, theoretical dead ends cut off, and quantitative engagement can grow alongside qualitative arguments.

68. Compare already D.A. CARSON, Exegetical Fallacies, Grand Rapids, MI, Baker, 1984, p. 142, on a related issue 40 years ago: “word frequency statistics are normally calculated on the basis of the null hypothesis … [which] figures out how likely various occurrences would be in comparison with a random drawing of words out of a barrel. But writers do not choose words that way. There may be convincing contextual or topical reasons why some words are chosen in one context and not in another” (emphasis added). J.S. KLOPPENBORG, The Farrer/Mark without Q Hypothesis: A Response, in POIRIER – PETERSON (eds.), Marcan Priority without Q (n. 26), 226-244, p. 232, rightly notes that probabilities for paraphrase are not equal for all words. Attending to multiple dimensions of paraphrase lessens the impact of these shifting probabilities. Kloppenborg uses the feeding of 5,000 (Mark 6,30-44 and parallels) as an example case, claiming the story could not be told without certain words in common. Yet one need only compare John’s version (6,1-15) to see that the same story can be told with far fewer lexical agreements, much less agreements against Mark. Still, ABAKUKS, Statistical Time Series Approach (n. 26), p. 121, acknowledges the issue already in his contribution to the same volume: “in examining the individual words that make up the text of Mark, as they are used by Matthew or Luke, it is readily seen that the words are not used independently of each other. Words retained unchanged are not randomly distributed in the text but tend to come in clusters of varying sizes … As a result, it is not valid to use elementary statistical techniques” as if a significant result intrinsically proves our case (although see also ibid., p. 124). His attempts to show the influence of these clusters warrant further development and integration to mitigate these issues.

THE (MINOR) MINOR AGREEMENTS

443

Appendix: Methods of Counting (1) Subtraction If a word from the hypotext is absent in the hypertext (either exactly or in substituted form), then that word is counted as a subtraction. In the overwhelming majority of cases, this is straightforward. Matthew’s adaptation of a saying in Mark is a clear example subtracting six words: Mark 6,4: Οὐκ ἔστιν προφήτης ἄτιμος εἰ μὴ ἐν τῇ πατρίδι αὐτοῦ καὶ ἐν τοῖς συγγενεῦσιν αὐτοῦ καὶ ἐν τῇ οἰκίᾳ αὐτοῦ. Matt 13,57: Οὐκ ἔστιν προφήτης ἄτιμος εἰ μὴ ἐν τῇ πατρίδι καὶ ἐν τῇ οἰκίᾳ αὐτοῦ. Luke’s paraphrase (4,24) includes transposition and substitution, but as it ends with ἐν τῇ πατρίδι αὐτοῦ, the ten words that follow are easily counted as subtractions. Still, methodological choices include: • Articles are counted as individual words, so ὁ Ἰησοῦς ἐστιν → ἐστιν counts two words subtracted. Either ὁ ἐστιν or Ἰησοῦς ἐστιν counts just one69. • Regarding substitutions, εἶπεν τοῖς μαθηταῖς αὐτοῦ → εἶπεν τοῖς μαθηταῖς counts one subtraction, but εἶπεν αὐτοῖς is not counted as a subtraction but as a (shorter) substitution. • In doublets of Markan material, if the word or phrase appears in either passage it is not counted as a subtraction: the hypertext adapted the words, it just did not adapt them twice.

69. Seemingly against FRIEDRICHSEN, Minor Agreements (n. 23), p. 401, who counters Vinson by claiming, “When a noun is omitted, so too is its definite article”. This statement is broader than he probably intends since it is not always the case: nouns of course drop without subtracting the article, e.g., ὁ Ἰησοῦς in Mark 10,27.29, ὁ in Luke 18,27.29 (the article now substantive). Matthew 19,26.28 retain Ἰησοῦς, so there is no MA here, but the possibility stands. In my way of counting, moving for example from ὁ Ἰησοῦς ἐστιν to just ἐστιν in both texts counts as two words mutually subtracted, while alternatively, say, ὁ ἐστιν in Matthew and ἐστιν in Luke count as one word mutually subtracted (Ἰησοῦς) and one word subtracted by Luke but not by Matthew (ὁ). Both the agreement and the disagreement factor in the calculation of expected values, in this case actually lowering the impact of the single MA, not inflating MAs. When counting by choices, however, both situations count as one negative MA that happen to be qualitatively different (ὁ Ἰησοῦς in one case, Ἰησοῦς in the other). It is unclear to me how one would determine a baseline for the number of choices an adapter would make since the counts effectively become ordinal: first choice, second choice, etc. Furthermore, agreeing to delete an αὐτοῦ seems vastly different to me from agreeing to delete an entire 150-word passage; yet both would be counted as one editorial choice. Counting by words reflects the difference in the magnitude of these subtractions (one word versus 150). To be fair to Friedrichsen, however, he may object that counting by words cannot differentiate subtracting one 150-word passage from subtracting 150 individual words spread out over multiple chapters – each situation carries the same weight in the data despite their qualitative differences. Hence the need for these methodological discussions to continue.

444

J.M. TRIPP

(2) Transposition Once a mapping was done of Markan words onto exact matches or substituted words or phrases in the hypertext, transpositions measure deviations in sequencing of these elements using the following rules: • For AB → BA, the shorter of A or B was chosen as the transposed text. • If A and B have the same word-count, then if both hypertexts transpose (elements of) A and B, the transposition that maximizes agreement is chosen. A simple case happens with Mark 3,31, where both Matt 12,46 and Luke 8,19 invert Mark’s ἔξω στήκοντες. It does not matter which word counts as transposed, but it seems clear that the same word should be counted. • If the substituted word or its semantic role determines its position, then no transposition is counted. The most frequent example is the substitution of καί and δέ at the beginning of a sentence, with only the latter postpositive. • The transposition of whole sentences within a passage or of whole passages within the larger text were not counted. • In doublets of Markan material, if a word or phrase was transposed in either passage, it counts as a transposition. When transposed twice, if it is transposed in either case to the same place as in the other hypertext, it counts as an equivalent transposition70. (3) Substitution Substitution introduces the most difficulties in getting an exact word-count, necessitating the following methodological choices: • Words are counted based on Mark’s phrasing, not the substituted phrase (which may be longer or shorter), e.g., αὐτῷ → πρὸς αὐτόν counts as one. • Changes in grammatical form are counted, including articles. If for example a masculine singular noun with article is substituted with another masculine singular noun with article so that both carry ὁ, the article is still counted as a substitution despite no alteration in form since it results from the choice of noun. • Changes in the form of prepositions based on syntax (e.g., ἀπό/ἀπ’, ἐκ/ἐξ, οὐ/ οὐκ) are not counted. • Due to inconsistent spellings within texts and frequent textual variants, differences in the spelling of Ἰωάν(ν)ης were not counted, but other variations of proper nouns were (e.g., Ἱεροσόλυμα/Ἱερουσαλήμ, Μαρία/Μαριάμ [cf. Matt 27,61])71. • In doublets of Markan material, if a word or phrase is substituted in either passage, it is counted as a substitution. If it is substituted with the same phrasing as the other hypertext in either passage, it counts as an equivalent substitution72. 70. This method adds only six additional transpositions, with one additional equivalent transposition, substituted forms of ἀνεχώρησεν (Mark 3,7) in Luke 6,17//Matt 4,24 (cf. Matt 12,15). 71. “Jerusalem” adds two substitutions (Luke 6,17; 18,31), “Mary” adds three (Matt 13,55; 27,56; 28,1). On the former, see J.M. ROSS, The Spelling of Jerusalem in Acts, in NTS 38 (1992) 474-476. 72. This method adds nine additional substitutions, with one additional equivalent substitution, βασιλέων (Mark 13,9) as βασιλεῖς in Matt 10,18//Luke 21,12 (cf. Luke 12,11).

445

THE (MINOR) MINOR AGREEMENTS

Figure 1 Mappings of Mark 6,4 onto Synoptic Parallels

Rock Valley College United States of America

Jeffrey M. TRIPP [email protected]

LIST OF CONTRIBUTORS

Olegs Andrejevs (PhD 2013, Loyola University Chicago) is an Instructor with the Department of Theology at Loyola University Chicago. His recent publications include Reception of the Twelve in Matthew and Luke: Comparing the Current Synoptic Hypotheses (ExpT, 2022); Mark’s Framing of the Walking on the Sea Incident (Mark 6,45.53) (ETL, 2022); and ΠΡΩΙ ΕΝΥΧΑ ΛΙΑΝ ΑΝΑΣΤΑΣ ΕΞΗΛΘΕΝ (Mark 1,35; 16,2.9a) (Revue Biblique, 2023). Robert Derrenbacker (PhD 2001, University of St Michael’s College, University of Toronto) is Dean and Associate Professor of New Testament at Trinity College Theological School/University of Divinity (Melbourne). His recent publications include The Enduring Impact of the Gospel of John (co-edited volume, Wipf & Stock, 2022). Eric Eve (DPhil 2000, University of Oxford) is Emeritus Fellow of Harris Manchester College, Oxford. His recent publications include Writing the Gospels: Composition and Memory (SPCK, 2016); Solving the Synoptic Problem: Introducing the Case for the Farrer Hypothesis (Cascade, 2021); and Relating the Gospels: Memory, Imitation and the Farrer Hypothesis (LNTS 592, Bloomsbury T&T Clark, 2021). Paul Foster (DPhil 2003, University of Oxford) is Professor of New Testament and Early Christianity at the University of Edinburgh. His recent publications include The Gospel of Peter: Introduction, Critical Edition and Commentary (TENTS 4, Brill, 2010) and Colossians (Black’s New Testament Commentaries, Bloomsbury T&T Clark, 2016). Alan Garrow (DPhil 2000, University of Oxford) is vicar at St Peter’s Harrogate and member of the Sheffield Centre for Interdisciplinary Biblical Studies. His recent publications include Streeter’s ‘Other’ Synoptic Solution (NTS, 2016); An Extant Instance of ‘Q’ (NTS, 2016); and Gnats, Camels and Matthew’s use of Luke (JSNT, 2023). Ronald V. Huggins (ThD 1996, Wycliffe College, Toronto School of Theology, University of Toronto) is an independent scholar. His recent publications include On the Secondary Nature of Kaṁsa’s ‘Massacre of the Innocents’ in Bhāgavata Purāṇa 10 (Journal of Hindu-Christian Studies, 2020); The Generation of the Deportation to Babylon: Matthew 1,17 as the Key to Counting Matthew’s Genealogy (ETL, 2021); and Lighthouse: Jerald & Sandra Tanner: Despised and Beloved Critics of Mormonism (Signature Press, 2022).

448

LIST OF CONTRIBUTORS

Simon J. Joseph (PhD 2010, Claremont Graduate University) is Lecturer in Early Christianity at the University of California, Los Angeles. His recent publications include Jesus and the Temple (SNTS MS 165; Cambridge University Press, 2016); Jesus, the Essenes, and Christian Origins (Baylor University Press, 2018); and A Social History of Christian Origins (Routledge, 2023). Alan Kirk (PhD 1996, University of Toronto) is Professor of Religion at James Madison University, Virginia. His recent publications include Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS 564, Bloomsbury T&T Clark, 2016) and Jesus Tradition, Early Christian Memory, and Gospel Writing: The Long Search for the Authentic Source (Eerdmans, 2023). Matthias Klinghardt (PhD 1988, University of Heidelberg) is Professor of Biblical Studies at TU Dresden. His recent publications include Das Neue Testament und sein Text im 2. Jahrhundert (co-edited volume, Francke, 2018) and The Oldest Gospel and the Formation of the Canonical Gospels (BiTS 41, Peeters, 2021). Edmondo Lupieri (License at SNS and LLC 1973, University of Pisa) is John Card. Cody Endowed Chair in Theology and Professor of New Testament and Early Christianity at Loyola University Chicago. His recent publications include Mary Magdalene from the New Testament to the New Age and Beyond (co-edited volume, Brill, 2020); Cronache dal Trumpistan. Diario di un teologo italiano in America (Di Girolamo, 2020); and Non uno itinere. Ebraismi, cristianesimi, modernità. Studi in onore di Mauro Pesce in occasione del suo ottantesimo compleanno (co-edited volume, Morcelliana, 2021). Robert K. MacEwen (PhD 2010, Dallas Theological Seminary) is Associate Professor of New Testament language and literature and Academic Dean at Tyndale Theological Seminary (the Netherlands). His recent publications include Matthean Posteriority: An Exploration of Matthew’s Use of Mark and Luke as a Solution to the Synoptic Problem (LNTS 501, Bloomsbury T&T Clark, 2015). John C. Poirier (DHL 2005, Jewish Theological Seminary of America) is an independent scholar. His recent publications include Psalm 16:10 and the Resurrection of Jesus ‘on the Third Day’ (1 Cor 15:4) (Journal for the Study of Paul and His Letters, 2014); Delbert Burkett’s Defense of Q (essay, Bloomsbury T&T Clark, 2015); and The Invention of the Inspired Text: Philological Windows on the Theopneustia of Scripture (LNTS 640, Bloomsbury T&T Clark, 2021). Chakrita Saulina (PhD 2022, University of Cambridge) is Assistant Professor of Biblical studies (New Testament) at Seattle Pacific University. Her recent publications include Envisioning the Church through the ‘Eyes’ of the Woman Clothed with the Sun and the Bride in John’s Apocalypse (Veritas, 2022) and Salvation through Jesus’s Sahala: A Batak Reading of Jesus’s Victory over the Devil and Death in the Letter to the Hebrews (essay, Bloomsbury T&T Clark, 2023).

LIST OF CONTRIBUTORS

449

Hildegard Scherer (ThD 2009, University of Münster; Habilitation 2015, University of Bonn) is Professor of Biblical Theology and Didactics (New Testament) at the University Duisburg-Essen. Her recent publications include Königsvolk und Gotteskinder. Der Entwurf der sozialen Welt im Material der Traditio duplex (BBB 180, V&R Unipress, 2016); Coherence and Distinctness. Exploring the Social Matrix of the Double Tradition (essay, Bloomsbury T&T Clark, 2018); Gott und die Feinde. Traditionen und neutestamentliche Vernetzung von Q 6,28 (essay, Mohr Siebeck, 2019); and Learning Lessons on Q. The 2DH and Q in Academic Teaching (essay, Kohlhammer, 2020). David B. Sloan (PhD 2012, Trinity Evangelical Divinity School) is an adjunct professor at Ashland Theological Seminary, Ohio. His recent publications include The τίς ἐξ ὑμῶν Similitudes and the Extent of Q (JSNT, 2016) and Matthean Conflations and Luke’s Utilization Procedure under the Farrer Hypothesis: In Defense of F. Gerald Downing (co-authored with O. Andrejevs, in JBL 141 [2022]). Jeffrey M. Tripp (PhD 2016, Loyola University Chicago) is a developmental math instructor at Rock Valley College, Illinois. His recent publications include Direct Internal Quotation in the Gospel of John (WUNT II/493, Mohr Siebeck, 2019); Sacred Texts & Sacred Figures: The Reception and Use of Inherited Traditions in Early Christian Literature, A Festschrift in Honor of Edmondo F. Lupieri (co-edited volume, Brepols, 2022); and The Eyewitnesses in Their Own Words: Testing Richard Bauckham’s Model Using Verifiable Quotations (JSNT, 2022). Joseph Verheyden (STD 1987, KU Leuven) is professor of New Testament studies at the University of Leuven and director of the Leuven Centre for the Study of the Gospels. Recent publications include Jews and Christians – Parting Ways in the First Two Centuries CE? (BZNW 253, De Gruyter, 2021; co-edited with J. Schröter and B.A. Edsall); Polemics and Networking in Graeco-Roman Antiquity (Lectio 12, Brepols, 2021; co-edited with P. d’Hoine, G. Roskam, and S. Schorn); Early Christian Commentators of the New Testament: Essays on Their Aims, Methods and Strategies (BiTS 42, Peeters, 2021; co-edited with T. Nicklas); The Composition, Theology, and Early Reception of Matthew’s Gospel (WUNT 477, Mohr Siebeck, 2022; co-edited with J. Schröter and D.C. Sim); On Using Sources in Graeco-Roman, Jewish and Early Christian Literature (BETL 327, Peeters, 2022; co-edited with J.S. Kloppenborg, G. Roskam, S. Schorn).

INDEX OF AUTHORS

ABAKUKS, A. 418 423-425 430-431 433 441-442 ABOGUNRIN, S.O. 68 ABRAHAMS, P. 352 ADAMCZEWSKI, B. 132-133 144 162 240 AHEARNE-KROLL, S.P. 279 ALAND, K. 429 ALEXANDER, L. 189 234 337 ALKIER, S. 132 151 169 ALLEN, D. 208 ALLISON, D.C. 27 32 43 132 150-152 158 161-163 245 248-249 300 314 317 ANDERSON, G.P. 413 ANDERSON, M.C. 197 ANDREJEVS, O. 2 6-7 12 52 58 65 132-134 149 159 163-165 176 239 287 298 403 407 ANDRIANOU, D. 340-341 343 ARCHER, J. 132 264 ARGYLE, A.W. 78 ARNAL, W.E. 50 54 56-59 61 142 151 374 ASCOUGH, R.S. 54 374 ASGEIRSSON, J.M. 120 ATTRIDGE, H.W. 200 AUNE, D.E. 49 AUSLANDER, P. 19 BACON, B.W. 86 409-410 BADDELEY, A. 197 BADENAS, R. 49 BAGNALL, R.S. 335 347 BAHRDT, K.F. 67 BAILEY, K.E. 408 410 412 BARKER, J.W. 4 273 279-280 284 368 BARR, A. 328 BARRETT, C.K. 49 249 BAUER, B. 97 BAUER, W. 64 BAUM, A.D. 381 BAZZANA, G. 50 53 57-58

BEASLEY-MURRAY, G.R. 301 BECKER, E.-M. 121 200 204 382 BEDUHN, J.D. 360 364 372 BEGG, C. 283 BELL, R.H. 251 BELLINZONI, A.J. 239 BENNY LIEW, T. 337 BERGER, K. 379 BERGSMA, J. 67 BIRD, M.F. 180-181 BLACK, S.D. 151 157 420 BLAIR, G.A. 240 BOCCACCINI, G. 65 BOCK, D.L. 65 BOND, H.K. 38 180 194 200 BONNER, C. 244 BORG, M.J. 75 82 BORGEN, P. 207 209 BORING, M.E. 138 238 254-255 259 269 419 440 BORMANN, L. 55 BOVON, F. 162 189 243 245-246 249 251 253 260 405-406 408-409 411 BRANDEN, R.C. 257-258 BRANDENBURGER, S.H. 34 BRANDT, S. 379 BRAUNER, U. 34 BROADHEAD, E.K. 76 79 BRODIE, T.L. 287 405 BROOKE, G.J. 221-222 BROOKINS, T.A. 425 BROWN, J.K. 315 BROWN, R.E. 37 39 76 BRUNT, P.A. 345 BÜLOW-JACOBSEN, A. 335 BULTMANN, R. 74 90 245 287 402 BURKETT, D. 180-181 297 BURNET, J. 349 BURRIDGE, R.A. 194

452

INDEX OF AUTHORS

BUTH, R. 289 BUTLER, B.C. 78 249 256 BUTTS, J.R. 76 CADBURY, H.J. 82 87 94 402 CALLAWAY, P.R. 222 CAMERON, R. 50-51 CAMPBELL, W.S. 49 CAREY, G. 68-69 CARRIER, R. 52 CARROLL, J.T. 406-407 CARRUTHERS, M. 196-197 281 CARSON, D.A. 442 CARTER, J. 66 CHANG, K.-H. 4 68 268 273-274 CHARLES, R. 55 CHARLESWORTH, J.H. 65 CHILTON, B. 249 CHRISTIE, A. 130 CLIVAZ, C. 396 COHEN, S.J.D. 234 COLGRAVE, B. 350 COOGAN, J. 286 COPE, L. 417 CORSTJENS, R. 162 252 COTTER, W.J. 74 CRAWFORD, B.S. 54 CRAWFORD, M.R. 142 284-286 CREED, J.M. 287 CRIBIORE, R. 335 426 CROMHOUT, M. 59 CROSSAN, J.D. 75 245 CROSSLEY, J.G. 51 60 DAMM, A. 1 214 255-256 DAVIES, G.N. 49 DAVIES, M. 78 239 DAVIES, W.D. 27 32 43 150-152 158 161163 245 248-249 300 314 317 DEFOE, D. 384 DELGADO, R. 68 DENAUX, A. 162 252 263 265-266 DERICO, T.M. 289 DERRENBACKER, R.A., JR. 1 3-5 7-8 10 54 74 141 193-194 196 227 233 268 273 279-280 302 319 326 328 336-341 348349 353-354 374 401-402 412 414-415 DERSHOWIT, I. 355 DE SOLAGES, B. 150 152 DE TROYER, K. 121 DETTWILER, A. 396

DEVILLERS, L. 396 DEWEY, J. 38 376 DEWITT BURTON, E. 152 DIANGELO, R. 60 DI VITO, R.A. 11 DONALDSON, T.L. 60 DOOLE, J.A. 8 182-183 190 192-194 DORMEYER, D. 414 DOWNING, F.G. 3-4 10 131 233-234 238240 282 287 292 336-339 341-351 372 414 425 DRURY, J. 78 DUNGAN, D.L. 68 240 268 DUNN, J.D.G. 17-18 143 DYER, B.R. 1 131-132 297 311 DYER, R. 68 EDWARDS, J.R. 249 405-410 EDWARDS, R.A. 90 ELDER, N.A. 103 ENNULAT, A. 136 157 180 421 423 EPELBOIM, J. 343 EVANS, C.A. 127 131 163 248-249 297 EVE, E. 3-5 7-8 51 53 132-133 138-142 149 151-153 156-160 163 169 195-197 205-207 211-212 215 217 225 227-229 239-240 243 245 247-249 251-256 265 277-278 287 290 292 299 336 352 381 426 EYSENCK, M.W. 197 FARMER, W.R. 78 366 372 FARNES, A.T. 133 162 FARRER, A.M. 51 73-74 78 80-81 83 85 91 116 142 160 219-232 287 346 417 371 FEIERSTEIN, M.M. 110 FERDA, T.S. 62 FISHBANE, M. 234 FITZMYER, J.A. 87 91 158 162-163 243 245 247 249 265 300 316 325 328 406 408-409 411 FIX, U. 236 FLEDDERMANN, H.T. 29 32 34 36 39 42 44 59 95 133 160-162 246-248 251-252 263 414 420 FOCANT, C. 241-243 FOSTER, P. 1 3 6 42 45 131-133 138 142 151 161 163 166 168 196-197 229 233234 241 268 279 302 343 374 401 419 FRANCE, R.T. 241-243

INDEX OF AUTHORS

FRANKENBERG, R. 68 FRANKLIN, E. 345 FRAYSSE, L. 66 FREIBERGER, O. 66 FRENSCHKOWSKI, M. 55 121 123-126 FREVEL, C. 118 FRIEDRICHSEN, T.A. 136 150-152 156-157 421-424 440 443 FUCHS, A. 180 FUNK, R.W. 28 31 75-76 82 GABOURY, A. 169 GÄCHTER, M.M. 118 GAGNON, R.A.J. 298 300 315 GAMBLE, H.Y. 205 GARCÍA, J.P. 127 GARLAND, D.E. 405 407 GARRETT, S.R. 244 GARROW, A. 3-5 7 9 53 132-133 138-139 146-147 151 160-161 240 246 253 263 272 278 294-295 307 328 336 346 415 GARSKÝ, Z. 423 441 GATHERCOLE, S. 199-203 205 GAVRILOV, A.K. 22 GERHARDSSON, B. 15-16 GILLIARD, F.D. 22 GNILKA, J. 162 GOLDMAN, A.L. 82 GOMÁ CIVIT, I. 158 GOODACRE, M.S. 1-2 4 51 53 55 74 78 132-133 135 137-139 141-143 146 149151 156 158 159-161 163 169 199-200 202-205 212 214 219-220 237 239 253 259 264 266 268 277 288-290 293 299 307 310-311 313-314 317 325 327 342 414 419-420 GOULDER, M.D. 10 31 74 78 80 83 85 91 141-142 150-151 156 159-160 163 167 193 215 220 224-229 231 246-247 252 257 265-266 268 282 289 310-311 318-319 324 335-340 342 346-347 417 420-422 430 GOUSOPOULOS, C. 58 GRAMAGLIA, P.A. 364 GRANT, J.N. 281 GREEN, H.B. 253 257 GREEN, J.B. 251 315 404 408 410 GREGORY, A. 1 42 131 151 196 202 229 233-234 279 302 343 374 401 419 GRIESBACH, J. 103 390

453

GRÜNSTÄUDL, W. 133 165-166 168 GRUNDMANN, W. 64 GUENTHER, H.O. 6 74 76-77 79 82 90 99 GUIJARRO, S. 112 GUNDRY, R.H. 78 107 158 162 249 317 321 410 421-422 GUTHRIE, D. 301 GUTTENBERGER, G. 241-243 HÄFNER, G. 131 136 141 156-158 163 HÄGERLAND, T. 139 259 310 313-314 HAGNER, D.A. 317 HAMILTON, C.S. 56 HARB, G. 4 54 131 372 414 HARLAND, P.A. 54 374 HARNACK, A. VON 64 90-91 360-361 399 HARRIS, S. 49 HARRIS, W.V. 205 HARVEY, A.E. 337 HAWKINS, J.C. 86-87 91 134 150 152-153 252 256 265 414 HAYES, C. 234 HAYS, C.M. 373 HEAD, P. 64 HEDRICK, C.W. 131 HEIL, C. 4 34 36 42 54 65 76 90 120 131132 372 414 HEIL, J.P. 49 HEILIG, C. 433 HEILMANN, J. 343 347 364 HELBIG, J. 117-120 HENGEL, M. 63 237 240 256 413 415 HESCHEL, S. 64 HIEKE, T. 34 120 HILL, C.E. 284 HOCKEY, K.M. 68 HOFFMANN, P. 34 36 59 76 120 137 249 298 403 HOLTZMANN, H.J. 134-135 161-162 165 241-242 253 HONORÉ, A.M. 423 425 431 441 HOOKER, M.D. 49 179 HOOVER, R.W. 28 31 82 HORNBLOWER, S. 185 HORRELL, D.G. 60 68 HORSLEY, R.A. 61-62 337 HORT, F.J.A. 391 HOWARD, G.E. 49 HOWES, L. 58

454

INDEX OF AUTHORS

HUGGINS, R.V. 3 5-6 10 73-74 96 98-99 132 240 306 396-400 414-415 HULTGREN, S. 407 414 HUMPHRIES, M.L. 242-243 245 HUNKIN, J.W. 287 HURTADO, L.W. 21-22 45 205 338 IVERSON, K.R. 45 JACKSON, J. 345 JACKSON-MCCABE, M. 59 JACOBSON, A.D. 115 JAMESON, H.G. 78 JEREMIAS, J. 87 91 94 162 249 402 JEWETT, R. 49 JOHNSON, L.T. 75 82 JOHNSON, S.R. 299 311 317 JOHNSON, W.A. 21 347 JOHNSTON, J.J. 127 JOSEPH, S.J. 2 6 12 52 58 60 62 65 67 71 134 148 JOSHI, K.Y. 68 JUNGBLUT, F.S. 343 KAHL, W. 70 132-136 138-140 143-144 149-156 158-159 164 168 237 239 KASTER, R.A. 426 KEENER, C.S. 194 KEITH, C. 8-9 70 182 188-194 202 204 235 KELBER, W. 16 18 KENNEDY, G. 190 KIMBER BUELL, D. 68 KINGSBURY, J.D. 86 KIRK, A. 2-5 7-8 10 52 74 131-132 134 138 140-143 148 161 169 190 196-197 219 230 232 234-235 262 268 272-274 280 327 336 349 351-354 401-402 KLEIN, H. 136 249 368 KLEIN, J. 236 KLINGHARDT, M. 1-3 10 132 138 142 151-152 154 156-161 169 249 253 359 364 366 371 378 380 383 385 387-398 400 KLOPPENBORG, J.S. 1 4 30 32 36 42 44 50 53-57 59-61 64 74 76-77 79 98 131-132 135 137 139 143 150 152 160-161 163 196 199-200 205 229 233 237 249 256 263 265 273 279 298 302 306 327 340 343 346 360 371 374 401 403-404 409 419 427 442 KLOSTERMANN, E. 287

KNIBB, M.A. 65 KNOX, B.M.W. 22 KNOX, J. 395 399 KOESTER, H. 59 199 203-204 KONRADT, M. 122 249 KRAUSS, A. 343 KRENTZ, E. 86 KÜMMEL, W.G. 90 LABAHN, M. 121 192 381 414 LACHMANN, C. 103-104 106 LACHMANN, R. 236 LAGRANGE, M.-J. 152 LALOU, É. 344 349-352 LANDRY, D.T. 337 LANGE, A. 117-118 120 LANGE, J. 86 LANGLOIS, M. 66 LARSEN, M.D.C. 8 181-187 190-191 193194 LAUBMANN, G. 379 LAUFEN, R. 160 249 LAYTON, B. 198 LE DONNE, A. 70 LEIPZIGER, J. 343 LEVISON, W. 350 LEWIS, N.D. 199 LICONA, M.R. 315 LIENHARD, J.T. 396 LINDEMANN, A. 50 76 79 137 LINNEMANN, E. 37 LIVINGSTONE, E.A. 49 LOADER, W.R.G. 131 LÖFSTEDT, T. 259 LOFTUS, J.W. 49 LOHR, C. 16 LOISY, A. 287 LONGENECKER, R.N. 74-75 LONGSTAFF, T.R.W. 431 441 LORD, A. 20 LÜHRMANN, D. 59 61 116 LULL, D.J. 61 LUMMIS, E.W. 78 LUPIERI, E. 2 5-7 12 101 111-112 114 132 LUZ, U. 118 128-129 162 192 248-249 258-259 266 MACEWEN, R.K. 3 5 7 9-10 132-133 138 145-148 151 168 240 253 257 267-268 307 312 314 318-319 327-330 409-410 415

INDEX OF AUTHORS

MACK, B.L. 51 61 MACKIE, S.D. 133 MAKARYK, I. 19 MANICARDI, E. 111 MANSFELD, J. 347 MANSON, T.W. 94 MARCHADOUR, A. 116 MARCUS, J. 26 40 67 241-243 249 264 396 MARDAGA, H. 162 252 MARJANEN, A. 204 MARSHALL, I.H. 249 315 406 408-409 411-412 420 MARTIN, D.B. 66 MARTIN, J.N. 68 MARTIN, L.H. 51 MASON, E.F. 222 MATSON, M.A. 214 266 368 MATTHEWS, S. 372 MATTILA, S.L. 233-234 280 284 MAZZINGHI, L. 111 MCARTHUR, H.K. 379 MCGEE, C.E. 19 MCLOUGHLIN, M. 136 149-153 158-159 420 428 430 440 MCNICHOL, A.J. 240 268 417 MEALAND, D.L. 418 425 MEEK, R.L. 117 MEIER, J.P. 75-76 249 METZGER, B.M. 43 MEYER, M.W. 121 MIGNE, J.-P. 379 MILLER, M.P. 50-51 54 MILLER, R.J. 82 439 MILLER, S. 220 MILLS, I. 287 MONAGHAN, C.J. 4 132 151 MONTEFIORE, C. 287 MORGAN, T. 426 MORGENTHALER, R. 23-24 26 423 MOSBØ, T.J. 345 MOSS, C. 180 MOULTON, J.H. 87 MOURNET, T.C. 18 143 MOYISE, S. 117 MÜLLER, M. 1 4 53-54 66 70 116 131135 137-142 237 239 264 277 372 409 MUIRHEAD, L.A. 354-355 NAKAYAMA, T.K. 68 NEILL, S. 75

455

NEIRYNCK, F. 5 76 86 133-134 136-138 145-146 149-160 168-169 246 251 258259 419 421 NEWPORT, K.G.C. 44 NEWTON, R. 60 NEYREY, J. 337 NICKLAS, T. 364 NIELSEN, J.T. 70 132-133 135 138 141142 169 239 264 277 409 NINEHAM, D.E. 51 74 116 371 NOLLAND, J. 162 256 258 298 406 408409 NORDENFALK, C. 144 NORELLI, E. 396 NOTLEY, R.S. 127 289 OLSON, K. 239 253 255 273 287-288 OMERZU, H. 1 4 53-54 66 70 116 131134 137-139 237 239 372 ONG, H.T. 381 ONG, W.J. 18 OROPEZA, B.J. 117 OTTEN, J.D. 404 PAINCHAUD, L. 114 PARDEE, C.G. 102 PARKER, H.N. 21 PARKER, P. 7 101 PARKS, S. 55 58 66 69 405 413 PARRY, A. 20 PARRY, M. 20 PARSONS, M.C. 410 PASTORELLI, D. 396 PATILLON, M. 426 435 PATTERSON, S.J. 50 198-204 210 PAULSON, G.S. 154 PEABODY, D.B. 240 268 417 PEDLEY, K.G. 340 PELLING, C.B.R. 205 282 290 PERRIN, N. 45 142 151 214 239 266 284285 288 299 315 PERTZ, G.H. 349 PESCH, R. 25 PETER, H. 109 PETERSEN, W.L. 280 283-285 PETERSON, J. 131-132 146 240 253 307 338 346 351 424 442 PFISTER, M. 129 PIERCE, M.N. 132 PIETERSMA, A. 316 PIPER, R.A. 51

456

INDEX OF AUTHORS

PITTS, A.W. 83 PLEVNIK, J. 77 PLONER, M.T. 118 123 POIRIER, J.C. 2-3 5 7 10 131-133 146 164 240 253 307 335 338 340-342 344 346 351 355 423-424 442 PORTER, S.E. 1 131-132 297 311 381 POWELL, E. 307 415 POYNTON, J.B. 352 PRETTY, R.A. 394 PUECH, É. 67 RÄISÄNEN, H. 245-246 250 REBER, K. 341 REED, J.L. 50 REID, D.G. 1 RENAN, E. 64 RHOADS, D.M. 20 238 RICHTER, G.M.A.v341 RIDER, J. 349-350 ROBBINS, V.K. 337 ROBINSON, J.A.T. 109 ROBINSON, J.M. 36 50 59 75-76 90 95 98 137 243 249 298 403 ROEDIGER, D.R. 68 ROLLENS, S.E. 55 58 62 69 ROLLSTON, C.A. 227 335 ROPES, J.H. 78 ROSS, J.M. 444 ROTH, D.T. 121 192 236-237 391 ROTHSCHILD, C.K. 66 ROUSE, M.A. 335 348-349 352 ROUSE, R.H. 335 348-349 352 RUBIN, D.C. 196 RUNIA, D.T. 133 207-210 214 347 RUSSELL, D.A. 233 RUSSELL, D.S. 282 SACKUR, E. 379 SAND, A. 162 SANDAY, W. 86 93 136 414 SANDERS, E.P. 49 78 239 SANDERS, J. 249 SAULINA, C.M. 3 5 7 9 68 132 275 SAVRAN, G. 439 SCHENCK, K. 208 SCHENK, W. 136 158 SCHERER, H. 3 6 7 115-117 130-131 137 161 373 SCHIRONI, F. 338 347

SCHLEIERMACHER, F. 86 SCHLERITT, F. 368 SCHLOSSER, J. 116 SCHMIDT, K.L. 265 287 SCHMITHALS, W. 37 SCHNABEL, E.J. 315 SCHREINER, T.R. 49 SCHRÖTER, J. 53 55 364 SCHÜRMANN, H. 61 406 408-411 SCHULZ, S. 59 61 162 249 SCHWEITZER, A. 97 SCORNAIENCHI, L. 121 SCOTT, B.B. 76 SELAND, T. 49 SELKIRK, A. 384 SELLEW, P.H. 38 SEMLER, J.S. 363 SENIOR, D. 131 SHARPE III, J.L. 349 SHELLARD, B. 233 257 266 345 368 SHELTON, J. 404 SHINER, W. 21 SIEGERT, F. 132 138 151 SIM, D.C. 241 243 SIM, M.G. 431 SIMPSON, R.T. 239 SKINNER, S. 244 SLATER, T.B. 86 SLOAN, D.B. 2 10-11 165 403 407 415 430 SMALL, J.P. 340 354 401 SMITH, D.A. 4 54 65 74 131 134 364-365 372 414 SMITH, D.E. 50 SMITH, J.Z. 66 SMITH, S. 208 SMITH, W.C. 49 SOARDS, M.L. 37-38 SORENSEN, E. 251 SPAWFORTH, A. 185 STANDHARTINGER, A. 375 STANLEY, C.D. 117 119 STANTON, G.N. 376 STECK, O.H. 59 61 STEFANCIC, J. 68 STEGEMANN, H. 221 STEIN, R.H. 83 252 299 STEINHAUSER, M.G. 76 STEMBERGER, G. 211

INDEX OF AUTHORS

STOLDT, H.H. 103 STRECKER, G. 64 131 136 420-421 STREETER, B.H. 42 80 93-96 137 150 160-161 165 263 287 367 402 428 SWANSON, D.D. 220-224 230 SYREENI, K. 238 266-268 TAUSSIG, H. 82 TAYLOR, V. 39 92 94-95 179 287 TAYLOR, W. 384 THEISSEN, G. 55 TILLIETTE, J.-Y. 352 TIWALD, M. 54-55 59-60 62 67 69 134 TÖDT, H.E. 59 115 TOV, E. 118 TRIPP, J.M. 2 5 10-11 132 423 426 428 435 437 441 445 TROBISCH, D. 370 385 TUCKETT, C.M. 3-4 36 55 59 61 73 117 124-125 131-132 134-135 141 160-161 163 199-200 226 238-240 249 253 263 268 335 421 TUPAMAHU, E. 60 67 69 TURNER, N. 87 TWELFTREE, G. 251 TYSON, J.B. 239 431 441 ULRICHSEN, J.H. 49 URO, R. 203-205 245 VAAGE, L.E. 51 56-58 76 VAN BELLE, G. 59 131 238 421 VAN DEN HOEK, J.L. 335 338 VAN DER GEEST, J.E.L. 338 VAN OYEN, G. 133 164-165 167 VAN SEGBROECK, F. 59 86 136-137 238 258 421 VASSILIADIS, P. 42 VATRI, A. 343 VENTURINI, K.H. 67 VERHEYDEN, J. 1 11-12 36 42 55 59 76 90 131-132 135 196 229 233 238 265 279 302 343 364 374 401 404 419 421 VINSON, R.B. 423-424 432 443 VINZENT, M. 364 395 VORLÄNDER, H. 383 VROOM, J. 221 223 224 WAGNER, S. 67 WAHLEN, C.L. 257 259 269 WALKER, W.O., Jr. 190 239 WALLACE, D.B. 25

457

WALSH, R.F. 54 WALTERS, P. 102 WASSÉN, C. 67 WATSON, F. 4 19 51 53 56-57 81 132-133 135 138-139 142 146 162 166 169 285286 335 WEIGOLD, M. 117 120 WEIMA, J. 77 WEISE, G. 236 WEISS, B. 134-135 148 161 163 241-242 265 WEISS, J. 154 287 WEISSE, H. 134 165 WELLHAUSEN, J. 287 WENDT, H.H. 404 409 WEREN, W.J.C. 117 120 WERNLE, P. 165 354-355 WEST, D. 233 WEST, H.P. 97 405 413 WESTCOTT, B.F. 391 WHITTAKER, J. 281 WIEBE, D. 56 WILKE, C.G. 97 WILLIAMS, C.H. 200 WILLIAMS, J.G. 414 WILLITTS, J. 56 WINK, W. 110 WIRE, A.C. 20 WISCHMEYER, O. 121 WITHERINGTON III, B. 180 WITTKOWSKY, V. 70 132 135 138 151 WOJTKOWIAK, H. 132 148 157-158 164 WOLTER, M. 136 189 243 245 263 265 368 417 WOODE, E. 384 WOODMAN, T. 233 WRIGHT, B.G. 316 WRIGHT, B.J. 21 41 76-77 WRIGHT, N.T. 75 WRIGHT, T. 75 YANCY, G. 68 YARBRO COLLINS, A. 40 241-243 ZAHN, M.M. 220-224 231 ZELLER, E. 138 ZENGER, E. 118 ZIESLER, J.A. 257 ZIMMERMANN, R. 121 192 ZOLA, N.J. 142 284-285

INDEX OF REFERENCES

HEBREW BIBLE Genesis 1,2 1,31 2,9 3,19 6,1 6,3 6,4 6,5 6,8 6,11-12 12,1-3 12,1 13,9 15,6 24 26,2 28,17 31,3 32,4-6 39,7 39,8-9 39,12 39,21 40,8 40,15 41,41 42,18 45,7-8 45,28 50,19 50,24-25

140 207 207 207 207 209 207 207 207 126 209 126 207 209 207 207 207 222 207 207 209 207 316 207 207 207 209 207 207 207 207 207 207 207 207

Exodus 1,9 2,23 3,6 8,14-15

207 207 223 122 250

8,15-19 8,19 12 12,11 15,25 18,14 20,18 20,21 20,22 23,20 24,12 29,40 31,2-3 31,18 33,7 33,17 34 34,28

247 250 226 207 225 207 207 207 207 207 317 207 210 224 207 250 207 209 221 108

Leviticus 9,3 12,14 14,21 16 16,2 16,33 16,34 17 17,7 17,8 18,6 19,18 23 23,10-15 23,12 23,13 23,14 26,39

224 224 224 224 207 224 207 224 224 224 224 207 122 224 230 224 224 224 224 223

460

INDEX OF REFERENCES

Numbers 6,2-12 11,7 14,44 15 15,4 15,5 15,11 15,15 15,24 15,25 20,14-20 28–29 28 28,26-30 28,27 28,30 28,31

209 207 207 224 224 224 224 224 224 224 316 221 224 230 224 224 224 224

Deuteronomy 4,12 5,9 5,31 6,4 6,13 6,16 8,3 9–10 9,9 9,10 10,20 17,14-20 17,17 23,1 24,1 24,3 25,5-6 34,4 34,7

207 121 207 122 121 121 121 250 108 250 121 222 222 221 122 122 122 207 207

Joshua 7,22

16,19 16,23 19,11-21

316 245 316

2 Samuel 2,5-7 3,12-14 3,12-13 3,26 11,4 12,27-28

316 316 316 316 316 316

1 Kings 10,1-13 10,1 17,4 17,17-24 20,2-6

125 125 316 404 316

2 Kings 1,10 1,12 6,32 14,8-10 16,7 18,14 19,9-13

109 109 317 316 316 316 316

1 Chronicles 19,2 19,16

316 316

2 Chronicles 9,1-12 24,21 35,21

125 125 316

Ezra 8,35

224

316

Nehemiah 6,3

316

Judges 11,12-27

316

Tobit 8,1-3

140 245

1 Samuel 9,9 16,14

207 245

Judith 3,1-4

316

461

INDEX OF REFERENCES

1 Maccabees 7,10

316

Psalms 91,11-12 110,1 117,22-23 LXX Proverbs 31 Ben Sirach 48,10

Daniel 7,25 12,7

379 379

121 122 123

Hosea 6,1 6,6 6,9

215 215 215

222

Zechariah 7,13 13,7

223 123

Malachi 3,1 3,23-24 4,5

317 407 127 407

127

Isaiah 7,12 29,13 37,9-13 56,7

121 123 323 316 122

Jeremiah 11,11 21,12//26,3

223 223

NEW TESTAMENT Q 3–11 3,2-3 3,7-8 3,7 3,8 3,16-17 3,21-22 4 4,1-13 4,3-8 4,4 4,8 4,10 4,12 6,9 6,39 7,1-10 7,2 7,6-9 7,9

414 137 126 231 60 125 61 137 298 121 123 120 298 414 120 121 121 121 121 95 231 61 302-304 414 405 305 299

7,10 7,12 7,18-35 7,18-23 7,18-19 7,18 7,22 7,26-28 7,27 7,28 7,29-30 7,29 7,31-35 7,31 7,32 7,44-47 9,57-62 9,57-60 10 10,3 10,10-12

299 302 405 406-409 414 298 407 407 298 406-407 407 120 124 128-129 407-408 137 405 408 410 409 61 61 405 408 407 298 406 414 126 61 60-61

462 10,12-15 10,12 10,13-15 11,2 11,4 11,14-26 11,14-20 11,14-15 11,16 11,17-23 11,17-20 11,17 11,21-22 11,29-30 11,30-32 11,30 11,31-32 11,31 11,39-47 11,39 11,41 11,42 11,43-44 11,46 11,47-48 11,49-51 11,51 11,52 12–22 12–19 12,27 12,28 12,39-40 12,42-46 12,53 13,18-21 13,28-29 13,28 13,29 13,30 13,34-35 14,5 14,11 14,16-23 14,26 15,4-10 15,4-9

INDEX OF REFERENCES

125 126 60-61 95 410 414 298 298 61 122 298 137 137 61 125 125 126 405 125 122 60 87 87 60 87 60 87 60 60 61 117 125 60 414 414 125 231 61 61 405 405 62 126 60-61 125-126 231 302305 60-61 302-305 60 61-62 137 60 60 405 410 405

15,8-10 16,16 17 17,2 17,20-21 17,26-30 17,28-29 17,33 17,34-35 22,28-30

137 122 126 137 137 125-126 137 137 405 62

Matthew 1–4 1–2 1 1,1 1,6 1,16 1,20 1,22-23 2,1 2,2 2,3 2,4 2,5-6 2,15 2,16-18 2,17 2,23 3–11 3–4 3,2 3,3 3,5 3,7-12 3,7-10 3,7 3,9-10 3,11-12 3,11 3,13-17 3,17 4,1-11 4,1 4,2 4,3 4,5 4,7

248 123 248 367 73 129 188 235 73 73 129 123 158 112 158 112 123 123 107 128 110 123 182 248 84 171 312 171 407 313 29-30 106 231 310 314 308 314 145 174-175 247 436 321 171 106 107 108 156 107-108 154 170 107 109 155 173

INDEX OF REFERENCES

4,8 4,9 4,11 4,12-17 4,12 4,13 4,14-16 4,15 4,17 4,23-25 4,23 4,24 5–7 5 5,1–7,28 5,1 5,2 5,3 5,10 5,11 5,13-16 5,13-15 5,13 5,14-15 5,14 5,15 5,16 5,17 5,19 5,20 5,21-48 5,25 5,32 5,38 5,39-40 5,39 5,40 5,41 5,42 5,45 6 6,1-18 6,1 6,2-16 6,2 6,5 6,9-13 6,9

107-108 174 107 106 109-110 110 123 128 108 84-86 312 259 321 113 444 196 267 320 94 226 108 155 171 84 203 388 84 89 203 91-93 93 145 214 92 203 167 84 394 84 84 42 122 174 167 388 42 174 174 42 94 42 84 93 225 201 203 84 201 321 321 213 262 266 82-85 95

6,14-15 6,19-21 6,22-23 6,22 6,24 6,25-34 6,26 6,30 6,31-33 6,31 7–26 7,1-5 7,2 7,3-5 7,7-11 7,7-8 7,8-29 7,10 7,11 7,13 7,16 7,18 7,21 7,22-23 7,23 7,28-29 7,28 8–13 8–12 8–9 8,1-5 8,1-4 8,2 8,3 8,5-13 8,6 8,8-10 8,8 8,9-10 8,9 8,10 8,11-12 8,11 8,12 8,13 8,14-17

463 318 225 93 262 214 225 413 215 84 162 231 310 162 155 171 319 228 145 228 213 262 266 74 167 259 94 174 84 248 318 225 174 84 155 171 318 155 319 259 9 86 267 297 299 303304 309 330 319 230 267 330 259 154 170-171 171 9 226 297 299 303-304 155 171 300 310 305 155 173 300 310 312 308 313 305 323 215 297 305 311 317318 323 325 84 231 309 324 299-300 310 312 260

464 8,14-15 8,14 8,17 8,19-22 8,19 8,25 8,26 8,27 8,28–9,1 8,28-34 8,29 8,31 9 9,1-17 9,1-8 9,11 9,13 9,14 9,18-26 9,18 9,20 9,22 9,26 9,27-34 9,27-32 9,27-31 9,27 9,29 9,32-34 9,32-33 9,33 9,34 9,35–11,30 9,35 9,36 10 10,5 10,7 10,8 10,9-14 10,12-13 10,17 10,18 10,19-20 10,20 10,21 10,24-25 10,25

INDEX OF REFERENCES

260 427 123 128 322 154 170 154 170-171 231 310 156 171 174-175 256 26 26 84 171 171 256 254 156 320 313 122 172 299 405 172 321 417 154 170 427 407 256 231 256 129 172 172 166 255-256 269-272 314 255-256 155 172 257 166 255 312-314 256 318 321 318 196 228 267 291 320 172 198 84 155 172 174 259 167 167 321 444 167 248 166 228 314

10,26 10,27-32 10,27 10,29 10,32-33 10,32 10,33 10,34 10,35-36 10,38 10,39 10,40 11–14 11 11,1-7 11,1 11,2-13 11,2-6 11,3-4 11,3 11,10 11,11 11,12 11,13-14 11,14 11,18 11,23-24 11,25-27 11,25 11,26 11,27-28 11,29 11,30 11,31 11,32 12 12,1-7 12,1 12,2 12,5 12,6 12,7 12,9-14 12,9 12,10 12,11 12,12-21 12,14

167-168 201 168 203 174 176 84 84 89 84 166 201 166 166 166 166 93 291 215 86 267 309 226 109 260 155 145 84 84 128 128 226 127 215 248 258 275 275 275 275 275 275 214 254 256 262 274 198 254 313 122 215 261 122 321-322 321 172 32 254 313

465

INDEX OF REFERENCES

12,15-45 12,15 12,17-21 12,18 12,22-45 12,22-37 12,22-32 12,22-30 12,22-24 12,22-23 12,22 12,23 12,24-26 12,24 12,25-26 12,25 12,26-27 12,27-28 12,27 12,28 12,30 12,31-32 12,31 12,32 12,33-42 12,33-37 12,33-35 12,33 12,34 12,38-42 12,38-39 12,38 12,39 12,40 12,41 12,42 12,43-45 12,43 12,44 12,45 12,46-50 12,46-49 12,46 12,47-48 12,47 12,50 13 13,1-50

261 156 444 123 128 248 214-215 238 262 238 254 269-272 275 314 166 255-256 256 166 246 256 129 257 166 145 154 214 312-314 145 239 246 258-259 244 246 259-260 258 84 246-249 261 167 248-249 261 167 167 261 214 214 225 231 310 90-91 255 262 166 313 155 172 214 32 90 261 261 31-32 255 35 35 261 255 262 254 154 444 154 154 84 202-203 214 267 291 320 214

13,2-8 13,3 13,7-8 13,10 13,11 13,12 13,14-15 13,14 13,23 13,24-30 13,24 13,28-29 13,28 13,29 13,30 13,31-33 13,31-32 13,31 13,33 13,36-42 13,38 13,42 13,44 13,45-46 13,45 13,47-48 13,47 13,50 13,52 13,53-58 13,53 13,54-58 13,54 13,55 13,57 14,1-14 14,5 14,8 14,9 14,12 14,13 14,14 14,15 14,22-33 14,31 15,1-20 15,1-14 15,1-6

201 172 201 154 170 84-85 166 123 128 427 201 215 84 330 155 155 215 214 201 84-85 172 84 201 215 324 231 309 324 84 201 201 84 201 84 215 231 309 324 84 321 86 267 309 110 156 259 321 444 443 445 318 427 155 158 154 156 85 172 321 231 310 228 322 313 320

466 15,3-9 15,4 15,11 15,12 15,14 15,22 15,27 15,33 16–25 16 16,1-12 16,1 16,2 16,4 16,6 16,8 16,11 16,12 16,13-22 16,13-20 16,13 16,17 16,19 16,21 16,24 16,25 16,26 16,27 17,5 17,9 17,15 17,19 17,20 17,23 17,26 18 18,1 18,3 18,4 18,5 18,9 18,10 18,12-14 18,14 18,18 18,21-22 18,23-35 18,23-27

INDEX OF REFERENCES

323 128 201 203 154 228 231 322 105 129 105 427 339 93 225 313 106 154 166 320 166 174 176 105-106 231 310 105-106 427 106 201 321 90 172 84 84 88 86 89 156 378 421 166 166 225 175-176 166 225-226 172 377 91 172 154 167 231 378 155 196 267 291 320 84 172 84 318 84 166 85 84 42 84 322 155 410 322

18,23 19,1 19,3 19,4-8 19,9 19,12 19,14 19,16 19,18 19,21 19,23 19,24 19,25 19,26 19,27-30 19,28 19,29 19,30 20,1-15 20,1 20,16 20,19 20,28-32 20,29-34 20,30 21,2 21,4-5 21,9 21,13 21,16 21,21 21,22 21,23 21,25 21,27 21,31 21,33 21,37 21,43 21,45-46 22,1-14 22,1 22,2 22,13 22,15 22,16 22,23-35 22,24

84 86 267 309 313 323 167 84 84-85 154 171 155 84 155 173 84-85 84 172 443 33 34 443 85 167 201 28 84 167 318 378 413 379 172 172 123 129 172 175-176 123 167 167 172 320 172 155 173 85 226 172 84 313 33 155 172 84 231 309 324 313 172 313 172

467

INDEX OF REFERENCES

22,27 22,31 22,33 22,34-35 22,35 22,37 22,42 22,43 23–25 23 23,1-36 23,1-12 23,2-3 23,2 23,4 23,5 23,6-7 23,6 23,7-10 23,11 23,13-33 23,13 23,14 23,15 23,16 23,23 23,24-26 23,24 23,25 23,27 23,29 23,33 23,34-39 23,34-36 23,34-35 23,34 23,37-39 23,39 24–25 24 24,3 24,5 24,17-18 24,23 24,26 24,33 24,34-36 24,35

156 419 128 172 259 321 88 155 173 172 247-248 196 43 86-89 92 113 313 43 43 172 43 43 43 167 388 87 321 43 43 43 84 86 87-88 43 113 231 87 214 231 87 87 87 231 310 44 44 318 321 44 318 175 267 291 320 225 172 173 167 167 167 85 24 25 201

24,42 24,44 24,46 24,51 25 25,1-13 25,1 25,10-12 25,10 25,11 25,14-30 25,20 25,21 25,22 25,23 25,24 25,29 25,30 26,1-2 26,1 26,3-5 26,8 26,17 26,27 26,29 26,34 26,35 26,36 26,39 26,47 26,49 26,50 26,57 26,60 26,61 26,64 26,67-68 26,68 26,69 26,70 26,73 26,75 27,12 27,3-4 27,9 27,11 27,12 27,17

166 166 225 231 309 324 225-226 225 84 318 318 155-156 173 33 104 154-155 171 173 155 173 154-155 171 155 173 154-155 171 166 231 309 324 318 86 267-268 309 320 173 173 173 85 155 173 25 427 173 156 320 171 90 154 320 154 155 155 157 157 173 173 173 154 171 427 104 157 320 320 128 155 173 320 155

468

INDEX OF REFERENCES

27,20 27,21-23 27,21 27,22 27,23 27,26 27,41-43 27,45-54 27,46-47 27,46 27,47 27,49 27,54 27,56 27,57 27,58 27,61 27,62–28,15 27,62-64 27,65 28,1 28,11-15 28,16 28,17 28,19-20 28,19 28,20

320 155 155 155 155 173 175-176 175-176 320 322 109 378 154 109 173 444 85 321 153-154 171 444 322 313 155 444 320 108 112 108 248 314

Mark 1,1–16,8 1,2 1,3 1,4 1,6 1,7-8 1,8 1,9-11 1,10 1,12-13 1,12 1,13 1,14-15 1,14 1,15 1,16-20 1,21-28 1,23-28 1,23

23 124 126 128 145 186 84 109 145 436 156 247 436 321 247 106 107 247 107-108 106 109 84-85 243 312 288 319 326 259 392 432-433 247 321

1,26 1,27 1,29–6,13 1,29-34 1,29-31 1,30 1,39 1,40-42 1,44 1,45 2 2,1–3,6 2,1-12 2,1-4 2,4 2,5-7 2,5 2,6-7 2,8 2,13 2,15 2,16 2,23 2,25-26 2,26 2,28 3,1-6 3,7-12 3,7 3,9 3,10 3,13-19 3,15 3,17 3,19-21 3,19 3,20-35 3,20-30 3,20-29

3,20 3,22-30 3,22-27 3,22-26 3,22 3,23-30 3,23-24 3,24 3,26

247 247 319 260 260 298 427 23 321 147 121 126 128 23 127 38 254 263 156 298 320 320 241 411 412 412 247 241 313 313 274 121 126 427 261 322 259 444 241 156 241 242 39-40 241 241 243 241 319 326 238 254 241 269-272 314 145 166 23 166 255 257 312 242 245 239 245 245

INDEX OF REFERENCES

3,27 3,28-30 3,28-29 3,28 3,29 3,30 3,31-35 3,31 3,32-33 3,34-35 3,35 4 4,1-34 4,1-20 4,1-9 4,1-2 4,3-9 4,11 4,20 4,21 4,22 4,24 4,25 4,26 4,30 4,35-41 4,38 4,41 5,1-21 5,1-20 5,7 5,8 5,21-43 5,21 5,22-48 5,22-23 5,22 5,27 5,34 5,41 6,1-6 6,2 6,4 6,7-8 6,7 6,8-11 6,10

242 251 242 239 167 247 247 241 252 255 243 443 154 242 84 293 38 226 243 241 288 201 241 84-85 427 91-92 167 167-168 201 145 166 85 84-85 322 154 156 23 256 26 84 247 192 298-299 26 23 321 417 154 420 411-412 427 39-40 110 288 298 315 319 326 363 392 156 443 445 84 242 259 167 427

6,13 6,15 6,17 6,20 6,30-44 6,33 6,34 6,45–8,26 6,45-52 7 7,1-23 7,6-13 7,6 7,10 7,11 7,24-30 7,26 7,28 7,31 7,34 8,4 8,11-12 8,12 8,14-21 8,15 8,21 8,27-30 8,27 8,28 8,31 8,34–9,1 8,34 8,35 8,36 8,38 9,4-6 9,4-5 9,7 9,9-10 9,11-13 9,12-13 9,12 9,14-29 9,17 9,20 9,25 9,27-29

469 242 259 126-128 23 427 442 156 85 318 365-366 390 321 122 128 23 322 323 123 126 128 121 126 128 39-40 298 105 105 23 39-40 427 166 255 156 318 105 427 321 90 126-127 23 89 156 378 380 417 421 430 147 166 147 166 156 166 109 126-127 377 91 126-128 123-124 121 298 244 247 247 247 244

470 9,31 9,34-35 9,35 9,36-37 9,37 9,40 9,42 9,47 9,49-50 9,50 10 10,1-12 10,2-12 10,3-9 10,3-4 10,11-12 10,14 10,15 10,19 10,21 10,23 10,25 10,27 10,29-31 10,29 10,31 10,32-34 10,34 10,41-45 10,46-52 10,47-48 10,48 11–16 11,9-10 11,9 11,10 11,11 11,12-14 11,17 11,23-26 11,23 11,24 11,25 12 12,1 12,6 12,10-11 12,11

INDEX OF REFERENCES

378 84 43 419 84 166 165 167 23 85 91 145 122 128 318 323 122 126 128 167 84 85 318 122 84 84-85 84 443 33 35 85 443 167 201 419 379 378 288 256 128 126 403 39 40 126 128-129 23 319 326 122 156 102 167 167 318 127 226 419 121 123 84 124

12,12 12,16 12,17 12,19 12,22 12,26 12,28-34 12,28-31 12,34 12,35-36 12,36-37 12,36 12,37-40 12,38-39 12,38 12,40 12,41-44 13,9-10 13,9 13,11 13,12-13 13,12 13,14 13,15-16 13,21 13,28-32 13,29 13,30-32 13,31 13,32 13,35 14,3-9 14,22-25 14,25 14,26-31 14,26 14,27 14,32-42 14,32 14,36 14,43-52 14,43 14,45 14,53-65 14,65 14,70 14,72 15,1

23 126 28 122 126 128 156 417 419-420 122 126 128 427 288 321 122 215 85 88 126 128 122 247 86 88-89 43 167 88 405 405 410 165 444 167 247 165 166 21 41 189 167 167 147 85 24-25 25 201 25 166 288 319 326 409 288 85 288 23 121 123 288 427 39-40 288 156 90 288 23 157 420 427 104 157 23

INDEX OF REFERENCES

15,2-5 15,2 15,5 15,6-14 15,9 15,11 15,12-14 15,12 15,15 15,22-26 15,22 15,27-32 15,33-39 15,33 15,34ff. 15,34 15,35-36 15,35 15,38 15,40-41 15,43 16,1-8 16,2 16,8 16,9-20 Luke 1–2 1,1-4 1,1 1,2 1,3 1,5–2,25 1,5-25 1,5-23 1,5 1,17 1,26 1,27 1,32-33 1,32 1,33 1,35 1,39-56 1,51 1,53-55 1,55

288 155 288 288 155 155 155-156 155 156 288 39-40 288 288 322 23 109 39-40 40 127 126 128 378 288 85 153-154 321 288 322 378 179 370 433

367 102 185 193 236 362 377 388 183 185 383 265 362 289 405 129 128 247 110 129 257 84 129 129 84 395-396 289 247 257 129

1,57-80 1,66 1,68 1,73 1,76 1,77 2,4 2,11 2,13-14 2,22 2,25-38 2,25-26 2,27 2,28 2,32 2,36 2,39 2,41-50 2,42 2,51 3 3,1–24,11 3,1–9,50 3,1–6,19 3,1–4,30 3,1-38 3,1-18 3,1-17 3,1-6 3,1 3,2-4 3,3-6 3,3 3,7-18 3,7-16 3,7-14 3,7-9 3,7 3,8-9 3,8 3,10-15 3,10-14 3,16-17 3,16 3,17-18 3,17 3,18-20

471 289 247 129 129 84 88 110 129 257 257 396 128 405 257 247 396 257 129 396 110 108 396 110 73 414 348 289 11 403 414 362 144-146 403 403 157 146 144 84 407 313 145 289 29-30 145 403 310 308 129 403 257 403 137 144-145 174-175 247-248 251 313 436 145 145 403

472 3,19-20 3,19 3,21-22 3,22 3,23-38 3,23 3,31 4 4,1-14 4,1-13 4,1 4,2 4,3 4,5-7 4,5 4,9 4,12 4,13 4,14-30 4,14-15 4,14 4,15-16 4,15 4,16-30 4,16 4,17-22 4,17-19 4,17 4,22 4,23 4,24 4,25-27 4,25-26 4,27 4,31–6,19 4,31-44 4,31-37 4,31 4,33-37 4,33 4,34 4,35 4,36 4,38-39 4,39 4,41 4,43

INDEX OF REFERENCES

109 157 403 247-248 403 73 73 129 248 363 248 106 403 107 247-248 108 156 154 260 107 107-108 121 155 107-108 403 106 109 248 407 156 84 321 110 288 319 326 362 392 413 110 394 363 124 128 394 314 362 392 394 443 445 127 405 363 363 402 293 246 259 362 392 315 392 393 432 247 394 247 247 260 315 260 246 260 246

5,1-11 5,1-8 5,12–6,11 5,12-14 5,12 5,13 5,17-26 5,20-21 5,21 5,22 5,26 5,29-30 5,30 5,39 6,1-5 6,1 6,4 6,6-11 6,6 6,12-16 6,17-49 6,17 6,19 6,20–8,3 6,20–7,35 6,20-23 6,20 6,21 6,22 6,24-26 6,26-27 6,27-29 6,27 6,29-30 6,29 6,32-39 6,33-35 6,35 6,36 6,37-38 6,38 6,39-40 6,39 6,42 6,43-45 6,44 6,46

288 293 315 319 326 406 402-403 293 158 158 171 171 156-157 320 411 171 247 171 313 171 402 158 215 158 158 427 158 257 322 158 293 293 444 156 11 403 414 406 403 84 155 203 388 203 89 403 93 388 174 42 174 146 403 84 83 403 145 228 228 322 174 238 254 174 225 155

INDEX OF REFERENCES

7 7,1-10 7,1 7,2 7,3-6 7,3 7,5 7,6-9 7,6-8 7,6 7,7 7,8-9 7,8 7,9 7,10 7,11-17 7,14 7,16 7,17 7,18-35 7,18-23 7,19 7,20-22 7,20-21 7,21-22 7,21 7,22 7,24-35 7,27 7,28 7,29-30 7,30 7,36-50 7,36-47 7,36 7,39 7,40 7,41-43 7,44 7,47 7,48-49 7,48 7,49 7,50 8,1–9,50 8,1-3

311 412 9 11 297 299 303-304 313 330 309 300 310 312-313 315 403 155 300 316 321 305 315 155 171 300 316 300 310 312-313 329 308 313 174 305 299 310 312-313 11 403-407 409 414 406 407 407 11 146 403 109 289 155 172 260 403 246 251 260 251 289 145 316 84 403 89 11 319 326 403-404 407412 414 409 288 173 174 410 173-174 412 411 412 412 411-412 293 293 403

8,1 8,4–9,50 8,5-8 8,10 8,13 8,16 8,17 8,18 8,19-21 8,19 8,20-21 8,20 8,21 8,22-25 8,24 8,25 8,26-39 8,28 8,29 8,40-56 8,41 8,44 8,48 9,2 9,4-5 9,4 9,5 9,7 9,9 9,10-17 9,11 9,12 9,15–18,14 9,18 9,22 9,23 9,24-25 9,24 9,25 9,26 9,30-33 9,31-32 9,31 9,35 9,36 9,38 9,39 9,41-47

473 226 402 201 84 154 174 176 91-92 167 167-168 201-202 166 252 255 444 154 154 84 252 158 322 154 170-171 156 158 171 26 84 247 299 406 417 154 170 411 84 167 427 174-175 157 244 420 85 156 170 390 90 172 89 156 378-379 421 166 203 203 147 154 166 175-176 166 109 402 91 172 175-176 377 91 172 247 158

474 9,41 9,42-43 9,42 9,44 9,46 9,48 9,50 9,51–18,14 9,51-56 9,51 9,52-56 9,52 9,54-55 9,55 9,57-62 9,57 10–13 10,1-20 10,1 10,2 10,4-11 10,5-6 10,8 10,9 10,16 10,17-19 10,20 10,21-24 10,21 10,25-37 10,25-28 10,25 10,26 10,30-35 10,38-42 11 11,1-13 11,1-4 11,1 11,2-4 11,2 11,5-13 11,5-8 11,9-13 11,9-10 11,11 11,13 11,14-36

INDEX OF REFERENCES

158 402 154 247 379 84 84 166 167 11 263 365 398 403 414 293 264-265 198 316 109 248 257 322 154 262 274 246 317 174 167 167 201 203 155 165-166 260 84-85 215 247-248 198 215 413 288 88-89 172 321 215 28 257 413 86-89 122 275 213 243 262 266 243 163 82-85 95 248 213 243 405 213 243 262 266 74 167 174 84 248 252 248

11,14-32 11,14-28 11,14-26 11,14-23 11,14-20 11,14-15 11,14 11,15 11,16 11,17-23 11,17-18 11,17 11,18-19 11,18 11,19-20 11,19 11,20-22 11,20 11,21-28 11,21-22 11,21 11,23-26 11,23 11,24-26 11,24 11,27-28 11,29-54 11,29-36 11,29-32 11,29 11,33-36 11,33-34 11,33 11,34-36 11,37–18,14 11,37-54 11,37-52 11,37 11,39-52 11,39-40 11,39 11,42 11,43 11,44 11,46

215 244 238 146 254 269-272 314 319 326 244-246 166 256 314 155 166 213 243-244 246 255 257 137 139 145 166 244 257 155 166 244 250 257-258 262 244 145 166 239 246 258-259 244 245-247 251 245-246 259-260 245-247 251 84 163 245-250 251-253 146 242 250-252 251 252 167 252 31 244 255 35 244 252-253 255 262 413 214 244 244 255 262 90-91 166 174 176 244 262 87 91-93 167 174 214 93 214 414 244 257-258 87 87 214 214 87 175 87 87 167 388 87 43 87

INDEX OF REFERENCES

11,47 11,49-51 11,52 12–19 12,1–13,29 12,1-2 12,1 12,2-8 12,2 12,6 12,8-9 12,8 12,9 12,10 12,11-12 12,11 12,12 12,13-21 12,22-32 12,22 12,23-40 12,24 12,28 12,29 12,30 12,31 12,32-34 12,33 12,35–13,9 12,35-40 12,35 12,40 12,49-53 12,49 12,51 12,52-53 12,59 13–14 13 13,1-17 13,6-9 13,10-17 13,16 13,18-21 13,18-19 13,18 13,20

87 44 84 87-88 274 414 214 318 105 146 168 167-168 201 174 176 249 84 89 166 167 238-239 247-248 254 261 269-272 167 444 248 162 215 162 225 84 310 155 162 162 162 225 84-85 215 225 225 166 201 201 174 166 174 328 297 318 328 319 326 257 413-414 129 214 328 157 84 84

13,22–18,18 13,22–18,8 13,22-35 13,22-30 13,22-23 13,22 13,24-30 13,24 13,25 13,26-27 13,27 13,28-29 13,28 13,29 13,30 13,31-33 13,31 13,32 13,33 13,34-35 14–17 14,1-10 14,1-6 14,3 14,5 14,11 14,12-15 14,15 14,16–15,7 14,16-24 14,16-23 14,16 14,23 14,24-25 14,26-27 14,27 14,28-33 14,34-35 15,3-32 15,4-7 15,8-10 15,11-32 15,17 16,1-13 16,1-8 16,13 16,16-18

475 339 268 318-319 215 328 265 268 318 328 318 155-156 173 318 318 155 171 9 297 299 303-305 311 317-318 323-325 215 309 84 167 201 318 328 318 315 318 318 44 318 328 263 328 257 89 172 322 328 328 84 146 33 328 155 226 328 203 328 166 328 91 93 145 328 257 42 42 413 174 413 28 413 122

476 16,16 16,18 16,19-31 16,21 16,23-31 16,24 16,29 16,31 17,2 17,4 17,6 17,11 17,21 17,23 17,28-29 17,31 17,32 17,33 18,1-8 18,2-5 18,15–24,53 18,15-43 18,16 18,18 18,22 18,24 18,25 18,27 18,29-30 18,29 18,30 18,31-33 18,31 18,33 18,37 18,38 19,1-27 19,1-10 19,9 19,11-27 19,11 19,12-27 19,12-14 19,15 19,16-26 19,16 19,17 19,18

INDEX OF REFERENCES

84 167 318 325 105 127 129 128 128 146 155 172 167 265 168 167 127 146 167 127 166-167 413 405 348 402 84 172 172 84-85 84 84 443 33 34 443 85 124 265 444 378-379 394 172 11 403 414 403 414 129 104 403 33 403 403 403 154-155 173 155 154-155 173

19,19 19,20 19,26 19,27 19,28–24,11 19,28 19,29 19,30 19,36-40 19,37 19,38 19,39-44 19,41 19,46 20,1-8 20,1 20,2 20,5 20,9 20,11 20,13 20,14 20,17 20,18 20,21 20,22 20,27 20,28 20,32 20,37 20,42 20,44 20,45-47 20,46 21,7 21,8 21,12 21,14-15 21,15 21,16 21,18 21,21 21,22 21,24 21,31 21,32-33 21,33 21,34-36

155 154-155 173 166 403 402 265 265 172 158 265 403 172 402-403 265 172 175-176 158 158 172 172 226 175-176 175-176 419 172 85 403 172 175-176 154 171 172 156 419 427 247 175-176 88 167 172 173 444 167 247 166 403 167 403 403 85 24 201 403

477

INDEX OF REFERENCES

21,36 22,3 22,14-38 22,15-20 22,15-16 22,15 22,19-20 22,19 22,20 22,21-23 22,24-30 22,28-38 22,28-30 22,28 22,29 22,30 22,31-34 22,31 22,37 22,39-46 22,42 22,43-44 22,47-53 22,47 22,48 22,51 22,53 22,54-71 22,57 22,58 22,59 22,61-63 22,62 22,64 22,67 22,70 23,2-17 23,2-5 23,3 23,17-23 23,17-22 23,18 23,20 23,21 23,22 23,24-25 23,25 23,27-32

166 107 260 11 288 403 414 288 288 173 173 288 288 403 414 33-34 36 35 36 175 288 107 260 124 158 288 173 403 288 156 90 175-176 403 403 260 288 173 173-174 154 173 157 104 157 173 173 155 174 403 288 155 173-174 288 155 155 155 173 154 175-176 288 175-176 403

23,33-34 23,35-43 23,35 23,36 23,39-43 23,40 23,44-48 23,45 23,47 23,48 23,49 23,51 23,52 24,1-12 24,7-8 24,7 24,12 24,19 24,44 24,46

288 288 173 154 403 174 288 378 173 403 288 154 153-154 171 288 403 380 368-370 391 414 394 128 124

John 1,2-3 2,12 3,22-30 4,46-54 4,46 4,47 4,49 4,50-51 4,50 4,51 4,53 6,1-15 12,20-22 13,36 16,12 16,13 20,3-10 20,30-31 21,24-25

123 393 109 297 299 329 300 300 300 300 300-301 300 300-301 442 315 419 383 383 368-369 391 236 236

Acts 1,5 1,15-26 1,18 2 2,34-35

247 112 154 248 124

478 2,38 3,18 3,22-23 7,2 7,19 7,36 7,48 8,28 8,32-33 8,36 9,32-43 10 10,6 10,28 10,30-31 11,16 12,3 13,7 14,9 16,17 16,30 16,37 17,2-3 17,22 18,25-26 19,35 21,37 22,2 22,17-21 22,27-28 23,5 23,17-18 23,35 25,5 25,22 25,24

INDEX OF REFERENCES

174 124 124 174 154 154 84 22 124 174 405 311 154 174 174 247 109 154 154 84 174 174 124 174 154 174 174 174 108 174 174 174 174 174 174 174

26,1 26,18 26,22-23 26,24-25 26,32 29,13

174 260 124 174 174 245

Romans 3,10-18 8,15 10,4

197 39 49

1 Corinthians 15,4 15,8

380 419

2 Corinthians 11,13

395

Galatians 1,7 2,4 2,14 4,6 4,30

395 395 395 39 107

Ephesians 6,11 6,13

252 252

2 Peter 1,17

377 377 384

Revelation 11,2

109

OTHER REFERENCES Adamantius Dialogue 1,24 2,14 2,15 2,18 5,9 5,12

362 394 393 393 394 394 396 378

Adso Antichrist

379

Aelius Theon Progymnasmata 139-140 140-142 140 141

426 435 426 435 437 428 426

479

INDEX OF REFERENCES

Alcuin

349

Ammonius of Alexandria

56–59 59,5b-7a 8HevXIIgr

222 223 273

Didache 1,2-5

294

Diodorus Siculus

9 279

9 285-286 292 Anselm Proslogion

349 349

Antonius Melissa

141

Apollonius Sophista Lexicon Homericum

141

Arrian of Nicomedia Anabasis of Alexander 1.praef.1 Ascension of Isaiah 11,7-10

282

Eadmer

349

1 Enoch 37–44 56,5-7 67,4-13

64-65 65 65

345

396 387

Basilides

395

Baudry of Bourgueil Life of Saint Cuthbert

352 350

Cassius Dio

282

Cicero Second Philippic

341 354

Dead Sea Scrolls 4Q41 4Q521 11Q19 3–49 18,2-10 22,8

Dionysius 279 345

Augustine Harmony of the Gospels 1,2,3 1,2,4

Clement of Alexandria Stromateis 2,5,3–6,4 3,86-95 7,16,93-94

Diogenes Laertius Lives of Eminent Philosophers 3,37 349

387 387

10 338-339 335 396 396

273 65 8 219-232 221 224 231

Epiphanius Panarion 42,11,3 42,11,6(16) 42,11,6(52) Eusebius

361-362 361 378 379 143 145-146 165 168 179 285-286

Church History I,7 II,15 III,4,7 III,24 III,39,15-16 III,39,15 VI,12,2-6

389 144 144 144 144 179 396

Epistula ad Carpianum

143

Gospel of Peter

396

Gospel of Thomas

8 28 50 69 195217 201 202-203 202-203 202

4–16 5 6 7

480 8 9 12 14 15 17 18 20 32 33 36 37 43,3–45,4 47,3-5 54 55 56 57 63–66 65–66 68–69 76 96 109

INDEX OF REFERENCES

201-203 201-203 202 203 202 203 203 201 203 203 203 203 202 202 203 203 203 201 202 202 203 201 201 201

1,7,2 1,8,4 1,15,1 3,10,3 3,11,3 3,15,1-2

396 396 396 396 396 364

Jerome Letters 18,16

348

Joannes Stobaeus Anthologion

141

Josephus Antiquities of the Jews 8.45-49 17.6.5.171-173 Jewish War 1.33.5.657-658 7.185

9 279 282-283 294 282-283 245 65 65 245

Julius Africanus

389 399

Julius Cassianus

396

Guibert of Nogent Life 1,17

350

Justin Martyr

284-285 338 360

Heracleon

393

Juvenal Satires 10,126

341 354

Lactantius Divine Institutes 7,14,3

379

Livy

9 235 282

Lull of Mainz

349

Hippolytus Refutation of All Heresies 6,35,3 7,26,9

395 395

Homer Iliad Odyssey

140-141 427 20 20

Hugh of Saint Victor

350

Ignatius of Antioch Ephesians 18,2 Smyrneans 1,1 Trallians 9,1

396 396 396

Irenaeus Against Heresies 1,3,2

360-363 385 393 396 Antitheses 395 Gospel of Marcion / Evangelium 1 10 387-400 *Luke *3,1 362 *4,16 362 Marcion

360 396

481

INDEX OF REFERENCES

*4,23-24 *4,28-30 *4,31-37 *4,42-43 *9,22 *9,44 *9,51–18,14 *17,18 *18,33 *Mark *8,31 *9,31 *10,34 *16,7 *16,8 *Matthew *16,21 *17,23 *20,19 *28,19-20

362 362 362 362 378-379 379 365 363 378-379

Megingoz of Würzburg

350

379 379 379 382 382 379 379 379 383

Melissa Augustana (= Melissa Barocciana; Florilegium Monacense; Florilegium Baroccianum) 141 Menander Kekryphalos fr. 209

341

Mishnah m. Ber. 5,5 m. Kel. 24,6

315 340

Origen Against Celsus 2,22 2,27 4,74 Commentary on John 10,9 Commentary on Matthew 10,17 Hexapla Homilies on Luke 7,4 16,5 20,2 33,1

393 396 364 394 393 393 396 286 396 364 396 364 393

Ovid Metamorphoses 9,522-570

335

Papias

179

Philo of Alexandria

8 10 141-142 195-217 On the Giants 208 1–5 207 20–58 207 Life of Moses 197 212 229 Migration of Abraham 197 209 1–13 213 1 207 3–54 207 5 209 8 210 13 210 On the Posterity of Cain 335 338-339 5–20 339 Special Laws 197 229 On the Unchangeableness of God 208-209 1–4 207 86–90 209 86 209 104 209 109–116 209 Plato Lakonians fr. 71 The Laws Protagoras 342a-343b Republic 372d

341 349 428 340

Pliny the Younger Letters 9,36,1-3

205

Plutarch Lives Cat. Min. 48,4 Pomp. 55,5

9 282 290 294 281-282 315 315

Pollio

282

482

INDEX OF REFERENCES

Polybius History

235

Protevangelium of James 396 11–17 396 20,1-3 396 Ptolemy

396

Quintilian Institutio Oratoria 1,9,1-3 1,9,2 2,7,2-4 10,5,4-11 10,5,9

426 426 435 426 426 426

Soranus of Ephesus

338

Strabo

279

Suetonius Lives of Eminent Grammarians 24,4 426 4,5 426 4 426 Testament of Moses 10,1

243 282 345

Tacitus Annals 13,20

345

Tatian Diatessaron

9 283 292 294 279 283-286 396

Tertullian

360-363 393-395 399-400

PRINTED ON PERMANENT PAPER

• IMPRIME

Against the Jews 8,13 13,5 Against Marcion 1,1,5 4,3,1-4 4,4,1 4,4,3 4,4,4 4,4,5 4,5,2 4,5,5 4,5,7 4,7,1 4,7,4 4,8,2 4,9,15 4,12,14 4,21,7 4,43,7 4,43,9 On the Flesh of Christ 23,2-6

338 338 361 395 361 395 396 395 395 396 394 394 393 394 393 394 394 378 362 361 396

Testimony of Truth

396

Valentinus

395

Virgil Aeneid

140-141 140

Willibald Life of St. Boniface

350 349

Xenophon Symposium 2,1

341

SUR PAPIER PERMANENT

N.V. PEETERS S.A., WAROTSTRAAT

• GEDRUKT

OP DUURZAAM PAPIER

50, B-3020 HERENT

- ISO 9706